Skip to content

News

Original source briefings, practical implications and questions worth asking before you change your stack.

Search stories, models or topics

15 stories

Nov 11, 2026

Models & releasesGoogle

Gemini Decisions API: confidence scores need calibration checks

Published: Nov 11, 2026

A lavender glass orb at a copper track junction leading to a toolbox, library and review desk, beside a confidence gauge.

Nov 9, 2026

Coding & agentsOpenAI

Agent operations need a working stop button

Published: Nov 11, 2026

Small green mechanical helpers move translucent code blocks along copper rails beside a manual review lever.

Nov 9, 2026

Models & releasesAnthropic

Claude Fable 5.2 and Mythos 5.2: read the access distinction

Published: Nov 11, 2026

Two purple model doorways: one open toward a glowing processor, the other closed with a large copper access key.

Nov 4, 2026

Models & releasesOpenAI

Images 2.6: build an asset workflow around the generator

Published: Nov 11, 2026

A framed clay landscape, smaller image plates, copper scissors and a green brush on a purple artist’s workbench.

Nov 2, 2026

Coding & agentsOpenAI

When agents write more code, Git becomes a product concern

Published: Nov 11, 2026

Branching copper repository rails carry green packages toward a lavender merge station on a purple workbench.

Oct 30, 2026

Research & evidenceOpenAI

AI crawling preferences deserve a deliberate publishing policy

Published: Nov 11, 2026

A small mechanical spider beside a copper web gate, with separate paths toward a public library and an adjustable shutter.

Oct 13, 2026

Coding & agentsGitHub

AI secret detection: the useful metric is a safer repository

Published: Nov 11, 2026

A copper inspection lamp reveals one glowing green key among organized purple drawers.

Oct 7, 2026

Models & releasesAnthropic

Claude Haiku 5.5: smaller models need explicit boundaries

Published: Nov 11, 2026

A small green messenger robot carries a lavender task cube over a bridge toward a larger elevated workshop.

Oct 5, 2026

Research & evidenceGitHub

ReviewBench: what a code-review benchmark should answer

Published: Nov 11, 2026

A lavender inspection lens highlights a misaligned green connector on a copper schematic in an open folio.

Oct 2, 2026

Research & evidenceGoogle

September’s Google AI recap: separate releases from adoption

Published: Nov 11, 2026

A copper orrery surrounds an open archive globe with glass satellites containing a processor, microphone and weather vane.

Sep 29, 2026

Models & releasesOpenAI

GPT-6.1 Sol: evaluate the work, then the price

Published: Nov 11, 2026

A copper balance weighs a lavender processor against green completed-task blocks beside an hourglass and a review checklist.

Sep 22, 2026

Coding & agentsOpenAI

Prompt caching: make repeated context earn its keep

Published: Nov 11, 2026

A copper carousel reuses lavender context plates between a green archive drawer and a small purple terminal.

Sep 22, 2026

Models & releasesAnthropic

Claude Opus 5.5 puts migration discipline in focus

Published: Nov 11, 2026

A moss-green orc hand carefully fits a glowing lavender processor into a copper and plum workstation.

Sep 18, 2026

Coding & agentsGitHub

RAG, MCP and skills: ask which problem each one solves

Published: Nov 11, 2026

A green library cabinet, copper bridge connector and lavender instruction scroll feed separate paths into one task station.

Sep 15, 2026

Research & evidenceIBM Research / Hugging Face

An agent that succeeds once still needs a reliability test

Published: Nov 11, 2026

A copper magnifying glass examines a connected lattice of green nodes and lavender links on a purple research desk.