@RichardSocher@Recursive_SI Congratulations on the release! Looks quite promising.
Would you say this would be useful in long context scenarios on something like an M1 Max 64gb?
OpenBot locks the model selector while tasks are queued. The only way I could change model was to delete the whole queue, and model choice should not be gated behind destroying queued work. Filed: https://t.co/UgFX3fMe6I @OpenBot_@norbertbodziony
@zachmoskow@concentrateai Agreed with this direction. Primary focus on the following would be great -
1) Flash moe models only, with system 1 models
2) minimum tps/tops - 100
3) maximum ttft - 10 ms
@superalesha There is something deeply untrustworthy, perhaps even conniving about this guy, Theo. I can't specifically point my finger to what it actually is, but I try to avoid him as much as I can.
Waiting for a native apps of @DeepSeekHarness for Mac OS, Linux, and Windows - the current electron MacOS app is a disgrace and extremely unreliable, truly.
@deepseek_ai
๐ DeepSeek-V4.1-Flash-1017 is here!
Stronger benchmark performance, faster task completion, and shorter outputs than the previous versionโa Pareto improvement in performance and cost. Now live on our website, app, and API.
๐ ๏ธ Optimized for @DeepSeekHarness, Tencent CodeBuddy, WorkBuddy, and more. Upgrade to DeepSeek Harness v0.2.1-rc.1 or later for the best experience with the new model.
1/4
@XunzhuoLiu Looking forward to the new models!
Big fan of Decision 3.0 + the free API - I've been testing the multimodal ones and would be happy to write a hands-on review if useful. Can share drafts or focus on whatever helps the team. No pressure!
TensorFold now has a "Beta" macOS app to launch the engine, so you don't need to jump into the CLI.
It's also under /app in the git repo if you want to drop any PR's against it.
https://t.co/KQLNskCovr
@superalesha There is something deeply untrustworthy, perhaps even conniving about this guy, Theo. I can't specifically point my finger to what it actually is, but I try to avoid him as much as I can.