As always Melanie Mitchell has is very insightful and precise, and decomposes difficult questions into pieces that can be analyzed. "Can Large Language Models Reason?".
@rodneyabrooks
-
Cruise Autonomous Vehicles Block Ambulance, Patient Dies
By
–
First report I have seen of this kind: two @cruise blocking egress for an ambulance contributing to death of patient. [I have personally seen many @cruise blocking roads and other drivers from moving.] This incident was on Aug 14th in SF.
-
Programming Efficiency: Learning as Alternative to Manual Labor
By
–
Page 455: programming: "about sixty workers, working steadily through the fifty years might accomplish the job, if nothing went into the waste-paper basket. Some more expeditious method seems desirable."-then suggests using learning to speed it up.
-
AI Engineers, Hype-Driven Funding, and the Need for Patience
By
–
"true bitter lesson"-nice allusion. I sometimes despair that some modern AI engineers were born on third base and they think the triple was just invented. But there are many serious people trying innovative research while having to navigate hype driven funding fads. Patience!
-
AI Vision Project MIT 66 and Intelligence without Reason review
By
–
3/3 AI Vision Project MIT 66. Or see my IJCAI91 long history review, "Intelligence without Reason". You can download some of these (but not all) at http://
people.csail.mit.edu/brooks/documen
ts.html
… -
Foundational AI and Cognitive Science Papers: Historical References and Key Works
By
–
2/3 All papers in 63 Computers&Thought, especially Minsky's overiew. Minsky 68 book Semantic Info Processing. Moravec 77. Lenat 77 (automated mathematician). Simon IJCAI79. Shakey project papers, 60's. Shannon Sci Am 1950. Rosenblatt 58 (see two lines at bottom cover below).
-
Historical Perspectives on AI: Turing, McCarthy, and Shannon
By
–
1/3 Read some history. "Machines Who Think", McCorduck 79 is a good overview. Turing '50 estimates 3,000 man-years of effort to get to human level, McCarthy&Shannon's 55 proposal for Dartmouth AI Conf explicitly says that computers can do it; also see their 55 "Automata Studies".
-
40 Years at Stanford: AGI Predictions That Never Materialized
By
–
Next week, 40 years since I joined the tenure track CS faculty at Stanford AI Lab; had finished my PhD there in 1981. 40 years??? We all thought we would have AGI by 1983. Did that happen? No. Has it happened yet? No. In 300 years we'll look back: What were they thinking?!?
-
Waymo and Cruise AV Crash Data San Francisco 2022-2023
By
–
AV crashes, reported by the companies, in SF from Jan 2022 to mid August 2023. 241 total. Waymo: 107, Cruise: 64, Zoox & others: 70. (FWIW I never see Zoox or others in the areas I usually traverse.) No fatalities.
-
Daniel Dennett Interview on Knowledge and Consciousness in NYT
By
–
A great interview with philosopher Daniel Dennett, friend of over 30 years, about what "knowledge" means, and how our souls are made from lots of tiny robots. And, as befitting Dan, it is the only mainstream media story I have seen with academic footnotes. https://
nytimes.com/interactive/20
23/08/27/magazine/daniel-dennett-interview.html
…