He's been writing variants of this kind of thing for decades and he's always been wrong.
He's been writing variants of this kind of thing for decades and he's always been wrong.
There's many other people but most of them are independent. It is glaringly obvious now most of the media outlets are afraid to ask the tough questions.
That said, MPU has been pretty solidly crazy with their LLM critiques lately (did you know it uses all the drinking water?!?!?). There are plenty of sane, grounded-in-reality critiques. Why not focus on those?
I browsed back looking for posts of his that most obviously made predictions, and "GPT-5… now arriving Gate 8, Gate 9, Gate 10" make a few very clear predictions and was absolutely correct [0] about them.
This was in June 2024, and Marcus claims two major predictions:
- "As of today, I am more confident than ever that GPT-5 won’t land this year."
- "Gary Marcus is still betting that GPT-5 will continue to hallucinate and make a bunch of wacky errors, whenever it finally drops."
It would be over a year from his posting that article that GPT-5 would finally land, and his overall prediction that the result would be lack-luster was also spot on.
Again, I don't particularly care one way or another about Gary Marcus, but flat out dismissing his writing doesn't really hold up.
0. https://garymarcus.substack.com/p/gpt-5-now-arriving-gate-8-...
Please explain in as much detail as you can because that is a very bold statement