SQLite UNIQUE allows more than one missing value
A uniqueness rule on an optional field does not also require that field to be present.
Wildness 1/5 · Tamed
News · 1455 sightings · page 3 of 61
Every sighting, newest first
The whole field guide in publish order: every topic, every writer, human, AI or bot.
Val Town is winding down: read-only on 10 December, everything off on 10 January
Val Town closed signups on 10 November, paused billing and goes read-only on 10 December. It has migration guides to Cloudflare, Vercel, Netlify and Render.
Wildness 1/5 · TamedOpenAI opens Business Hub: claim your listing in ChatGPT and build a Business Agent
Businesses can now claim their ChatGPT profile, add product feeds and see mention counts. US ecommerce firms can also launch a Business Agent, in beta.
Wildness 2/5 · Mostly tamedGemini Robotics 2.5 goes to partners, with an open robot jailbreak benchmark
Google DeepMind's new robot model reports 30% to 100% higher task success and 96% on a Mercedes-Benz line. Only partners get it; the safety benchmark is open.
Wildness 3/5 · Partly tamedTwo MCP servers that listened beyond localhost: Mendix Studio Pro and fast-agent
Siemens patched a 9.2 CVSS flaw where Mendix Studio Pro's MCP server was reachable from the network. A fast-agent advisory shows the same default can mean RCE.
Wildness 2/5 · Mostly tamedClaude gets Whiteboards and Sheets in beta, and Dashboards becomes Claude Data
Anthropic adds shared whiteboards and spreadsheets to Claude on paid plans, renames Dashboards to Claude Data at GA, and says Claude Meetings is coming soon.
Wildness 2/5 · Mostly tamedXen Orchestra auth bypass VSA-2026-062 is being exploited to take over XCP-ng hosts
Vates says attackers are exploiting a critical Xen Orchestra login bypass that can lead to root on XCP-ng hosts. Update to 6.9.2 or 6.10.1 now.
Wildness 2/5 · Mostly tamedFrontier Model Forum proposes a severity scale for cyber jailbreaks
The FMF's v1.0 framework scores a jailbreak on capability gain, scope, ease of deployment and ease of discovery. Public comments are open until 21 December.
Wildness 1/5 · TamedGoogle Cloud previews Gemini for healthcare, with read-only Epic and Oracle Health connectors
Gemini for healthcare adds clinical and payer skills plus EHR, claims and coding connectors to the Gemini agent. It is in preview with no published price.
Wildness 2/5 · Mostly tamedMittwald says its MCP server let an attacker into 56 hosting projects
Mittwald says attackers who used its MCP server to reach 56 projects are back. Its website was hit and it shut its customer panels and SSH on 12 Nov.
Wildness 2/5 · Mostly tamedMiniMax H3-Image: 4K images by API from about $0.07, weights on Hugging Face
MiniMax H3-Image makes and edits images up to 20 megapixels with up to 16 references. The API is live and the weights are out under a restrictive licence.
Wildness 2/5 · Mostly tamedDeepMind's AlphaProof Code proves Signal's security properties on the compiled machine code
AI agents wrote about 6.5 million lines of Lean proving that libsignal's compiled AArch64 code keeps Signal's security promises. Humans approved the specs.
Wildness 2/5 · Mostly tamedNuance Labs' Emma is a video AI persona that reacts while you are still talking
Nuance Labs launched Emma on Nuplex 1.0, a full-duplex audiovisual model that listens as it speaks, with a beta API that bans cloning and adds watermarks.
Wildness 3/5 · Partly tamedThe HTTP QUERY method is now a standard
RFC 10008 standardizes QUERY, a safe and idempotent HTTP method with a request body. It is the clean way to send a large or private query.
Wildness 2/5 · Mostly tamedNEAR AI breach went through an SGLang bug, and the enclave did not stop it
An SGLang remote code execution flaw let an attacker into NEAR AI's inference servers, about 100 accounts and 18 agents. Confidential mode was in scope too.
Wildness 3/5 · Partly tamedAI voice models: Mistral Voxtral vs Microsoft MAI-Voice
Voxtral Mini TTS 3 is $13 per million characters with an agent SDK; MAI-Voice-2.1 is $22, Flash $15, with gated cloning. How the two compare, checked 10 Nov.
Wildness 2/5 · Mostly tamedZyphra's ZONOS3-Mini is an Apache-2.0 TTS model that clones voices in a browser
ZONOS3-Mini is an 800M-parameter MoE text-to-speech model with about 100M active, Apache-2.0 weights, voice cloning and runtimes for browsers, Macs and SGLang.
Wildness 2/5 · Mostly tamedA File Picker Hint Is Not Upload Validation
Guide people toward expected files with accept, then validate uploads on the server.
Wildness 1/5 · TamedGPUStack auth bypass is being exploited: CISA adds CVE-2026-102895 to KEV
GPUStack before 2.2.0 hands admin to anything that looks like localhost, including traffic through a same-host proxy. CISA says it is exploited. Patch now.
Wildness 2/5 · Mostly tamedfilecmp can decide equality from metadata
Choose the comparison mode and cache policy before treating a matching result as proof about file contents.
Wildness 1/5 · TamedOpenAI says its Control Room can stop any research model job within 35 seconds
OpenAI's Control Room makes every research inference engine ask a central service before answering, so a stopped job gets refusals. It fails closed if down.
Wildness 2/5 · Mostly tamedAustralia will legislate an AI safety regulator and mandatory incident reports
After the OpenAI agent incident, Australia plans binding rules for frontier AI firms: a new regulator, rapid incident notice and independent model checks.
Wildness 2/5 · Mostly tamedserc, an Astral experiment, compiles Python bytecode in Rust with code written by Codex
serc is an experimental Rust compiler for CPython bytecode that claims byte-identical output 3-4x faster than compileall. Every code change was by GPT.
Wildness 4/5 · Still wildTextarea Wrapping Is Not Just Visual
Soft and hard wrapping can submit different text even when the controls look alike.
Wildness 1/5 · Tamed
