people with advanced degrees who can’t distinguish between pure LLMs (which is what I critiqued in 2022) and LLMs enhanced with neurosymbolic techniques (which is what I championed in 2022) disappoint me. i get that many tech bros don’t really get it, but scientists should take
RESEARCH
-
Neurosymbolic AI surpasses pure LLMs as predicted
By
–
1. pure LLMs did in fact hit a wall and then as i predicted (2022) neurosymbolic AI provided a way past the LLMs; it’s very evident if you look at all the tools, harnesses, loops etc in Codex, Claude Code etc as a scientist i would have expected you to be more attuned to that
-
Longevity escape velocity with only GPT 5.5 and Claude 4.8
By
–
Imagine we achieve longevity escape velocity, and all of us live forever, but the best AI models we’ll ever be allowed to use are GPT 5.5 and Claude 4.8.
-
Gary Marcus warns hyperscaling is a financial blunder for AI
By
–
“Scale cannot solve AI’s fundamental problem with accuracy” If my argument @financialtimes
, excerpted below, is remotely correct, hyperscaling will prove to be among the biggest financial blunders in history. We must seek alternative foundations for AI. -
Gary Marcus: AI scaling is capital misallocation, but early days
By
–
“Gary Marcus … says that investments into scaling AI is the “greatest capital misallocation in history” which leaves everyone “on the hook”. However, he also says that it is still early days for the AI industry and further exploration into different model architectures and the
-
Scale cannot solve AI’s fundamental accuracy problem
By
–
source: “How much compute does the world really need? Scale cannot solve AI’s fundamental problem with accuracy”
-
HumanEgo: robot learns skills from human egocentric video
By
–
Your robot could learn a new skill just by watching a few minutes of a human wearing smart glasses! University of Maryland presents HumanEgo: a framework that turns 30 minutes of human egocentric video into a zero-shot robot policy. Instead of needing robot data, it extracts
-
AI 5.5 fails deep report and forgets it
By
–
Yeah, 5.5 is pretty dumb. Asked to generate a deep research report and it spent way too little time on it. Asked it to tell me the word count of the report and it was like "what report? I only see your prompt!" wtf
-
Grok’s reports improve by learning from uploaded videos
By
–
Grok's reports are getting better about what it learned by watching the videos I upload:
-
Huge step for agentic AI: Qwen-AgentWorld teaches environment understanding
By
–
A huge step for agentic AI…teaching models to understand environments, not just act in them. Qwen-AgentWorld is a fascinating look at where general agents may be heading.