SuperQwen3.6 35b might be better than sonnet 4.6 🤔

It kind of feels like og crypto recently

Working on getting cleaner benchmark data atm but TLDR is GPT 5.4 > GLM 5.1 > Opus in bedrock and another harness for $20M per token > Qwen 3.6 35B > gemini 3.1 > Minimax > grok > > > > opus in CC w/ sub wouldn't surprise me if qwen models overtake glm next few weeks locally

sucks another anon account was actually just a handsome chad. i guess i'm really the only fat retard on twitter

i'll give the ai doomers credit they'll at least have a conversation with you even if it is second year cs student just read yud takes. the ai haters/deniers/luddites will straight up refuse to use it and threaten to kill you at the extreme

I think nvidia will continue to grow but their margins will definitely be smaller with more volume in the future

with some dubious extrapolation qwen and glm will be better than mythos within a month or two anthropic is lying out their ass

BoC is going to become cool again and my pfp is gonna look lame and less esoteric

it's going to take a while for most people to get over their fear but I genuinely see no reason to hire a lawyer or accountant again unless i'm at trial/mediation

I hope people realize the codex automation stuff was all already possible with osascript and they gave it a neat gui and tool call

the more people I interact with on this site and in SF the more I realize I may be blessed with the perfect amount of autism to understand nuance and social incentives while still being able to obsess over pedantic bs. could just be life experience over zoomers though.

Jensen is dumb. What is electrical engineering and managing a $5T corporation before the crushing intellect of an eloquent podcaster. Thus spoke a F30U30 griftocrat…

the jensen-dwarkesh interview is this week's iq test mythos was last week's

how the fuck do people do sales without copious amounts of stimulants

tmux is all you need

recruiters are literally a phone menu to get to the hiring manager and make sure you have work authorization. 10 10 10 yes $180-200k base operator

the hardware availability in Austin is attrocious

I take all of my video calls as voice calls now and just tab out so I don't have to look at myself and obsess over facial expressions

meetings are secretly an excuse for everyone to not actually do anything most of the day while still looking busy. don't believe middle management's lies. everyone does everything 5 minutes before if there's a meeting scheduled.

I have a huge feeling qwen 3.6 27B when it comes out is going to be an absolutely massive wake up call for anthropic. People are already abandoning ship it's only a matter of time before big corporate customers start to realize they're getting rugpulled on inference