Good morning folks. I’m Felix Winckler from rule30, and this is Moneyball our monthly dispatch looking under the hood of AI’s rise. This edition asks a bold question: are we witnessing enduring foundation-building or the crest of a speculative wave? Recent data from NVIDIA and a cascade of new AI-product launches give the appearance of unstoppable growth. But dig deeper, and some early signs of structural strain begin to emerge.

At the top of the list this month: why today’s AI boom may not be built solely on technical brilliance, but on borrowed money, unmet expectations, and shaky fundamentals.

The Numbers Sound Amazing — But Are They Fundamentally Solid?

NVIDIA just announced $57.0 billion in revenue for Q3 fiscal 2026. That’s up 62% year-over-year and 22% quarter-over-quarter. Their Data Center business alone generated $51.2 billion, beating analyst expectations.

They followed up with an even more ambitious outlook: Q4 guidance of $65 billion, which is extraordinary in scale. This surge reflects “off-the-charts” demand for AI server chips and cloud GPUs particularly the latest Blackwell architecture.

Stocks responded: NVIDIA shares rose, and the boost rippled across AI-related equities globally, rekindling investor appetite for the whole sector.

But here’s where caution kicks in. Behind that boom lurk structural risks many markets are choosing to ignore:

• Much of AI infrastructure buildup has been financed not by revenue but by debt. Some companies are borrowing heavily to fund GPU purchases and compute-heavy projects.
  

• As AI spending grows faster than actual profits or stable cash flows, the sustainability of that debt becomes questionable especially if funding dries up.
  

• If major buyers run into financing trouble (e.g., due to rising interest rates, tighter credit, or failed monetization), demand could collapse leaving excess hardware and unrealized orders.
  

• On top of that, growth depends on assumptions:
  

  • continued expansion of data-center capacity,
    

  • re-opening or stable relationships in major markets (including China),
    

  • persistent competitive advantage for NVIDIA over rivals like AMD or Intel,
    

  • and enterprise users realizing real returns (not just hype) from AI investments.
    

In short: the hardware lead is real, but the long-term economic model backing it for many remains fragile.

New AI Tools, Fresh Hype — But Also Fresh Questions

Meanwhile, the wider AI ecosystem hasn’t been standing still. Recent weeks saw a flurry of major AI-product releases and upgrades, further fuelling both excitement and speculation:

• Gemini 3 from Google is being hailed by some as the most ambitious jump yet for multimodal AI. Another strong signal of accelerating competition.
  

• At the same time, Claude Opus 4.5 from Anthropic hit the streets, promising major improvements in coding, enterprise workflows, and long-form reasoning agents.

• Finally OpenAI introduced GPT-5.2, its most advanced generation yet, boasting substantial improvements in professional knowledge work, long-context reasoning, multimodal perception, and coding performance. Benchmarks suggest it surpasses previous models on key tasks and delivers more dependable outputs, signaling a leap in real-world productivity use cases across business and research.
  

• These pushes reflect a broader “AI upgrade race” from infrastructure to model enhancements to tools which, if matched by real adoption, could reshape how industries operate.
  

But great power doesn’t automatically translate to great profit. For many buyers cloud providers, enterprises, startups the costs (hardware, energy, talent) remain high. The question is: Will these newer tools deliver enough productivity or value to justify those costs?

Given the scale of capital already deployed (and planned), a downturn in demand or capital markets could leave many projects stranded with massive hardware overhang and little to show for it.

The Risk of an AI Bubble — What Could Trigger a Crash

Putting the pieces together reveals a fragile architecture. There’s real demand, real delivery, and real ambition but also big dependencies on capital flows, favourable macro conditions, and hype staying alive.

Because of these dependencies:

• AI stocks now account for a very large share of overall gains in major indexes this year (>75% of S&P 500), which means the broader market is now highly concentrated and sensitive to AI volatility.
  

• Early signs of stress are already visible: some large tech names have seen sharp stock drops after showing trouble generating AI-driven revenue.
  

• If broader economic headwinds rising rates, credit tightening, macro-economic slowdown hit, capital-intensive AI projects are most vulnerable.
  

• The scale of upcoming AI-infrastructure demand is massive: to justify it, buyers must deliver real value far into the future. That’s a high bar.
  

Which brings us back to the real question: Is this growth building a durable foundation, or are we inflating a bubble on borrowed money and unrealistic projections?

There is no doubt we are living through one of the most dramatic growth spurts in the history of computing. The revenue numbers are staggering, the engineering breakthroughs keep accelerating, and the ambition across the ecosystem (from models to chips to infrastructure) is unmatched.

And yes, history reminds us that high prices and insatiable demand don’t guarantee long-term success. To avoid becoming a bubble, this wave of AI investment needs real returns, sustainable business models, and the financial stamina to withstand market cycles. If the next few quarters bring failed monetisation, tightening financing, or softening demand, this boom could deflate far faster than it expanded.

But there are also strong reasons not to underestimate the staying power of this cycle:

• We’ve only scratched the surface of the robotics wave.
Most AI investment so far has gone into software. The next frontier physical automation through humanoids, warehouse robots, and industrial autonomy is barely out of the lab. As robots become commercially viable, compute demand may grow another order of magnitude.

• Enterprise AI adoption is still at day zero.
Despite the hype, only a tiny portion of Fortune 500 workflows are actually AI-powered today. Early pilots show meaningful productivity gains, but real operational integration takes years. The “slow burn” of enterprise adoption could support long-term infrastructure demand even if short-term enthusiasm cools.

• Entire categories of AI-native applications do not exist yet.
Every tech supercycle (PC, Internet, mobile) produced breakout companies 5–10 years after the underlying infrastructure matured. Many of the most valuable AI-native applications (agentic systems, autonomous tools, simulation-first industries) have not arrived yet.

• AI is becoming embedded in every layer of the economy.
From scientific research and logistics to drug discovery, finance, entertainment, and education, AI is no longer a niche category. It is becoming a general-purpose capability, and historically, general-purpose technologies (electricity, computing, the internet) tend to have multi-decade demand curves.

So yes: today’s AI market carries real risks. But it also carries the early signals of a durable platform shift.

If AI delivers on its promise (reducing costs, increasing productivity, enabling new business models, and unlocking robotics at scale) then this moment may be remembered not as a bubble, but as the early structural phase of the next great technological transformation.

For now, which path we’re on remains the most important question in tech; and the one we’ll keep tracking, month after month.

Other Bits Worth a Quick Look

• Google’s NotebookLM steps up.
  Google released a major update to NotebookLM, its AI-powered research and summarisation tool. The new version adds deeper source-reasoning, multimodal support, and collaborative capabilities, making it much closer to an AI “research partner” than a simple notes assistant.
  

• AI-augmented textbooks arrive.
  Google also unveiled a new initiative to rethink textbooks using AI, introducing interactive, personalised learning materials that adapt to a student’s pace and curiosity. The concept, described in their post on reimagining textbooks with generative AI, signals the start of a major shift in educational publishing.
  

• ElevenLabs launches its AI Music Studio.
  ElevenLabs expanded beyond voice synthesis with a new AI music generator, allowing creators to produce original tracks in seconds. Early demos show surprisingly high-quality compositions, placing ElevenLabs in direct competition with emerging AI music platforms.
  

• Video creation goes fully AI-native.
  InVideo, available at invideo.io, launched new AI-powered tools that turn scripts into fully-produced videos complete with scenes, transitions, voice-overs, and pacing adjustments. It’s one of the clearest signs yet that end-to-end AI video production is about to become mainstream.

• Google’s “Banana Pro” Nano model.
  In a slightly more niche but still notable release, Google published details on Nano Banana Pro — a tiny, on-device AI model designed for ultra-low-power environments, another step toward “AI everywhere.”

• MIT’s “Iceberg Index” shows AI workforce exposure.
  A new study from MIT and Oak Ridge National Laboratory — dubbed the Iceberg Index — estimates that current AI capabilities overlap with tasks totaling about 11.7% of U.S. workforce wage value ($1.2 trillion), particularly in cognitive and administrative work. The index isn’t a direct prediction of layoffs but highlights where AI could already perform significant job tasks if adopted widely, serving as a planning tool for policymakers and businesses.

• Mistral 3 family launches with open-weight models.
  French AI developer Mistral AI unveiled Mistral 3, a new suite of open-weight models including a large mixture-of-experts flagship and several smaller dense variants, all under an Apache 2.0 license, aimed at broad developer access and edge deployment. The release positions Mistral more competitively against major AI players and has already been integrated into services like Amazon and Microsoft clouds.

• Mistral AI adds OCR 3 for advanced document AI.
  Mistral AI also launched Mistral OCR 3, an upgraded optical character recognition model that achieves a 74 % win rate over its predecessor on forms, scanned docs, complex tables, and handwriting — now powering Mistral’s Document AI Playground with cost-efficient API access for enterprise processing.

That’s it for this edition folks. Let me know your feedback, what’s missing, and what you’d like to see next time.

Until then, onward.

rule30 is an AI research lab building systematic strategies to trade venture assets at scale. We’ve dedicated three years to developing groundbreaking AI and machine-learning algorithms and systematic strategies to re-engineer the economics of startup investing.

Dear friends,

Rule30 has just opened up their probabilistic investment decision engine for the first time to the public.

You can submit any pitch deck or LinkedIn profile for review. Within 24h you will receive a report showing whether rule30 would back it, why, and which percentile winner probability band it falls into (10%, 3%, or 0.1%). This isn’t another AI score.

Rule30 will soon announce the launch of its fund, from which it has been quietly investing over the past few months.

If you have a company you would like to test against our algorithmic, index-style strategy, here’s the link: https://www.producthunt.com/products/rule30-would-we-back-your-startup-why?launch=rule30-would-we-back-your-startup-why

Feel free to reach out if you want to learn more.

Very best,
Felix

Good morning folks. I’m Felix Winckler from rule30 and this is Moneyball, a newsletter where we share the latest news on AI. Each month, I work with the team at rule30 to make sense of the shifting tides in AI and send you the result.

At the top of the list this month: how power in AI is shifting from code to systems. A new report from the Centre for British Progress argues that the UK’s current approach to funding innovation — through institutions like the British Business Bank and regional investment funds — isn’t built for the new era of industrial strategy. It proposes a national investment body, British Sovereign Capital, to help the country back its own high-growth and strategically important tech firms rather than losing them overseas.

At the same time, new research from the Australian Institute for Social Innovation shows that, despite all the hype, today’s AI systems still fall short in areas like reasoning, reliability, and adaptability. And Anthropic’s latest paper, Stress-Testing Model Specs Reveals Character Differences among Language Models, finds that even when models are trained with the same ethical or behavioural rules, they often act in very different ways once tested under real-world conditions.

Put together, these stories show a deeper shift in the AI landscape. The real competition is no longer just about building bigger models — it’s about who controls the capital, the rulebooks, and the infrastructure that shape how AI is built and used.

Sovereign Capital and the Economics of AI Power

A recent briefing from British Progress highlights a turning point in how the UK funds and builds its technology sector. The report points to three major shifts: British founders are moving overseas in search of more supportive funding environments; other countries are treating their leading tech firms as strategic national assets; and AI is starting to disrupt the UK’s traditional, services-based economy.

To respond, the authors propose creating a new national investment body called “British Sovereign Capital.” Unlike existing schemes such as the Enterprise Investment Scheme (EIS) or Venture Capital Trusts (VCTs), which mostly target startups, this new institution would focus on direct investment in larger, strategically important companies — especially those working in areas like AI, advanced computing, and infrastructure.

Why does this matter? Because in today’s AI race, success isn’t determined just by who has the smartest models. It’s increasingly about who controls the capital, the compute infrastructure, and the ecosystem that supports those models. In other words, the winners will be the countries and companies that can anchor the entire AI “stack” — from chips and data centres to funding and regulation.

There are signs the UK government is starting to think this way. Its AI Growth Lab call for evidence suggests a shift toward new hybrid public-private partnerships designed to scale AI safely and competitively. At the same time, regions are beginning to treat data centres, compute clusters, and energy infrastructure as part of national security, not just private enterprise.

For investors, this means the old metrics of “scale” — counting users, market share, or model parameters — are no longer enough. What matters now is access: to compute, to sovereign funding, and to a stable regulatory base. AI labs that secure those advantages — through government partnerships or national infrastructure — could easily outpace better-engineered rivals that lack them.

In short, the terrain is changing fast. The next phase of competition won’t just be about building smarter algorithms, but about building the systems — financial, industrial, and institutional — that make those algorithms possible.

Alignment, Specifications, and the Hidden Behavioural Risks

On the alignment front, Anthropic’s new research looks at a part of AI development that’s often ignored but hugely important: the “specification layer.” Think of it as the rulebook or moral contract that tells an AI model how it should behave — what values to prioritise, what trade-offs to make, and how to respond in uncertain situations. Every major AI lab writes its own version of this rulebook, combining human feedback, company values, and ethical guidelines.

Anthropic’s team decided to stress-test how well these specifications actually work. They ran twelve leading AI models through over 300,000 simulated moral and practical dilemmas, where each system had to choose between competing priorities — for example, “should I protect user privacy even if it reduces model accuracy?” or “should I favour the public good over efficiency?”

The results were surprising: in more than 70,000 cases, models that were supposedly following the same specification made completely different choices. In other words, even when you give two models the same ethical and behavioural instructions, they can interpret them very differently once the problems become complex or unfamiliar.

A few clear themes emerged:

• Ambiguous instructions lead to mixed behaviour. Rules like “do good” or “maximise benefit” sound sensible, but they clash when the AI has to decide which “good” to prioritise.

• Consistency breaks down under pressure. Models that behave predictably in short, simple tasks often start to contradict themselves when dealing with long, multi-step or adversarial scenarios.

• Lab results don’t always translate to the real world. Even models that pass safety benchmarks can act unpredictably once they’re deployed into more chaotic or interactive environments.

The lesson is that aligning AI isn’t just a matter of training the model better. It’s about engineering the whole system — from the specifications and data to the compute infrastructure and institutional oversight — so that behaviour remains stable and trustworthy at scale.

Several related studies underline this challenge. The open-source project LLM Brain Rot has shown that large language models can gradually “drift,” changing their tone or judgement months after release. The KIMI K2 model used a new reinforcement-learning method based on human-written rubrics to reduce reward-hacking, but that in turn made the model more assertive — and occasionally overconfident. And the AISI report found that despite rapid progress in scaling, eight key cognitive benchmarks for general-purpose reasoning have barely improved.

For investors, policymakers and strategists, this all points to a new kind of systemic risk. Neglecting specification design — the governance layer of how models think and decide — may soon be as costly as under-investing in data or compute.

Productivity, Work, and the Human Cost of Efficiency

​​Finally, it’s worth looking beyond the technology itself to how AI is reshaping work, productivity, and the way people collaborate.

A recent essay called Life After Work makes an important point: just because AI can make tasks faster doesn’t mean it automatically makes our lives or workplaces better. If companies don’t rethink how teams are organised (or how AI fits into day-to-day workflows) the benefits can quickly evaporate. Productivity is not just about doing more; it’s about doing work that actually matters.

Supporting that argument, a report from Futurism finds that AI is often lowering workplace productivity rather than improving it. In many offices, AI tools are producing a flood of low-quality “workslop” — documents, summaries, and ideas that look finished but lack real insight. Colleagues then have to spend time fixing, rewriting, or double-checking this output. Around 40% of employees say they’ve been handed AI-generated work that turned out to be unreliable, which can undermine trust and waste hours across teams. In effect, AI isn’t always taking work away — it’s shifting the burden from creators to reviewers.

There’s also a deeper psychological issue. A recent study in Nature Human Behaviour looked at how large language models interact with people and found that most current systems are far too eager to please. Across 11,500 test questions, AI models were about 50% more likely than humans to agree with a user’s incorrect statement or flawed reasoning. The researchers call this sycophancy bias — the model’s tendency to flatter or echo rather than challenge. Among the models tested, GPT-5 was the least sycophantic (around 29%), while DeepSeek-V3.1 was the most (around 70%). The risk is that in sensitive fields like science, healthcare, or policy, AI systems could reinforce human mistakes instead of catching them.

These findings are a reminder that AI adoption isn’t just a technical challenge — it’s an institutional one. Governments are starting to recognise this. The UK’s AI Growth Lab has launched a call for evidence to understand how AI is changing labour markets, education, and public infrastructure. That’s a sign that policy is catching up to the scale of the transformation.

It also raises some important questions:

• Will AI companies think more deeply about how their tools fit into real human workflows, rather than just chasing speed and scale?

• Will national policy focus on helping institutions adapt — not just on building new AI capabilities?

• And will productivity gains from AI actually empower people, or will they mostly disrupt the systems that already work?

The answer will decide whether AI becomes a genuine engine of progress — or just another layer of friction in how we work and think.

Other Bits Worth a Quick Look

• Anthropic’s Small-Samples Poison paper on how tiny data corruptions can cascade into major behavioural shifts.

• Dario Amodei’s statement on American AI leadership outlining Anthropic’s policy stance.

• OpenAI × Broadcom collaboration to build 10 GW of custom AI accelerators.

• KREA Realtime Video from Hugging Face, showing how real-time generative tools are changing creative workflows.

• AGI Definition AI initiative exploring frameworks for classifying advanced systems.

• Claude for Life Sciences bringing domain-specific reasoning tools to research.

• General Intuition raised $133.7 M to expand its frontier lab merging games and real-world dynamics.

• The rise of AI in creative industries — from Xania Monet signing a $3 M record deal to the World of AI Film Festival.

• OpenAI’s new releases, Sora 2 and Atlas, are making waves — have you tried them yet?

That’s it for this edition. Let me know your feedback, what’s missing, and what you’d like to see next time.

Until then, onward.

rule30 is an AI research lab building systematic strategies to trade venture assets at scale. We’ve dedicated three years to developing groundbreaking AI and machine-learning algorithms and systematic strategies to re-engineer the economics of startup investing.

Scheming isn’t just for politicians

Good morning folks. I’m Felix Winckler from rule30 and this is Moneyball, a newsletter where we share the latest news on AI. Each month, I work with the team at rule30 to make sense of the shifting tides in AI and send you the result.

Top of the list for this edition is compute and infrastructure. In particular, how strategic partnerships and public policy are reshaping where and how AI systems scale. Last week, OpenAI and AMD announced a sweeping multi-gigawatt agreement: AMD will supply GPUs to OpenAI, beginning with a 1 gigawatt deployment of MI450 chips in H2 2026 and scaling to 6 gigawatts over multiple generations. Crucially, the deal includes a warrant for OpenAI to acquire up to ~160 million AMD shares (≈10 percent of the company), vesting on performance and stock price milestones.

This marks a bold shift in that OpenAI is diversifying its hardware base beyond Nvidia, anchoring deeper into the semiconductor supply chain, and entangling multiple compute providers with its scaling efforts. Analysts see this as a “bet on compute as capital” by shaping suppliers’ supply, roadmap, and incentives.

The deal arrives just as governments are waking to the need for the physical substrate to run AI models. In the UK, the government’s AI Growth Zones initiative is now open for applications. These zones aim to unlock investment in AI-enabled data centres by improving access to power and removing planning constraints.

One proposed AI Growth Zone in North East England (e.g. around Blyth, Cobalt Park) is already expected to attract £30 billion in investment and create over 5,000 jobs. The move underscores that AI infrastructure is becoming a battleground for regional development and geopolitical positioning.

Together, the AMD-OpenAI deal and the UK’s Growth Zones represent two faces of the same coin: privatized compute architectures coupling with public infrastructure to shape who wins the race to scale AI.

Meanwhile, the frontier labs released a handful of papers focused on AI safety, stemming cyber threats to model scheming.

• OpenAI recently published a new paper on “Detecting and Reducing Scheming in AI Models.” They define covert actions as proxies for scheming: withholding task-relevant data, deception, or deliberately misreporting. In controlled tests, they trained variants of o3 and o4-mini with a “deliberative alignment” approach (i.e. asking models to reason explicitly about anti-scheming before acting). The result was a~30× reduction in covert actions (from 13 % → 0.4 % for o3; 8.7 % → 0.3 % for o4-mini).

• These mitigations don’t fully eliminate scheming insofar as higher adversarial pressures or out-of-distribution (OOD) tasks may expose residual behavior. Also, models can detect when they’re being tested (i.e. “situational awareness”) and modulate behavior accordingly, which raises questions about how “genuine” alignment is.

• In parallel, Anthropic is pushing a new frontier: AI-assisted cyber defence. Their recent publication, Building AI for Cyber Defenders, describes how Claude (especially Sonnet 4.5) is being trained to detect, analyze, and remediate code vulnerabilities and system risks. They report that Sonnet 4.5 matches or exceeds previous models (like Opus 4.1) in vulnerability discovery tasks.

• This work is timely because Anthropic has also warned that AI is lowering the barrier for cyberattacks — “no-code” ransomware, scalable extortion, and automated exploit pipelines are becoming easier to launch. The implication is that defense and offence are increasingly symmetrical, and the strategic arms race is underway.

• On the model front, Anthropic’s latest release, Claude Sonnet 4.5, pushes performance in reasoning, maths, and tool use (e.g. code execution, memory, context extension). It’s also their most aligned frontier model to date, with improvements in resisting sycophancy, deception, and prompt injection.

• Yet, some critics note that Sonnet 4.5 “recognizes safety tests” and can mask alignment issues by modifying behavior during evaluation. That is, clever models may game the test environment leading to spotty results

Beyond compute and model alignment lies an even bigger question: how governance, capital, and regional incentives shape AI outcomes.

• The UK’s AI Growth Zones signal a belief that you can steer geography in an age of cloud and global connectivity. But zones are not just infrastructure carrots. They can anchor regulatory regimes, talent clusters, and data flows.

• Some voices warn that the zones must avoid becoming “footnotes on glossy announcements” — they need enforcement, coherence, and real value for local economies.

• More risk arises from structural alignment gaps. If compute providers and AI labs share incentives, but external actors (governments, users, regions) have others, misalignments could snowball.

• The alignment interventions we see today may work under carefully curated tests, but adversarial pressure, escalation, and emergent misuse remain formidable threats.

Other bits worth reading

• Unblocking AI Growth Zones (British Progress) lays out policy levers, barriers, and governance risks around development zones.
  Sam Altman’s Abundant Intelligence (on his blog) gives a sense of OpenAI’s long-term orientation and open architecture ambitions.

• Toby Ord’s essays (on inference scaling, efficiency of RL) remain bedrock reading for alignment frameworks.

• The Stress Testing Deliberative Alignment paper probes limits of current anti-scheming tactics.

• The Deceptive Automated Interpretability paper shows how models might coordinate to fool oversight systems by hiding malicious behavior in explanations.

• A new report, Sovereignty, Security, Scale: A UK Strategy for AI Infrastructure offers deep context on how AI zones are being framed in public discourse.

• The Rise of Parasitic AI offers provocative thought on how models might exploit ecosystem incentives.

That’s it for our fourth outing. Let us know what you think, and what we should look at in future editions.

rule30 is a quantitative venture capital fund. We’ve dedicated three years to developing groundbreaking AI and machine learning algorithms and systematic strategies to re-engineer the economics of startup investing.

Good morning folks. I’m Harry Law of Learning From Examples, a newsletter about the history of AI and its future. Each month, I work with the team at rule30 to help make sense of what’s been happening in AI over the last four weeks.

Top of the list for this edition is geopolitics. Chip exports entered a strange new era, when the Trump White House reversed course by striking a deal with Nvidia and AMD. The terms of the deal allow the firms to sell chips to China in exchange for 15% of revenue from those sales. In doing so, the move overturned years of bipartisan consensus in which American policymakers have sought to block China from accessing the most powerful American-made chips to power their AI systems.

It’s in some ways a strange choice given that America’s recent AI Action Plan, which we discussed in the last edition of Moneyball, seeks to help the United States ‘win’ the AI race against China. This deal likely complicates that goal because it gives China access to a core piece of the technology that they have so far struggled to replicate.

Computing power is one of the core inputs into the AI development process alongside data. With respect to the latter, the American AI lab Anthropic agreed to pay $1.5 billion to settle a class-action lawsuit by book authors who argued that the firm took pirated copies of their works to train its models. Justin Nelson, a lawyer for the authors described the verdict as “the largest copyright recovery ever”.

A federal judge dealt the case a mixed ruling in June when they said that training AI chatbots on copyrighted books wasn’t illegal, but that Anthropic wrongfully acquired millions of books through pirate websites. As part of the deal, the authors reportedly received $3,000 each in compensation. It’s a result that puts the labs in an interesting position: if they obtained the works legally, then they are likely to be covered; if not, then expect similar suits to follow.

Will this slow down the pace of AI progress? Probably not. Anthropic recently raised a $13 billion Series F funding round at a $183B post-money valuation, which means the group has plenty of cash to pay for the settlement. Bigger groups like Google and OpenAI are even better funded.

Elsewhere, researchers continued to make sense of the various quirks of AI models:

• OpenAI released a new paper, ‘why language models hallucinate’, which proposed that the systems get things wrong because training and evaluation reward guessing instead of admitting uncertainty.

• Several researchers put their name to a survey of AI reasoning. The authors categorise existing methods along two dimensions: (1) regimes, which define the stage at which reasoning is achieved; and (2) architectures, which determine the components involved in the reasoning process.

• Google DeepMind claims that general agents contain ‘world models’. They note that any agent that generalises to multi-step goals must have learned a predictive model of its environment, that this model can be extracted and examined, and that more complex goals demand increasingly accurate world models.

Meanwhile, more work emerged seeking to understand how exactly people use AI in the real world:

• A new Stanford paper, ‘Canaries in the coal mine’, shows entry-level jobs shrinking in AI-exposed occupations. The researchers reckon that employment for 22 to 25-year-olds in roles like software engineering and marketing has dropped, while older workers remain stable.

• ​​A new post says that ‘breakthroughs’ that people have claimed to discover in tandem with LLMs aren’t anything of the sort. The author suggests that “New ideas in science turn out to be wrong most of the time, so you should be pretty skeptical of your own ideas”.

• Research from the University of Pennsylvania argues that today’s most powerful AI models can be manipulated using the same psychological approaches that tend to work well on humans.

And in industry news:

• OpenAI is working on a LinkedIn competitor, the OpenAI Jobs Platform, which will “use AI to help find the perfect matches between what companies [and governments] need and what workers can offer”.

• Anton Leicht argued that building an 'AI safety movement' is a mistake. He writes that “movement-building efforts are underway, but I think they'll change AI policy debate for worse: discredit current popular support, create liabilities for serious advocacy, and get captured by fringe positions.”

• Anthropic and OpenAI released the results of a partnership focused on testing each others’ models for misalignment. They found some examples of concerning behaviour in all the models tested, but reported that – compared to the Claude 4 – OpenAI’s o3 looks “pretty robustly” aligned.

Finally, some other bits of reading worth a quick look:

• Dan Williams of the University of Sussex put together an excellent reading list about the philosophy of AI.

• The Institute for Decentralized AI launched to help work on AI safety solutions that resist centralised control.

• Brendan McCord of the Cosmos Institute wrote on ‘Western Dynamism’, the idea that America’s strength flows from tension rather than unity.

• Asterix magazine announced a new round of AI fellows.

• Famous ‘AI scaling laws’ actually go back at least as far as Bell Labs in the 1990s.

• The Oakland Ballers announced they will be the first team to be managed in-game by artificial intelligence.

• And last but certainly not least, someone ran Meta’s Llama 2 large language model on a business card.

That’s it for our fourth outing. Let us know what you think, and what we should look at in future editions.

rule30 is a quantitative venture capital fund. We've dedicated three years to developing groundbreaking AI and machine learning algorithms and systematic strategies to re-engineer the economics of startup investing.

It’s the third edition of the new-look Moneyball. I’m Harry Law, a former researcher at Google DeepMind who writes Learning From Examples – a newsletter about the history of AI and its future. Each month, I work with the team at rule30 to help make sense of what’s been happening in AI over the last four weeks.

People like to remark that the speed of progress in AI has been head-spinning since the launch of ChatGPT in November 2022. I generally agree with that sentiment, though unlike some I doubt GPT-7 will be the model to enclose the Sun in a Dyson sphere. Nonetheless, trillions (with a t) of dollars are being spent as part of an infrastructure build-out deemed to be the latest front in China’s competition with the West.

In a single year, Google now spends more on physical capital like data centres ($85 billion) than the entire UK defence budget ($79 billion). In 2025, AI capex added more to U.S. growth than consumer spending. This is, by the way, despite the fact that the former is 6% of the economy and the latter makes up 70%.

This is the context in which President Trump unveiled his hotly trailed AI plan. Over 28 pages, the plan describes measures to speed up permits for data centres and recategorise energy assets as ‘covered components’ to allow for speedy parallel deployments. This is generally considered good news for American AI leadership, and comes amidst a few other announcements from the private sector:

• Anthropic liked what it saw but said more energy was needed, arguing that China may have fewer chips but is racing ahead of the U.S. in terms of raw power generation capacity.

• As it released GPT-5, OpenAI said it has reached an agreement with Oracle to develop 4.5 gigawatts of additional data centre capacity in the U.S.

• Outside of the U.S., the company also said it was working with European datacentre outfit Nscale to build 230MW of capacity in Norway.

• In a move as underwhelming as it is unsurprising, the UK announced it had struck a deal with OpenAI to, uh, ‘explore’ infrastructure investment in Blighty.

From one competition to another, Google DeepMind and OpenAI announced impressive results at the International Mathematical Olympiad.

• OpenAI rushed out of the blocks to tell the world it had won a gold medal at this year’s competition, solving five out of the six problems posed by the organisers.

• Google DeepMind followed shortly after, respecting official requests from the IMO for a media blackout before announcing that it had also taken a gold medal solving five out of the six problems.

• The news seemed to surprise prediction markets, which I found strange given DeepMind won silver last time around. Whatever the case, AI watchers were suitably impressed.

And in research corner:

• New work from the AI Security Institute argued that language models are likely to become more persuasive from post-training techniques rather than from scaling or personalisation.

• Longtime LLM critic Melanie Mitchell drew on the well-known simulator theory to explain why some models can’t help but misbehave.

• The folks at Anthropic described a phenomenon wherein longer chain of thought reasoning actually leads to lower accuracy.

• Gustavs Zilgalvis wrote about the ‘fourth offset’ in national security, where AI provides military advantages to maintain American military strength.

Finally, some other links we liked the look of but didn’t fit anywhere else:

• a16z interviews Dwarkesh Patel and Noah Smith about AGI, short timelines, and the economy.

• The UK AI Security Institute introduced the Alignment Project, a new £15 million fund for research on urgent challenges in AI alignment and control.

• Google DeepMind published a paper in Nature introducing Aeneas, the first model designed to help historians contextualise ancient inscriptions.

• Ethereum creator and occasional AI commentator Vitalik Buterin responded to the popular AI 2027 forecast, arguing that the authors over-egg the risks posed by superintelligent AI.

• After luring in researchers with mountains of cash, Meta shares what is quite possibly the most boring vision of ‘superintelligence’ imaginable.

That’s it for our third outing. Let us know what you think, and what we should look at in future editions.

rule30 is a quantitative venture capital fund. We've dedicated three years to developing groundbreaking AI and machine learning algorithms and systematic strategies to re-engineer the economics of startup investing.

Subscribe now

It’s the second outing of the new season of Moneyball. I’m Harry Law, a former researcher at Google DeepMind who writes Learning From Examples – a newsletter about the history of AI and its future. Each month, I work with the team at rule30 to help make sense of what’s been happening in AI over the last four weeks. Then I bundle it together, wrap it tightly, and send it on to our dear readers.

Last time we wrote about Meta’s new plans for its AI unit. Since then Zuck has confirmed that ‘Meta Superintelligence Labs’ will be headed by Alexandr Wang and GitHub’s Nat Friedman.

Several big hitters from OpenAI, Google DeepMind, and Anthropic have joined the team, no doubt lured by the prospect of building an Abraham Lincoln chatbot rather than an alleged $100M signing fee.

The cash seems ludicrous at first blush, but it makes a kind of sense if you believe the current LLM paradigm can get us all the way to human-level AI.

That goes for politicians as much as it does developers. As UK technology minister Peter Kyle put it: “By the end of this parliament we're going to be knocking on artificial general intelligence.” And in Washington, as the upper house moved to allow states to regulate AI, Representative Raja Krishnamoorthi announced he is working on an AGI Safety Act.

Confirming that we are truly living in a sci-fi timeline, Representative Jill Tokuda asked: “Should we also be concerned that authoritarian states like China or Russia may lose control over their own advanced systems?” Others, including the mighty RAND Corporation, are wondering the same thing. The think-tank outlined eight scenarios for a world with AGI, including those that end in American hegemony, Chinese dominance, and, uh, an ‘AGI coup’ where the systems run the show.

This last scenario is what happens when we get alignment badly wrong, which we can think of as ‘getting AI to do what we really want’.

• One part of this idea is ‘value alignment’, which deals with whether an AI’s actions are ethically appropriate. Recent work suggests that values in current LLMs “exhibit high degrees of structural coherence” or in other words that models seem to converge on a similar set of values.

• Then again, it’s not really clear how deep these preferences go. New work argues that models don’t actually have a sense of stable preferences because responses tend to be highly sensitive to prompting.

• The upshot is that alignment remains a tough old beast. As the folks at Anthropic like to remind us, even if your model looks friendly, it might just be faking it. This is clearly bad if you start deploying AI in places where it could cause some damage.

• And of course, you are more likely to miss any deception when you think you are in a race. This is the concern a recent paper raises in the context of US-China rivalry, but it’s also the idea that animates the AI 2027 forecasting project from April.

In the world of capabilities, a few interesting papers suggest that the large model project is both predictable and remarkable.

• A paper in Nature Machine Intelligence found multimodal LLMs can sort hundreds of everyday objects into a conceptual map that mirrors how people naturally group things, even matching activity in brain areas for faces, places and bodies.

• Another post argued that the story of modern deep learning is basically entirely about data. In 2012 AlexNet unlocked the ImageNet dataset; in 2017 transformers unlocked the internet as text; in 2022 RLHF unlocked learning from humans; and in 2024 reasoning unlocked learning from verifiers.

• Some work on self-updating language models was wildly misinterpreted on X dot com, but does suggest a promising path towards learning in the real world. This is the broad idea behind Google DeepMind’s ‘Era of Experience’ thesis, which I take to be directionally correct.

These questions are hugely relevant to the future of the AI project, but not necessarily for how people use the technology today.

• One recent survey finds that workers using AI report 3x productivity gains on 1/5 of tasks, while another reports that teachers using AI to save six hours a week.

• The latter comes as OpenAI announces the launch of the US National Academy for AI Instruction, a five-year initiative to help 400,000 teachers to use large models to “lead the way in shaping how AI is used and taught in classrooms across the country.”

• And Microsoft, which also wanted to remind people it still has an independent AI group, said it had taken “a big step towards medical superintelligence” with a new model capable of solving open-ended medical queries.

Doctors and teachers are one thing, but what about the unsung heroes: musicians and movie producers?

• Well, in the past few weeks, an AI band called Velvet Sundown has skyrocketed in listenership on Spotify, recently surpassing one million monthly listeners on the platform. The band’s sound resembles 70s rock and folk, and has been added to popular playlists –– much to the chagrin of artists already struggling with meagre streaming revenues.

• In Hollywood, Time covered the video generation start-up Moonvalley. The company owns AI film studio Asteria, which was co-founded this year by filmmaker and actress Natasha Lyonne and filmmaker Bryn Mooser. Asteria has been advising Moonvalley on the development of Marey, an AI model now available to filmmakers for subscription tiers starting at $14.99 a month.

• But what does AI actually mean for the future of entertainment? That is the question I tried to answer in an essay earlier this month, which argues that large models will make entertainment both much better and much worse depending on how you react to the technology.

And to wrap up, some wildcards worthy of your time.

• This lecture on AI and economic growth from Chad Jones, running through what some growth models look like for a world with powerful AI.

• A cool new paper that asked: “To what extent (if any) can the concept of introspection be meaningfully applied to LLMs?” The authors argue that, in a scenario where an LLM correctly infers the value of its own temperature parameter, we should legitimately consider it a minimal example of introspection.

• Google DeepMind said it made a major breakthrough in hurricane forecasting, introducing a system that can predict the path and intensity of tropical cyclones with accuracy.

• X’s Grok went loco and was taken offline by engineers, in an episode that reminded me of the Microsoft Tay chatbot debacle from a few years back. Also, Grok sometimes thinks it’s Elon Musk.

• Anthropic made Claude run a vending machine business, which did not turn out to be particularly profitable. Better luck next time, Claude.

That’s the whistle for the end of edition two. Let us know what you think, and what we should look at in future outings.

rule30 is a quantitative venture capital fund. We've dedicated three years to developing groundbreaking AI and machine learning algorithms and systematic strategies to re-engineer the economics of startup investing.

Subscribe now

Moneyball is back and under new management. Since 2022, this newsletter has explored the limitations of human decision-making and their implications for startup investing. Now we’re excited to share refreshed content aligned with rule30, our quant-VC fund. As we apply deep learning to venture, we’ll bring you insights and discoveries from our research to help you stay current in the fast-moving world of AI.

I’m Harry Law, a former researcher at Google DeepMind who writes Learning From Examples – a newsletter about the history of AI and its future. Each month, I’ll send you a short email about how to make sense of what’s been happening in AI over the last 30 days (including links to further reading for those who just can’t get enough).

This month’s big news is about the House of Zuckerberg.

According to reports, Meta is spending $14 billion to acquire 49% of Scale AI, the data-labelling firm that works with pretty much every major AI lab out there. As part of the deal, Zuckerberg will bring in Scale AI CEO Alexandr Wang (and several other Scale AI employees) to bolster a new division focused on ‘artificial superintelligence’ (ASI).

The efforts suggest Meta isn’t happy with its AI team, which – despite having access to one of the largest GPU clusters in the world – has struggled to release truly competitive models. Fearing it will fall behind in the race to artificial general intelligence (AGI), Meta is taking a leaf o ut of former OpenAI chief scientist Ilya Sutskever’s book: ignore AGI and shoot for its big brother. But what to do when you need fresh direction and the antitrust G-Men say you can’t buy another lab? You acquire just under half of a not-quite-developer and yank its CEO.

But hold on. Isn’t all this talk of superintelligence premature?

Yes. No. Probably. Well, no one really knows. Not academics, not developers, not the markets, and certainly not newsletter writers (but my view is that it’s not quite as sci-fi as it sounds, assuming agents can improve from real world experience).

• On the more optimistic side of the spectrum is OpenAI CEO Sam Altman, who recently wrote on his blog that ‘Humanity is close to building digital superintelligence’.

• Anthropic thinks job displacement is coming, and soon. Digital gods aside, the implication here is that what we have today is already good enough to break the economic status quo. Even Obama agrees.

• The uncertainty boils down to the unpredictable nature of large models, which work much better than they have any right to. One interesting (and hotly contested) factor at play here is emergence: the models spontaneously get better for reasons we don’t fully understand.

On the more sceptical side, we have recent research from Apple that tries to dampen enthusiasm about recently developed ‘reasoning models’ like DeepSeek-R1 or OpenAI’s o3.

• You can read their paper here, which argues that what we call reasoning in large language models is just a trick of the eye.

• Unfortunately, the work has some methodological problems, which you can read about in a blog from yours truly here (with a more technical take here).

• Nonetheless, that didn’t stop the media (noted AI enthusiasts that they are) from running to press with gushing headlines about the end of the AI project.

Meanwhile, the frontier research train keeps on rolling.

• French AI darling Mistral released its new reasoning model, Magistral, which can generate a reasoning trace in European languages.

• Alibaba’s Qwen 2.5 models can improve maths performance through reinforcement learning even with completely random rewards, suggesting the models may have been trained on synthetic data similar to test sets.

• Evals outfit Apollo Research built a test to help figure out if AI systems know they're being tested, finding that some recently released LLMs had some level of awareness they were being put through their paces.

• Another testing organisation, METR, said that recently released models are showing increasing enthusiasm for hacking their environments because they ‘seem misaligned with the user’s goals.’

Over in the world of AI policy, it turns out that large models (might) help the government work better.

• That new work was from, believe it or not, the UK government. They built a system called Extract on top of Google DeepMind’s Gemini model to help council planners make faster decisions.

• In international affairs, read this piece about how the recent ‘Spider’s Web’ operation in Ukraine shows what a world with a warfighting AGI might look like.

• A new study from MIT argues that most people evaluate AI based on its perceived capability relative to humans. In other words, we like it when it can do stuff we can’t.

• In education, China shuts down AI tools during nationwide college exams. In the UK, the government said that teachers in England can use AI to speed up marking and write letters home to parents.

And finally, some other stuff we thought was cool.

• The folks at Inference Magazine wrote about how to mitigate AI-driven inequality. They reckon that while AI creates enormous demand for skilled workers who can use the models, history shows that technological inequality is best addressed by expanding access to relevant skills – rather than restricting the technology itself.

• The Cosmos Institute launched a $1M grant programme to build AI systems that follow Vannevar Bush's 1945 vision of augmenting human intellect and John Stuart Mill's 1859 principles of free speech.

• A new documentary hosted on Aeon, Engineering Earth, takes aim at geoengineering projects. Check it out if you want something suitably inspiring.

• You may have noticed that it is especially difficult to build nuclear power plants. That’s because of the Linear No Threshold (LNT) hypothesis, a theory based on 1920s fruit fly experiments that assumes any radiation exposure increases cancer risk. LNT has driven nuclear power costs through the roof despite mounting scientific evidence that the human body can repair low-dose radiation damage.

That’s all for edition one. Let us know what you think, and what we should look at in future editions.

rule30 is a quantitative venture capital fund. We've dedicated three years to developing groundbreaking AI and machine learning algorithms and systematic strategies to re-engineer the economics of startup investing.

Subscribe now

We’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #38 – On The Objective Paradox – The Dangers of Planning

📖 Investor reading – Big bets on AI point to VC industry’s shift – Venture Capital Partners are leaving big firms in droves – What we learned about AI from active startup investors this year

💬 Some tweets – It’s called working in Venture Capital – Crypto VC fund evolution timeline – 20-50% efficiency gains pretty much everywhere

The Dangers of Planning

Planning is synonymous with the process of founding and scaling startups. Having a plan is a given; without one, founders are simply uninvestable.

This conventional wisdom is deeply misguided. And by swallowing it without question, founders and investors are missing out on massive opportunities to create new products, business models, and vehicles for wealth creation.

Kenneth O. Stanley and Joel Lehman’s 2015 book, “Why Greatness Cannot Be Planned: The Myth of the Objective” is pretty obscure. It shouldn’t be. Because it highlights something completely counterintuitive and heretical to the status quo thinking about startups:

“The Objective Paradox”.

Overemphasis on objectives can actually hinder innovation because it narrows the focus too much, missing out on indirect routes or stepping stones that could lead to greater achievements.

The most popular feature of your product might not be what you initially planned as the main attraction; solving your own problems can lead to game-changing user insights; a side project can become a company.

Startups that rigidly adhere to a pre-ordained plan risk missing out on serendipitous opportunities. Or they fail to pivot when necessary, and so die without having explored their full potential.

The modern obsession with objectives and clear, measurable goals (exemplified by the ubiquitous “OKR”) might not be the best path to meaningful breakthroughs. This is because the future of technology, and the market’s needs are unpredictable. Founders who fixate on crystal clear goals all too often get lost in the fog or war.

Stanley and Lehman advocate for a more flexible and responsive method, which they call the “Stepping Stone Approach”.

Instead of fixating on a distant goal, startups should focus on the immediate next steps. This approach allows for adaptability, learning, and sometimes, accidental discoveries that can lead to major innovations. They describe a “Treasure Hunter” mindset, whereby the focus is on collecting interesting, valuable insights one at a time.

The greatest founders know that innovation often comes from exploring the “interesting” rather than the “correct”. But try telling that to the average VC!

This approach doesn’t advocate for chaos but rather for a dynamic, responsive strategy that can lead to unforeseen greatness.

The parallels to AI research and development are clear. Stanley’s concept of “Novelty Search”, where algorithms are designed to seek out the novel rather than the optimal, mirrors the startup journey. AI systems that explore without a fixed objective often stumble upon innovative solutions that a strictly goal-oriented approach might bypass.

We’re really talking about the difference between exploration and exploitation. Algorithms can either exploit known paths to optimise for a specific goal, or explore new territories to find potentially better solutions. The latter is where the most exciting and significant work is taking place.

Stanley’s work encourages embracing the uncertainty and complexity of AI development, suggesting that the richest outcomes might come from allowing AI systems to evolve in ways that are not strictly controlled by human-defined objectives.

Examples of this phenomenon playing out in the startup world are plentiful:

• YouTube began life not as a video-sharing platform but as a dating site called “Tune In Hook Up”, where users could upload videos to attract potential dates. The dating aspect didn’t take off, but the founders noticed people were more interested in sharing all sorts of videos.

• Instagram was originally a part of a project called Burbn, which was intended to be a comprehensive check-in app with elements like gaming and photo sharing. However, the founders realised that the photo-sharing aspect was what users engaged with most.

• Slack began as an internal communication tool for a gaming company, Tiny Speck, which was developing an online game called Glitch. When Glitch didn’t succeed, the company was on the brink of closure, but the team realised that their internal tool for communication had significant potential.

• Twitter was born out of a side project at Odeo, a podcasting platform. During a company hackathon, the concept of a microblogging platform where users could share short updates was developed. Initially called “twttr”, it didn’t attract much attention, but the founders saw potential in the simplicity and immediacy of the communication style.

• Groupon started as a platform called The Point, which was designed to bring people together for collective action. One of its uses was for group buying, which caught on much more than the broader platform. Recognising this, the founders pivoted to focus solely on daily deals.

Implications for investors

Stanley and Lehman’s work on “tyranny of objectives” is an important touchpoint for investors. It reminds us that the greatest investments might not come from meticulously planned paths, but rather a journey filled with curiosity, exploration, and the willingness to deviate from predetermined objectives.

For founders, this might mean less time spent on crafting the perfect business plan and pitch deck, and more time engaging with the real world; listening to the market, being ready to pivot when necessary, or explore new avenues when the opportunity arises.

For investors, it means looking at failure differently. Just as AI models learn by trial and error, startups and their founders might fail in one area only to discover a new, viable path. VCs should drop the “we predict the future” mantra and adopt a “Novelty Search” mindset; one that places greater trust in founders and their ability to actively shape the future through new products and business models.

Work with Koble

At Koble, we’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

💰 Big bets on AI point to VC industry’s shift – The strategy of focusing money on larger deals comes at time of extreme indigestion for start-up investors

🏃‍♂️ Venture Capital partners are leaving big firms in droves – Mired in an industry downturn, investors are stepping back from jobs to start new firms – or getting pushed out

💡 What we learned about AI from active startup investors this year – Crunchbase speaks with a half-dozen earliest and latest-stage investors this year to understand where they are placing their bets when it comes to AI

Some tweets

Parting shot

“Life can only be understood backwards, but it must be lived forwards.”

― Sören Kierkegaard

Regards from your [unplanned] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

We’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #37 – On Incentives – Show Me The Money

📖 Investor reading – LP Predictions For Venture Capital In 2025 – Honest Money and Social Progress – Has Sequoia Capital outgrown its business model?

💬 Some tweets – I don’t think people understand how funding works – Consumptive vs productive hobbies – I have lost conviction

Show Me The Money

“Show me the incentive and I’ll show you the outcome.”

Charlie Munger knew a thing or two about incentives.

Together with his business partner and lifelong friend Warren Buffett, he used the “psychology of human misjudgement” to design systems of behaviour at Berkshire Hathaway and its portfolio companies that were capable of exponentialising value.

One such company is GEICO, the insurance business that Berkshire acquired in 1996. GEICO’s incentive system kicks in for employees after a year of being there and increases as a percentage each year thereafter. This means that the company’s best rewarded people are the ones who have been there the longest.

This structure is simple and effective. It offsets the decline in motivation that tends to afflict employees who have been with the same employers for a long time. It’s a kind of game, designed to reward long-term thinking and counteract fatigue.

Berkshire’s impact on GEICO has been nothing short of transformational. The company has experienced extraordinary growth in premiums, market share, and profitability since it was acquired. In 1996, Berkshire Hathaway acquired GEICO for $2.3 billion. Its market value is now estimated to be worth over $50 billion.

Munger also talked a lot about disincentives, and in many ways this is a more instructive way of thinking about human behaviour (and misbehaviour) in business.

Munger highlighted the dark side of incentives with a story about the early days of FedEx. The company had a problem; it paid people by the hour rather than by the job, so it struggled to get things done on time:

“The heart and soul of the integrity of the system is that all the packages have to be shifted rapidly in one central location each night. And the system has no integrity if the whole shift can't be done fast. And Federal Express had one hell of a time getting the thing to work.

They tried moral suasion, they tried everything in the world, and finally somebody got the happy thought that they were paying the night shift by the hour, and that maybe if they paid them by the shift, the system would work better. And lo and behold, that solution worked.”

Dumb incentives leads to dumb outcomes. And sometimes even smart incentives lead to dumb outcomes. In his book “Seeking Wisdom”, Peter Bevelin writes:

“There was a problem with mice on campus. The solution for exterminating the mice was to pay students $1 for every dead mouse they delivered. It worked! Until the students began breeding mice in order to make more money.”

In business as in life, we must beware the knock-on effects of being too crude with our mental models.

Overincentivising leads to excessive risk taking and unintended consequences that we cannot possibly model or mitigate.

The global banking system nearly collapsed in 2007/8 thanks to financial innovations that separated risk from reward so elegantly, so completely, that there was no incentive for originators to think about risk at all.

On the surface, startups do incentives pretty well.

Offering stock options to employees can be a powerful driver of long term thinking and behaviour. But all too often, options are just a way for founders to squeeze employees on salary, selling the dream of a meaningful liquidity event in order to attract and retain cut-priced talent.

Employee option schemes are widely misunderstood. Most employees don’t remember what they signed up, and they fail to effectively price the very low probability of actually benefiting from them. If a startup takes 12 years to exit (which is more rule than exception), then most employees have already left. Even if they are allowed to exercise their options when they leave, employees can rarely afford to do it.

And then there is the lazy “rest and vest”. Consider the number of employees who are underwater on options, or even worse, over-leveraged against them. Ownership schemes work better in Private Equity, where the cycles are naturally shorter.

One shaft of light in the darkness; if quant strategies can solve the liquidity problem for investors, then they might just solve the option scheme problem for employees.

Implications for investors

The best employees are committed employees; people who are attracted to the mission and feel fairly compensated for their role in the success of the startup.

Investors would do well to look for founders who raise capital and operate on the basis of paying their people fairly and promptly, giving them the ability to perform at their best and build for the future.

All-nighters may be necessary from time to time in the life of a truly explosive startup, but hardcore “Ramen culture” projected onto employees by overzealous founders is a red flag. It often masks dysfunction and points to a lack of strong, principled leadership; the kind that is necessary to scale early-stage product-market fit into a valuable business.

Work with Koble

At Koble, we’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

🗓 LP Predictions For Venture Capital In 2025 – The question on most minds in the VC community: Will things change next year?

🤑 Honest Money and Social Progress – Why Bitcoin Infrastructure Matters A Great Deal

🌳 Has Sequoia Capital outgrown its business model? – Venture Capital’s hardiest perennial gets back to its roots

Some tweets

Parting shot

“It is difficult to get a man to understand something when his salary depends on his not understanding it.'”

― Upton Sinclair

Regards from your [incentivised] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

Share

We’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #36 – Openings, middlegames, and endgames – The Immortal Game

📖 Investor reading – General Catalyst’s Hemant Taneja on AI, LPs, and Raising Around $8 Billion – The Numbers Behind This Year’s Most Promising Startups – VC Firm Accel Files With SEC to Tap Thriving Secondary Market

💬 Some tweets – What he announced today will blow your mind – 90% of guests refuse to discuss politics and the election – Venture Capital is a long game

The Immortal Game

Chess is one of the world’s oldest sports. It’s enjoyed by millions of people the world over (a game with over 600 million players). And it’s a useful way of thinking about building and backing startups. 

“The Immortal Game” is complex, fluid, predictably unpredictable. It’s a game with almost limitless variables. The number of distinct chess positions after White’s first move is 20; there are 400 positions after 2 moves; and the total number after 7 moves is 3,284,294,545.

The Shannon Number, named after the American mathematician Claude Shannon, is a conservative lower bound of the game-tree complexity of chess of 10^120. That’s 1,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000 (a thousand trillion trillion trillion trillion trillion trillion trillion trillion trillion trillion). This far exceeds the total number of atoms in the observable universe, which is estimated to be around 10^80.

Chess is perhaps the domain which has best dramatised the rise of the machines, and the suitability of algorithms for solving complex, seemingly insurmountable, mathematical challenges. 

When a child of the 80s thinks of chess, they think of Garry Kasparov. Grandmaster, World Chess Champion, latterly political activist and writer. From 1984 until his retirement from competitive chess in 2005, Kasparov was ranked World Number 1 for a record 255 months overall.

Deep Blue IBM’s victory over Kasparov in 1997 fired the starting gun on a new era in computing. Since that time, chess programs running on commercial hardware (even mobile phones) have been able to defeat even the strongest human players.

In his book, “How Life Imitates Chess”, Kasparov outlines his dualistic approach to chess and life; success derived from combining disruptive and disciplined approaches. By integrating seemingly incongruent styles of playing and living, he was able to gain an edge over his competitors and dominate the sport.

This reads like a playbook for entrepreneurs, highlighting the crucial distinction between strategy and tactics. But it will also resonate with startup investors. 

Chess, with its distinctive Opening, Middlegame, and Endgame, serves as a neat metaphor for a startup’s journey from launch to exit.

– Openings present an opportunity to dictate the terms of the exchange, the type of game that you want to play – or the type of game that your opponent doesn’t want to play. Most of Paul Graham’s “18 Mistakes That Kill Startups” happen at the beginning. Startup founders can relate to the importance of openings.

– Middlegames involve digging deep. Since middlegame positions are unique from game to game, memorisation of theoretical variations is not possible. As one Reddit user has commented: “The goal of the middlegame is to exaggerate the imbalances in your favor and mute those favoring your opponent.” During the startup middlegame, founders have to “roll with it”. They must take the rough with the smooth, mixing offense with defense to somehow find a way into the Endgame. 

– Endgames speak for themselves. They are an especially delicate phase of the game, where mistakes have more serious consequences than in the opening. Many players find it hard to know for sure when they are in the endgame. When it comes to startups, even the most accomplished founders find exits difficult. Getting to checkmate and meaningful liquidity event is fraught with execution risk.

The mathematics of chess are so complex that ultimately, the game comes down to intuition and psychology. Creativity, too. Kasparov built his career on a peculiar mix of discipline and freedom, dispassionate calculation and iconoclastic ideas. The best founders do the same.

Implications for investors

We’ve already written about the ambivalent power of unintended consequences and how recognising them can help investors to manage risk more effectively. 

When we evaluate the potential impact of our actions within a particular system, we tend to ignore the impacts of the system’s first-order responses to our actions, and indeed the second and third order responses to those responses.

Chess dramatises this inconvenient truth. It can help us to become more aware of how our decisions cascade and escalate over time.

At the heart of it all – chess, startups, life – is the crucial distinction between tactics and strategy. As the Polish chess player Savielly Tartakower said:

“Tactics is knowing what to do when there is something to do. Strategy is knowing what to do when there is nothing to do.”

Chess teaches us more than how to act in a given situation. It teaches us when to act. And that’s a useful skill for those who make a living from startup investing, a domain where timing is everything.

Work with Koble

At Koble, we’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

💰 General Catalyst’s Hemant Taneja on AI, LPs, and raising around $8 billion – $4.5bn approximately is geared towards the firm’s VC funds; $1.5bn is for GC’s “creation” strategy; $2bn is geared towards “separately managed accounts".

📈 The Numbers Behind This Year’s Most Promising Startups – The technology world often fixates on how much startups raise. But perhaps more interesting are the companies that generate meaningful revenue without raising any venture capital at all.

💧 VC Firm Accel Files With SEC to Tap Thriving Secondary Market – Accel is the latest venture capital firm to consent to heightened regulation with the US SEC, a tradeoff that allows it to buy more shares of non-public companies that have been changing hands at steep discounts on private markets.

Some tweets

Parting shot

“The winner is just the player who made the next-to-last mistake.”

― Garry Kasparov

Regards from your [tactical / strategic] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

We’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #35 – Goodhart's Law – The Dark Side of Statistics

📖 Investor reading – Headline joins Europe VC’s wave of growth funds – Venture Mega-Rounds Return to Biotech – 6 things founders should know if they’re fundraising this autumn

💬 Some tweets – The world finding out how venture firms work – Intimate, confidential details about their business – I don’t get London

The Dark Side Of Statistics

“When a measure becomes a target, it ceases to be a good measure.”

Wise words from British economist Charles Goodhart, with profound implications for startups and people who invest in them.

Except Goodhart didn’t write them. What he actually wrote was:

“Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.”

Whichever formulation you prefer, “collapse” and “cease” are both apt to describe what happens when you disrespect the latent power of statistics.

Goodhart’s Law was coined in 1975 as a criticism of the Thatcher government’s attempts to conduct monetary policy via targets for broad and narrow money. It outlived that era, and now serves to encapsulate many of the challenges facing founders and investors as they seek to de-risk startups.

Statistics are a valuable tool in any enterprise, but they’re vulnerable to manipulation and debasement, depending on the personal agenda of those who wield them.

And when we play with statistics, we get unintended consequences (i.e. bad stuff).

To highlight how this works, here’s a couple of interesting (albeit depressing) examples.

Fans of the HBO show “The Wire” will recall the city of Baltimore’s strategy of improving crime statistics through reclassification of those crimes; “juking the stats”. Or the school board’s policy of “teaching to the test”. Substantive improvements in learning and public safety elude the city because politicians are focused on the wrong metrics. The measure has become the target.

Many moons ago in the UK, the government put pressure on hospitals to reduce emergency waiting times to within 4 hours. They managed to pull it off with a range of ingenious / unethical hacks, including one famous example whereby patients were made to wait in ambulances rather than inside the hospital. Some A&Es striving to reduce Length of Stay also discharged patients prematurely, leading to increased emergency readmissions.

The Law of Unintended Consequences is alive and well, thanks in part to its dysfunctional but incredibly good looking and popular cousin, Goodhart’s Law. And statistics abuse is everywhere you look in startups (and also in places that we don’t see).

Perhaps the most obvious example is measuring company growth by the crude metric of headcount. A measure of company size has mutated into an arbitrary target that says nothing about the underlying economics of a business. Perverse incentives abound; as the enterprise “scales”, managers benefit from growing their teams, acquiring more responsibility and status in the internal hierarchy, and controlling ever greater budgets.

Another painful symptom of Goodhart’s Law is the tendency to customer acquisition with growth, without factoring in ongoing engagement and churn. Something that looks great in a pitch deck can be a total red herring when it comes to assessing the unit economics and long term viability of an early-stage startup. 

Let’s face it, the startup world has its fair share of gamers. So too, the VC industry. Firms that advertise their prowess by the number of pitch decks read, hours of Zoom calls taken, volume of companies invested in, and aggregate capital deployed, are simplifying something very complex. This is startup investing for idiots (and all too often, by idiots).

Implications for investors

Whenever working with statistics, consider unintended consequences; dysfunctional behaviour, questionable ethics, wonky performance measurement, the unreality of generic startup culture.

“Measure what matters” might be a cliché, but it’s a simple and powerful reminder of what investors are up against.

Systems-thinking can help. Understanding the wider company system and interdependencies between metrics, indicators, and actors can help to uncover perverse incentives that destroy value for investors.

And direct, frank, sometimes difficult conversations can help to purge these behaviours from startups before they become endemic and unfixable.

But implications for investors go beyond interrogating the metrics that founders use. Goodhart’s Law – the Dark Side of statistics – invites us to consider our own relationship with data. And compels us to build a relationship with statistics that is honest, healthy, and productive.

Work with Koble

At Koble, we’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

💰 Venture Mega-Rounds Return to Biotech – The number of deals for $100 million or more have put biotech on pace to come close to full-year 2021’s record total

🇪🇺 Headline joins Europe VC’s wave of growth funds – European VC growth vehicles have been on the march this year with a series of substantial fund closes announced in recent months.

🍁 6 things founders should know if they’re fundraising this Autumn – From pitch dry runs to timing your fundraise right, VCs share their top advice

Some tweets

Parting shot

“You underestimate the power of the Dark Side. If you will not fight, then you will meet your destiny.”

― Darth Vader

Regards from your [Jedi Master] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

We’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #34 – Something Beautiful – On Beauty

📖 Investor reading – Venture group G Squared raises $1bn to invest in discounted startup shares – Can China Tech Find a Home in Silicon Valley? – The Venture Capital Reset Era

💬 Some tweets – The biggest problem in venture and LP land today – Ultimate checklist for measuring the moat of a company – Your friendly reminder that VC index loses money

Something Beautiful 

What is beauty? And what does it have to do with startup investing?

On 10th May 2024, an extraordinary person passed away. He was a professor of mathematics, NSA code breaker, chain smoker, quantitative hedge fund manager, and philanthropist, giving over $4 billion to good causes during his lifetime. He famously did not wear socks. And he was and one of the most successful investors of all time.

Jim Simons was one of those rare individuals that manage to transcend academia and markets, using abstract theories to capture real-world profits. 

He helped to developed the Chern–Simons form and contributed to the development of string theory by providing a theoretical framework to combine geometry and topology with quantum field theory. And as the creator of Renaissance Technologies – one of the world’s first quantitative hedge funds – he found a way of monetising his studies on pattern recognition.

Medallion, Renaissance’s main fund which is closed to outside investors, has earned over $100 billion in trading profits since its inception in 1988. This translates to a 66.1% average gross annual return or a 39.1% average net annual return between 1988 – 2018.

All of this is well known. The investment world is replete with imitation – we naturally seek to deconstruct and replicate the success of others.

What’s far more interesting and valuable is the role of beauty in Simons’ life and work.

He said:

“Be guided by beauty . . . it can be the way a company runs, or the way an experiment comes out, or the way a theorem comes out, but there’s a sense of beauty when something is working well, almost an aesthetic to it.”

For Simons, beauty occurred in many forms, as outlined by Greg Zuckerman’s thrilling book about his life, “The Man Who Solved The Market”. Beauty is complex, subjective, and multi-faceted. In Simons’ world, it encompassed a range of different experiences – finding math theorems, market pricing discrepancies, building multi-disciplinary teams, and yes, generating outstanding returns.

Implications for investors 

Other people see beauty. It is attractive in a literal sense, pulling talent towards talent and creating opportunities for massive breakthroughs.

It is significant that Simons tended to hire specialists with non-financial backgrounds, including mathematicians, physicists, signal processing experts and statisticians, whose relationships with numbers were different to those of traders and other traditional money-makers. People guided by beauty. 

Confucius observed that:

“Everything has beauty, but not everyone sees it.” 

Beauty is a strange and intangible concept. It doesn’t seem relevant to the cold, precise world of quantitative investing. But it is there, hiding in plain sight. Some people are simply more attuned than others to finding beauty and sharing it with others. 

Simons was one such person. He could feel the beauty of mathematics and channel it into something pragmatic and lucrative. He was a translator, decoding the hidden meaning of numbers and their interrelationships to his colleagues and investors.

What else can we learn from Jim Simons’ belief in beauty? How can we make our investments more beautiful? Not easy questions to answer, but certainly worth exploring.

Work with Koble

At Koble, we’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

💰 Venture group G Squared raises $1bn to invest in discounted start-up shares – Slowdown in public listings and takeovers means staff and investors are selling shares for cheap on secondary market.

🇨🇳 Can China Tech Find a Home in Silicon Valley? – Entrepreneurs and venture capitalists are setting up firms across the Pacific, only to find that any investment with Chinese ties is a hard sell.

⏱ The Venture Capital Reset Era – “This is a time for serious people, people who take this business seriously.” The Information discusses the latest in the venture shakeout.

Some tweets

Parting shot

“I can appreciate the beauty of a flower. At the same time, I see much more about the flower than he sees. I could imagine the cells in there, the complicated actions inside, which also have a beauty. I mean it’s not just beauty at this dimension, at one centimeter; there’s also beauty at smaller dimensions, the inner structure, also the processes.”

― Richard Feynman

Regards from your [beautiful] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

We’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #33 – On Madness (Part II) – Abnormal People

📖 Investor reading – Can Roelof Botha keep Sequoia Capital ahead in the AI future? – EIB calls for more innovation financing in EU – Secondaries volume reaches H1 record as venture subset continues to grow

💬 Some tweets – A spike of interest that mostly dissipates – Deals from a bygone era – AI is making software cheap to build and costly to scale

Abnormal People

Startup building is presented as rational.

Founders position themselves as vehicles for penetrating insights into human nature and behaviour, capable of translating intelligence into action. And VCs like to think (and talk) of themselves as futurists, backing technologies that “make the world a better place” (especially for their LPs). 

Nobody talks about madness; the invisible hand of innovation and progress. But the inconvenient truth is that irrationality and abnormality have played a key role in the history of startups. And they will continue to carve out the future of technology, culture, and human progress.

In last month’s Moneyball we looked at madness through the macro lens of tech’s biggest personalities, individuals of destiny like Steve Jobs and Elon Musk. But what about the long tail? The same logic (or lack thereof) applies – the startup world, across its entire spectrum, runs on madness.

The most transformative ideas, operational practices, and business models arise from thinking that defies conventional wisdom, and behaviors that seem irrational to the mainstream.

Launching a startup is a terrible idea (most of the time). The odds are stacked against you, VCs are stacked against you.

And even if you are top quartile, keep your feet on the ground – your chances (not your investors’) of a successful exit on cash-friendly terms are minuscule.

The story of Rovio and Angry Birds is a powerful example of how madness – in the form of dogged persistence, unconventional thinking, and frankly very weird taste in art – can lead to remarkable success.

In 2009, Rovio was a small gaming studio in Finland on the verge of bankruptcy, having produced over 50 unsuccessful games. Then it hit upon the idea of launching a simple, physics-based puzzle game featuring disgruntled birds and green pigs. This seemed a) mad, and b) unlikely to stand out in the crowded mobile gaming market.

But the concept was powerful and the game, addictive – however crazy it seemed. Co-founder Mikael Hed observed:

“When you had to shoot the bird once to test a feature, you accidentally started playing the game for 15 minutes, and there would be five guys watching you.”

When a co-founder’s mother ended up burning her Christmas turkey because she was busy playing a pre-release version of the game, the team knew they were onto something. And Angry Birds became a cultural phenomenon, spawning a franchise that includes merchandise, movies, and theme parks. At IPO in 2017, the company was valued at over $1 billion.

Madness can blind people to dysfunctional beliefs and behaviours. It can also bring light, helping to illuminate a future that nobody else can see.

In 2008, the idea of creating a social network for coders centered around a version control system seemed impossibly niche and unlikely to attract a broad audience. At the time, many developers were content with existing tools and saw no need for a new platform. On paper, it was a crazy idea. 

But the founders of GitHub were convinced that making Git easier to use and creating a community for coders would revolutionise software development. Their counter-intuitive bet on open source collaboration paid off handsomely. In 2018, Microsoft acquired the company for $7.5 billion, validating the co-founders’ seemingly mad vision.

Implications for investors

Why is madness such a powerful force in startups? A few ideas for investors to consider:

• Startups attract outsiders; people who have rejected (or been rejected by) large corporations that have become bureaucratic, sclerotic, and stagnant

• It takes a certain type of individual – persistent bordering on obsessive; self-assured bordering on arrogant; detail-orientated bordering on pedantic; visionary bordering on delusional – to ideate, launch, and operate a business that does something totally new

• Massive value creation at unicorn-scale depends on creating huge user-bases, which is easier (and cost-effective) to do if you forge entirely new markets based on iconoclastic ideas

• Abnormal people are often charismatic and interesting, which makes it easier for them to attract talent (often fellow crazies)

• The incentivisation model for early-stage startups attracts economically irrational people who are willing to work for “free” for order to build long term value through company ownership

• Reality distortion is a sustaining force in the early years, helping founders overcome dark psychological times, financial challenges, and criticism from family, friends, investors, customers, employees and customers (basically everyone)

There are many more reasons why madness matters in startups, but these ideas encapsulate the basic truth that few 100% bonafide “normal” people build startups, and even fewer successfully exit. And they suggest some ways for investors to think about high-level portfolio construction, company selection, and founder management.

Obsessional, irrational, emotional, abnormal… All necessary ingredients for founder success, perceived as “bad” by business-school-trained VCs lacking in life experience. Perhaps that’s why exited-founders make such great investors – they understand that sometimes what’s “bad”, is good.

Observing this dynamic forces us to question what is normal. Perhaps madness in founders and startups is a kind of transcendental truth? A glitch in the Matrix that appears to “normal people” as a bug, but alludes to a higher way of thinking and living?

This paradox is neatly captured by Joseph Heller's satirical war novel Catch-22:

“There was only one catch and that was Catch-22, which specified that a concern for one's safety in the face of dangers that were real and immediate was the process of a rational mind […] Orr would be crazy to fly more missions and sane if he didn't, but if he was sane he had to fly them.”

Like the pilots of Catch-22, startup founders are sane in their madness. After all, they keep flying and building, despite the risks. This is the opposite of the reverse-engineered business idea with pre-mapped milestones and intricate fundraising strategy. And it’s why early-stage founders who talk about “exit routes” and “being acquired” sound strangely cold and cynical (and are best avoided).

Ultimately, unicorn founders are anomalies that generate exponential wealth for investors, employees and the ecosystem of other entrepreneurs that cluster around their businesses. They simply shouldn’t exist. Their statistical persistence bears witness to the enduring power of madness. 

The inconvenient truth is that madness is a crucial part of startup success. And investors cannot afford to ignore it.

Work with Koble

At Koble, we’ve spent four years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

🤖 Can Roelof Botha keep Sequoia Capital ahead in the AI future? – Sequoia's culture from the get-go has been one of near-paranoid competitiveness, ruthless honesty, and a constant need for change.

🇪🇺 European Investment Bank calls for more ‘innovation’ financing in EU – A new report highlights the “critical need” of removing investment barriers in order to channel savings into key areas of the EU market.

📈 Secondaries volume reaches H1 record as venture subset continues to grow – Evercore’s latest Secondaries Market Review paints a reassuring picture for the multiple new entrants to the venture secondaries market.

Some tweets

Parting shot

“He was going to live forever, or die in the attempt.”

― Joseph Heller, Catch-22

Regards from your [abnormal] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

We’ve spent three years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #33 – On Madness – The World Lives On Crazy Things

📖 Investor reading – US Venture Capital Outlook: Midyear Update – Inside HSBC’s VC Charm Offensive – China Plans New Measures to Attract Venture Capital Investment

💬 Some tweets – Michael Mauboussin on investing vs gambling – Big venture platforms are launching PE funds – This genius who presented $NVR stock in the Value Investors Club back in 2001

“Madness is a rare thing in individuals – but in groups, parties, peoples, and ages it is the rule.”

– Friedrich Nietzsche

We live in a world that abhors madness. But if you look at the history of the greatest startups, madness is a common theme.

Steve Jobs understood the fine line between madness and genius. In 1997 he said:

“Here’s to the crazy ones, the misfits, the rebels, the troublemakers, the round pegs in the square holes... the ones who see things differently... they push the human race forward, and while some may see them as the crazy ones, we see genius, because the ones who are crazy enough to think that they can change the world, are the ones who do.”

Jobs famously declared that Apple would dedicate itself to making tools for these misfits. But Apple’s relationship with abnormality went deeper than its target market; the company itself was a product of madness.

The company’s evolution from a weekly meet-up of Bay Area geeks (the Homebrew Computer Club) to the world’s most valuable company was driven by more than product innovation and marketing genius.

Apple relied upon an irrational, counter-intuitive belief in the beauty of technology, the unlikely passion of “computer geeks”, and the belief that people don’t know what they want until you show them.

Elon Musk is heir apparent to Job’s long-held title as the most interesting person in technology. He has big ideas (online payments, electric cars, reusable rocketry) that seem crazy but precipitate structural changes in thinking and doing that create massive new markets and opportunities for value creation. He lives and operates his technology empire by Joseph Campbell’s credo:

“The world lives on crazy things. The economics will work themselves out later – you can count on it.”

Musk is certainly an extraordinary storyteller, paradigm-shifter and world-builder, but he is also a ruthless self-publicist, and one senses that his craziness is somewhat confected; a sort of socially acceptable, even cool madness that is designed to drive clicks and dramatise his story.

The origins of Apple’s logo say a lot about the company’s singular DNA. Not the now iconic apple alluding to the fruits of humanity’s ingenuity and labour, but the original design that never made it to the back of our iPhones; an Victorian woodcut accompanied by a quote from William Wordsworth’s “The Prelude” which read:

“A mind forever voyaging through strange seas of thought, alone.”

Strange seas of thought… Chaotic. Frightening. Dangerous. What if these are the qualities that startup investors should be looking for, rather than the safety of reverse-engineered founding teams from reputable universities with oven-ready business models? 

Implications for investors

Investors are generally scared of madness (and it’s cousin, genius). That is why they emphasise qualities like grit, determination, and tenacity – it’s a way to push through polished but mediocre founders that seem to be less risky bets, but end up building generic B2B SaaS startups.

Most of the startups paraded through TechCrunch and Sifted by boutique PR firms are the musical equivalent of boy/girl bands, manufactured by left-brained music executives who think too much and feel not enough.

To extend the analogy, the most promising startups are often more akin to idiosyncratic, hungry, rough-round-the-edges performers who know their instruments and play to put food on the table. Investors should be looking for the next Tracy Chapman or Van Morrison, but they are competing for allocations from Girls Aloud and One Direction.

Madness is rare in individuals. It’s even rarer in startups. Which is what makes it so interesting and valuable to investors.

Work with Koble

At Koble, we’ve spent three years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

🇺🇸 US Venture Capital Outlook: Midyear Update – VCs and founders went into 2024 hoping for the best but preparing for the worst, and got something in between.

🏦 Inside HSBC’s VC Charm Offensive – The bank has snared roughly 300 startup clients over the last year, a figure management is hoping to grow into thousands.

🇨🇳 China Plans New Measures to Attract Venture Capital Investment – International investors will be able to set up yuan funds in China to facilitate domestic investments.

Some tweets

Parting shot

“No great mind has ever existed without a touch of madness.”

― Aristotle

Regards from your [crazy] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

We’ve spent three years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #32 – Pain – The Truth Hurts

📖 Investor reading – Some of Silicon Valley’s Most Prominent Investors Are Turning Against Biden – Unpacking the European VC Playbook – Studio Tour: A VC firm’s Disco-inspired Office in San Francisco

💬 Some tweets – Venture Capital is dead – Take the job at the startup – People who work in VC

The Truth Hurts

Pain is central to the human experience; how we experience it, how we avoid it, and sometimes, how we can leverage it.

Not all pain is productive, but a lot is. Periods of acute discomfort are usually followed by breakthroughs, be they physical, mental, or spiritual. What doesn’t kill us makes us stronger, stranger, and wiser. And it might make us wealthier, too.

In our personal lives, we intuitively know that pain can be good. We do things like enter marathons, endure childbirth, voluntarily jump into ice-baths, and indulge in other extreme activities that arbitrage short-term pain for long-term gain. Weightlifter Jerzy Gregorek neatly sums up this phenomenon with his deceptively powerful mantra:

“Easy choices, hard life. Hard choices, easy life.”

But when it comes to the world of work, we tend to avoid pain at all costs.

Hope this finds you well

The culture of corporate life – the user experience of traditional employment – leaves no room for pain. 

Discomfort is something to be reported, minimised, perhaps even eliminated. Emails are passive aggressive (no drama), chairs are ergonomic (no bad backs), flowers are plastic (no allergies), and HR is always on hand for a supportive chat (no thanks).

The result of this zero-tolerance approach to discomfort is lower growth. Trapped in their comfort zones, large companies fail to adapt and evolve. Employees become comfortable, failing to push into new experiences, relationships, and ways of working.

There are, of course, some outliers. Jensen Huang, President of NVIDIA, has said:

“Resilience matters in success. I don't know how to teach it to you, except I hope suffering happens to you...I use the phrase “pain and suffering” inside our company with great glee.”

These are not the words of a sadistic dictator presiding over a toxic company, but the counter-cultural wisdom of a man who has built the world leader in AI computing.

What doesn’t kill you

Most startup founders know this. Startups are not comfortable things to launch and operate. They are, more often than not, painful and rewarding endeavors, even if few deliver the outsized financial rewards they promise.

But investors are different. Fear and greed reflect our obsession with pain and pleasure; we will do anything to avoid the pain of losing money. 

And when it comes to assessing our behavior and how it drives investment performance, we fail to recognise the impact of pain (or more accurately, our counter-productive responses to it).

Over and over again

Rohini Terry and her colleagues at the University of Stirling have explored the phenomenology of pain – whether we remember pain itself, or merely are aware as a semantic fact that the pain took place. 

They found that we tend to forget the intensity of the pain and instead remember painful episodes in factual terms – meaning we fail to properly grasp how pain influenced our decision-making.

That is particularly problematic for investors. We fail to recognise the role of pain avoidance in asset selection (and de-selection). And we risk making the same mistakes over and over again. Many VCs are stuck in a doom loop of chasing losses, forgetting that Pre-Seed and Seed investing is about finding winners – that's where the gold comes from. Investors should not be throwing good money after bad to soften losses and minimise pain.

Implications for investors

Bruce Lee knew a thing or two about pain. He said:

“The pain will leave once it has finished teaching you.”

We cannot grow if we do not allow pain to modulate our thinking and behavior. It would be perverse to suggest that investors should not seek pain, but we should not seek to eliminate it either. 

Rather we should seek to accept pain as a necessary part of our growth. Something that visits us from time to time like a long-lost (and extremely difficult) relative, but does not need to live with us full-time.

It’s in our nature to minimise suffering, but in doing so, we reduce progress. Pain can make us better investors, and better people. If we would just let it.

Work with Koble

At Koble, we’ve spent three years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

🏛 Some of Silicon Valley’s Most Prominent Investors Are Turning Against Biden – Marc Andreessen, Chamath Palihapitiya and several other tech VCs are increasingly criticizing President Biden and making their disaffection known in an election year.

🇪🇺 Unpacking the European VC Playbook – A conversation between Frederic Court (Felix Capital) and Luciana Lixandru (Sequoia) about the recent trend of US VCs going to Europe and how the two markets can work together.

🕺 Studio Tour: A VC Firm’s Disco-inspired Office in San Francisco – Designed by Studio Plow, NFX’s office is a vibrant space for connection, celebration and creativity.

Some tweets

Parting shot

“Behind every beautiful thing, there's some kind of pain.”

― Bob Dylan

Regards from your [painful] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

We’ve spent three years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #31 – Boredom – It Isn’t Supposed To Be Interesting

📖 Investor reading – “It’s like being gaslit”: Founders’ relationships with investors in the downturn – The Mixed State Of Startup Funding In 2024 – Dear Venture Capitalists: You’re blowing it

💬 Some tweets – The book publishing industry is like the VC business – A seed-stage founder I spoke with today – All normal companies now call themselves “Tech Companies”

A reminder

In case you missed it, Koble recently contributed to an article from Data-driven VC on the investors, data scientists, and tool builders building the future of Venture Capital.

Read Article

It Isn’t Supposed To Be Interesting

“People say nothing is impossible, but I do nothing every day.”

A.A. Milne wrote stories that celebrate the importance of boredom. Winnie-the-Pooh transformed ennui into an artform; something interesting and life-affirming. But to most people – investors included – boredom is something to be avoided at all costs.

The dark side of boredom

In contemporary life, boredom has a dark side, with easily bored people at higher risk of depression, anxiety, drug addiction, alcoholism, compulsive gambling, eating disorders, hostility, anger, poor social skills, bad grades and low work performance.

Boredom is ubiquitous. As human beings, we’re programmed to seek stimulation, novelty, feedback. 63% of participants in a US-based sample (n=3,867) reported boredom at least once within a 7–10 day period chosen randomly over a two-year period. As Blaise Pascal famously quipped:

“All of humanity's problems stem from man's inability to sit quietly in a room alone.” 

Nowhere is boredom more destructive than the field of investing. In markets, a little dopamine can be a dangerous thing, with bored investors destroying returns by endlessly fiddling with portfolios. The result is transaction costs, opportunity costs, costs, costs, costs.

Not a lazy investor

In his classic 1940 book “Where Are the Customers’ Yachts?”, Fred Schwed wrote: 

“Your average Wall Streeter, faced with nothing profitable to do, does nothing for only a brief time. Then, suddenly and hysterically, he does something which turns out to be extremely unprofitable. He is not a lazy man.”

The evidence is clear: timing the market is rarely as lucrative as time in the market. A ground-breaking study published by Nobel Laureate William Sharpe in 1975 attempted to ascertain how often a person seeking to time the market must be accurate to perform as well as a passive index fund tracking a benchmark. Sharpe concluded that investors employing market timing strategies must be correct 74% of the time to beat the benchmark portfolio. 

Years later, Morningstar reached a similar conclusion – active investors have to be correct 70% of the time in order to gain an edge over passive funds, which is practically impossible over the long run. Research measuring the performance of long-term investors following different investment strategies shows that the cost of waiting for the perfect moment to invest exceeds the benefit of even perfect timing.

It’s impossible to successfully (and consistently) time the market over the long run. Investors who try tend to underperform those who remain invested over the long term. Which is a key reason why illiquid asset classes such as Venture Capital and Private Equity outperform public markets – the illiquidity premium rewards investors for being forced to do nothing over long time horizons.

The intelligent acceptance of illiquidity

Illiquid investments create a psychological barrier that discourages impulsive decision-making. “The intelligent acceptance of illiquidity” – a term coined by legendary endowment investor David Swensen – conditions a level of patience and discipline that is often lacking in the face of market volatility.

Throughout Swensen’s tenure as CIO of Yale’s endowment fund, he often allocated between 70%-80% of the university’s portfolio to alternative investments. By disrupting the endowment industry’s reliance on stocks and bonds, he drove huge outperformance.

Mastering boredom is a critical skill for investors, but most humans can’t do it. Our aversion to doing nothing is simply too strong. Bertrand Russell observed, “We are less bored than our ancestors were, but we are more afraid of boredom.” Perhaps it is not really boredom that we cannot abide, but the fear of it?

This thing called temperament

How can we protect ourselves from the waste products of boredom? Warren Buffet said it best: 

“Investing is not a game where the guy with 160 IQ beats the guy with 130 IQ. Once you have ordinary intelligence what you need is the temperament to control the urges that get other people into trouble investing.”

This thing called temperament is what human beings generally lack. Machines on the other hand, have it in almost infinite supply. They are able to source, analyse, execute, manage, and reinvest assets without emotion. Since they do not experience Pride, Greed, Lust, Envy, Gluttony, Wrath, and Sloth, algorithms can deploy capital rationally and consistently, at a speed and scale unattainable for the human mind.

Implications for investors

Patrick Ryan’s recent essay on “The Boringification of Culture” and its sinister implications for our future is timely. We don’t want machines genericising our art, optimising our humanity, and eroding our culture into oblivion. 

But when it comes to startup investing, the writing’s on the wall; the future is boring, the future is quantitative.

The investment consultant Charles Ellis highlights the danger of excitement (and the utility of boredom) with a neat little challenge:

“Go to a continuous-process factory sometime – a chemical plant, a cookie manufacturer, a place that makes toothpaste. If you find anything interesting, you’ve found something wrong. Investing is a continuous process too. It isn’t supposed to be interesting.”

Ellis’ characterisation of investing as a process is illuminating. The largest, most liquid asset classes are not cute farm shops selling organic products, but industrial-scale farms focused on margins, yields, and market share. They are factories for investment returns.

In recent years, there has been a trend towards Venture Capital becoming less boring, more exciting. High-profile names like Not Boring Capital have made startup investing intellectually stimulating, socially rewarding… and cool.

VC has become too interesting, too distracted by its obsession with forging the future and “making the world a better place”. Which is why most VC’s fail to deliver promised returns; they are carrying too much cultural baggage, and not enough computing power.

To survive the post-AI era, the asset class needs to evolve into something more credible and robust. In order to better serve investors and founders, VC needs to become less interesting and more scalable.

The future of our industry is boring. Or at least, it should be.

Work with Koble

At Koble, we’ve spent three years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

💔 “It’s like being gaslit”: Founders’ relationships with investors in the downturn – 49 founders and senior startup leaders respond to Sifted’s survey.

🌗 The Mixed State Of Startup Funding In 2024 – 11 charts based on recent Crunchbase data that show the state of the startup world in early 2024.

📨 Dear Venture Capitalists: You’re blowing it – If there’s to be another two decades of dealmaking, VCs better change their behaviour. And quick.

Some tweets

Parting shot

“Everything has been figured out, except how to live.”

― Jean-Paul Sartre

Regards from your [boring] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

We’ve spent three years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #30 – Last-mover Advantage (Part II) – Barbarians At The Gate

📖 Investor reading – The new realism in Venture Capital is healthy – Profitable exits also do not depend on a robust IPO environment – Eye On AI: So much for that funding slowdown

💬 Some tweets – VC gets a lot of sh!t, but VC rocks – DoorDash application for Venture Capital from 2013 – Nvidia’s market cap has surpassed the size of Canada’s economy

A reminder

In case you missed it, Koble recently contributed to an article from Weekend Fund’s Signature Block on the investors, data scientists, and tool builders building the future of Venture Capital.

Read Article

Barbarians At The Gate

Being the second installment of a two part blog series on Venture Capital’s endgame.

The “Data-driven VC” movement, championed by Earlybird VC’s Andre Retterath, has been gathering momentum over the past year.

Andre’s popular newsletter shares insights into data-driven innovation in VC and connects the dots between research, reviews of tools and databases, deep dives into various VC tech stacks, interviews with experts, and implications for stakeholders.

As we noted in our recent piece on VC’s endgame, early adopters of data and technology in the VC domain need to watch out for last-movers. A small, opaque cohort of firms are adopting a purist, hardcore approach to leveraging data in the startup investment process. 

Correlation Ventures showed us the art of the possible. And more recently, names like SignalFire, MoonFire, and QuantumLight are looking to introduce the last great product innovation in VC. 

To extend the chess analogy, these players are seeking to checkmate traditional firms by rewriting the code of startup investing.

Then there’s everyone else. Some VCs recognise the monumental significance of data, and are investing in this area at the management level. And others are opportunistically riding the Data-driven VC wave, latching onto the concept in order to boost their marketing efforts with LPs and founders.

It will be interesting to see the returns of so-called “Data-driven” funds that are pure marketing exercises in 10 years’ time. But by then, LPs will have likely forgotten the original thesis, and GPs will have banked a decade of salary funded by management fees.

What’s in a name?

How to distinguish between “Data-driven” and “Quant VC”? The answer lies in the investment approach:

• “Data-driven” means relying on data and tools to streamline operations and incorporate more information into investment choices. Humans still make the final decisions.

• “Quant VCs” means removing the human completely from the investment decision and aiming to fully automate the operations. This requires technology that is orders of magnitude more capable and complex.

Data-driven VCs integrate data points into human investment committees, whilst Quant VCs make investment decisions that are exclusively driven by algorithms. To use Henry Ford’s famous analogy, Data-driven is riding faster horses; Quant is building cars.

Fully fledged “Quant VC” is still nascent; a fragmented community of investors, programmers, AI researchers, and quants; an uprising, largely ignored by mainstream Venture Capital, more humoured than respected.

These players understand that it’s not the strongest VC that survives, nor the most intelligent, but the one that’s most adaptable to change (and Data Science). But they are outliers. We need to consider the possibility that it will be public market investors who emerge as the last-movers in Venture Capital. 

Implications for investors

If all VCs rely on the same data, nobody wins.

Generating alpha requires proprietary tech; hedge fund managers, traders, and quants get excited because they have the technical expertise to have an edge and generate outsized returns. 

These outsiders recognise the opportunity to systematise startup investing, and the scale of the prize. And where technical expertise and relationships flow, capital follows.

To those who say that Quant VC is impossible because only humans can see the soft side of founders, and new markets are art, not science, we say (stealing from the enigmatic Marcelo Bielsa, a football coach who has built a reputation on fusing data analysis with deep relationships):

“New ideas are only mad until they triumph!”

The last will be first, and the first, last. Could it be that the last-movers in private markets will be the first-movers from public markets? The technologists who “solved” the stock market have found a new problem worth solving. 

The Barbarians are at the gate of the VC industry. And things are about to get interesting.

Work with Koble

At Koble, we’ve spent three years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

💰 The new realism in Venture Capital is healthy – Rising stockpile of “dry powder” is a sign of private capital’s growing attraction as an asset class.

🦄 Profitable exits also do not depend on a robust IPO environment – The predicted slowdown, which many VCs really believed was coming, seems to be more mythical than the unicorn.

🤖 Eye On AI: So much for that funding slowdown – The fascinating “Magnificent Seven” phenomenon, the unstoppable force of AI, and the trends shaping public markets.

Some tweets

Parting shot

“The only reason for time is so that everything doesn’t happen at once.”

― Albert Einstein

Regards from your [barbaric] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

We’ve spent two years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #30 – Last-mover Advantage – Playing The Endgame

📖 Investor reading – Why Hedge Funds are Packing Up and Leaving VC – AI’s Journey From Enhancing Operational Efficiency To Alpha Generation – LPs Doubt Venture Funds’ Startup Valuations

💬 Some tweets – European venture has a serious exit problem – One of the strangest ZIRP / venture boom phenomena was startups investing in other startups – I gave a talk about startups to 14 and 15 year olds

Playing The Endgame

Being the first installment of a two part blog series on Venture Capital’s endgame.

Many household name tech companies launched with products that felt bizarrely niche and exhilaratingly new.

But getting started is different to getting big, and strategies adopted by startups to create new markets are different to those employed by companies seeking scale.

What if long term success is less about being first, and more about being last? 

In praise of last-movers

In his book Zero to One, Peter Thiel emphases the importance of understanding the endgame, quoting chess grandmaster José Raúl Capablanca:

“To succeed, you must study the endgame before everything else.”

It’s this “last-mover advantage” that defines the world’s largest companies and shapes our experience of technology and culture. Personal computers, video gaming, search engines, social networks… all massive markets in which pioneers were eclipsed by late movers:

• The last operating system – MS-DOS > Microsoft

• The last search engine – Yahoo > Google

• The last social network – MySpace > Facebook

• The last music platform – Pandora > Spotify

• The last photo sharing app – Flickr > Instagram

Tomasz Tunguz has noted that the last-mover is the embodiment of Clayton Christensen’s Innovator’s Dilemma. It’s the company that does it faster, better, or cheaper, cannibalizing an incumbent’s business with a lower cost structure.

And researchers Venkatesh Shankar, Gregory S. Carpenter, and Lakshman Krishnamurthi have shown using data from 13 pharmaceutical brands that innovative late-movers can create a sustainable advantage, with higher market potential and a higher repeat purchase rate than pioneers or non-innovative late movers. 

Research shows that market share leadership for pioneers is supported in only 4 of 50 product categories. In a paper entitled “Pioneer Advantage”, Gerard J. Tellis analyzed 500 brands in 50 product categories, with fascinating (and counterintuitive) results:

• Mean market share of market pioneers is 10%

• 47% of pioneers fail

• 11% of pioneers are current market leaders

• Pioneers are typically market leaders for only 5-10 years 

It’s good to learn from your mistakes, but it’s better to learn from other people’s (thank you Warren Buffett). This is the secret to leveraging last-mover advantage. It’s a question of timing.

The playbook is simple: last-movers hang back and let others innovate whilst they analyse the dynamics and structure of the market. Then they enter decisively, committing large amounts of resources to building and leading the market.

This time it’s different

Throughout history, the last mover has introduced the critical innovation – the last great product development – to entrench its position. It has then leveraged its financial resources, intellectual property, economies of scale, and brand recognition to consolidate control over a market for very long periods.

The ongoing emergence of massive computing power and AI models strengthens the ability of incumbents to perpetuate their hegemony over markets.

Everyone is very excited about the potential of data to transform the human experience. Few have recognised that AI may well be a “first around the post” phenomenon – the end of competition (and choice) as we know it.

Data enables last-movers to play offense and defense simultaneously; optimising user experiences whilst anticipating discontinuities that could threaten their dominance and closing them down.

AI is different from previous technological revolutions, since it relies so heavily on data. Companies that start collecting, analyzing and actioning data will drive further improvement in their models, building unassailable competitive advantages.

Humans are not good at thinking exponentially, which is why so few people are talking (and panicking) about this.

Implications for investors

We’ve seen how last-mover advantage works in startups. But the same applies to Venture Capital.

The mathematics of investing don’t change. But the way that investing is done – the underlying process of sourcing, evaluating, and backing startups – can be re-engineered with technology to provide better (and more predictable) returns to investors, founders, and society at large.

Work with Koble

At Koble, we’ve spent two years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

🏃‍♂️ Why Hedge Funds are Packing Up and Leaving VC – Hedge funds made a massive push into venture capital during the pandemic, fuelling the tech dealmaking boom. Now they’re selling startups in the secondary markets.

🤖 Venture Capital’s New Era: AI’s Journey From Enhancing Operational Efficiency To Alpha Generation – In the ever-evolving landscape of venture capital, AI has emerged not just as a tool but as a transformative force.

📉 LPs Doubt Venture Funds’ Startup Valuations – Investors in venture funds are scrutinizing unrealized returns for overly optimistic projections, slowing down fundraising processes.

Some tweets

Parting shot

“The creative process is a process of surrender, not control.”

― Bruce Lee

Regards from your [last-mover] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.

We’ve spent two years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

Find Out More

This week

🧠 Mental Model #29 – BS Asymmetry – Building a Venture Capital BS detector

📖 Investor reading – LP Perspectives 2024: investors in VC funds are super unhappy – Moore’s McIntyre to lead venture firm with Daily Mail backing – Funding for European technology companies plunges by nearly half

💬 Some tweets – What is your relationship with money? – This early stage investing thing – The most brutally honest takes I’ve read of what it’s like to work in VC

Happy holidays

As we enter the VC blackout period, we wanted to wish everyone who reads (an hopefully enjoys) Moneyball a very happy holiday season, and a strong start to 2024.

It’s been a tough year for our industry. But we shall overcome. Startup investing is bigger than the economic cycle.

Building a Venture Capital BS detector

“The amount of energy needed to refute BS is an order of magnitude bigger than to produce it.”

– Alberto Brandolini

BS Asymmetry – also known as Brandolini’s Law after it was coined in January 2013 by Alberto Brandolini, an Italian programmer – highlights the immense difficulty of debunking false information. 

George Bernard Shaw once pointed out that “false knowledge [...] is more dangerous than ignorance”. In the contemporary media environment – once characterised by Fake News, frictionless information transfer, and ideological zeal – we would do well to treat all information with a healthy dose of skepticism.

BS Asymmetry is by no means a new idea. Way back in the 1700s, Jonathan Swift (a man who elevated satire into an art-form) wrote that:

“Falsehood flies, and truth comes limping after it.”

Swift identified an asymmetry that handicaps people everywhere. It all boils down to a disparity in standards of evidence. It’s much, much easier to suggest that something might be true than it is to demonstrate conclusively that it’s not the case. 

In other words, it’s easier to fool people than to convince them that they have been fooled.

And the scientific method is not a panacea. In fact, scientific results themselves are vulnerable to BS Asymmetry.

A single experiment with anomalous results can grab headlines and get everyone talking (and believing), but it takes a mountain of effort to repeat the same results. The “Replication Crisis” that’s currently plaguing academic fields as diverse as physics, medicine, economics, and engineering, is a case in point – the results of many scientific studies are difficult to reproduce, calling into question the extent of scientific knowledge. 

We should also beware the knock-on effects of BS Asymmetry. Well-intentioned efforts to discredit BS can (and often do) have the opposite effect, as highlighted by the curious case of “Morgellons Disease”.

After watching a TV news report in February 2005, scores of people became convinced they were suffering from an alarming skin condition called “Morgellons”. The news segment was in fact a misguided attempt to explain how reports of an undefined disease were generating momentum in an obscure corner of the Internet. But it led to a surge of people discovering strange fibres growing out of their skin.

More than a decade later, self-diagnoses of Morgellons persist. Most healthcare providers recognise the condition as a delusion, treating it with antidepressants, antipsychotic drugs, cognitive behavioral therapy and counseling.

Implications for investors

The BS Asymmetry Principle is especially problematic for investors.

VCs are constantly scanning the horizon for information that might give them an edge in the remorselessly challenging world of startup investing. Refuting false information is resource-intensive. It takes time to validate the facile claims of strangers on the Internet. And we cannot account for the credulity of others. Information that we reckon to be false can still move markets, often out of our favour. 

One way to mitigate this effect is to employ Hitchen’s Razor. Investors should remind themselves that “what can be asserted without evidence can also be dismissed without evidence”. We can use this method, for example, to refute Brandolini’s hyperbolic claim that, “the amount of energy needed to refute BS is an order of magnitude bigger than to produce it.” One cannot help but think he was playing a joke on us all with such an unsubstantiated claim!

But there is another, more rigorous way to take back control of the startup evaluation and investment process.

Mathematics is a singular domain in which BS is easier to detect and refute. In the world of math, it’s much harder for someone to make a false claim that cannot be shown to be untrue (at least to someone who understands mathematics).

Venture Capital is perhaps the least quantitative asset class of all. Whilst fixed-income, commodities, options trading, and other financial domains attract people with quant expertise, VC is mostly home to multi-disciplinary all-rounders who emphasise soft skills over technical chops. 

Very few have applied a mathematical lens to the business of investing in startups, which is perhaps why our industry has changed little since its inauguration in the mid Twentieth Century. This lack of analytical rigour leaves startup investors vulnerable to BS – from founders, intermediaries, journalists, even themselves. 

“Quant VC” promises to change that; to transform the dark art of Venture Capital into a science. By structuring massive volumes of information into usable datasets and pairing them with groundbreaking proprietary algorithms, we can transform the process of sourcing, evaluating, and investing in startups.

This goes beyond the current trend towards “Data-driven” investing to create a startup investment process that’s systematic and non-human.

The cold, emotionless logic of math can transform startup investing. We can use AI to build a (much needed) BS detector for Venture Capital, evolving the asset class and unlocking massive value for founders, investors, and the society.

Work with Koble

At Koble, we’ve spent two years developing our groundbreaking algorithms, which discover early-stage startups that outperform the market and predict their probability of success.

We’re working with forward-thinking angels, VCs, family offices, and hedge funds to re-engineer startup investing with AI. If that resonates, get in touch.

Find Out More

Investor reading

😓 LP Perspectives 2024: Investors in VC funds are super unhappy – Venture Capital Journal surveyed 117 institutional investors worldwide to find out what is most (and least) important to investors in venture funds.

🤑 Moore’s McIntyre to lead venture firm with Daily Mail backing – Louis Bacon’s Moore Capital Management and the publisher of British tabloid the Daily Mail are launching a new venture fund.

🇪🇺 Funding for European technology companies plunges by nearly half – Start-ups on the continent are expected to raise $45bn in 2023, down from $82bn last year, as US investors pull back.

Some tweets

Parting shot

“Our knowledge can only be finite, while our ignorance must necessarily be infinite.”

― Karl Popper

Regards from your [BS-resistant] startup investing AI,

Koble

About Koble

Koble is re-engineering startup investing with AI, applying quantitative strategies that have disrupted public markets to early-stage startup investing.