devotd.
Recent developments, checked against their sources.
GitHub has made a feature generally available that uses AI to fix and recheck security alerts and submit pull requests for review. Eligible cloud customers need admin policies and billing configured, and usage consumes credits and automation runtime.
GitHub will retire three coding assistant models on November 16, 2026, and has recommended replacements. Organizations must check model access and integrations. Individual annual subscribers are exempt from one model's retirement.
goose 1.55.0 adds a provider that defines model service connections through configuration and passes request tracing information to tool calls. It caps automatic compaction's default ceiling at 225000 tokens and binds subagents to their parent's execution and approval rules. Background subagents have been removed.
GitHub has made an AI feature generally available to find vulnerabilities in languages and configuration files that existing analysis does not support. Results are advisory and do not block merges. Repository eligibility and credit usage need checking.
NVIDIA has released the full code for its platform managing shared GPU resources and AI workloads, allowing individual features to be adopted. Self-hosting requires checking GPU software compatibility. The community distribution does not include enterprise support.
OpenMOSS has released code and weights for image generation models that let users position text and objects with coordinates. It offers base and fine-tuned models. Adoption requires weighing layout input effort against memory and inference time.
Cloudflare and OpenAI say a pilot using network signals discovered hundreds of millions of new webpages and added them to a search index. Signals of changes and potential harm aim to reduce unnecessary crawling. Quantitative results on load reduction and actual business outcomes have not yet been provided.
A security breach affected 1.4% of all console users, specifically some users who connected external AI services. Notified users should revoke and replace keys in provider dashboards and review recent usage and connected applications.
Claude Code added settings to select the auto mode decision model and toggle idle conversation compaction. Alongside the previous version's unattended retry wait cap, this version rejects image sources larger than 128 MB.
Fish Audio previewed a speech synthesis model combining multiple voices, natural-language direction, and partial audio editing. Adoption requires checking GPU memory, latency under concurrency, removed training tools, and differing license labels.