Today morning in San Francisco we had an awesome curated gathering of exceptional founders building the future of Financial Services, Healthcare, and Commerce through applied AI. The discussion was curated by @gai_ventures_ - myself and Vijay ​There were no slides. No panels.

AI isn’t “good at engineering” and “bad at design” by accident. It’s structural. AI does core work only when: - Inputs are deterministic - Outputs are testable - The medium is machine-native Code? Yes. Spreadsheets? Yes. Visual taste? Not yet. AI kills execution before it

Hot takes on AI & jobs are everywhere. Actual data has been rare. A survey of 1,750 tech workers by @lennysan and team shows AI does core work when 3 things are true: The work is formalizable + quality can be objectively verified + medium is machine-native (text, code, tables)

"The reward for great work is more work, and I find that saying that maxim to the right person, the kind of person I want to spend time with, their eyes go wide and they understand it immediately. That the reward for great work is not money, power, fame. It is the privilege to

Google just announced real-time translation for video-calls.

Bill Gurley (@bgurley) CLEARLY outlines his AI thesis in 14 minutes to Tim Ferriss (@tferriss) 1. Every real tech shift that creates wealth also attracts speculation, and the two usually show up together 2. AI looks like an industrial wave where real companies and workflows get

The -1 to 1 journey of a startup is literally just this

Agency > Intelligence I had this intuitively wrong for decades, I think due to a pervasive cultural veneration of intelligence, various entertainment/media, obsession with IQ etc. Agency is significantly more powerful and significantly more scarce. Are you hiring for agency? Are

world's best M&A market

Books, essays, podcasts, talks ... all welcome.

What’s one thing you read, heard, or watched this year that meaningfully changed how you think -- abt business, leadership, or life? Content that made you pause… reconsider assumptions… and operate differently afterward.

New Build AI podcast by gAI Ventures is out. YouTube: https://t.co/y3EodZyUhp Spotify: https://t.co/rdJqmF9VYx Apple:

Just recorded a pod w/ @sawickipedia (serial founder + investor). He’s building @GumshoeAI, backed by venture studio @psl Big idea from the convo 👇 AI visibility will become the #1 customer acquisition channel. Some sharp insights: • AI is the new search layer. If LLMs

Holiday reading

At Prove we had 80+ saas subscriptions. A typical tech growth stage company in the US would be similar. I see the future where they will cancel most of the subscriptions but still get the services. Each provider like workday or Navan or booking . Com or Spotify will create

In the AI space, do we all agree now on the following? 1) AGI is still a decade away (per Andrej Karpathy), and today’s AI excels only in narrow, well-defined tasks, lacking human reward systems like curiosity or intrinsic motivation. 2) Vertical AI is picking up momentum

There is too much consensus building up -- everyone is long Alphabet $Goog Google now is there in almost every critical layer of the AI stack. It has its own chips (TPUs), the world’s largest datasets in video (YouTube), images (Google Images), and the open web (via Search),

https://t.co/jfEQQdDvsq

Prompt engineering is a phase, not a moat. The best AI won’t wait to be told what to do. It’ll read context, infer intent, and act at the right moment. The real winners won’t feel like “AI products” at all. They’ll feel obvious. And impossible to replace.

Everyone is talking about AI models. But Larry Ellison said something that cuts deeper: --> The real value is shifting from public internet data to proprietary, permissioned, workflow-native data. Some examples of companies with privileged data: ~HIMS with patient interaction +

My twitter 2025 wrapped.