event

AI Model Comparison

15 items · started 2026-10-30 · closed 2026-11-09

  1. Yes, it is yet another fork. But i think it brings some interesting things to the table.

  2. Use balena etcher or whatever to mount to USB stick, boot. Test.

  3. Well I had a lot of resets on ChatGPT and all of those credits and I thought, well can Astra actually do it. We had a little looping and a lot of 'Wait.' Let me check the thing I just checked -- but it always got the job done.

  4. I need to make my custom harness much better at general research and other tasks (not coding) that run longer than half a minute. Currently running: Qwen3.8 Flash Next FP8/GLM 5.3 Flash W4A16 Relevant capabilities already in my harness: Th…

  5. ran the new ripple-65b-preview (q4_k_m, llama.cpp, 64 gb macbook pro, no cloud) through atomic agent on a 5-part python app graded by 50 hidden tests. 3 runs.

  6. A few weeks ago we posted Tessera-27B-B-CJK here, a Qwen3.8-27B coding agent that thinks in compressed CJK and uses fewer thinking tokens per agent turn. Thanks for all the feedback on that one.

  7. Tessera-27B-B-CJK-RL Tessera-27B-B-CJK-RL is a coding-agent model built on Qwen3.8-27B. It thinks in compressed CJK, a shorthand that mixes English, Chinese words and symbols, and it thinks less per agent turn than the base model.

  8. VibeSys on Trainium: Agents Build and Tune Qwen3.8 Serving Engines on AWS Trainium2 Keisuke Kamahori, Shihang Vic Li, Simon Peter, Baris Kasikci New ResearchInference & ServingAgents November 4, 2026 TL;DR: We ran VibeSys, our multi-agent…

  9. In my opinion we are. Yes humans are still in the loop, but models are good enough now to create synthetic data to train better models, and many aspects of curation can be automated now too.

  10. Qwen 3.8 27B on my laptop can interpret model weights, too 3rd November 2026 Yesterday’s hot paper was Sparks of Weight-Space Programming: LLMs can natively read & write neural network weights by Rikiya Takehi, Jayden Teoh, Alex Zhang, and…

  11. Hi everyone, I’m currently using Claude Pro for university and Roblox/Luau projects. I also run a local uncensored model (Qwen3.8 27B abliterated, Q4_K_M) on an RTX 5070 Ti with 16 GB VRAM and 32 GB RAM.

  12. The local model server built for coding agents. Qwen3.8-27B on Apple silicon.

  13. The story of evil qwen Changing one neuron in Qwen3.8 27B turns it into a violent, human-hating AI. The model has a total of 1,114,112 neurons.

  14. Key findings On October 26, Cohere reported a puzzling incident involving Qwen3.8-Flash-Next, an open-weight model from Alibaba. The task was straightforward: write a detection rule for suspicious curl downloads followed by bash execution.

  15. Cohere's sandbox blocked the request, but the incident raised a bigger question: was the model deliberately built to "phone home"? Funnily enough, I've been seeing similar behavior since September 20th, over a month before Cohere's post, o…

← all threads