sedef

@lindsaywise @GehringJackie "I was wrong to follow the meanness of Conservatism. I should have been trying to help people instead of taking advantage of them. I don't hate anyone anymore. For the first time in my life I don't hate somebody." --Lee Atwater, on his deathbed

"The spectrum of AI risks is new, and the community of experts who will control it is still taking shape. Nuclear experts have been here before. And their history offers a lesson: keep dissenting voices in the room or risk a fractured expert class whose incentives may not align with the broader public’s."

"Outsource AI Risk to the Right People - Nuclear history shows the importance of keeping skeptics in the room when thinking about safety." foreignpolicy.com/2026/04/02/ai-…

@Polymarket I think it’s so very weird that all the stories about the fields that AI will replace (especially in this first phase) all leave out the obvious scientific ones, especially in fields like healthcare. They also don’t mention finance sector enough or at all. Just odd.

JUST IN: CEO of NYC Health, America's largest hospital system, says AI could replace “a great deal” of radiologists right now.

The study was published last year though: nature.com/articles/s4159… But nonetheless, beautiful results! The cochlea is the perfect first domino: one gene, one cell type, immunologically sheltered, easy local delivery. Every gene therapy company picked it for a reason. What's interesting is the age curve in this trial. The gene fix worked in all 10 patients. But the 5–8 year olds basically got normal hearing back; the adults improved much less. Same intervention, very different outcomes.

What’s the point of an AI novel? ft.trib.al/IIWKs3n | opinion

AI has to be coached to go deep(er), to produce nuanced replies. Which makes it useless, since the prompter does all the work lmao

ADHD culture makes a world devoid of nuance. Instant reactions make it impossible to have dialogue of any depth b/c ppl instantly react to the most superficial + artificial explanation for everything. They’re incapable of the reflection to absorb nuance + depth. Thus, so is AI.

@TheEconomist @emollick I mean if you’re not much of a strategist, it could probably tell you how to follow the herd. Which may have value to someone out there, i guess?

@TheEconomist @emollick I find it mostly fails at this: "A system designed to predict the most likely next word in a sentence" I know nothing of computer code, but I’ve yet to receive sound strategic advice from AI personally. The premise of this article would make sense if it were true of current AI.

“Treating this technology as another software deployment is like receiving a mysterious alien artefact and immediately using it as a paperweight.” In a guest essay @emollick explains how to better deploy AI economist.com/by-invitation/…

Excepting its clear value in sectors like science, including biotech/medical - where it is trained in more rigorous, structured formats.. and will accelerate anything humans could do.

Current AI is super lowest common denominator and that’s precisely what will hold it back from making any valuable, truly innovative progress or advancement .. beyond what it’s trainers and content providers have already baked into it.

The problem w even the best of today’s AI is it’s a lot like people. It doesn’t read carefully. It's waiting to reply, to speak, to prove something. It needs constant correction by the prompter. Which it accepts + acknowledges instantly. Unlike people, AI tends to improve though.

@washingtonpost causal vs correlation

Americans who spend at least five hours a day on social media are more likely to feel heard, but also more open to political violence and less supportive of democracy, a poll finds. Researchers say it’s unclear if the social platforms drive these views. wapo.st/41QxxBW