crazy talk. don't read.
i lost. had a nightmare in 2019, something went extremely wrong that could have prevented. family blamed me, said you could have worked harder. the board said 2027/9. felt like a premonition. i am nowhere near i should be. chip is getting heavier everyday
i hate how inefficient the world actually is. we could become unimaginably better. vector addition of the cumulated efforts is so low to its total potential
thats exactly what i think of inference engineering. unless you are training your own models and serving them, it's pretty much useless to any startup. you will get the cheapest for the god making companies themselves and they will be optimised by the model itself.
the big discount on intelligence
https://t.co/o9mGqnLhsJ
“We stopped the individual engineering subscriptions to Claude and Codex on the engineers’ Macs, and routed all engineering work through the API. The team and backlog didn’t change, and we didn’t ask the engineers to work more.
Before this, five engineers collectively paid $800 to $1000 per month for Max and Pro plans.
After this, the bill was approximately $1000 per day for the same work. So in one fell swoop, the cost went up thirty times.”
it feels like that the world itself is nothing but a vibe coded project with a lot of different agents of different strengths and weakness making a future they think is going to happen and try to live up to that
@WhiskyToad true that, but always take prediction for next 3-5 years implying that ai would pretty much not do that. we are already meandering in superhuman territory with cybersec. and it's all personal software so all data could just be local and so no need for large database
first date, bill comes, i said i’m paying, she said yeah as you should. no thank you, no next one’s on me, just that. and now she wants a second date lol. nah, that was the whole date for me