Gabriel

Studying mathematics @QMUL.

[Gabriel @Gabik_B]

@suhailrazzak @JustJake Not sure you can use Claude on new papers efficiently to synthesise new breakthroughs. Only possible in formalised fields like areas of maths that covered by LEAN

[Gabriel @Gabik_B]

@DurovPD Hey 👋

[Gabriel @Gabik_B]

@shadcn FOMO marketing , been out there for centuries

[Gabriel @Gabik_B]

@GlibKryz Both explore legacy

[Gabriel @Gabik_B]

@InvestingAddict Industrial Enslopification of the web from local searches to X

[Gabriel @Gabik_B]

@sota_web3 @superteam Congrats 🙌
Following the development 👀

[Andy @prompt_Tunes]

👀

[Gabriel @Gabik_B]

@Ohjayy Hello there from Hackney

[Pedro Domingos @pmddomingos]

If you figure out how a large multi-agent system can autonomously coordinate to maximum effect, you’ll win a Nobel Prize, a Turing Award and a trillion-dollar fortune all in one.

[Gabriel @Gabik_B]

@SuccinctLabs This gotta unlock tons of web3 use cases, ZCAM!

[Gabriel @Gabik_B]

@beffjezos Yep, can eat compute, drink compute, live in compute and commute on compute. Such a great idea, need to push it further!

[Gabriel @Gabik_B]

@WhereHasThisCol @leecronin Problem is LLMs do not support logic, they support only correlation

[Gabriel @Gabik_B]

@christiangori96 If your product is great it markets itself

[ThePrimeagen @ThePrimeagen]

I don't think it is as big of a deal as people are making it.

Hype is annoying, anthropic please just IPO already.

[Gabriel @Gabik_B]

@CoinMarketCap Yep

[Gabriel @Gabik_B]

@xiaonweb Bug finding got an upgrade, programming is not solved until we have causation models

[hlib Zakrevskyi @Hlibzakrevskyi]

Anthropic on whether Mythos can bypass monitoring:

we asked the model to guess our weaknesses, it didn't mention them, so we think it probably can't exploit them.

From the formal risk assessment, §5.5.1.1.

[Gabriel @Gabik_B]

@MacroMate8 For yield to be high it needs to come from somewhere. Maybe hyper liquid native lending would offer decent rates

[Gabriel @Gabik_B]

@0xblacklight I feel like this is absurd in age of ADRs, Design Docs and sprint cycles. Everyone who’s measuring at LoC/tokens spent is just trolling IMO.

[Gabriel @Gabik_B]

@tomhacks @vercel The 200 means it failed 200 times. Lots of failing