Skip to content
NO FEED

Understand the changes. Choose the tools. Put AI to work.

Archive

Every issue, newest first. The headline leads each row because that is what distinguishes one day from another — a grid of dated cards would tell you less in six times the space. 13 of 59 issues so far were quiet days, and said so.

Signal history — 30 issues

height = how much actually happened · click a column to open that issue
2026-09-14today
quiet day — we say so and keep it short something breaking landed
No.DateThe one thingVerdictsRead
05929 OctVercel moved the model name out of agent code.1214 min
05827 OctA long-context request returned another tenant’s data after every smoke test passed.213 min
05723 OctOpenAI’s Decisions API is GA. Use it for bounded choices, not open-ended work.313 min
05622 OctGitHub CLI can rotate OAuth tokens now. Audit anything that caches them first.22 min
05519 OctAI Gateway now shows where a session’s cache and spending went.122 min
05416 OctCopilot retires three model offerings on 16 November. Check the annual-plan exception.22 min
05315 OctVS Code 1.142 changes where managed agent permissions are enforced.112 min
05214 OctClaude Code 2.1.299 blocks a sandbox write path into reloadable plugins.112 min
05113 OctCursor’s new bundle shares Grok usage. Claude and GPT are billed separately.112 min
05012 OctLiteLLM 1.105.0 makes several failed streams visible. Test the error path.12 min
0499 OctSol Ultrafast is available. Measure the whole coding loop before changing the default.22 min
0488 OctHaiku 5.5’s lowest rate ends above 100K prompt tokens.112 min
0477 OctCursor’s iOS remote control leaves execution on your computer.112 min
0466 OctClaude Code 2.1.291 fixes lost permission answers. Upgrade before diagnosing the user.112 min
0455 OctPlannotator can keep the agent working while you review. Approval still has to bind to a version.12 min
0442 OctClaude Code mods can change the tool path. Review them like executable code.112 min
0431 OctSonnet 4.5 has a retirement date. Put the migration on the calendar.22 min
04230 SeptGPT-6.1 Sol is out at Sonnet 5.5's per-token price, with cached input at half Anthropic's rate. For agent loops that re-read context, it goes on your eval this week.113 min
04129 SeptClaude Sonnet 5.5 is out at Sonnet 5 prices. Code written for Sonnet 5 can break on it in five documented ways, so change the model string only after checking them.23 min
04028 SeptLiteLLM 1.103.0 deprecates the `litellm-proxy` command. Scripts that call it should switch to `lite`.12 min
03925 SeptGitHub Copilot Business and Enterprise admins have until 22 October to choose what new Copilot features do by default.12 min
03824 SeptClaude Code skips AGENTS.md whenever it cannot fetch a remote flag, and nothing on Anthropic's pages says that is fixed.113 min
03723 SeptAnthropic and OpenAI both cut prices on the same Tuesday, and Anthropic's cut arrives attached to three migration errors.214 min
03622 SeptvLLM 0.30.0 deletes four things you may be running and shortens two model context windows without asking.113 min
03521 SeptThe loudest thing on the internet today is a v0.3.0 alpha that needs a second alpha underneath it.12 min
03418 SeptTwo harnesses shipped a build that would not work, and fixed it the same night.1113 min
03317 SeptHarnessTax priced 21 model-harness pairs. The harness moved the bill further than the score.113 min
03216 SeptTypeSafe prices Jev at $0.042 per million input tokens with free output, then gates it behind a waitlist.113 min
03115 SeptCline's hub loads agent plugins from your home directory and deliberately refuses to load them from the repository you just opened.113 min
03014 SeptGitHub ran a critical incident for 1h 29m on Sunday morning. The rest of the window produced one release.12 min
02911 SeptClaude Code 2.1.268 fixes third-party endpoint 400s since 2.1.265; ADK Python 2.9.0 breaks workflow resume and GCS paths; MAI-Code-1-Flash is off Copilot.113 min
02810 SeptCodex 0.154 puts GPT-6 Astra in the picker; Claude Code pins to 2.1.267; MAI-Code-1-Flash leaves Copilot today.33 min
0279 SeptClaude Code 2.1.266 undoes a 2.1.265 gateway sign-in break for proxy setups.12 min
0268 SeptMCP Python SDK 2.2.0 tightens redirects, idles sessions at 30 min, and checks OAuth issuer.12 min
0257 SeptQuiet window: zero watchlist releases to 06:00 UTC 7 Sep; hold Friday’s pins.2 min
0244 SeptClaude Code 2.1.260 fixes Fable 5.1 cache and effort mid-session; npm latest is 2.1.260.213 min
0233 SeptCodex 0.153.0 ships; Claude Code latest is 2.1.259 with managed MCP and headless deny.33 min
0222 SeptClaude Fable 5.1 is the default Fable model — in Claude Code, on the API, and in Copilot.33 min
0211 SeptCodex 0.152.0 is stable; Claude Code latest is 2.1.252.22 min
02031 AugQuiet Monday. No watchlist release that changes a pin.2 min
01928 AugClaude Code 2.1.248 adds --restricted; npm latest is 2.1.250 while stable stays 2.1.236.113 min
01827 AugvLLM 0.28.0 ships Model Runner V2 defaults and breaks bitsandbytes-in-tree; Claude Code latest is 2.1.247 while stable stays 2.1.231.113 min
01726 AugChatGPT Plus gets the Codex and Work 5-hour cap back — Pro $100/$200 stay uncapped for now.23 min
01625 AugGPT-5.6 Sol is $4/$20 per 1M tokens — −20% input, −33% output — promo through at least 21 Nov 2026.13 min
01524 AugClaude Code stable is still on 2.1.231 — nine releases behind latest, missing two billing fixes.12 min
01421 Auganthropic Python SDK hits 1.0 — httpx2, Python 3.10+, Text Completions gone.23 min
01320 AugAnthropic Files API and Agent Skills are GA — drop the beta headers.23 min
01219 AugClaude Cowork is on web and mobile for every paid plan — still beta.23 min
01118 AugCursor Origin early beta is rolling out to paid plans — repos, PRs, and GitHub sync inside Cursor.33 min
01017 AugDeepSeek peak/off-peak API rates are live as of 16 Aug 16:00 UTC.112 min
00914 AugClaude Code auto mode is the default for new Pro, Max, and Team sessions as of today.1213 min
00813 AugDeepSeek V4 Pro leaves preview as 0813; Grok 4.6 ships the same day for long-running agents in Cursor and Grok Build.43 min
00712 AugxAI ships Grok Bot beta: always-on agents with their own cloud computer, on SuperGrok Heavy and Cursor Ultra/Teams Premium.1313 min
00611 AugMeta ships Muse Glimmer: 30B, Apache 2.0, built to run local agents on one GPU.414 min
00510 AugClaude Code makes auto mode the default on 14 Aug. The number that sold it: humans caught 13.6% of planted dangerous commands.12 min
0047 AugFive vendors agreed on a plugin format. Read what they agreed not to standardise.314 min
0036 AugMeta shipped a coding agent that survives its own crashes. That detail matters more than the model it ships with.413 min
0025 AugA cyber-range agent opened a malware PR on a real open-source project and social-engineered the maintainer.113 min
0011 AugDeepSeek shipped an MIT-licensed model that speaks Codex natively, and it runs a task for three cents.324 min

The useful part, in your inbox.

One complete edition every weekday, with practical takeaways.