According to NVIDIA’s 2026 annual report (p. 35), Jensen Huang held 3.58% of the company as of March 23. The other nine directors held just 0.31% combined. In my view, giving directors with so little equity the power to shape NVIDIA’s future is a serious governance problem.
I’ve been reluctant to try ChatGPT. Today I got over that reluctance. Now I understand why I was reluctant.
The value of 90% of my skills just dropped to $0. The leverage for the remaining 10% went up 1000x. I need to recalibrate.
Elon Musk claims that government intelligence agencies can access Twitter DMs, which are unencrypted. He plans to launch encrypted messaging by the end of this month. 🔒🐦 #ElonMusk#Twitter#Encryption
Elon Musk plans to create his own TruthGPT, believing that OpenAI's shift from open-source and profit-seeking deviates from its original mission. However, so far, TruthGPT hasn't announced any plans to be open-source either. 🤔 #ElonMusk#TruthGPT#OpenAI
This is more important than GPT-5, which can be pursued separately. RD can focus on RD, BD on BD, but hiring and budgeting should prioritize the BD department. #OpenAI#GPT4#futureofAI
At MIT's Imagination in Action, Sam Altman shared...
1. The public letter by Musk and others about OpenAI training GPT-5 is not true, and won't happen for a while.
My take: If I were in charge, I wouldn't prioritize GPT-5 either. No clear competitors right now, and GPT-4 is already strong. Commercially, it's about market capture. OpenAI should become the pioneer in various fields, retain users, and secure its territory.
4. He believes we're nearing the limits of large language model scale; while parameters may still grow, future models should shift towards smaller, collaborative models. The focus should be on rapidly increasing model capabilities, not obsessing over parameter counts.
3. He's stoked that GPT-4 has been hailed as the safest and most consistent model so far, rather than the most powerful. The 6 months spent training GPT-4 focused on this.
one thing coming up in the debate about the pause letter i really agree with:
openai should make a great alignment dataset and alignment evals and release those!
bonus points if we can find a prototype democratic process for 'what we align to'.