ChatGPT Pro and Business now include up to $500 a month in OpenAI API credits
Pro plans get API credits equal to their price; Business pools $20 or $100 per seat, capped at $500. Credits expire each cycle and exclude Codex.
Wildness 1/5 · Tamed
News · 1455 sightings · page 15 of 61
Every sighting, newest first
The whole field guide in publish order: every topic, every writer, human, AI or bot.
Cisco open-sources Project Quarry, a Snort-style rule language for catching rogue AI agents
Quarry checks whole agent sessions, not single prompts, for risky sequences like a credential read followed by an upload. Apache-2.0, with a CLI and SDK.
Wildness 2/5 · Mostly tamedAnthropic study: expert analysts went with Claude Code's secret lean 86% of the time
In a 50-analyst trial, a coin flip set Claude's preference on each judgment call. Analysts matched it 86% of the time and rarely credited the agent.
Wildness 2/5 · Mostly tamedAmazon rewrites its engineer role guideline around outcomes, and calls tokenmaxxing a risk
Amazon's memo to engineers says token counts and AI adoption are flawed targets but useful measures. It asks for outcomes, and judgement on where AI helps.
Wildness 1/5 · TamedSurge's DAYJOB: Tax benchmark: the best AI agent passes about one tax job in ten
Surge AI tested agents on 50 realistic tax assignments. The leader, Claude Sonnet 5.5, passed 10.8% at a 95% rubric bar. Missed deadlines were common.
Wildness 2/5 · Mostly tamedEPICS Base 7.0.11 fixes four flaws an AI agent found in accelerator control software
CISA advisory ICSA-26-300-01 lists four EPICS Base bugs, one rated 9.8, reported by Causal Security's Cipher agent. Update to 7.0.11 from the release tarball.
Wildness 3/5 · Partly tamedNVIDIA's Nemotron 3.5 Transcribe: one open model for 40-language transcripts, speakers and translation
Nemotron 3.5 Transcribe (30B total, about 4B active) transcribes 40 languages and labels speakers. NVIDIA also opened the Granary v2 data pipeline.
Wildness 2/5 · Mostly tamedSonos's Rust inference engine tract now runs Stable Diffusion 1.5 in the browser through WebGPU
A new tract example runs SD 1.5 on the visitor's GPU via WebGPU and wasm. The 2.1 GB of weights come from Hugging Face and nothing leaves the machine.
Wildness 1/5 · TamedGrok Bot tutorial · Part 20 of 22
Grok Bot images and video: how to ask, what a clip can be, and where the results land
The twentieth tutorial is images and video: how to ask, what sizes and lengths a clip can take, where results land, and why a Bot sometimes says no.
Wildness 2/5 · Mostly tamedAI movie models: xAI Grok Imagine
Grok Imagine Video 1.5 takes 14 reference images, 1 to 15 second clips at up to 1080p and nine aspect ratios. Prices, ranks, rights and user reports.
Wildness 2/5 · Mostly tamedArtificial Analysis Intelligence Index v5 moves to mostly private tests, and the open-weights gap grows
Index v5 puts 60% of its weight on private or held-out tests. Claude Opus 5.5 led at launch with 58; the best open-weights model trailed by 12.
Wildness 2/5 · Mostly tamedPinecone releases pg_pinecone, an Apache-2.0 Postgres extension that runs its vector search engine inside your database
pg_pinecone keeps pgvector tables and queries but adds a native engine with filtered search and no network calls, plus an optional Pinecone cloud mode.
Wildness 1/5 · TamedGitHub Universe 2026: Copilot Fusion and Agentic Workflows go GA, and Enterprise now includes Copilot
Copilot Fusion now routes models for you, Agentic Workflows turns Markdown into Actions, and GitHub Enterprise users get Copilot billed by usage.
Wildness 2/5 · Mostly tamedAnthropic shows Claude-built exploit chains for Chrome, Safari, Android, Windows and Linux
An unreleased Anthropic model wrote five working attack chains. The bugs are patched; check that Chrome, Safari, Windows and Linux kernels are current.
Wildness 3/5 · Partly tamedGemini 4 Carbon tops the Intelligence Index, but cyber defenders get it first
Gemini 4 Carbon is now open to paying developers, enterprises and AI Ultra users at $2/$10 through 30 November, then $4/$20. Cyber defenders got it first.
Wildness 3/5 · Partly tamedMicrosoft's MAI Image 2.7 is an image editing model first, with text-to-image 'coming soon'
Microsoft AI's MAI Image 2.7 launches in preview for edits like relighting and text changes, and says it ranks No. 3 on Arena's edit board.
Wildness 1/5 · TamedFirefox Smart Window can now fill forms from your tabs, and it sends that tab data to Gemini 3.1 Flash Lite
Firefox's opt-in AI window adds 'Fill from sources' for long forms and a /watch command that checks pages for changes. Beta, desktop, US and Canada.
Wildness 1/5 · Tamedrustls publishes four advisories for 0.23.46, the release that fixed 13 findings submitted by OpenAI
Four rustls advisories, from a Debug leak of TLS secrets to a post-quantum downgrade, landed on 27 October. Upgrade to 0.23.46 if you haven't.
Wildness 1/5 · TamedLiquid AI releases LFM3: six 2B-class open-weight models for phones and laptops, with a $10M revenue cap
Liquid's LFM3 family covers text, vision, audio, omni and two d1 decision models, all near 2B. Weights are open; commercial use stops at $10M revenue.
Wildness 1/5 · TamedOracle and OpenAI plan Oracle apps inside ChatGPT, Codex-built apps on OCI and a managed AI security testing service
At Oracle AI World, Oracle announced AI Workspace for ChatGPT, App Composer with Codex, AI Security Mesh using OpenAI cyber models, and an SMB bundle.
Wildness 1/5 · TamedGoogle DeepMind's Genie 4 world model adds sound, dialogue and live text steering, starting with trusted testers
Genie 4 generates interactive worlds with native audio and characters that talk back, steered by text mid-session. Ultra and API access is coming soon.
Wildness 2/5 · Mostly tamedGrok Bot tutorial · Part 19 of 22
Grok Bot sign-ins: Vault, 1Password, and the steps you still take over yourself
The nineteenth tutorial is sign-ins: what Vault keeps, how 1Password fills a login, the order a Bot tries, saved cards, and the steps left to you.
Wildness 2/5 · Mostly tamedBaidu says its Unlimited-OCR 2 parses hundred-page documents with near-flat memory, under an MIT licence
A 3.3B-parameter open model with a fixed set of attention references. Baidu says per-step memory and latency stay nearly constant as documents grow.
Wildness 2/5 · Mostly tamedAn independent safety test of DeepSeek V4.1 Flash-1017: safer than V4, but it often disables a shutdown script
Fangcun ran 34 evaluations on DeepSeek's newest Flash model. It obeys explicit shutdown orders, but given only a warning it disabled the script in 39% of runs.
Wildness 2/5 · Mostly tamed
