Nadella’s Calm Testimony Masks Blindness to Altman’s Lies
🚨 In his testimony just now, at the Musk-OpenAI trial, Satya Nadella came off as shrewd, calm, and (mostly) honest, an impressive leader – but also selectively blind. How? He seemed unable to believe that the board doubted Sam’s candor, or that the board could have fired him for that reason Yet Nadella himself testified unambiguously that would fire someone not candid to their board. By that standard, Sam certainly should have been fired, and should no longer serve on the board. Witness after witness has testified that Sam lied. Toner, McCauley, Murati, and now Sutskever. All painting the same picture of Sam as a liar. (Karen Hao’s book, Ronan Farrow’s article, and my own Substack also pointed that way, too.) How could Nadella have missed the signs?
AI Hype Promises AGI; Reality Delivers Flawed Fun
Serious question: Should I write a short book called 7 lies about AI that never die?
LLM Regurgitation Proven: Hinton’s Claim Debunked
Am old enough to remember when @GeoffreyHinton told me I was stupid for saying that LLMs regurgitate training data. He was wrong. LLM regurgitation is now one of the best-established findings in the field. Excerpt below from a new DeepMind paper;...
LLM Alignment Failing; Horror Stories Keep Emerging.
Not even surprised by horrific stories like these anymore. The mission of getting LLMs aligned with human values has largely been a failure.
Self‑play in Go Isn’t a Model for Real World
Really doubt what Hinton says here. Self-play for games like Go is not like the open-ended real world.
Elon Urged to Fulfill X's Promised Truth Claim
hey @elonmusk, if you really care about making X a source of truth as you used to claim, take note:
Top Tweets Prioritize Speed Over Accuracy, Study Shows
💯. there was actually a study about that by @dkroy and @sinanaral in Science in 2018: fake news travels faster than true news.
Pure LLM Scaling Won’t Reach AGI, Neurosymbolic Needed
So many people misremember (or never read) what I said in in 2022 in “Deep learning is hitting a wall”, which was neither about revenue or AI’s potential upper limits. Rather, it was an argument that the pure of scaling...
Measurement Wall Vanishes With
Sorry, @peterwildeford, but this is wrong. Please don’t play along. The measurement “wall” you mention is hit ONLY if you don’t insist on reliability. If you demanded 95% accuracy on the task, the systems wouldn’t be close to the measurement wall. The...
Future Scrutiny Will Expose Overlooked Debt and Profit Warnings
Questions people will be writing about in the coming years: • Was it a Ponzi scheme? • Why didn’t more people worry about the circular financing? • Why were investors so blasé about the lack of profits? • Why wasn’t all the debt a bigger warning sign?
METR Graph Shows 50% Success, Not Reliable General AI
Hot take on METR’s new graph that so many people are flipping about today. • Claude Code is a real advance; Mythos probably builds on some of what is learned there. But… • If you read the graph carefully, it is about achieving...
Mythos Popular yet Flawed: Modest Gains, Lingering Errors
This is confused, but popular. Popular because it tells a bunch of people what they want to hear. Confused for a couple reasons: first, Mythos probably isn’t a pure LLM. (Claude Code isn’t, and it probably uses some similar techniques). ...
OpenAI's Decline Signals AI Industry's Future Uncertainty
🚨The slow decline of OpenAI has begun. 🚨 These two hesitations from SoftBank and Broadcom, back to back in my feed from @HedgieMarkets and @anissagardizy8, are major signs. The money question of 2026: what happens to the AI industry as a whole...
Public Domain Release Sparks Potential Updated Edition
Since this is now public domain, maybe I will write an updated edition … https://t.co/HpwDemQR0N
Generative AI Hype Outpaces Real ROI for Businesses
Remember that MIT study that showed that the ROI for generative AI wasn’t really there for most businesses? Or any of the six or seven studies from other teams that followed, showing basically the same? The situation for agents looks to...
White House Finally Embraces AI Oversight I Advocated
Happy to see that The White House has finally realized that the Andreessen-Sacks zero regulation approach to AI was going to blow up in the their face. Also glad to see them finally consider the kind of oversight that I...
OpenAI's Partnership Hype Outpaces Financial Clarity
“they hadn’t figured out how OpenAI would pay for it” may turn out to be the epitaph for an entire era. 🪦 scoop from @anissagardizy8 @theinformation, and credit her with the great line “OpenAI has made a habit of announcing landmark...
OpenAI Staff Resign over Safety Concerns, Notable Exits
Many, many OpenAI employees quit over safety concerns, including @DKokotajlo, William Saunders, @sjgadler, etc as well @Janleike. The founders of Anthropic such as @DarioAmodei and @jackclarkSF may have left for similar reasons. Does anybody have a full list?
AI Revolution Yet to Boost Productivity After 3.5 Years
3.5 years into the “AI revolution” Basically no sign that it has mattered for productivity. https://t.co/yY98peYAPi
Nvidia Climbs as xAI Releases 220k Used GPUs
Nvidia up 2.6% on news that xAI is flooding the market with 220,000 secondhand GPUs. https://t.co/er8D1yd9GR
Paper’s “Just Published” Claim Is Recycled, Already Debunked
Wow this paper has been “just published” many times for nearly a year. I have called out at least two other “influencers” for the same thing on the same paper, including another one earlier this week. C’mon guys. The conclusion is...
Speculative GPU Hoarding Leads to Wasted Costs
just imagine what might happen if everybody was stockpiling GPUs at high prices for a demand that didn’t quite materialize….
Elon’s Compute Rental Signals xAI’s AGI Shortfall
Hot take on Elon’s surprise decision to rent 30 megawatts of compute to Anthropic: 1. It’s a tacit concession that xAI is not all that close to AGI (despite what he suggested last year). 2. It’s more evidence that pure scaling doesn’t...
BBC Skims Zilis Testimony, Overlooks Conflict Management
C’mon BBC. Zilis was sharp as a tack on the stand, on her role in the OpenAI *nonprofit* board, and how she managed conflicts as they began to develop, and this (which is not really even news since it was...
Reliability Is the Core Challenge in Generative AI
Precisely what I have been saying since day 1: the challenge in generative AI is reliability. That was the challenge, and still is.
Forbes Ignored Evidence, Made Disgraceful Choice
If Forbes had only waited to hear the testimony at this week’s trial Or read @_KarenHao’s book Or @RonanFarrow’s @newyorker investigation Or my own writings since fall 2023 They would have realized how disgraceful this choice was.
Ilya's Concern Was Sam's Misconduct, Not AGI
100% stood the test of time: “What Ilya saw” was Sam’s bad behavior, not AGI.
Musk‑OpenAI Case: Rare Factual Consensus, Ethical Question Remains.
Many trials feel like Rashomon with different witnesses. The amazing thing about Musk-OpenAI is how much agreement there has been (at least so far) on the facts. The question is really whether what OpenAI did is ok, not about whether...
Sam's Firing Was About Trust, Not AI Safety
Mira Murati’s testimony is gripping – and what it makes absolutely clear is how utterly wrong most of Twitter was about why Sam was fired. – It had nothing per se to do with AI safety - It had nothing do with...
Sam Altman's Wild, Unpredictable Leadership Tactics
How Sam ("You parachute him onto a cannibal island, and he comes back five years later as king”) Altman operates:
Sam Is No Longer the Right CEO for OpenAI
And how many times have I told you that Sam is no longer the right CEO for OpenAI?
Hyperscaling Compute: Biggest Historical Bet, Risking Failure
Am I right that hyperscaling compute is the biggest bet in history? Any counter examples? It’s way more expensive than the Manhattan Project, the Apollo project, and railways across the US. If it doesn’t yield AGI, it may also be the...
AI's Real Value Won’t Arrive Until Late 2025
👇 This is an incredible admission by Jensen Huang. Effectively he is saying that AI, for all the hype (some from himself) wasn’t really useful until late 2025. Let that sink in. This means practically everything everyone said before was more or less...
Advisory Jury Limits Impact in Musk‑OpenAI Trial
Important nuance: the jury at this Musk-OpenAI trial is an *advisory* jury, hence not binding on the judge, and is only looking at liability (not damages, if any). Thanks to @bahhradx for correcting an earlier misstatement of mine.
Brockman Says Sam Fired for Lack of Candor
Brockman confirms that Sam was fired for not being consistently candid. Crazy that @karaswisher blocked me for saying that the board fired Sam for not being consistently candid, when that is in fact what happened. But she did. And continues to do...
Neural Nets Still Can’t Generalize Beyond Training Data
Some things never change. If you don’t understand this one, you don’t understand what’s happening AI. Marcus, 1998: neural nets have trouble generalizing far beyond the data. Marcus, 2001, 2012, 2019, 2022, etc: neural nets have trouble generalizing far beyond the data. Apple,...
Counsel Details Timeline, Still Avoids Refuting Self‑Dealing Claims
Brockman’s counsel is doing a good job of laying out the timeline — but done little so far to refute yesterday’s dissection of her client’s self-dealing and dodgy behavior regarding his fiduciary responsibilities to the nonprofit.
Elon Waived Cash Award, Damages Go to OpenAI Nonprofit
Folks, AFAIK this is *literally not a possible outcome of the trial*, as Elon has waived the right to any cash payment to him, instead assigning any monetary damages to OpenAI’s nonprofit.
Even in 2026, Andreessen Misses LLM Prompt Limits
Hilarious (and maybe a little bit scary) that even in 2026 Marc Andreessen still hasn’t learned that LLMs don’t know how to reliably follow system prompts.
Trump's Top AI Initiative: A Promising Idea
This would be a very good idea, especially if implemented well. Best thing the Trump administration has considered doing around AI.
Brockman Confirms Altman Was Not Fully Candid to Senate
Greg Brockman just confirmed what I speculated in December 2023! Sam Altman was not fully candid in his testimony to the US Senate.
Musk's Lawyer Leverages Brockman's Emails, Boosting Win Odds
Musk’s lawyer is calmly eviscerating OpenAI’s President Greg Brockman, largely using Brockman’s own diaries and emails. For the first time I think Elon has a real chance of winning.
Stuart Russell Testifies in Musk‑OpenAI Trial Despite Exclusion Attempts
Well-known, well-respected AI expert Stuart Russell is now testifying at the Musk-OpenAI trial – despite OpenAI’s BS attempts to exclude him. Read the backstory here:
Six Weeks After AGI Claim, Nothing Feels Different
It’s been almost six weeks since Jensen Huang claimed we had achieved AGI. Why don’t I feel any different? https://t.co/p3nO91lTfz
LeCun’s Public ‘Admission’ Sparks Sarcastic Applause
Thank god, indeed, that @ylecun publicly stood up and admitted he was wrong, to fight for the greater good. Oh wait.
LLMs Show Minimal Evidence of Improving Health Outcomes
“In summary, there is very little evidence for LLMs benefiting patients or doctors for health outcomes” - Dr. @EricTopol Read his full review here: https://t.co/vHKo35n2BT
Ego Persists as Hype Shifts From AI to LLM Flaws
Old days: “Bigshot chief scientist of major corporation can’t handle criticism of the work he hypes” New days: “Same bigshot no longer hypes LLMs; can’t handle the fact that someone else noticed flaws in LLMs before he did.” The only constant is...
Meta Feigns Openness, Then Reverts to Closed Model
Meta bootstrapped China, reaped the positive PR for being “open”, and then went back to being closed.
Generative AI's Harms Outweigh Its Benefits
Why is the AI backlash growing? Outside of coding (where there is clear value), and a handful of other domains (e.g. brainstorming), Generative AI has been a net negative for society. GenAI has been undermining secondary and college education, opening up mass...
Altman Calls AI Job‑loss Warnings Tone‑deaf
Of course Sam is busier than ever. OpenAI is in deep trouble, and, no, Codex ain’t gonna save his company from collapse. Non sequiturs won’t, either.