06-reference

not boring wdoo 206

2026-08-14·reference·source: Not Boring (Weekly Dose of Optimism)·by Packy McCormick
physical-aihard-tech-optimismfusion-energynot-boring-capital-portfoliorobotics

Why this is in the vault

Issue #206 of the standing Friday optimism series — kept mainly for the Fuse Energy neutron-yield item, a concrete engineering-efficiency data point from a disclosed Not Boring Capital portfolio company, plus the pattern-continuation of the series' physical-AI/hard-tech throughline already well-tracked in prior WDoO notes.

Curation section

  1. Avidrone wins DARPA Heavy Lift Challenge. Canadian drone maker Avidrone's 29.3 lb "Katana" uncrewed aircraft carried a 112.4 lb payload (3.84:1 ratio, just under the 4:1 target) to win DARPA's heavy-lift drone competition; a final attempt at the full 4:1 ratio crashed. Packy frames heavy-lift drone economics (more weight, further, cheaper) as a category that "will reshape the world like nothing since the car."
  2. Josh Kushner/Bob Iger buy the Lakers for $12.5B, a new NBA-franchise record. Packy re-quotes his own April "Scarce Assets" essay (Thrive Eternal's permanent-capital thesis around iconic, non-replicable franchises) — self-referential, not new analysis, but ties to a recurring series theme about capital chasing genuinely scarce assets vs. the fungible $10B-neolab-per-week pattern (he name-checks Jeff Dean's "Disco Loop" raise as the counterexample).
  3. Recast Systems comes out of stealth — a second weather-modification/cloud-seeding startup (after Rainmaker, covered previously in "The Great Differentiation"), claiming a cheaper seeding material, a new hyper-resolution AI weather model, and 100+ cloud-seeding flights for Texas and New Mexico governments. Packy is enthusiastic about founder Olivia Li's "more science-forward, more handmade" brand positioning. Verification check performed (per the WDoO portfolio-disclosure gotcha): searched public sources for a Not Boring Capital / Recast Systems investment relationship — found no confirmation either way; the search surfaced only an unrelated "Recast Capital" (VC-manager platform) and an unrelated fintech "Recast." Could not verify Not Boring Capital is NOT an investor here — treat Packy's framing as messenger, not confirmed-neutral.
  4. Fuse Energy's Faeton-X hits 1.27×10¹² neutron yield — explicitly disclosed as a Not Boring Capital portfolio company. The result: highest published current efficiency (4.5 MA/MJ) of any pulsed-power z-pinch device, beating Livermore's MJOLNIR shot on efficiency (1 MJ vs. 1.3 MJ for a comparable neutron yield). Packy's framing: because z-pinch yield scales as drive current to the fourth power, a 40% current-efficiency gain compounds to ~4x more fusion output, not a linear 40% — and Fuse's near-term revenue model (selling neutron shots to defense/satellite makers for radiation-hardness testing) funds iteration toward the harder TITAN fusion-power path.
  5. Matic now supports voice control — "Hey Matic, go clean the living room." Packy frames this as validating Matic's original bet (be excellent at one narrow task before generalizing) over humanoid-first approaches. Sponsor-block ambiguity: the issue's "brought to you by Matic" slot explicitly states Matic did NOT pay for or request the placement — Packy says he just likes the product (received one free unit at launch). That disclosure is clean on the paid-sponsor question, but a public search found no confirmation either way of a Not Boring Capital equity stake in Matic; combined with item 5 being a second, full write-up of the same company in the same issue, this reads as sustained enthusiasm for one product rather than neutral curation, disclosed-paid or not.

Zero deep-fetches triggered this issue: all five main items are drone racing, sports-team M&A, weather-tech stealth launch, fusion-lab physics, and a robot-vacuum firmware update — none crossed the RDCO relevance threshold (agentic AI, data infra, harness engineering, chip/memory capital cycle) tightly enough with a specific enough hook to justify following a link past what's summarized above.

⚠️ Sponsorship

No paid sponsor this issue (Packy states directly there wasn't one). The nominal "brought to you by Matic" slot is unpaid, unsolicited enthusiasm for a product he owns and received a free unit of — closer to a testimonial than an ad, but functionally occupies the sponsor slot and gets a second full item (#5) later in the same issue. Flagged as structural bias (repeat, uncompensated brand amplification) rather than a disclosed paid relationship; sponsored: false is accurate for the paid-ad question but the reader should treat the Matic coverage as enthusiast messaging, not neutral reporting.

Mapping against Ray Data Co

The concrete connection is item 4 (Fuse Energy): the "current efficiency is the compounding variable of which yield is the derivative" framing — small efficiency gains compounding non-linearly because of a power-law relationship — is a clean cross-domain analogy for RDCO's harness-engineering thesis, where marginal gains in agent-loop efficiency (fewer wasted tool calls, tighter context management per CLAUDE.md hard rule 4) are argued to compound rather than add linearly across a long session. It's an analogy, not a data point for an active RDCO project, so this stays a conceptual reinforcement rather than a thesis-changing signal. The Recast/Matic items are weaker: interesting as pattern-continuation of the series' physical-AI optimism throughline already covered by prior WDoO notes, but neither adds new RDCO-specific signal.

Related