AI researchers on X describe pursuing frontier model work through unemployment and uncertain pay, stating that understanding of optimization methods, architectures, and training objectives remains immature and often overstated by major providers

@samsja19 My past https://youtu.be/PhReQgyprLg?si=oHxHriTYA-xbc_7F

More videos in my alt account

@_arohan_ we did research when having no compute during free time or school

there's a few things which are really funny about this entire situation: - the concept was always true despite the ai gold rush. some people are just experiencing it newly now. - people who maximize research fun and love of the craft often just end up maturing into the exact high net worth research roles

but it's still not easy for most people because most people have not found something that they really enjoy/they have baseline stability worries/have massive comparisonitis and tbh the way out of it has always been perseverance

@_arohan_ Jack Clark believes AI will be able to automate the entire AI research workflow by end of 2028. Might not be early for some

@_arohan_ > One way is to figure out what to work on is to work on things that matter and not think of rewards

reward hacking is one of the worst things you can do for your intellectual growth

@big_algocracy Even the neural networks wants to give us good life advice

@spicey_lemonade Its a good goal

@_arohan_ this feeling of just building stuff💪

@_arohan_ you can just enjoy things

@_arohan_ @aryaman2020 So well said. One also needs to genuinely enjoy working in a field (regardless of rewards above a certain threshold), and not continually chase hype trains.

@_arohan_ we did research when we were fun-employed :))

@_arohan_ @CoreAutoAI Haha I was doing my PhD hoping to join google brain. Wish the good old days still existed.

@_arohan_ > Big token wants to convince you otherwise

hmm, this is based.

@_arohan_ Big token 🤌

@_arohan_ @aryaman2020 Did u just coin the term…. “big token”?

@_arohan_ I am gonna steal the Big token phrase.

@_arohan_ I’ve met very few people that optimized for wealth and actually got wealthy. Likely chance is if you do something that you are passionate about you will get rich

@GauravML @_arohan_ @aryaman2020 Also maybe I am living in a stone but Anthropic and OpenAI aren’t even public yet right? How can people make 20M+ there? Is it largely from payouts/estimation based on valuation?

@_arohan_ big token, mannnn….

@_arohan_ This only works when you don't have responsibilities