K
Kraus Labs
← Back to Daily Brief

Daily Brief — 2026-10-08

X

@cursor_ai

Claude Haiku 5.5 in Cursor: the model is now available in Cursor Settings > Models; on shorter requests it costs 10x less than Claude Haiku 4.5 ($0.10/M input, $0.50/M output tokens), and Claude Sonnet 5.5 cache reads drop to $0.10/M. post (https://x.com/cursor_ai/status/2107897245282799864) · 2,451❤ · 110 RT

@poteto

Starting with Grok Bot: advises connecting the apps you use daily (calendar, Slack, issue tracker, CRM, Drive) and doing everything with one bot before adding specialist bots. post (https://x.com/poteto/status/2107830403549831186) · 1,505❤ · 74 RT

Time to rewrite (ttr): proposes a rough heuristic for how agent-ready a codebase is: how long a fully automated rewrite would take, and whether agents could verify identical behavior, performance, and maintainability. post (https://x.com/poteto/status/2107913381751730352) · 1,412❤ · 41 RT

Monitor X for feedback: quotes the news that Grok Bot can now search, read, and monitor X; suggests routing feature requests to an issue tracker and bug reports to a Cursor cloud agent to triage and fix. post (https://x.com/poteto/status/2107963437154435182) · 1,403❤ · 49 RT

@rauchg

Knowing when to stop: every program can be hardened and optimized almost forever, with real costs in time and attention; even with unlimited tokens you must choose tradeoffs, because agents will keep drilling. post (https://x.com/rauchg/status/2107962327169675566) · 725❤ · 44 RT

@simonw

Haiku 5.5 notes: published pelican-on-a-bicycle tests and notes on the pricing of the new Claude Haiku 5.5; says it draws far better than Haiku 4.5 at a tenth of the cost. post (https://x.com/simonw/status/2107938675405586713) · 210❤ · 11 RT

@hwchase17

Deep Agents tool loading: deepagents can now load specific tools only when a skill that needs them is loaded, without breaking the prompt cache on OpenAI and Anthropic models. post (https://x.com/hwchase17/status/2107908436965179896) · 58❤ · 10 RT

Interrupt London: the LangChain conference is next Tuesday, with talks from Orange and OVO on agents in production and workshops on Deep Agents and Engine. post (https://x.com/hwchase17/status/2107921716605517953) · 34❤ · 5 RT

@HamelHusain

Ready-to-use eval metrics: answers that using off-the-shelf evaluation metrics is a bad idea in most situations. post (https://x.com/HamelHusain/status/2107848442177364155) · 45❤ · 8 RT

Silent

@martinfowler, @karpathy, @swyx: no original posts in window

Gaps

@xAI: API not authorized to read this account's posts (skipped)

GitHub trending

  1. morluto/rea (https://github.com/morluto/rea): Reverse engineer anything with agents, from app behavior down to native binaries. (TypeScript, +4,655★ today)

  2. mattpocock/skills (https://github.com/mattpocock/skills): Skills for Real Engineers. Straight from my .agents directory. (Shell, +1,403★ today)

  3. boykopovar/AnyPS5 (https://github.com/boykopovar/AnyPS5): Tool for automatic PS5 executables porting to Linux and Windows. (C++, +2,716★ today)

  4. ayghri/i-have-adhd (https://github.com/ayghri/i-have-adhd): A skill to stop your coding agent from burying the answer. ADHD-friendly output. (Python, +619★ today)

  5. cathrynlavery/diagram-design (https://github.com/cathrynlavery/diagram-design): Editorial diagram design for Claude Code, Codex, GitHub Copilot, Factory Droid, and Pi. 42 diagram types. Self-contained HTML + SVG. No shadows. No Mermaid slop. (HTML, +825★ today)