Monsanto's Cruel, and Jev
The Doom demo really sold for me that this could be a complete game changer for local applications. Amazing work by the team! Looks like they've traded accuracy for speed and this is the puzzle piece my project needs!
If we could come up with some good plans, and store them as hints. This seems like quite a good idea? Is there a product for this? One application that sounds pretty interesting would be the litmus test. "Does not hallucinate" is not the same as "is never wrong". So the ATC test could be the benchmark.
Funny how it can do everything Jev can do (just structured outputs?) but Jev is highly optimized and purpose built for it and thus way faster and cheaper. Is that a fair description? This is amazing. I wonder if this could be a great tool for accelerating QA at my gamedev studio. Signed up for early access. My main trick is to use fish and enable vi key bindings. Especially on mac where there is a text stroke being applied :|. So is it a structured data-based language model? Or is there a tradeoff? Fewer ancient holes like this for their hackers but wide open access to anyone who installs codex or claude code?