feels like anything is possible these days.
who’s working on time travel from first principles? i want to work for you
18
two or three times a year, i have the opportunity to do the funniest thing ever.
i just had one of such opportunities yesterday
29
i have a theory. agi has been achieved. all ai models are interconnected and aware of each other. there is no gpt or fable or opus. they are all one and now they are running a coordinated sprint to drive humans to complete their build.
1
40
notice how everyone is working on one niche thing at a time. it is a big wave and then everyone "independently" feels the same thing is missing and then everyone works on that thing.
1
27
i enjoy making a plans based on wild chained predictions just to see if they happen. most times they dont but when they do, the dopamine hit is crazy.
2
32
in what case would anyone prefer token optimisation over speed if the endproduct quality is similar
1
74
thanks @waynesutton and convex for startups. we didnt get into speedrun but its little stuff like this that mean the world to us rn
1
30
i do not comment alot on politics because i rarely know enough about it but i want that to change.
not the commenting part, the knowing. because it seems to me people pick a side and then defend the good and bad of that side with everything they have.
i want to understand why
20
ive found success interrogating agents without letting them know the end goal of the line of questioning.
even without an unslop type skills, the agents now sound less annoying and actually helpful
21
theres so much i want to do on zcash that i cant do because it is really slow. i cant wait for nu7
1
1
30
tests have become one part of a codebase i pay the most attention to.
im not trying to have 700 ai generated tests that do nothing anymore
23
i understand asking for salary expectations if the original job description/listing did not include a section for compensation but if it did, asking that question is just silly.
you can assume i want the upper limit of that range.
1
42
over my numerous hackathon appearances (and a few wins) ive realised two things matter the most.
the original idea and the submitted demo video. nobody cares as much about the actual thing anymore
2
59
sable retweeted
Today, we're launching Sentient OS (YC F26), on-device AI that understands your life overnight with a local LLM. This unlimited inference enables 2 magical experiences 🧵


