OPENAI
30 days · UTC
Synchronizing with global intelligence nodes...
Reports of OpenAI Codex instability and silent GPT-5.5→5.4 downgrades
OpenAI Codex is showing regressions and silent model downgrades that can break automated code workflows. Developers report GPT-5.5 xhigh requests bei...
Z.ai open-sources GLM-5.2: 1M‑context coding model built for long runs, with cheaper long‑context compute
Z.ai released GLM-5.2, an MIT-licensed 1M-context open-weight coding model aimed at long-horizon, repo-scale engineering tasks. GLM-5.2 pairs a solid...
AI agents need real identities: AppViewX launches PKI-driven control plane as guardrail latency and shadow use bite
AppViewX launched a PKI-based identity and access layer for AI agents, pushing enterprises to treat agents like service accounts with least-privilege ...
OpenAI sunsets GPT-5.2/5.3-Codex in Codex; expect migrations and sporadic capacity pain
OpenAI removed GPT-5.2 and GPT-5.3-Codex from Codex for ChatGPT subscribers, pushing teams to swap models while capacity errors surface. Users report...
SpaceX to acquire Cursor; agent orchestration becomes the new battleground for your SDLC
SpaceX is buying Cursor, and vendors are converging on an “agent orchestration” layer your team will have to pick. IDC says SpaceX is acquiring Curso...
Gemini now speaks the OpenAI SDK — plan for a single client, many backends
Google’s Gemini Enterprise Agent Platform now works with the OpenAI SDK, making model swapping and multi-provider routing much easier. Per Google’s d...
ChatGPT 5.5 mode shift triggers real-world regressions; OpenAI SDK adds spend alerts
OpenAI shifted ChatGPT to 5.5 modes, retired older Codex models, and teams are seeing operational side effects. A practical explainer walks through h...
Local-first AI jumps to the browser: fewer servers, less lock-in
In-browser AI is crossing the line from demo to deploy, shifting workload off your backend and onto users’ devices. A hands-on build shows end-to-end...
Cheap intelligence is here. Build the harness.
LLM compute is getting cheap, but the bottleneck is the harness that turns it into permissioned, auditable decisions. A founder cut model spend 97% b...
OpenAI Responses API adds conversation state, simplifying multi-turn chat backends
OpenAI's Responses API now includes built-in conversation state that replaces thread-like handling for multi-turn chats. OpenAI’s new [conversation s...
Coding LLMs: leaderboard winners vs cost-per-fix reality
Leaderboards crown Claude Fable 5, but real repo runs show cheaper models can hit parity on fixes if you route smartly. The latest [LLM Reference](ht...
LangChain adds versioned tracing, tightens Anthropic file search, and fixes OpenAI streaming tool-calls
LangChain’s latest releases boost observability and harden provider behavior, with a few changes that may affect streaming and file-search integration...
APIs become the execution layer as agents emerge as the UI — and Microsoft’s marketplace is leaning in
SaaS consumption is shifting from dashboards to API-first systems driven by AI agents, with Microsoft’s marketplace pushing agents as sellable apps. ...
OpenAI’s Responses API quietly rewrites how you build long-running agents
OpenAI’s Responses API moves agent workflows to server-side sessions and webhooks, cutting glue code and changing how you run long tasks. A developer...
MCP is making databases first-class tools for LLM agents
Vendors are shipping official MCP servers for major databases, making agentic, schema-aware DB access viable without custom glue. An [InfoWorld round...
Agents got safer: Antigravity patches skill security; WindsurfAPI fixes WebFetch completion semantics
Antigravity hardened its popular skill library against path traversal and unsafe plugin behaviors, and related agent stacks tightened up tool-call han...
Agents are now a real workload: Uber’s bill and 1,000 deploys/month force ops, cost, and control rewrites
AI agents moved from experiments to production-scale work, exposing gaps in pipelines, cost controls, and access governance. Teams are hitting deploy...
Copilot switches to metered use and opens BYO models — you now own the cost curve
GitHub Copilot moved to usage-based billing and now lets enterprises route VS Code Chat to their own model endpoints. GitHub confirmed that all Copil...
AI agents are forcing a real trust and cost layer
Teams are running into agent reliability and cost spikes while vendors add partial governance features. Engineers reported GitHub Copilot Agent Mode ...
Agentic CLIs harden up: open, permissioned, and local-first
Agentic coding CLIs are quietly shifting toward safer, open, local-first workflows—and the ecosystem just shipped changes that make this concrete. A ...
Run research spikes with an "AI intern": real lessons from pairing with ChatGPT
A startup used ChatGPT as an "AI intern" to run a month-long data spike and scaffold a cross-branch query API. In a real experiment, a startup paired...
DeepSeek V4 Flash resets price/perf expectations; start routing on live pricing data
DeepSeek V4 Flash now delivers near GPT-4o quality at a fraction of the cost, and live pricing feeds make cost-aware routing practical. [Riley Kim’s ...
OpenAI Python SDK adds Bedrock support; AWS exposes OpenAI models via an OpenAI‑compatible endpoint
OpenAI’s Python SDK now supports Amazon Bedrock’s OpenAI-compatible Responses API, making it much easier to run OpenAI-style apps on Bedrock. openai/...