Spinal Tap meets John Malkovich
What’s so funny about the need for world models?| Marcus on AI
Some things never change| Marcus on AI
Hard to say, but definitely not out of the question| garymarcus.substack.com
Promises about AI replacing employees often don’t work out — and may not be such a hot idea.| garymarcus.substack.com
Strangely enough, I have seen this movie before| garymarcus.substack.com
Post GPT-5, the narrative has flipped — and that could hurt them| Marcus on AI
Unpacking a popular new example from an AI researcher at Google DeepMind| Marcus on AI
Last year, big tech couldn’t stop talking about how AI “agents” would be the next big thing in 2025. It hasn’t quite turned out that way.| garymarcus.substack.com
Maybe OpenAI’s board had a point, after all| garymarcus.substack.com
A new release botched … and a breaking research new paper that spells trouble| garymarcus.substack.com
It’s here, finally, but not everything people dreamt it would be| garymarcus.substack.com
What we know, what we would like to know, and what it may take years to know| Marcus on AI
A strange new world, presented mostly without comment| Marcus on AI
What could happen if power, recklessness, incompetence, and indifference came together| garymarcus.substack.com
Neurosymbolic AI is quietly winning. Here’s what that means – and why it took so long| Marcus on AI
Coding has been the strongest use case. But a new study from METR just dropped.| Marcus on AI
A compassionate but skeptical letter that Douglas Hofstadter wrote to one of his readers| Marcus on AI
The downsides of generative AI are growing like wildfire| garymarcus.substack.com
Is AI so valuable to society that we should place this bet?| garymarcus.substack.com
LLM failures to reason, as documented in Apple’s Illusion of Thinking paper, are really only part of a much deeper problem| garymarcus.substack.com
The most popular rebuttal was … a joke. Literally.| garymarcus.substack.com
Breathless predictions about Artificial General Intelligence (AGI) are nothing new; they still aren’t true| garymarcus.substack.com
Also: another paper that seals the deal| garymarcus.substack.com
LLM “reasoning” is so cooked they turned my name into a verb| garymarcus.substack.com
Everyone’s been talking about it. JD Vance read it. What does it actually tell us?| garymarcus.substack.com
61% in a recent survey report no ROI or disappointing ROI.| garymarcus.substack.com
I hope the media will look into this| garymarcus.substack.com
Probably so. A response to Scott Alexander’s essay, “Somewhat Contra Marcus On AI Scaling”| garymarcus.substack.com
Nightmare on LLM Street| garymarcus.substack.com
Decades of government kowtowing to Big Tech has thus far failed to produce a decisive victory| garymarcus.substack.com
What Casey Newton’s new essay skewering AI skepticism does—and does not—get right| garymarcus.substack.com
Science, sociology, and the likely financial collapse of the Generative AI bubble| garymarcus.substack.com
Important new study from Apple| garymarcus.substack.com
An update from the person who first called the bubble| garymarcus.substack.com
What comes after chatbots?| garymarcus.substack.com
Even on little things, Sam is not consistently candid.| garymarcus.substack.com
The conventional wisdom, well captured recently by Ethan Mollick, is that LLMs are advancing exponentially. A few days ago, in very popular blog post, Mollick claimed that “the current best estimates of the rate of improvement in Large Language models show capabilities doubling every 5 to 14 months”:| garymarcus.substack.com
Some possible economic and geopolitical implications| garymarcus.substack.com
Shocking new paper with potentially serious implications| garymarcus.substack.com
And why it is unlikely to get much better anytime soon| garymarcus.substack.com
Monkey see, monkey confabulate| garymarcus.substack.com
Some thoughts on what it all means for AGI| garymarcus.substack.com
Copyright infringement issues could sink their business, but why should they have to pay licensing fees like everybody else?| garymarcus.substack.com
A full of spectrum of infringment| garymarcus.substack.com
The Errors They MakeWhy We Need to Document Them, and What We Have Decided to Do About it| garymarcus.substack.com
No, LaMDA is not sentient. Not even slightly.| garymarcus.substack.com
No, we probably won’t all die anytime soon, but there is still a lot to be worried about.| garymarcus.substack.com
Edge cases remain a serious, unsolved problem, a hundred billion dollars later.| garymarcus.substack.com
An essay on technology and policy, co-authored with Canadian Parliament Member Michelle Rempel Garner.| garymarcus.substack.com
And how did some prominent journalists utterly miss this initially?| garymarcus.substack.com
What comes after ChatGPT? 7 predictions for 2023| garymarcus.substack.com
New systems like chatGPT are enormously entertaining, and even mind-boggling, but also unreliable, and potentially dangerous.| garymarcus.substack.com
It's not just monkeys and typewriters. It's more interesting than that.| garymarcus.substack.com
Hint: It’s not all about scaling| garymarcus.substack.com
How MetaAI’s Galactica just jumped the shark| garymarcus.substack.com
