Do some research (with Opus 4.8 :( )

comparing composer 2.5 and opus 4.8 with a small cursor plan #saasmakers

Benchmark bunch of AI models, since I've been using Opus 4.5 since a long time now, and wonder if most recent models are really that better #crisp

optimize 3D cache system to test Opus 4.8 #remix

My Cursor AI token usage resets every month on the 18th, and I'd only used 60%. So I had Opus 4.7 one shot a modular test suite for Kit and write tests for EVERYTHING. Unit tests. Integration tests. API tests. Opus banged it out in 30 minutes. And after an hour of tweaking, every test passed. My token usage after it was all complete, 65% 🤦‍♂️ Unbelievable.

testing cursor planned prompt with opus 4.7 to compare the results #thecompaniesapi #saasmakers

My #openclaw is much less interesting with a Gemini model that Opus was.

refactored all tests with opus, gemini and grok #thecompaniesapi

website audit for copy (B+), design (B-), and product (C+) by Opus 4.6 (using cursor) and ask it to make all A+ #paperstick

Monster #kit refactors everywhere. Used 50% of my Cursor Pro+ account on Opus 4.6 tokens (worth it)

I got a local LLM working in #openclaw. Wow is it a step down from Opus.

testing out opus 4.6

#productivity installing a hook to notify me with a sound or macos notification when claude code needs my input. looking into using opus - not clear on when claude uses what model.