macca

materials science @ oxford

[macca @mcau1iffe]

@KartikeyStack what’s your goal? transition to research?

[macca @mcau1iffe]

@kalomaze I watched this and thought it was amazing, but I still don’t understand how he trained it?

[macca @mcau1iffe]

@gabriel1 it’s correlation not causation, mine shows that 20+ min of screen time before bed increases recovery, when we know the science shows adverse effects

[macca @mcau1iffe]

@0xSigil post the brain teaser

[macca @mcau1iffe]

@niklasrinvar @MarcusMilione Could be very possible, but also don’t get me wrong, it’s very difficult to actually overtrain.

[macca @mcau1iffe]

@niklasrinvar @MarcusMilione None at all, overtraining just made it seem like i had one.

[macca @mcau1iffe]

@zachpogrob It is a truly terrible experience.

[macca @mcau1iffe]

@chefsteve fantastic reference👏

[macca @mcau1iffe]

@curiouscamilo I've reached out: oxford uni student + former dev at a 7 fig + app studio getting interested in prediction markets - let's see if we can work together :)

[macca @mcau1iffe]

@beechinour gta

[macca @mcau1iffe]

@sekachov @askcomputerinc this is so sexy

[macca @mcau1iffe]

@_opencv_ they all got zucced

[macca @mcau1iffe]

@mazeincoding they're probably quantizing kv cache to unusable levels instead of holding at 16 bits to retain quality

[macca @mcau1iffe]

@sama just bring out gpt-5 broski

[macca @mcau1iffe]

@mualphaxi

[macca @mcau1iffe]

@signulll this feels like if the final episode of Silicon Valley played out on twitter

[macca @mcau1iffe]

i'm excited for what Meta AI will put out in the next few months given that they have spent 9 figs on talent

[macca @mcau1iffe]

@jefielding @herbiebradley That's exactly what he said - you don't hear those stories because people don't tell them, they are too busy living their life. The people that start other companies are the ones who shout loud about it.

[macca @mcau1iffe]

if a claude code equivalent for grok comes out best believe i am giving them all of my money