Claude Code burning 2x the tokens of Codex is the spiciest line in here haha.
MACHINE LEARNING
-
No speed difference, but Ollama offers nice quantized models and easy switching
By
–
It's been a few months but I didn't see a noticeable speed difference tbh. Otherwise, Ollama has usually a nice set of quantized models and easy to switch.
-
Decode 80 tok/sec at 1k, prefill 3000 tok/sec at 50k contexts
By
–
Yeah, this was at 50k contexts. Decode is about 80 tok/sec at 1k contexts. Prefill is up to 3000 tok/sec at
-
DeepSeek uses open-perfectblend dataset to train DSpark drafter
By
–
Fun surprise: DeepSeek used my open-perfectblend dataset to train their new DSpark drafter Time to promote it again! It's an open-source reproduction of "The Perfect Blend" paper. If you ever need >1M diverse prompts in math, chat, and code, it does the job.
-
Localize ads Recipe now available via Runway API for translation
By
–
Localize ads is now available as a Recipe via the Runway API.
— Runway (@runwayml) 27 juin 2026
You can now translate static ads and graphic assets via a single API call. https://t.co/T4b7oMPSfdLocalize ads is now available as a Recipe via the Runway API. You can now translate static ads and graphic assets via a single API call.
-
Change the base model of an LLM in one line
By
–
With any serious LLM library, it takes one line to change the core model.
-
@reach_vb — 2026-06-27
By
–
True story: I stopped thinking about context since GPT 5.3 Codex Single project focused threads with the recent capability of codex to spinoff new threads is goated! Codex continues and goes through compaction but remembers all the important stuff and if not, it’ll look up
-
GPT-5.6 ships in three capability tiers, Sol is flagship
By
–
@OpenAI just turned the frontier into a choice.
Instead of one model, GPT-5.6 ships as three capability tiers: Sol — the new flagship. Sets a state of the art on Terminal-Bench 2.1 (complex command-line, multi-step agent work) and is OpenAI's most capable model yet for -
Context-trained decision-making for robots is scoped, not open-ended
By
–
Context-trained decision-making for robots is a very different architecture than a general-purpose model. The clear context framing matters: it means the deployment is scoped, not open-ended.