wow. GenAI is definitely not living up to the hype, for most people.
@garymarcus
-
Oracle CEO Change Triggers Stock Decline and Sexism Concerns
By
–
Oracle is down more than 50% since two men (jointly) took over Safra Catz’s CEO job.
— Gary Marcus (@GaryMarcus) 10 avril 2026
Sexism is up 500%? 5000%? https://t.co/dRZoE9xtR9Oracle is down more than 50% since two men (jointly) took over Safra Catz’s CEO job. Sexism is up 500%? 5000%?
-
Zuckerberg’s Moltbook Acquisition at Market Peak
By
–
Mark “Metaverse” Zuckerberg totally bought Moltbook at the peak of the market
-
Mythos Not AGI: Tuned Model Not General Intelligence Advance
By
–
I rest my case: Mythos isn’t AGI. It’s not even better at biology than the last model. It’s tuned to particular things, not a giant advance towards general intelligence. Same as it ever was.
-
New Science of Intelligence Needed Beyond Code
By
–
that’s nonsense. we need new science (of intelligence/cog sci) not just code
-
Domain Verifiability Economics Impact AI Implementation
By
–
it’s much more than that. you can only do the trick if your domain is verifiable (like logic and coding) and the economics make it worth doing a lot of specific work for that domain. it’s not just about expertise, it’s about the nature (and economics) of the domain.
-
Claude’s coding benchmark performance analysis and evaluation
By
–
possibly on coding benchmarks, but see the graph in here: https://
open.substack.com/pub/garymarcus
/p/three-reasons-to-think-that-the-claude?r=8tdk6&utm_medium=ios
… -
Claude’s Due Diligence Failures: Gary Marcus Analysis
By
–
they shoulda diligenced, you shoulda, too. https://
open.substack.com/pub/garymarcus
/p/three-reasons-to-think-that-the-claude?r=8tdk6&utm_medium=ios
… -
Three Reasons to Think That Claude Has Limitations
By
–
See: https://
open.substack.com/pub/garymarcus
/p/three-reasons-to-think-that-the-claude?r=8tdk6&utm_medium=ios
…