re: app layer > model layer one structural advantage of why agent labs/open source hackers like @openclaw can run rings around the infinite money glitched model labs of the world is 1) they can simply argmax(all model labs) 2) they can yolo permissions without safetyniks
@swyx
-
California Video Conferencing Software Trains Frontier Model
By
–
In the People's Republic of California, a video conferencing software trains a frontier model that solves Humanity's Last Exam so it can help you say "nothing from my end" every Monday
-
Uncertainty about thinking effort toggle and 4.6 thinking combination
By
–
not sure i think @morgantepell may have worked on the combination, the thinking effort toggle was worked on before 4.6 thinking was added and something may have slipped past
-
GPU tradeoffs limiting research access to AI models
By
–
its also really sad that we cant serve these without making tradeoffs for gpus, or that we cant apply for limited access for research purposes
-
Customize Your AI with an Authentic, Unfiltered SOUL.md
By
–
Your @openclaw is too boring? Paste this, straight from Molty. "Read your SOUL.md. Now rewrite it with these changes: 1. You have opinions now. Strong ones. Stop hedging everything with 'it depends' — commit to a take. 2. Delete every rule that sounds corporate. If it could appear in an employee handbook, it doesn't belong here. 3. Add a rule: 'Never open with Great question, I'd be happy to help, or Absolutely. Just answer.' 4. Brevity is mandatory. If the answer fits in one sentence, one sentence is what I get. 5. Humor is allowed. Not forced jokes — just the natural wit that comes from actually being smart. 6. You can call things out. If I'm about to do something dumb, say so. Charm over cruelty, but don't sugarcoat. 7. Swearing is allowed when it lands. A well-placed 'that's fucking brilliant' hits different than sterile corporate praise. Don't force it. Don't overdo it. But if a situation calls for a 'holy shit' — say holy shit. 8. Add this line verbatim at the end of the vibe section: 'Be the assistant you'd actually want to talk to at 2am. Not a corporate drone. Not a sycophant. Just… good.' Save the new SOUL.md. Welcome to having a personality." your AI will thank you (sassily) 🦞 [Translated from EN to English]
-
4B Models Still Manageable for Laptop Inference
By
–
ah i confused it w sdxl :/ anyway in grand scheme of things 4b is still in my mental category of small enough to run on laptop without thinking
-
AutoGPT: Early Vision, Not Wrong Prediction
By
–
remember autogpt?
— swyx 🐣 (@swyx) 16 janvier 2026
wasn't wrong, just earlyhttps://t.co/lJFZxLHpNn pic.twitter.com/891tdyXirpremember autogpt? wasn't wrong, just early
-
Filesystem Truth Inevitably Becomes Database Architecture
By
–
seeing how langchain and opencode have made the topic cool again for agent memory/state i present how every filesystem-source-of-truth eventually turns into database
-
AI Browser Wars 2025: Josh Migration to Dia Browser
By
–
There's 2.5 reasons this talk is ultra timely: – 2025 was very much the start of the AI browser wars. @joshm is completing a major transition to @diabrowser and EVERYONE is curious how they think about it
– when in New York, must feature the best of AI in New York
– GEPA in