If only there was a company that brought the tool call error rate to 0
@nebiustf @togethercompute @DeepInfra @GroqInc @novita_labs @GoogleAIStudio
Testing Hermes-agent, will report back. Gotta say it’s really easy to install.
"But you’re missing out on faster-than-light network I/O"
Pro tip: don’t write an agent runtime in Rust
Don't use Python when you can use CLI tools
remember when people paid for compilers
remember when programming was free
Pro tip: don’t write an agent runtime in Rust
Believe it or not, it was a contrarian idea in 2023 (!)
Many of my mutus who were early believers (2023) in structured outputs found jobs in really cool companies and I am absolutely not surprised. Kings and queens.
Many of my mutus who were early believers (2023) in structured outputs found jobs in really cool companies and I am absolutely not surprised. Kings and queens.
Doing the Lord’s work
I didn't expect DeepSeek v4 PRO (not Flash) to run well on the Mac Studio M3 Ultra with 512GB of RAM. This is 2 bit quantized with the same DwarfStar recipe used for Flash. 433GB GGUF file. 130 t/s prefill, 13 t/s generation. Prefill in the video is low because small prompt.
Show more
Under the Volcano is criminally underrated. It is the most beautiful book about self-destruction ever written. Mezcal, heat pressing down, the main character drifting. I’ve carried it around for almost two decades, and just started re-reading it. Recommend.
Show more
Writing a rebuttal to the structured-outputs-degrade-quality papers. The issues are embarrassingly basic, but unfortunately the citations keep coming. So here we go.
If only there was a company that brought the tool call error rate to 0
@nebiustf @togethercompute @DeepInfra @GroqInc @novita_labs @GoogleAIStudio
Following @julien_c’s tweet I bought a MacBook Pro with 128B unified memory, and started running Qwen3.6 as my daily driver. It’s baffling.
Apple outsmarted everyone in this industry.
Starting a company in a garage is boring so we started @dottxtai in a French castle instead
