06-reference

not boring wdoo 216

2026-10-23·reference·source: Not Boring·by Packy McCormick
nuclear-energyfusionmanufacturingbiotechnot-boringweekly-dose-of-optimism

Why this is in the vault

Filed for Marathon Fusion's byproduct-monetization strategy (the clearest bet-sequencing data point in the issue) and The Nuclear Company's Prosperity One / Radiant's factory-built-microreactor items, which are directionally on-theme for the founder's current hardware/manufacturing exposure interest; also logging a Reducto sponsor-recurrence data point that crosses a pattern threshold worth a README update.

Mapping against Ray Data Co

The load-bearing connection is item (5), Marathon Fusion: the company hasn't achieved net-positive fusion power, but it's already monetizing the byproducts of the infrastructure built toward that goal — fast neutrons that can transmute mercury into gold can also make medical isotopes (copper-64/67) and enrich lithium-6, so it's shipping revenue-generating output years before the core breakthrough lands. That's the same structural pattern as RDCO's own cert-escalator sequencing (project_phdata_cert_escalator_path): bank independently-valuable milestones (certs, in RDCO's case) on the way to the L4→L5 bet rather than treating the whole thesis as a single all-or-nothing payoff. It's a reusable test for any future RDCO bet: does the infrastructure you're building toward the big goal throw off monetizable byproducts now, or does it only pay off at the finish line?

Secondary and worth noting: item (2), The Nuclear Company's Prosperity One (6.6 GW, six AP1000 units) and the Navy's Radiant Kaleidos microreactor pick, both lean on the same lesson — Vogtle Unit 4 came in ~30% cheaper than Unit 3 purely from keeping the reactor design consistent across repeat builds, and Radiant's whole pitch is factory-built-then-trucked-to-site. That standardize-and-repeat-to-cut-cost playbook is the direct analog to the founder's recently-stated interest in hardware/"vibe manufacturing" exposure (home lab, SendCutSend, B2B SaaS + D2C) — if RDCO ever moves on a hardware-adjacent bet, this is the cost curve to plan around, not a bespoke-build-every-time model.

⚠️ Sponsorship

Sponsor: Reducto (document-parsing/OCR infra for AI pipelines) — standard "brought to you by" block, r-1 model pitch, migration-credit CTA. No bias risk to the curated content below the fold; disclosure is clean and separated from the editorial items.

Pattern-break worth flagging: Reducto has now sponsored #209, #210, #211, #213, #215, and #216 — 6 of the last 8 tracked WDoO issues, including back-to-back at #215→#216. The README's running note treats Reducto as "unknown-until-scanned, no standing relationship" after the #212 (Jack & Jill) break, with #213 and #215's non-consecutive resumptions each logged as insufficient evidence on their own. A second consecutive issue (215, 216) revives the original 3-consecutive-issue confirmation shape seen at #209-211, just not contiguous with it. Worth a founder/parent call on whether to re-promote Reducto to "default sponsor, verify per issue" given the 6-of-8 density, rather than continuing to treat each issue as a fresh coin flip.

Structural self-promo (not a paid sponsor, but disclose): Packy links his own prior work throughout — the "Take Weird Ideas Seriously" essay (alchemy framing for the Marathon Fusion item), the "Fuse" and "Small Steps" essays (framing devices for the same item), and a self-referencing callback to WDoO #177 and Dose #173. Also two "video created with Runway" credits (items 1 and 4) — consistent with the already-confirmed quarterly Runway video-production partnership (see [[2026-10-13-not-boring-runway-world-models]]), not a new relationship. Astera is referred to as "our friends at Astera" with no explicit investor disclosure — per the README's standing WDoO portfolio-disclosure-ambiguity pattern, treat enthusiasm here as a friendly-network signal, not verified-neutral; not independently checked this issue (no single item carried outsized enthusiasm that would clear the bar for a public-investor-list verification pass).

Curation section

  1. FDA approves dostarlimab (Jemperli) for dMMR rectal cancer — accelerated approval as standalone first-line treatment; MSK's 12-patient 2022 trial had a 100% response rate with no chemo/radiation/surgery. Coherence Neuro also raised a $45M Series A (Founders Fund, OpenAI) for BCI-based "cancer therapy that learns." Third-party domains (FDA.gov, Coherence Neuro), but biotech/oncology is outside RDCO's current relevance threshold — no deep-fetch.
  2. The Nuclear Company's Prosperity One (6.6 GW, six AP1000 reactors, Lewis County KY, targeting 2028 NRC approval) + Radiant's Kaleidos microreactor selected by the Navy for Naval Weapons Station Crane, Indiana. See Mapping above — the repeat-build cost-curve lesson is the hook, already fully covered in the blurb; no deep-fetch needed.
  3. Warhound (Ben Lamm/Colossal Biosciences + Palmer Luckey/Anduril, $200M raise, $1.2B valuation) — genetically-engineered + hardware-augmented working dogs, licensing Colossal's gene-editing and Anduril's sensor/vest tech. Novel corporate-structure pattern (equity-only JV combining two existing companies' licensed IP) but not RDCO-relevant enough to fetch past the blurb.
  4. Doris Tsao/Astera Neuro's consciousness finding (bioRxiv, Hesse et al.) — neurons in macaque IT/prefrontal cortex alternate every ~60ms between representing the consciously-perceived vs. suppressed face in binocular rivalry; a decodable signal exists for content outside awareness. Interesting adjacent to the vault's standing mechanistic-interpretability thread (see [[2026-10-16-not-boring-wdoo-215]]) but this is biological, not model, interpretability — noted, not deep-fetched.
  5. Electric Plant's Bombadil — foundation model reading bioelectric signals from plants (100 species, 4M+ hours of recordings) to estimate stress states; $11.2M raise, claims 20%+ farm water-use reduction. Third-party and specific, but agtech is outside the current RDCO relevance threshold.
  6. Marathon Fusion — mercury-to-gold transmutation via D-T fusion neutrons, 10x CERN's 2025 rate; near-term pivot to medical isotopes (Cu-64/67) and lithium-6 enrichment (beat a 40-year Soviet record). Load-bearing item — see Mapping above.

Zero deep-fetches this issue. None of the six items clear all three criteria together (third-party domain, RDCO relevance, specific hook beyond the blurb) — the biotech/nuclear/consciousness/agtech items are outside the current RDCO relevance threshold, and the nuclear/fusion items that are directionally relevant were already fully explained in the newsletter's own text with no additional hook to chase.

Related