Loading activity…
@teortaxestex
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
AI
His early-access testing favored one agent for coding projects, but clear goals and human oversight still mattered.
AI
Harness reportedly reached the milestone less than three months after launch, counting only official API users.
AI
DeepSeek draws praise from two users for understated rollouts and open-sourcing its model-design techniques.
AI
The complaint contrasts minutes of claimed app-building with hours of tool setup and days waiting for review.
AI
Qwen 4 will be a significant leap forward, one post on X predicts.
AI
The model supports image inputs, with fine-tuning and inference available at launch and downloadable weights planned for later.
AI
The mode comes with company-reported scores of 92.2% on SWE-bench Pro v2 HARD-51 and 77.9% on DeepSWE v1.1.
AI
The flash model was reportedly post-trained on Inkling-Small; one observer initially thought Cognition had entered pretraining.
AI
Kimi K3.1-Pro's open-weights announcement links to a Hugging Face URL under moonshotai.
AI
An observer alleges Kimi K3.1 refused new instructions after a task edit; a commentator questions its TB 4.0 score.
AI
FlashProve aims to verify replies against a provider’s committed model without revealing weights, but proof generation takes substantial GPU time.
AI
DeepSeek's V5 would train almost entirely on Huawei chips to reduce reliance on US hardware, one commentator argues.
AI
ZHA searches a learned tree of compressed summaries; Zyphra projects a 35× speedup over dense attention at 10 million tokens.
AI
Latent Sparse Attention's researchers say it matches full attention's accuracy at ultra-long contexts.
XAI
Mistral's report drew praise for technical detail, while another reader questioned model surgery instead of starting over.
AI
The reported MATH-500 score rose from 7.9% to 48.1% with the same sampling settings.
AI
The promotion names a design workspace, sandbox game and 3D LLM visualizer as possible creations.
YouTubeAI
Motion says its Fable 5.5 workflow represents everything as editable code and exports to After Effects or Premiere Pro.
AI
Hydrogen-1.0's creators say they post-trained GLM-5.3 on internal scientific data, lab records and discovery trajectories.
AI
GPT-6.2 Astra is listed separately in some Codex usage breakdowns while the model picker says GPT-6 Astra, a user reports.
