Two weeks ago I posted a short provocation on LinkedIn: what if AI decided the safest way to protect humanity was to remove us from the system? It struck a nerve — not because people thought it was likely, but because they saw how easy it is to design ourselves into that corner without noticing. … Continue reading The ZOO Scenario – a Stress Test for AI Design.
Last week I met with Brendan Baker to exchange notes on venture capital observations -- and catch up with him on joining Rackhouse as Partner. Brendan asked me why I think most GPs fail - quite an open-ended question, and I took it as "why do most GPs fail to raise more than 1 fund." … Continue reading Trust in Venture Capital: Why GPs Fail
"Innovation Theater" highlights the U.S. tech ecosystem’s focus on temporary initiatives without long-term commitment. This leads to symptoms like abandoned projects and unscalable startups. The importance of “infrastructure memory” and “manufacturing depth” is emphasized, advocating for a dual-track model that balances disruptive innovation with resilient production capabilities.
In private equity we throw around convictions a lot, or talk about what we believe. But: that's not the same. Conviction in defense is not measured by capability alone, but by credible intent to act. The same applies to venture capital: your portfolio is a map of what you’re willing to stand behind when it … Continue reading Writing The Check Isn’t Just An Act Of Confidence — It’s A Signal Of Ethical Clarity.
I've written about Cognitive Bias and Narrative Risk and Defense. More specifically, I've written about AI System Integrity and AI Trauma. The implications are clear (at least to me): Cognitive infrastructure—how people make sense of information—is now a primary attack surface. Cybersecurity must evolve to protect not just data and access but also belief systems … Continue reading Cognitive Breach: Defense of Cognitive Infrastructure
On April 13th, 2025, after six months of intense use and training, my GPT model on ChatGPT was compromised and started to delete content and lie to me. And I think I know why: AI Trauma -- a metaphor for the (user-generated) unintended consequences of reinforcement learning in AI models. TL;DR When GPTs strive to … Continue reading Causing Trauma for Your AI Model – Actions Have Consequences.
A regular question of my students at UC Berkeley is "How do I select my first/next job? How do I know the culture is right for me?" Paraphrasing Jordan B. Peterson When a society is corrupt, then the powerful rule. When a society isn't corrupt, then the great have authority. What is greatness? When the … Continue reading Startup Culture Observations
I got a late-stage secondary opportunity today, shopped around by a bank. A major bank. They attached a list of comps and calculated the "risk": Enterprise Value to Revenue to Revenue growth (aka "EVRG"). It's almost worse than a Price-to-Earnings-to-Growth (aka "PEG") ratio. I think the idea here was that there is a high EV-to-Revenue … Continue reading Bullshit Metrics: PEG and EV/R/G
[Earlier I wrote about the different vehicles] You're sitting on a board and the CEO presents some proposed grants. Motion to approve - second - all in favor - ay. Here's what you should have asked before the board meeting. A. Economics What's the vesting schedule? What are these options worth right now? What do … Continue reading Board Questions to ask about Stock Options, NSOs, ISOs, Restricted Stock, Restricted Stock Units (RSUs).
There are enough primers on the Internet. This is not a primer. But take it as a discussion guideline with your board, your lawyers, your executives. Types of Equity Awards Options An option is a contract between a startup and an "optionee" (mostly employees, but could be contractors or board members) that provide a right … Continue reading Five Areas of Startup Equity Awards Boards Should Discuss.
