@garymarcus

–

08 April 2026 19h13

correct: not publicly released, but see Project Glasswing for limited distribution

Pure LLMs Insufficient for AGI: Reasoning and Hallucination Challenges

By

–

08 April 2026 18h45

if the guy actually read my work he would see that what i actually said was
– Pure LLMs alone would not get us to AGI (even if they improved in some respects)
– Pure LLMs would continue to struggle with hallucinations and reasoning (they have)
– We need to incorporate elements

OpenAI and X: Concerns Over AI Safety Restraint

By

–

08 April 2026 18h36

the biggest fear as noted is that openai and x (and perhaps others) would not show the same restraint.

International AI Agency and Treaty Advocacy Since 2023

By

–

08 April 2026 18h30

i have been advocating an international agency and treary around this for years, see my 2023 TED talk and article in the economist.

Necessary but insufficient: AI’s next decade outlook

By

–

08 April 2026 17h48

it is necessary but not sufficient; see my article “next decade in AI”

Do Guns Cause Effects Equate to General Intelligence?

By

–

08 April 2026 17h47

guns cause reliable effects on a general basis. does that make general intelligence? sending you off to mute

AI Hype Cycle: Initial Enthusiasm Meeting Reality Check

By

–

08 April 2026 17h46

or a psychological effect of initial enthusiasm and hype gradually facing reality?

False dichotomy about LLM capabilities and limitations

By

–

08 April 2026 17h43

literally fourteen minutes after my last explanation of why this is a false dichotomy 🤦‍♂️ Michael Foster (@realmfoster) so do LLM’s not work or are they too capable? Which is it Gary? — https://nitter.net/realmfoster/status/2041903282730283192#m

AI Companies Intentionally Degrade Models Between Releases

By

–

08 April 2026 17h39

1. If true (and it does fit with my perceptions FWIW), this is an amazing and incredibly damning graph

2. Can anyone find the source on which it is based? https://t.co/wn93WshtVL
— Gary Marcus (@GaryMarcus) 8 avril 2026

1. If true (and it does fit with my perceptions FWIW), this is an amazing and incredibly damning graph 2. Can anyone find the source on which it is based? Marcin Krzyzanowski (@krzyzanowskim) "Anthropic, OpenAl and Google release their new models with high quality from day one then slowly nerf them until the next model, so when the next model hits, its perceived as a bigger jump than it actually is" sounds right what's happening — https://nitter.net/krzyzanowskim/status/2041793699223322657#m