← Archive

Mighty Brief

Friday, June 19, 2026

Start Here

  1. Ethan Mollick says managers are showing the highest success rates with Claude Code. The useful read is not “non-coders can code.” It is that clear specs, judgment, and review loops are turning management into an AI leverage skill.
  2. GLM-5.2 just took the top open-weights spot on Artificial Analysis. Open models are still not “winning,” but the gap is tightening fast enough that model optionality keeps getting more valuable.
  3. The US is reportedly pressing ASML over fears an EUV-class tool may have reached China. If true, this is not just export-control theater. It would mean the hardware moat is leaking at exactly the moment software is compressing.
  4. Garry Tan amplifies a simple go-to-market truth: nobody gets promoted for cutting AI spend. The better pitch is more output, faster cycles, and stronger operator leverage.
  5. Nikunj’s quiet SEO side project is a good reminder that Claude Code / Codex are not just developer tools. They are distribution tools if someone is willing to instrument and iterate the surface area.

BM

Unread scan for dan.wick@blankmetal.ai is blocked this morning. gog gmail search --account dan.wick@blankmetal.ai 'is:unread newer_than:1d' --max 10 --no-input returned invalid_grant / invalid_rapt, so there is no trustworthy BM inbox read for Friday, June 19, 2026.
MEMORY still says the BM bottleneck is sequencing, not demand. The live board is still crowded: Rally, HR&A, Tungsten, Great Hill, Sentinel, Orbis, Care Providers of Minnesota, and Edmunds Govtech.
Best BM angle today: sell workflow lift and optionality, not model religion. Brownfield clients do not need frontier rhetoric. They need a credible before/after path.

Munger Observer

Latest note is still dated May 1, 2026. The calendar context is stale, but the principle is still right.
Opportunity Cost: stale priorities are still charging rent. March-era loops are hanging around long after their leverage expired.
Margin of Safety: protect slack before the day fills up. A good day breaks when there is no room for variance.

Personal Thought

The easiest mistake today is to confuse signal intake with momentum. The better move is to cash one real operator chip early, then let the market noise compete for whatever attention is left.

Quick Scan