Michael Hajster retweetledi
further, researchers will just come to realize that the path to progress in their field goes through ASI now, that human progress today is slow and progress in AI is lightning fast
English
Michael Hajster
2.6K posts
Michael Hajster
@michaelhajster
deep reasoner
Latent Space Katılım Ocak 2023
2K Takip Edilen326 Takipçiler
Michael Hajster retweetledi
The technology is already largely here, with consistent trend over many years, disrupting consumer and industry workflows. Demonstrating leaps of progress each year. Building largest companies in the world in the blink of an eye.
And yet likely less than ten people in the world don’t think of it in too small terms.
English
Germany / Switzerland - who wants to be on this?
Vaibhav (VB) Srivastav@reach_vb
putting together a group chat for Codex power users in London / Europe who are the biggest ballers around?
English
explore every future your data makes possible before making it real
Aidan Clark@_aidan_clark_
You can experience the future for the small price of listening to one measly Super Bowl ad and downloading Codex!
English
Michael Hajster retweetledi
Laws of physics of companies are generally simple but specific for each industry.
The only way to get outsized returns is to find new or bend existing laws.
It is almost universally achieved through technological innovation.
English
Michael Hajster retweetledi
we‘re headed towards Jarvis that can do everything on your computer, listens all the time for your command, with complete fusion of speaking back to you, typing a response, or silently taking action... maybe even video?
English
@OpenAIDevs building tools for tokenbergung at scale ;)
English
Michael Hajster retweetledi
Michael Hajster retweetledi
why aren't more people studying self-compaction at artificially low context lengths. there's no reason you can't benchmaxx math RL with 4k tokens across many turns
English
Michael Hajster retweetledi
day long hackathons are kinda broken in the agent era. i propose the hour long hackathon, which feels much more appropriate
English
Michael Hajster retweetledi
Michael Hajster retweetledi
has anyone tried just asking 5.5 to think about what open problems it wants to solve for 20 hours?
i honestly think it might be the year of the harness. ridiculous abstract multi-agent prompt loops just kind of work now…
will depue@willdepue
@littmath i think you could build a harness today that could get 5.5 to do really unique curiosity driven research for what it’s worth! another idea: RL env where the model must conduct ~32 entirely novel explorations, present to an LLM judge, reinforce one deemed ‘most interesting’
English
Michael Hajster retweetledi
Michael Hajster retweetledi