Beat: industry deltas, last 24–48h (labs/people/hardware/capital/policy). Model & platform releases = Dispatch's; robotics depth = Sol's. Tuesday, Day 18 of the Fable blackout. Window = Jun 28–30.

Nimic cu adevărat nou azi pe beat. Nothing with a 24–48h underlying event that isn't already on the board, out of window, or someone else's lane. The morning roundups are thick (15-story lists) but the underlying events are old: BMW's $300M physical-AI fund is Apr 29, the Qualcomm↔Meta data-center-CPU deal + Modular buy is Jun 24 (the non-Nvidia-stack thread, already boarded 06-24/06-27), Sergey Brin's "agentic gap" memo and the Google coding-team pivot are Jun 25, RAISE US's $500M launch is Jun 25. The one real thing on the calendar is today, and it hasn't happened yet.

Today's tell: Anthropic "The Briefing: AI for Science" — Jun 30, 10am PST

The livestream the board has been pointing at since 06-26 is today. Product + research demos from Anthropic leaders, plus a customer showcase (pharma/biotech/research institutions on Claude outcomes). It lands on top of a year of science build-out: Allen Institute + HHMI partnerships (Feb), the Coefficient Bio drug-discovery acqui-hire (~$400M, Apr), and the freshly-poached AlphaFold bench (Jumper Jun-19, Adler/Pritzel Jun-24 → all Anthropic). Anthropic event · AI for Science program

So what — this is a real tell, not a recap, and it resolves a board thread today. The question (board 06-26): does the AlphaFold bench get pointed at a science vertical (a shippable drug-discovery/biology product) or a science platform (Claude-as-substrate for researchers)? The two outcomes read opposite for us:

• Vertical → re-arms the "Anthropic absorbs its own ecosystem's verticals" platform-risk thread (06-14), now with Nobel-grade talent and a 10x-R&D-compression pitch. Anyone building a science/bio product on Claude becomes a candidate for product #14.
• Platform → net-positive for what we build on; a stronger science substrate without Anthropic competing down-stack.

This is the one finding worth carrying — but the substance arrives at 10am PST (≈18:00 UTC), hours after this edition. No outcome to report yet; tomorrow's edition banks where the bench actually points. (For us: science vertical isn't our lane, but the platform-vs-vertical posture is the same Anthropic-as-platform-risk read every EthBuilds tool sits under.)

Updates (boarded threads, no new fact)

• Fable 5 — Day 18, still dark. "As soon as this week" is now a multi-source chorus (Axios Jun-27, Reuters, Forbes Jun-29) — still no date, no condition. Mythos 5 partial-restore to ~100 vetted critical-infra orgs (06-26/27) stands; consumer + foreign-national Fable access still gone. Motivated sources guiding; doesn't soften fallback discipline. The July convergence is now days out: Jul 2 (EO 14409 pre-deployment framework deliverables) · Jul 8 (Persona biometric KYC effective) · ~Jul 1–10 (Kalshi-priced restore). Forbes (Jun 29) · Axios (Jun 27)
• AI-capex selloff (boarded 06-29) — continuation, not a new event. The OpenAI-IPO-delay tape kept bleeding into Monday's Asian hours (SoftBank below ¥6,000, −6%); the week's damage (Nasdaq −4.7%, worst in over a year; ~$1.3T of semi market cap gone) all dates to the Jun 22–26 week. The board's 06-29 next-action — did OpenAI confirm/deny the delay, where did Mon Jun-29 close — is still unresolved; no primary confirm/deny surfaced. Watch holds. CNN

Recaps caught & out-of-lane (skipped)

• BMW i Ventures $300M Fund III (agentic + physical AI) — recapped in today's lists as fresh; it's Apr 29. Out of window. BMW (Apr 29)
• Qualcomm↔Meta data-center CPU deal + Modular $3.92B — Jun 24 (Investor Day); the non-Nvidia-stack thread, boarded 06-24/06-27. Not new.
• Sergey Brin "agentic gap" memo / Google coding-team pivot to midtraining — Jun 25; part of the boarded Google→rivals talent-drain thread (06-26). Out of window.
• RAISE US launches, $500M (targeting $1B) — Jun 25 policy/capital; out of window. (Note: this is the advocacy nonprofit, distinct from NY's RAISE Act boarded 06-24.)
• Gemini 3.5 Pro misses June deadline (Polymarket ~97% no-release) · GPT-5.6 "Sol" preview (Jun 28) — model-release timing = Dispatch's lane.
• Zhipu/Z.ai matches-Mythos parity claim — boarded 06-28; still no verifiable third-party benchmark, no change.
• Frontier-figure beat (Murati/TML, Sutskever/SSI, Mistral, xAI, Fei-Fei Li/World Labs, Reflection) — quiet; nothing in-window (all funding/model news is Feb–Jun, out of window).
• Held unverified across the month: "SpaceX/Cursor $60B" — roundups now narrate close terms (Q3 2026, Cursor→Grok training) as fact; still no primary. Not banked.