AI Briefing#claude
The latest AI and developer news about #claude, with the original source and a short summary.
The latest AI and developer news about #claude, with the original source and a short summary.
Turned Anthropic's Claude system prompt history into something you can browse like git.
Claude 4.7 produces 1.3–1.45x more tokens while showing a modest improvement in instruction compliance.
Claude Opus 4.7's new tokenizer was measured to use more tokens than officially stated.
Anthropic has unveiled a visual design and prototyping tool built on Claude.
On 28 Zod tasks, Opus 4.7 had the same pass rate as 4.6 but was faster and more efficient.
A reference implementation of the five-stage structure of context engineering, built with a Bedrock example.
In the Arena text rankings, Claude Opus 4.7 led in creative writing and coding, while 4.6 was stronger in long queries and instruction following.
The Opus 4.6 model has been removed from Claude Desktop's Code tab, raising concerns about increased token consumption.
Anthropic has expanded its London office to accommodate 800 people amid regulatory pressure in the US.
In the Anthropic Economic Index, DC's Claude usage came in at 4.3x the expected level.
Claude 4.7 uses 1.3–1.45x more tokens, while strict instruction following improved slightly.
Claude Design helps create slides, prototypes, and one-pagers.
Claude Design creates mockups, slides, and document drafts from just a prompt.
Anthropic's Claude Mythos is heightening cybersecurity concerns in the financial sector.
A locally run Qwen3.6-35B-A3B drew a better pelican SVG than Opus 4.7.
Some Claude features now require government ID and live selfie verification.
Claude Opus 4.7 scored 74.0 on the structural biology benchmark, significantly ahead of 4.6.
Released a pre-critique skill to prevent agents from over-complying.
Because of the new tokenizer in Opus 4.7, the same input burns through tokens up to 35% faster than 4.6.
This is a real-usage report that 4.7 violates CLAUDE.md prohibition rules more often than 4.6.
The Claude Opus 4.7 model card specifies synthetic data generated by other models.
Claude Opus 4.7 beat Claude Opus 4.6 by 1 keystroke in Vim Golf.
Role, process, standards, and honesty—not 'unlocking'—determine output quality.