We’ve entered “phase 3” of the AI era

3 trends, 2 theses and 1 tool from Shuo

Hello friends!

Welcome (back) to Shuo’s Snippets” where I share what’s new and next in startups and tech. As always, thank you for being someone who’s made me a better and smarter person.

This is my way of sharing notes and sparking discussion, so feel free to reply anytime – I’d love to hear what you’re seeing. No hurt feelings if you opt-out!

So, here’s what I’ve been seeing this past month investing in fractional founders* as well as teaching entrepreneurship at Berkeley and Stanford:

📈 3 trends in startups/tech/venture
🤔 2 theses on what’s next
🔧 1 tool I love

*a fractional founder is an entrepreneur who is transforming their part-time project into their full-time startup

💌 Someone shared this with you? Connect with me and join the email list here.

3 trends in startups/tech/venture

⚙️ We’ve entered “phase 3” of the AI era

  • While it may feel like just yesterday that “AI” became a buzzword, the field has already entered its 3rd phase of innovation:

    • Phase 1 (~2022-2023) was all about “prompt engineering” — figuring out what to say to AI to get a higher quality output.

    • Phase 2 (~2023-2025) was all about “context engineering” — figuring out what information to share with AI to get a more tailored output.

    • Phase 3 (~2025-?) is now all about “compound engineering” — getting AI to learn from its previous work and mistakes, so it can do things more efficiently the next time around.

  • As if startups weren’t already coming to market at breakneck speed, the rise of “compound engineering” is likely to make the pace of launching and growing a startup even faster.

  • To help our founders adopt best practices faster, our team at IOVC has built the largest crowdsourced prompt library, tailored for early stage founders and optimized for Phase 3 — it’s available here (updated on an ongoing basis).

🤖 More managers are using AI tools to make management decisions

  • A growing number of managers—currently 60%—are using AI to make critical business decisions, including determining raises, promotions, and firings.

  • What I call “AI-augmented management” is especially common among tech founders. For example, Google co-founder Sergey Brin has reportedly used AI to summarize large group chats to identify top performers. AI even even helped him identify and later promote a hard-working but quiet female engineer (after validating his/AI’s observations with her human manager).

💰 Venture capital is less attractive as more money goes to fewer companies

  • Venture capital is becoming even more concentrated. This year, 41% of all VC dollars have gone to just 10 startups—8 of which are AI companies. This is a 75% increase from last year — and greater than any year in the past decade.

  • With so much money going towards so few bets, LPs of VC firms are increasingly hesitant to put even more money into the asset class. Venture isn’t dead, of course. It’s just that VC firms increasingly need to offer a thesis that’s not just “we’ll invest in the same 10 companies as everybody else.”

2 theses on what’s next

📈 AI startups will increasingly start as consulting firms

  • No one wakes up in the morning thinking, I want to buy a new piece of software. And yet, this is exactly what tech startups try to do everyday: to convince someone to buy a widget.

  • The most successful startups sell solutions to existing problems—not solutions in search of problems. What’s the easiest way to identify a problem, and solve a problem? By working side-by-side with a client.

  • As the number of AI startups continues to explode—and as potential customers find it increasingly difficult to tell one AI startup from another—I believe more startups will start not with a product but with a service.

  • In classic Silicon Valley fashion, there’s a term for offering “engineers as a service”: it’s called “forward-deployed engineers” (popularized by Palantir).

🤔 The best investors are non-consensus, not necessarily contrarian

  • A contrarian view bets against the crowd (i.e., you think the popular opinion is wrong).

  • A non-consensus view bets away from the crowd (i.e., you think the crowd isn't focused on what's most important).

  • Given the “wisdom of the crowds,” it’s hard to be 100% contrarian. After all, how can everyone be wrong?

  • A better strategy is to not to be contrarian but simply non-consensus: instead of trying to prove everyone wrong, look where others are under-investing (e.g. a certain market, model, tech, execution, cultural trend).

1 tool I love

🎓 Smarter lesson creation tool for teachers

  • TeachShare is an evidence-based AI tool that helps educators create engaging, differentiated, and evidence-based classroom lessons.

  • It enables teachers, parents, and other educators to quickly turn ideas into ready-to-teach resources that adapt to any subject, grade level, or curriculum standard.

What’s top of mind for founders?

Founders have been asking me a lot about founder-market fit. You can hear my latest thoughts below 👇🏼

Please hit “reply” with any thoughts and reactions, and stay tuned for more on what’s new and next in the coming month!

Cheers,

Shuo