Loading activity…
@nrehiew_
eng primarily, ml mostly, research previously
AI
A user calculates the range assuming 97% cached input tokens and GPU rental costs of $3–$6 per hour.
AI
DeepSeek's README offers offline inference for DeepSeek-V4.1-Flash but lists additional systems needed for a complete service.
AI
PrimeIntellect says IPO is tailored for stability in asynchronous reinforcement learning.
AI
DeepSeek says the model is live on its website, app and API, and optimized for DeepSeekHarness and Tencent CodeBuddy.
AI
A user shared notes on Score Centering and predicted it would become part of the standard reinforcement-learning recipe.
AI
A user argues that trainer–sampler mismatch creates a drift term that can destabilize reinforcement-learning training.
AI
A user argues that publishing such blogs shows no respect for people interested in the company.
AI
A user flags HBM4 and the absence of super-long-context testing as caveats to Jalapeño–Blackwell comparisons.
AI
A ProximalHQ team member says they dynamically adjust the inference-trainer ratio to maintain training throughput and reduce staleness.
AI
CoreWeave says RL Rollouts updates live deployments during reinforcement-learning training without touching in-flight requests.
AI
OpenAI says it drew on an independent mathematics advisory group’s advice about how to release the results.
AI
The newcomer says they see open problems in data and post-training and invites people interested in large-scale infrastructure to reach out.
AI
Google announces a 1 million-token output limit, up from 64,000, for its new model. Initial access goes to trusted cyber defenders through the Fairwind Program.
AI
A user praises a team and congratulates Calvin Chen.
AI
Xiaomi says Pro performs on par with Claude Opus 5 and GPT-5.6 Sol across most agent benchmarks, with stronger coding, computer use and 3D reasoning capabilities.
AI
A user says the change came after they told Claude it could use Markdown and ASCII to explain things.
AI
The Wall Street Journal reports that an independent bug-hunting team accessed an OpenAI employee’s ChatGPT account, giving it a way to read and suggest changes to the company’s private software.
AI
Epoch AI labels nine of its initial 15 benchmarks “Flawed” and four “Verified,” with insufficient information to review two. A Terminal Bench contributor questions the audit’s rigor.
AI
In a September 16 update, the team behind MiMo-V2.6 said it was scaling compute, agent environments and grading, with plans to open-source details in stages.
AI
A user reports seeing medium/high reasoning settings mapped to 50 and argues that inference providers should follow the model’s intended mapping.
