💼 Sink or Swim

Good morning! It’s Wednesday, March 22nd. This is now our fourth morning briefing. I’d love your feedback at the end!

Now here’s what other vc-backed CEOs are talking about.
-Anuj

Crystal ball for startup exits

• What to know: PitchBook developed a predictive model to determine your startup’s fate: M&A? IPO? Bust?

• How does it work? The model leverages cap table data (e.g. who is in, how much raised, what stage), compares it to previous outcomes those data points were involved with, and makes a prediction for your company.

• How well does it work? It’s correlative. In a backtest of 26,000 companies with known outcomes, Pitchbook self-reports a 74% accuracy rate, and found those with top decile characteristics were ~3x more likely to have a successful exit (via M&A or public listing).

• What does this mean?: Naturally, your outcome is influenced by your own actions. But it does lend credibility that your cap table matters. It’s a good thing to have in mind the next time you fundraise!

Altman’s AI apprehension

• What happened: OpenAI CEO Sam Altman sat down with ABC to talk through what doesn’t — and very much does — worry him about the explosive advancements in AI.

• What to know: Altman’s worried how this tech can be used negatively: cyber attacks, misinformation, and society not shifting jobs quickly enough.

• Why this was interesting: ChatGPT has had exponential growth since its release. Given the existential nature of these implications, I’d expect policy and mainstream media to be all over this sooner rather than later.

• A tidbit: Interestingly he noted a big part of why GPT-4 is better than GPT-3.5 is because of its ability to reason better. OpenAI models are deductive reasoning engines, not memorization. I didn’t pick that up before!

Financial planning: survivor edition

• What to know: Sequoia Partner Ravi Gupta dropped an insightful piece, noting how long financial runways often lead us to ignore reality and lower performance standards.

• How to visualize? American children at much better at swimming than they are at math. While these teachers treat their kids the same, swimmers have an unforgiving performance standard: water. You sink or you swim.

• What’s the takeaway? Longer runway influences us to prioritize feelings over performance, and this is the dangerous path to mediocrity.

• What can I do? Embrace reality & hold yourself to a high standard. Does your 2023 financial plan represent excellence? Or is it a set of numbers your company uses to say its “beating plan?” As Ravi mentions, slightly outperforming a middling financial plan doesn’t matter. What matters is moving your company towards excellence.

AI flunks chartered financial analyst exam

• What to know: On the separate notes of explosive AI & financial planning, it’s best to not have ChatGPT lead your finance team!

• What happened: The Financial Times gave ChatGPT-3.5 sample questions from a CFA Institute Level III paper. It performed poorly, scoring 8 out of 24, when it needed at least 17 to pass.

• Why? As ChatGPT is a reasoning model (just learned that!) it struggles as finance has arbitrary rules and definitions that run counter to AI’s logical thought.

😅

Thinking of my Bay Area friends!

You’re all caught up.

See you tomorrow & make yourself proud today! 👊🏽