Gary Marcus

Rank 14 of 47
|
Score 83
5/18/2025, 12:55:15 PM
X
In reply to:
Brain4brain
@NotBrain4brain
路
246d
@toncijukic @GaryMarcus @SteveStricklan6
Ton膷i Juki膰
@toncijukic
路
246d
@GaryMarcus @NotBrain4brain @SteveStricklan6 Other images are so out of distribution that it can never happen. 馃槄
Gary Marcus
@GaryMarcus
路
247d
@NotBrain4brain @SteveStricklan6 Is that an excuse? I bet with the right prompt you can actually get it, so you might argue it isn鈥檛 properly prompting imagen :)
Brain4brain
@NotBrain4brain
路
247d
@SteveStricklan6 @GaryMarcus That's because it's doing tool calling to imagen, a separate tool, the model isn't doing the generation
Steve Strickland
@SteveStricklan6
路
247d
@GaryMarcus Was hoping Gemini 2.5 Pro would get this right - but nope. It鈥檚 Google Deep Mind鈥檚 flagship model trained on vast quantities of data and compute. But it still gets tripped up by this most basic of LLM gotchas. Generative AI is inherently unreliable technology.
Gary Marcus
@GaryMarcus
路
247d
So, let me get this straight, a bunch of LLMs that don鈥檛 understand clocks and time zones (to say nothing of chess boards and the rules of chess) are going to kill us all before the end of the decade?

I am all for studying AI safety but can we please decouple it from ludicrous

The conversation revolves around the capabilities and limitations of AI models in understanding and generating images related to time and clocks. It critiques the reliability of generative AI, highlighting its struggles with basic tasks. The tone is lighthearted and somewhat humorous, with a focus on the technical aspects of AI.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does no harm and is focused on technical critique.聽[+1]
  2. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The conversation promotes understanding of AI limitations, though in a humorous way.聽[+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.
    Engages in constructive criticism of AI technology without personal attacks.聽[+1]