Knowledge catalogue
in case you needed more evidence that tokenmaxxing was a stupid idea Meta burns 2.65B a year on AI tokens. at 300K for a Meta engineer, that's enough to pay ~9,000 engineers for a full year. now ask y
Should we be spending literally trillions on a speculative bet when we can’t even solidly document clear positive impact? Great analysis of what we don’t yet know below, by @bencasselman drawing on a
Maria Ressa and Yann Bengio, prominent figures in technology and AI governance, co-authored a significant letter addressing concerns related to the UN Independent International Scientific Panel on AI.
Reselling “excess AI processing capacity”: first SPCX, now META. In a rational world these moves would be seen as signs that we have *already* started to overbuild. Boom. $META is developing a cloud s
That day when @elonmusk starts to sound a bit like my 2023 TED Talk, which called for “a global, nonprofit organization to regulate the tech for the sake of democracy and our collective future.” Good
Another case of what I call The Klarna Effect. 🦔Ford admitted it had to rehire engineers after replacing them with AI systems that couldn't do the job. The company's VP of hardware engineering said th
Gary Marcus discusses 'The Klarna Effect,' examining how AI adoption by companies like Klarna has led to significant workforce reductions despite increased productivity. The post explores the broader
GLM 5.2 is the ultimate culmination of what i have argued here for the last three years. no technical moat and widely known formula -) convergence -) price wars -) small or negative margins. GLM-5.2 i
hard to see how anyone makes much money from all this in the long run. more like the airline industry; very small margins and big expenses pouring trillions in probably wasn’t wise China’s AI playbook
This article examines how coordinated AI systems could be weaponized to undermine democratic processes through large-scale disinformation campaigns and manipulation. Written by prominent AI researcher
wild. Google reportedly limited Meta’s use of Gemini due to a shortage of compute resources. — FT Google is in a position where it can’t sell Gemini to Meta as freely as it might want to. Compute rema
🚨 BREAKING: Anthropic and US Govt insiders expect the administration's limits on Fable 5 could be lifted as soon as this coming week, per Axios. The Pentagon and National Security Agency still have to
Gary’s analysis is pretty much spot on as ever. This is such a worrying development People are not getting how significant this news – The White House delaying GPT 5.6 – is. Here’s the context and a s
Ladies and gentlemen, I am pleased to announce that staplers & pencil erasers will once again be available for sale to some companies, as well as a few close personal friends and important donors. Thi
more arbitrariness, more favoritism. Semafor is reporting that The US government has lifted its block on Mythos 5 in a letter this afternoon from Commerce Secretary Howard Lutnick to Tom Brown. Fable
Thanks for running our open-source work on current frontier models “The results are: the most capable models today (GPT-5.5 Pro) did outperform the best models from before (79/100 vs 69/100), but did
This is from a popular inference provider GLM-5.2 plus the US banning the most capable new models means open source caught up to SOTA closed source coding models This could be v problematic for Anthro
This is very interesting. Coinbase seems to have lowered their token spend ($$) to about half, by 1) routing to cheap inference like GLM 5.2 and Kimi 2.7 that are still pretty performant 2) Smart rout
Adding my voice to the chorus who are saddened by @Om’s too-early passing. Just heartbroken to hear of the passing of the incredible Om Malik. @Om was always a pioneer, a deep thinker, and a truly ori
'AI Is Dumber Than Investors Think — ft. Gary Marcus' Ed Elson is joined by Gary Marcus to discuss why he’s concerned about the fact that we’re all-in on AI. They explore why he argues generative AI i
Although the tech bros and some pharma think AI will cure cancer, Nobel Laureate Jennifer Doudna, inventor of the Crispr gene-editing technology says that innovation is still really in the domain of h
Always playing checkers and never chess. President Trump on truth social: any country that imposes a Digital Services Tax on American companies will immediately be met with a 100% tariff on any and al
“America’s preoccupation with “winning” the AI race with China could well lead to unprecedented catastrophes, even catastrophes on a global scale. Not all games are zero-sum, and if this fact doesn’t
“An AI future where only a few US companies, subject to the whims of the US government, have frontier AI runs a high risk of dystopian outcomes.” An AI future where only a few US companies, subject to
Gary Marcus clarifies a distinction between profit and revenue in a discussion, apologizing for previous lack of clarity on this economic/financial concept. The post likely addresses common confusion
Gary Marcus argues that poor performance in military applications correlates with or reflects fundamental weaknesses in AI capabilities and development. The post likely critiques limitations in curren
Gary Marcus posted a correction regarding a New York Times report about a company contemplating delays to their IPO plans. The post appears to address inaccuracies or clarifications needed in the orig
Fantastic interview: Ed Elson is talking with Gary Marcus about AI tech, overrelience on LLMs, the business of AI, etc. I totally enjoyed this: https://youtu.be/_pSivPlRx5o?si=dzrJKpUGTHE0Y76I @edels0
from tokenmaxxing to this in less than three months. anyone who doesn’t see this as affecting Anthropic and OpenAI is not paying attention. UBS says 60% of companies now watching AI budgets are moving
“Gary Marcus … says that investments into scaling AI is the “greatest capital misallocation in history” which leaves everyone “on the hook”. However, he also says that it is still early days for the A
Gary Marcus discusses fundamental challenges and limitations in current AI systems during an interview, likely covering topics such as the gap between narrow and general intelligence, reasoning limita
This post highlights an apparent error where Google's Gemini AI attributed a film review to Roger Ebert for a 2016 movie, despite Ebert's death in 2013, making such a review impossible. The post, shar
“If there is a deflation of the AI bubble, the optimists say that the new infrastructure will remain even if the companies do not — just as railways survived the 19th-century railway bust. However, th
In a matter of weeks, U.S. federal AI policy has gone from implausibly libertarian to increasingly draconian and opaque. Today, over 35 distinct observations, I analyze how we got here and offer the m
Gary Marcus discusses arguments surrounding trillion-dollar implications, likely related to AI development, regulation, or economic impacts given his expertise in artificial intelligence and cognitive
June has been a dark month for the US generative AI industry, between hamfisted US policy and the transition to cheaper Chinese AI models. And then there is the delay of the OpenAI IPO. Is that the co
Gary Marcus argues that $SPCX exhibits high volatility because it functions as a meme stock tied to a rocket company, with its price movements driven by sentiment and social media enthusiasm rather th
This post likely discusses how Orwellian tactics—such as propaganda, doublespeak, and manipulation of language—are being employed in contemporary contexts, possibly relating to AI, technology, or poli
People want jobs, not subsidies: Taxing corporations to fund programs is usually controversial and partisan - but when it comes to AI job displacement voters are shockingly left wing. It's genuinely e
people with advanced degrees who can’t distinguish between pure LLMs (which is what I critiqued in 2022) and LLMs enhanced with neurosymbolic techniques (which is what I championed in 2022) disappoint
Gary Marcus shares an actual REI advertisement as a follow-up to previous discussion about flawed bicycle examples, crediting Oren Etzioni for the reference. The post appears to contrast a real-world
“Scale cannot solve AI’s fundamental problem with accuracy” If my argument @financialtimes, excerpted below, is remotely correct, hyperscaling will prove to be among the biggest financial blunders in
This article argues that increasing computational scale alone cannot address fundamental accuracy and reliability problems in AI systems, suggesting that AI development requires solutions beyond raw c
SPCX - SPACEX BOND SELLOFF DEEPENS SpaceX's 25 billion bond sale is suffering unusually steep losses, with paper losses exceeding $300 million. Traders say fast-money investors may be exiting, while c
Gary Marcus comments on significant losses in $SPCX stock, noting that investors who purchased at 225 experienced approximately one-third losses within two weeks as the stock fell below 150. The post
Talked with a few folks inside of AI labs (OpenAI, Anthropic) about what they think of the future of software engineering. The “closer” to shipping production code engineers are, the less they believe
Things came undone this week: Starmer resigned, SpaceX tanked, the Iran deal stumbled, and a housing bill that passed the Senate is dying on Trump’s desk. Today’s conversation with @GaryMarcus is a mu
This is what’s causing Anthropic to aggressively beg for govt protection (see below). Customers are finding cheaper alternatives. Keeping employees requires continuing ultra-rich secondaries ($$) that
THIS. The reason the US Government seems out of its depth now is because people Andreessen and Sacks steered them wrong before and the USG got caught out, utterly unprepared. It seems both foolish and
We stress tested many frontier AI models for multimodal medical reasoning (including GPT-5, Claude 3.5, Gemini 2.5 Pro). They’re not ready. Faulty reasoning, use of inappropriate shortcuts, hallucinat
Gary Marcus fact-checks a claim that WeWork had no revenue, indicating this is a debunking of a false statement about the coworking company's financial performance. The post clarifies that WeWork did
“WeWork’s IPO fell apart. I’ve been calling OpenAI the (possible) WeWork of AI for a long time. I will not be surprised if their IPO plans apart, too” - @garymarcus, June 10, two weeks before OpenAI d
ah, the creative S-1 For those unfamiliar, WeWork pioneered the creative S-1 and the subsequent failure to IPO, and Sam seems to be following a comparable path. /cc @GaryMarcus https://qz.com/1685919/
An almost perfect prediction of the current AI predicament — and what may ultimately unravel our entire economy — from almost three years ago (August 2023), at Marcus on AI: Nearly every word remains
An important point which Gary will break down on tomorrow’s @ProfGMarkets episode Tune in Imagine if i had a business that gave you 20 bills for 10 bills. As word spread, my customer base would grow.
Gary Marcus discusses potential implications for generative AI if Michael Saylor (MicroStrategy CEO and prominent Bitcoin advocate) experiences financial or reputational difficulties. The post likely
Big Breaking News: White House asks OpenAI to delay GPT- 5.6. NEW: Trump admin asks OpenAI to stagger GPT-5.6 release over cyber concerns. Will approve 'access customer by customer during this preview
Chart of increasing revenue, pre the decline of token-maxxing, not looking at the immense costs/lack of profitability, not considering threats from open-source models, price-war dynamics etc. @noahopi
Crazy how many times people have told me over the last five years that a solution to hallucinations was right around the corner — and yet here we still are. This study tests how often LLMs invent answ
Disagree with @mcuban’s opening sentence – I do think some people genuinely hate the noise, blight, impact on electric prices, etc—but he is quite right that data centers have become a proxy for hate