Browsing: Gary Marcus
“draw an image representing a benchmark result that might have been bogus”As a scientist, OpenAI’s widely-watched o3 livestream, December 20th,…
Breaking news, not fiction: OpenAI finally released o3 and Tyler Cowen claimed it was AGITyler Cowen has become the ultimate…
”Those claiming we’re mere months away from AI agents replacing most programmers” should think again
“This isn’t the first time we’ve seen outcomes that suggest some of the ambitious ideas about AI agents directly replacing…
Deep learning is indeed finally hitting a wall, in the sense of reaching a point of diminishing results. That’s been…
Some brief but important updates that very much support the themes of this newsletter:“Model and data size scaling are over.”…
The USA Math Olympiad is an extremely challenging math competition for the top US high school students; the top scorers…
Hot take: GPT 4.5 is mostly a nothing burger. GPT 5 is still a fantasy.• Scaling data and compute is not…
Early reviews of GPT 4.5 have been absolutely brutal. Here’s a sample:The comments from Howard, a pioneer in large language…
My April’s Fools joke – that I had a sneak preview of GPT-5 —obviously went too far for some people…
Kevin Roose’s latest essay in The New York Times is, like a lot of essays, dreamy, offering a vision of…