We beat Opus 4.7 & Gemini 3.1 Pro on MMLU-Pro Would you ask your lawyer to check you for cancer? Sansa selects models for you in real time, so you get frontier intelligence aligned to your task. 👇

things move so fast these days by the time im used to a model's quirks a new one come out and everything changes

i say "almost done" at every stage. ten percent done, almost there. ninety percent done, almost there.

is anyone else's entire business plan just: "keep going and find out" things change so fast it's wild

startup idea: copy something that works

I have to do outreach on LinkedIn so I’ve started building a tool to make me more efficient

touché claude

the new x algo sucks every post is identical

Anthropic’s Fable 5 actively sabotages ML/AI work Without telling the user I’ll never use an Anthropic model again for any coding Trust destroyed

if you saw your own landing page on someone else's timeline, would you click it?

you know those guys who was their car even when it’s clean because they just love it so much? that’s me, but with my landing page anyone else?

when building your startup you’ll run into people like Raj trust your gut & ignore them

X said communities were done so I stoped posting or replying there I loved posting there, but oh well. a few months later… they are still in my feed feels like I have been bamboozled the day I decide to post there again will probably be the day the remove them

i'll code for 12 hours straight no problem but one cold dm takes me 45 minutes why are we all this way?

not impressed with Fable 5 I will stick with composer 2.5 we are at a point that the benefit is so marginal using these frontier models yet the cost is astronomical

anyone else have a unshakable feeling this is going to be a big week for your business? big milestones coming up!

in elementary school, we learned about the gold rush I remember thinking that those guys must have been quite dumb betting everything on finding gold that might not exist yet… here I am

YC's position in Mercor would be worth >$1B today if they hadn't rejected us in 2023. They passed because a previous recruiting startup didn't work out. These days, YC seems more focused on funding clones of fast-growing companies than making contrarian bets. What's the largest company YC has invested in over the last 3 years? The hit rate seems surprisingly low.

Made some changes I like this concept a lot more Do you think this is an improvement?

working remote vacations === working from pretty places

Introducing model routing with Sansa Sansa Router picks the right model for every task, automatically EXCEEDING frontier performance with a 90% reduction in cost