Fresh stories
Study finds context compactors retain only 17% of standing agent rules
A University of Pennsylvania study found common compactors retained only 17% of standing session rules while preserving task information. Practitioners use verified Markdown handoffs and background compaction to manage long coding-agent runs.
GitHub users report outage disrupting commits and Actions
Users reported failures retrieving commits, running Actions, and syncing GitHub-hosted repositories. Some developers said the disruption blocked pull-request merges amid broader reliability complaints.
Study finds context compactors retain only 17% of standing agent rules
A University of Pennsylvania study found common compactors retained only 17% of standing session rules while preserving task information. Practitioners use verified Markdown handoffs and background compaction to manage long coding-agent runs.
Report: OpenAI reportedly commits to 8 GW of Ohio AI capacity
Reports say OpenAI agreed to secure roughly 8 GW of capacity at Ohio’s PORTS-Pike campus. The initial 800 MW is expected in 2028, with SB Energy funding grid and transmission upgrades.
Cursor opens Origin code hosting beta with GitHub mirroring
Cursor Origin adds Git repository hosting, GitHub mirroring, pull requests, reviews, and agents that can edit code and push branches. The service is rolling out in early beta for paid plans.
Top storiesthis week
Qwen releases Qwen3.8 27B multimodal model under Apache 2.0
Qwen released its open-weight Qwen3.8 27B vision-language model with 262K native context and adjustable reasoning. In a 484-sample test, enabled-thinking scores fell from above 92% through 64K to 74.3–81.8% at 128K.
Codex opens 1M-token GPT-5.6 Sol context for ChatGPT subscribers
Codex now lets ChatGPT subscribers enable a 1 million-token context window for GPT-5.6 Sol. Automatic compaction starts at 900,000 tokens, and tokens beyond the default window count double against limits.
OpenCode Go revises limits after DeepSeek price increase
OpenCode says it revised Go limits after DeepSeek raised prices. Its operator is testing hosting configurations intended to bring DeepSeek service closer to its prior price point.
Reports: GLM-5.3 post-training lifts Terminal-Bench from 4.6 to 28.3
Reports say GLM-5.3 retained GLM-5.2's base model while post-training raised Terminal-Bench from 4.6 to 28.3 and DeepSWE from 46.2 to 66.9. A technical account attributes the gains to RL infrastructure changes.
Anthropic plans Claude text-watermark detection API
Anthropic plans global Claude watermarking based on statistical word-choice patterns rather than hidden characters or user identifiers. A MIT-licensed removal repository reportedly reached roughly 11,000 GitHub stars shortly after the announcement.