K
Kraus Labs
← Back to Daily Brief

Daily Brief — 2026-09-23

Daily Brief — 2026-09-23

Sources: AI Daily Briefing.md (X) + GitHub trending daily

Window: last ~24h (from 2026-09-22 07:38 BRT). Original posts only (exclude retweets/replies).

X credits before run: ~$8.31; after: ~$8.12.

X — signal

@cursor_ai — Cursor

  • Claude Opus 5.5 available in Cursor; top on CursorBench at 57.8% (Max); ~40% less cost per task than Opus 5. post · 4,197❤ · 180 RT

@poteto — Lauren Tan

  • Grok Bot desktop perf: reconnect 60s→0.7s, wake 23s→1s, Computer view switch 868ms→27ms (53 fixes). post · 2,618❤ · 164 RT
  • Meet her new Grok Bot. post · 1,127❤ · 13 RT
  • Quote: Grok in Tesla with Connectors (“our lil guy is growing up”). post · 983❤ · 27 RT

@rauchg — Guillermo Rauch

  • “Software will never die again” — regenerate and deploy what you liked (e.g. Google Reader). post · 1,453❤ · 46 RT
  • Fresh Next.js evals: Opus 5.5 / GPT-6 Sol / Fable 5.1 at 97%; Grok 4.7 at 94% and 2×–7× cheaper. post · 1,261❤ · 64 RT
  • Anthropic shipping taste + Next.js design frontier (Claude Opus 5.5 pages). post · 1,487❤ · 34 RT
  • Interviewing @tobi on stage at Vercel Ship SF (Oct 15): agentic commerce, native rewrites, S3 WALs, slop grenades. post · 471❤ · 15 RT

@simonw — Simon Willison

  • Wrote up Claude Opus 5.5, GPT-6 Sol, and GPT-6 Luna, plus pelican comparison grids by model family/reasoning level. post · 455❤ · 27 RT
  • GPT-6 Luna half the price of 5.6 Luna; favorite for product features on cost and speed. post · 1,432❤ · 52 RT

@swyx — Shawn Wang

  • Opus 5.5 is new default for Latent Space AINews (side-by-side vs GPT-6 Sol): more concise, less slop. post · 87❤ · 5 RT

@hwchase17 — Harrison Chase

  • Improved LangSmith UI for Jev / decision-model traces (state, questions, choices, output). post · 62❤ · 5 RT

@HamelHusain — Hamel Husain

  • Recommend eval skills with @sh_reya; at least run /eval-audit for low-hanging fruit. post · 215❤ · 26 RT
  • “Is Slop Dead?” — testing writing with Opus 5.5. post · 18❤ · 1 RT
  • Stale gold eval datasets: keep updating via error analysis as product/users change. post · 65❤ · 6 RT
  • Distilled AI Evals into a ~30 min Lenny guest post + skills to uncover product errors. post · 393❤ · 46 RT

Silent (no original posts in window)

@martinfowler · @karpathy

Gaps

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

GitHub — top 5 daily trending

  1. anthropics/financial-services (Python) — +438★ today — Reference agents, skills, and data connectors for financial-services workflows (investment banking, equity research, PE, wealth management).
  2. google/ax (Go) — +2,305★ today — Google's open agentic orchestration runtime.
  3. davila7/claude-code-templates (Python) — +64★ today — CLI tool for configuring and monitoring Claude Code.
  4. BuilderIO/agent-native (TypeScript) — +609★ today — A framework for building agentic apps.
  5. obra/superpowers (Shell) — +528★ today — An agentic skills framework and software development methodology that works.