herbertroitblat.substack.com/p/how-to-suc...
AI copilots are everywhere. They are used frequently by a few people and occasionally by others. AI would provide more value if it reduced the cognitive load of using it. Make it invisible and useful.

NSF wants to use proposals to help predict—and strategically direct funding to—scientific discoveries and technological advancements. With all due respect I cannot see how this could be worthwhile. news.uchicago.edu/story/nsf-aw...

The US Food and Drug Administration (FDA) is seeking comments on draft guidance on Considerations for the Use of Artificial Intelligence To Support Regulatory Decision-Making for Drug and Biological Products.
www.fda.gov/media/184830...

Super-intelligence? Not even high school intelligence.
Students show characteristics of intelligence that are absent from GenAI but needed for general intelligence. #AI #GenAI #LLM #AGI
herbertroitblat.substack.com/p/super-inte...

The central question in AI is whether computation is enough for intelligence? I think that it is. A secondary question is how do we get from language models to cognition models? I think we need different kinds of models. #AI #AGI

A stochastic parrot, an infinite monkey, and a parasite walk into a doctor’s office. The NY Times said: A.I. Chatbots Defeated Doctors at Diagnosing Illness. It should have read that iteratively developed prompts are more effective than incompetent ones. herbertroitblat.substack.com/p/a-stochast...

State of thought on #GenAI and the fallacy of scaling.
I argue that the idea that language models could spntaneously become cognitive models of intelligence through scaling is a massive self-deception. www.linkedin.com/pulse/state-...