Tech News

JC

Top comment: @julien_c “@etnshow”

FB

Top comment: @FilipBaturan “@thsottiaux Hey Tibo, is it possible to set up a hybrid approach in Codex where we use GPT models as architects and open source models as executors? I'm trying to get it working, but it seems Codex expects the same tool-calling protocol from other models, which they don’t have.”

AS

Top comment: @alexstamos “@DavidSacks So what is your position? Should Fable remain banned for an unwritten standard that no other model is held to and for capabilities that are available from Chinese open-weight models? Should the USG kneecap a US champion the week GLM 5.2 pulls even with Opus?”

GR

Top comment: @rauchg “@kumard_3 Yes. In a very tangible way I can describe in this tweet. Deploying eve basically gives you an API backend. And then you can query it with a 𝚞𝚜𝚎𝙴𝚟𝚎𝙰𝚐𝚎𝚗𝚝() hook we produce, for Next, Vite, TanStack to consume. https://eve.dev/docs/guides/frontend/overview”

MB

Top comment: @bakkermichiel “Scenario: https://europe2031.ai Summary: https://europe2031.ai/summary Audio: https://europe2031.ai/audio/ The ⭐⭐⭐ team: Daan Juijn, Stan van Baarsen @DadaJudith Lily Stelling, @philip_fox_ @AlexTPet @TomChivers”

KA

Top comment: @yacineMTB “@suchenzang humour, or deep resentment?”

T(

Top comment: @teortaxesTex “@zephyr_z9 I am not very impressed by the paper granted, this is is "dog playing chess" but I'm sure Deli himself would design better experiments”

GM

Top comment: @gavinmckew “@garrytan One of my biggest annoyances in tech, mention the cost of everything it replaces but not yours.”

DE

Top comment: @DanClayEllis “@Jason @JoshuaBaer I just read this unfortunate news - a tragic loss for the Texas innovation ecosystem. Condolences to Joshua's family and friends on this loss.”

GB

Top comment: @GavinSBaker “Original post:”

HA

Top comment: @hallerite “It's the same principle behind ECHO (SFTing on tool-call outputs), if you understand RL in an information theoretic way. We simply need to get as much out of every rollout as possible. Naive GRPO is unbelievably wasteful.”

GK

Top comment: @thegautamkamath “@johnschulman2 I know people would be interested to see the reviews if you still have them!”

MB

Top comment: @mohitban47 “@peterbhase Congrats again @peterbhase on the new Program Lead role at Schmidt Sciences -- proud of you for creating these very timely and important grant opportunities for the wider academic/research community!! #ProudAdvisor 🤗”

CH

Top comment: @kimmonismus “https://artificialanalysis.ai/models/glm-5-2”

MR

Top comment: @MRashadnow “@natolambert People will read an experiment from an ai lab and start announcing the death of stuff”

RO

Top comment: @krishnanrohit “@MattMoynihan1 10x what she does today, at least”

CH

Top comment: @kimmonismus “3/ First alpha systems up before end of the year. Full production 2027. If Tensordyne indeed delivers 1000+ T/s on 1- or 2-Trillion param models at <10 USD/1MTs - they might actually make a big dent in the market and become a serious new vendor.”

N🍜

Top comment: @NeonNoodle22 “@growing_daniel when you can't copy Elon's homework anymore”

MA

Top comment: @MarcusSpillane “@ClementDelangue @huggingface Open datasets are table stakes now. Two years ago enterprise teams paid a lot for what's now a free download. The real moat is curation and lineage. Nobody's bottleneck is data anymore, it's trust in the data.”