🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

I recently returned to sarah' Consistently Candid podcast to discuss the most important AI developments in the ~9 months since my last appearance #1: (the obvious one) RL on LLMs works, and we'll soon see superhuman math & programming AIs, at least. The race is on!

some very real interest in my podcast from some very real potential guests

"How, exactly, could AI take over by 2027?" Introducing AI 2027: a deeply-researched scenario forecast I wrote alongside Scott Alexander, Eli Lifland, and Thomas Larsen

Takeways from AI 2027: a scenario focusing on superintelligence (ASI). (a) ASI could arrive soon (b) ASI will dictate humanity's future (c) We're not ready for ASI: misalignment and human power grabs are both threats. Race dynamics might be crippling for safety. 🧵

the other night I successfully AGI-pilled a man on a first date and then he ghosted me, what to make of this

Do scientists have a plan to make powerful AI safe? AI companies predict the arrival of AIs more intelligent than humans this decade. They admit these systems may cause catastrophe if uncontrolled. Some research agendas aim to prevent this, but none are silver bullets🧵

idk really what to do about AI anymore but I’m in Florence

relatedly, today I saw 5(!) policemen demand to check a busker’s documentation before he was allowed to resume his cover of coldplay’s Fix You

Bruh I think AI risk is literally real

Internal deployment was always the main threat model IMO. IIRC I tried to get the Preparedness Framework to cover internal deployment too, but was overruled. It's one of the ways in which this whole evals thing could end up being just safetywashing. (In the AI 2027 scenario, the

I am strongly in favour of making discussions of our dystopian AI future cosier for example, EAG fireside chats would benefit from actual fireplaces

AI is improving rapidly, but when will it get smarter than humans across the board? It depends who you ask. Here’s what the experts are saying 🧵

i'm getting insecure seeing all these posts about 4o being an extreme sycophant bc it isn't any nicer to me than claude like he must hate me

idk really what to do about AI anymore but it’s my birthday and I’m drinking out of this big paper cup