literally my whole career is about trying to get us to *not* lock in on asymbolic neural architectures.
@garymarcus
-
Evaluating LLM Reasoning: Rigorous Analysis Beyond Data Contamination
By
–
To be persuaded that LLM was reasoning I would want to see (a) an analysis that compared the output with training set in a more serious way than superficial examination of data contamination in the GPT-4 paper & (b) robustness across different formulations of test problems, such
-
GPT-4 Lacks Deep Comprehension Harvard Research Claims
By
–
Who wants to tell @Harvard that GPT-4 does not in fact offer deep comprehension and complex reasoning? Attn: @sapinker
-
Community Celebrates Doug Lenat’s Legacy in AI
By
–
So many people mobilizing and self-organizing around ways to celebrate Doug Lenat it brings tears to my eyes.
-
Autonomous Vehicle Safety: Premature Technology Rollout Risks
By
–
The saga of what happens when you roll out premature technology too quickly continues. Some will protest that human drivers make mistakes too, which is of course true, but given the tiny fleet size and the number of accidents, I doubt they are truly comparable.
-
Doug Lenat: Tenacious Pioneer in Artificial Intelligence, 1950-2023
By
–
A short essay in memory of the tenacious and brilliant Doug Lenat, 1950-2023 https://
open.substack.com/pub/garymarcus
/p/doug-lenat-1950-2023?utm_campaign=post&utm_medium=web
… -
Recent AI Research Paper from July 2023
By
–
His last paper, with a little help from me, July 31, 2023: https://
arxiv.org/abs/2308.04445 -
Scientists Scarce but Civil Society Well Represented at Event
By
–
Not a lot of scientists here but great to see at least a few outstanding representatives of civil society in the crowd — progress!
-
AI-Generated Sports Reports Criticized for Poor Quality and Lack Authenticity
By
–
“The reports were mocked on social media for being repetitive, lacking key details, using odd language and generally sounding like they’d been written by a computer with no actual knowledge of sports.” #enshittification
