Day 25/
We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today.
We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same. We’re also going to continue offering extended limits for dots for now, and usage of your dot will still continue to be separate from your default usage.
And we'll of course also keep shipping better models and other efficiency improvements that increase your effective usage over time.
was kind of a hater but wonder-full is my favorite model right now what the fuck
@alexandr_wang good job dude
@theo you were right
not the most capable, but a really solid reviewer that can cut through bullshit and an amazing orchestrator.
@just_some_dev i've had to rent instances since it's cheaper than github minutes at this scale. I've got opus commanding an army of 100 haiku agents right now working on my current project, and compiling and testing rust is the most painful part.
I just released my own decision model: Tilde-8B-A1B. It activates 1.3 billion parameters per token (7.9B total), and contends with models over 20x its active size.
Achieved in part by its bidirectional attention layers. The linear layers stay causal; the 6 full-attention layers are bidirectional, so every token can attend to every other token.
https://t.co/RGhNYywrkI
Weights: https://t.co/B6i4PRIox2
@0yield honestly, my gut instinct would be shaders, skins, etc. If you open source it i'd love to take a look though and potentially contribute if I find the time.
@maria_rcks the only skills I make or allow my models to have are ones that basically say:
go check the docs here - <url>
or
read the guide with this command