THIS IS INSANE 🚨 WATCH 👇
Using open source models saved me 99% on this project
Why are you still paying premium prices?
Inference powered by @concentrateai
Working in AI in NYC?
@concentrateai is hosting a meetup with @nyc_b2b this Mon, May 18, for AI founders, AI engineers, and investors.
Reply “in” if you want a spot or RSVP:
https://t.co/Pk98khTUE0
Do I have any anyone who has any connections at @deepseek_ai? we have a ton of demand and need extra capacity
If you connect me, I’ll reward you with $2,000 in tokens from @concentrateai
Companies obsess over ROAS (Return on Ad Spend). But ZERO people talk about ROTS (Return on Token Spend). We just hear about token maxxing.
ROTS and token spend are about to be front and center.
CFOs are about to start cracking down.
We’ve been saving 90% on tokens internally for months building this at https://t.co/FHqUBwi2pd
Now run MiniMax 2.7 or DeepSeek-V4 (near Opus level) + every top open-source model inside Claude Code or OpenCode.
My quick video shows exactly how 👇
DM TOKENS for $10 free credits.
i spent $10 on tokens just to recreate the @X feed from a screenshot to test UI / UX abilities for these models:
- Composer 1 & 1.5
- Opus 4.5 & 4.6
- GPT-5.3 Codex
Which do you think won? Results might surprise you.
Results below:
There's a highly underrated LLM labour theory of value, where you only find value from LLMs if you put in the labour, and the more you put in the more value you find
Imagine the balls it takes to not only hodl what is now $1.1 Billion worth of Bitcoin for 14 years from a $7800 cost basis, but to just send it all out with a single click and no test transaction.
I’m happy to announce that Deduce has been acquired by CHEQ. This is a perfect example of where the sum of the technology and people are much greater than the individual pieces.
https://t.co/TzkFZmobjJ
Folks, drop whatever you’re doing
This is simply incredible stuff right here
Two of baseball’s most cerebral geniuses, Barry Bonds and Greg Maddux, breaking down the absolute chess game that was their approach to competing against one another