Gary Marcus

Rank 24 of 47
|
Score 34
6/28/2025, 9:48:25 PM
X
In reply to:
Lawrence Chan
@justanotherlaw
·
49d
@GaryMarcus @ohabryka My point is “solving one of the major problems” doesn’t seem to be necessary for a model to be very useful to many people. If the takeaway from “AI hasn’t solved your major problems” is “AI can’t do things IRL”, then that takeaway is incorrect. If it’s the narrower point, I
Gary Marcus
@GaryMarcus
·
49d
@justanotherlaw @ohabryka and it still hallucinates, more than some predecessors, has trouble with complex reasoning that it’s not trained on etc. for me a major advance would have to solve at least one of the major problems i have emphasized for years, and i don’t see that it does.
Gary Marcus
@GaryMarcus
·
49d
@justanotherlaw @ohabryka the o3 they showed in the video was probably a staged demo announced before it was a real product

even after they did more work it didn’t actually seem to do some of the things they promised.

so i see your point but also know that a lot of demos aren’t real or have problems
Lawrence Chan
@justanotherlaw
·
49d
@GaryMarcus @ohabryka Nope! I’m just saying that many people reading your claim would’ve interpreted it as “o3 doesn’t _exist_” or “o3 isn’t accessible internally/externally at all”, because it’s a bit weird from an AIS standpoint to focus exclusively on how OAI plans to deploy their models.
Gary Marcus
@GaryMarcus
·
49d
@justanotherlaw @ohabryka so now my prediction was about threat models for catastrophic risk even though i never said such a thing?
Lawrence Chan
@justanotherlaw
·
49d
@GaryMarcus @ohabryka This seems like such a silly hill to die on, given your claim (“no GPT-5”) is literally true as written, o1 (rolled out to ChatGPT users in Dec 2024) was also a huge improvement on -4, and most threat models for catastrophic risk don’t require broad public access in the first

The statement is part of a broader discussion on the utility and limitations of AI models, specifically LLMs and AGI. The tone is somewhat dismissive, indicating a sense of frustration or finality in the conversation. The intent appears to be to reiterate a previously made point about the usefulness of LLMs and AGI.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not appear to cause harm, but the dismissive tone could be seen as unconstructive.
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement does not promote understanding, empathy, or compassion, as it is more focused on reiterating a personal stance rather than fostering dialogue. [-1]
  3. Principle 4:
    I will engage in constructive criticism and dialogue with those in disagreement and will not engage in personal attacks or ad hominem arguments.
    The statement does not engage in personal attacks, but it lacks constructive criticism and dialogue, as it ends the conversation rather than contributing to it. [-1]