@zengucapital What do you think about MU earrings call today? They have to pump it right? They all will work together to keep this going as long as they can :(
@VictorTaelin Coding is not important to our species. Even if what you is true, this doesn’t solve anything humans actually need for progress. This is just a tool that is good at coding, not “AGI”. No other field finds this useful. Software engineering/coding does because of structured data.
@bcherny Bro you work there and you had “fable” for a while now. At some point you should have some integrity and conscience. You can’t keep lying and marketing for the sake of it. I guess you need to keep the hype up or else your stock won’t be worth
@jachiam0 Both will end up as SAAS companies providing tools. For the engineering, people should really see how inefficient, non-deterministic (random), and inaccurate LLMs are. No amount of “scaling” will fix that and token costs are only going to go up because of the core nature.
@fleetingbits Why can’t they build it around Claude or ChatGPT? Harvey and Lagora aren’t doing anything special here. I bet it wouldn’t cost them more than $50m to build this. Data privacy is not the inly issue here, they probably see huge cost saving as well.
@AndrewSchrbr@blader Can you direct to get a setup like this? What harness helps you run multiple models on a single code base? Also, how does reviewing code work in this case? Thanks!
@SuvanshSanjeev Shouldn’t AI be easy for anyone to code and SWEs are not needed. If that’s the case, why hire L6 or above (more expensive) who know how to code?
@jarredsumner@zeeg I know my comment is funny. Reality is, Anthropic is a propaganda machine and a good one. Dario might not have directly asked and I obviously don’t know what happened internally. But it’s not a coincidence that this perfectly aligns with the fundraise. Good luck at Anthropic!