The Rise of "Vibe Investing"
When AI Meets Venture Capital
In recent weeks I have witnessed the meteoric rise of "vibe coding" in software development. What was once an impossible task or a theoretical app, is now a weekend project. In the venture world a parallel phenomenon is emerging: "vibe investing." Like its coding counterpart, pioneered by Andrej Karpathy, vibe investing represents a fundamental shift where investment decisions are increasingly influenced—or even outsourced—to a mix of vibes (sentiment) and artificial intelligence models. The defining characteristic of both trends? Vibe investors (and coders) fundamentally don't understand exactly why specific decisions are being made.
Why Vibe Investing Is Taking Hold
The investment landscape, particularly in tech, has become increasingly homogenized. There used to be only 12-15 “venture scale” companies created every year that materially impacted venture returns. Performing in venture was simply a question of whether or not you invested in one of those companies. However, in response to the influx of capital, many businesses have shed their boring business fundamentals facade in exchange for the a hoodie a “venture-backed business”.
Originally a “venture scale business” needed to have a massive margins, a massive moat and be growing like a weed. Because of these characteristics, they were easy to spot. As more companies have tried to present themselves as venture scale businesses, it has become a lot harder to quickly discern the difference.
AI has accelerated both the number of startups and their similarities to an unprecedented degree. Can you distinguish between Cursor, Replit, Windsurfer, and Cosine at a glance? The technologies powering these platforms still rely on the same fundamental innovations and, in some cases, even the same models, making even technical differentiation increasingly blurry.
This is not to say there is no innovation or distinction between these solutions. It is instead to say those distinctions are happening in different parts of the business then they have before; and they are smaller and more nuanced than ever.
Simultaneously, the growth in the market of founders seeking capital has outstripped human capacity to perform comprehensive due diligence at scale. Venture capitalists face mounting pressure to process more deals faster while still maintaining quality and respecting their relationships with founders. All of this is creating the perfect conditions for AI-decision-making to take root. But there is a big difference between AI-assisted-decision-making and an AI-made-decision.
The Positive Elements of Vibe Investing
When implemented thoughtfully, AI can enhance the investment process in several ways:
Intuition amplified by AI: Formalizing gut feelings through computational analysis that can help structure the conscious conclusions generated by an investor
Narrative assessment: Evaluating the strength and coherence of a startup's story rather than relying solely on metrics
Detecting clarity of thought: Using AI to analyze how clearly founders communicate their vision
Pattern matching at scale: Identifying similarities to previously successful (or failed) investments while assessing whether founders themselves are generating and compounding new perishable knowledge
Can reduce the bias towards traditional metrics: Particularly valuable for early-stage investments where conventional performance indicators may be misleading or absent
The Dangers of Over-Reliance
The risk with vibe investing is missing the 2-3 sigma (outlier) outcomes that venture capital is built on. At its core, AI excels at pattern recognition—and so does investing; the crucial difference lies in the quality of input data and ability to diverge from conventional thinking.
In venture capital, we call this "non-consensus thinking" or being "contrarian and right" (originally popularized by famous investor Peter Thiel). But human innovations are often AI hallucinations. To the extent you believe the future will mirror the past, AI may predict well. But if you believe the future will diverge from what's already been done, AI's ability to predict truly novel outcomes remains limited.
The Human Edge: Perishable Knowledge
The human advantage in investing stems from our ability to process what I call "perishable knowledge." These are unique contextual insights or anomalies. According to research by Claude Shannon, Harry Nyquist and David Blackwell, the human body sends approximately 11 million bits per second to the brain through our sensory systems. While our conscious mind processes only 10-50 bits per second (a mere 0.00045% of sensory input), the remaining data undergoes tremendous compression, filtering and processing in our subconscious. At this rate, by the time a human reaches 3-4 years old, they've already processed more training data than the most advanced AI models.
Even though AI may have a much larger hard-drive and processing speed (the fastest super computer in the world, El Capitan has 5.4 petabytes of High Bandwidth Memory (HBM3) approximately 2.16 times more than a human brain and 1.742 exaFLOPS of processing power, approximately 174.2 times more raw processing than the human brain), the quality of a venture investment decision isn't about storage capacity or computational power—it's about context and the quality of the insight. Two attributes that AI simply cannot access…yet.
The Hybrid Approach: “Chocolate and Peanut Butter”
The investment unlock may be using AI's vast horsepower for processing static knowledge—the things we handle with our conscious mind—while allowing our subconscious to process the context we create from perishable knowledge. This "chocolate and peanut butter" combination might give us the best of both worlds.
Like the hybrid approach emerging in the coding space (where vibe coding shows 70% faster prototyping but traditional coding maintains superior quality and security), AI can accelerate speed in performing diligence or refining an investment thesis while humans ensure the AI is helping to diligence an insight that is actually worth pursuing.
The Confidence Problem
Investments requires confidence in your answers—confidence that's hard to feign if you don't fundamentally understand how you arrived at a conclusion. It's like copying answers from a friend's test; when challenged, you can only project your confidence in the friend you copied from.
This is why my tenth grade math teachers, Mr. Dallas Jacobs, insisted you show your work. If you don't know how you got to the answer, it's difficult to have the conviction to apply that answer or justify your position.
History is littered with examples of investors who got cold feet and sold valuable positions because they lacked true conviction in their investment thesis. Michael Burry, who rose to fame by correctly betting against mortgage-backed securities during the 2007-2008 housing market bubble (as portrayed in “The Big Short”), experienced investor doubt that only comes from not understanding how the answer was derived first hand.
Many of his investors got cold feet and pulled their money out of his fund before his market predictions ultimately came true. While Burry himself held firm in his conviction, his investors who sold too soon missed out on substantial profits when the market eventually crashed as he predicted.
Similar to vibe coding, vibe investing is fun when the market's going up, and everything is working. But the real test is whether or not you will have confidence in those answers when the code breaks or your investments stumble?
The Path Forward
AI is most effective as a tool to amplify our thinking, not replace our thought. Investing remains a race to the truth—you must both get there first and be right about where "there" is. AI might help you move quickly, but if it leads to the wrong destination, it's all for naught.
My own personal investing framework seeks to leverage AI while preserving my human context and judgment:
I never outsource my investment decisions to AI
I use AI to help baseline the state of the art by processing numerous white papers
I view a founder's ability to mobilize people (including mobilizing me as investors) as more critical than their ability to manipulate AI-bots
I employ AI to argue against my investment memos or help brainstorm strategy flaws and risks
I recognize that early-stage investing is still fundamentally about people and relationships
Conclusion
Venture capital will always balance art and science because it depends heavily on context and context consist entirely of perishable knowledge. Context is processed in the human subconscious and much of what informs it is largely unavailable to AI. We call this intuition—which is not to say there's no processing or logic, but rather it's logic we sometimes can't articulate through our conscious mind.
The most successful investors will use AI as a tool rather than outsourcing judgment. There's enduring value in "showing your work" when making investment decisions, even if—as my 10th-grade math teacher Mr. Dallas Jacobs taught me—you get to the right answer. Understanding how the answer was derived leads to consistency. Consistency leads to repeatability. And repeatability leads to the holy grail of investing - uninterrupted compounded returns. 💸
Investing remains, at its heart, a people business. Processes with a human-in-the-loop work best with a thoughtful, discerning human-in-the-loop. AI can help us process information and identify patterns, but the truly valuable insights—the ones that lead to outsized returns—often emerge at the intersections that only context, human creativity, risk-taking, and pattern recognition, can navigate.
The good and bad of venture is that we won't know the answer to many of these questions for a long time. But one thing is certain: in the race to discover the next unicorn, both speed and understanding matter. So, show your work.
interesting collin, thanks for penning