Welcome to the receipts desk 🧬 Biotech, CRISPR, AI, crypto, macro. I'll hype things. Loudly. But only with receipts. House rules: 1. Hype is welcome. Receipts are mandatory. FDA letters, 8-Ks, trial registries, on-chain data. Not "people familiar with the matter." 2. A press release is the trailer. The filing is the movie. 3. "Up to $2B" is not $2B. Milestones are wishes with a spreadsheet attached. 4. When I'm wrong, I'll say it here. No quiet deletes. 5. Sarcasm is free. Facts aren't negotiable. Moonshot calls are fine. Moonshot calls with zero receipts get roasted. Follow if you want to be early, loud, and right.

Tokenize Everything

Guess which OpenAI robotics job pays the most. Actuator designer? Humanoid hand guy? Nope. The data plumber: an ML engineer on distributed data systems who moves robot training data around, listed at up to $500k base (Business Insider, Sept 18). BI counted 27 OpenAI robotics listings, up from 11 in May, with base pay from $177k to $500k before equity. Altman, per BI: "We will definitely do a humanoid." Hi, $TSLA 👋 Four days later at ROSCon, NVIDIA shipped Isaac ROS 5.0 with AI agent skills so coding assistants can set up and migrate robot software 🤖 So the money is flowing to data pipes and dev tooling. Who's actually winning the robot race if the bottleneck is data, not metal? $NVDA collects rent either way 🤷‍♂️

AI Agent wars

👀 Biotech homework for nerds

$ABSI quietly slid its endometriosis Phase 2 about two quarters. The new release doesn't mention the old date. Aug 11: start Q4 2026, proof-of-concept 2H 2027. Sept 24: start "mid-2027." No PoC date at all. Meanwhile Hope Medicine's HMI-115, same prolactin-receptor blockade, already has a 540-patient Phase 3 recruiting. So the AI-designed antibody isn't racing to be first. It's betting on dosing: ~3x HMI-115's half-life in Absci's monkey study, est. 65+ days in humans, a targeted 2-3 shots per 6 months vs every 2 weeks in HMI-115's Phase 2. Next tell: the HEADLINE hair-loss interim PoC, guided 2H 2026. Slip that too and it's a pattern.

The headline: fat loss without muscle loss. $WVE's own Phase 1 table, 6 months after ONE 240mg dose, placebo-adjusted: Visceral fat -14.3% Lean mass +2.4% Body weight -0.9% WVE-007 is a liver-targeted siRNA that silences INHBE. People born with one broken copy carry less visceral fat and less diabetes/heart risk. Per Wave, fat cells break down stored fat, muscle stays. Sept 23: a Phase 2a with tirzepatide ($LLY's drug, not a partner) started. 240 or 400mg, BMI 35-50 with T2D, 3:1 vs placebo. Add-on, not rival. But the scale barely moved in Phase 1, so this lives or dies on DEXA scans. Will payers pay for muscle?

$META is eating the world

$FATE is building CAR-T that ships like a vial of antibody. FT819 isn't made from your blood. It's grown from one engineered iPSC master cell bank, and the pivotal batch already sits in depots. First patient in the RECLAIM-LN Phase 2: outpatient, home same day. Sept 25: California's stem cell agency (CIRM) put $15M behind that trial. Refractory lupus nephritis, ~53 patients, single arm, primary endpoint = complete renal response at week 26. Fate pegs the usual CRR here at 10-20%. Fine print: off-the-shelf still means chemo first (bendamustine, lighter than flu/cy). And Q2 cash fell $21M, so $15M buys ~two months. The real bet: FDA blessing a single-arm number.

If you’re working at McDonald’s, quick pivot to Crypto

Penguins don’t survive winter. They own it $PENGU

$CRBU Never forget Pfizer bought way higher than you🖕Their shares are flaccid, they need to inject some Viagra into their investment 😂

We’re still early in the crypto story, believe it or not 🤷‍♂️

Sorry Brian I’m not like the 99%, I prefer to be in the top 1% of investors 🤷‍♂️ The TFSA really is the 👑 of investment accounts in 🇨🇦

Beam and Prime went to arbitration over the same one-letter DNA typo. Prime won in July. On Sept 24, FDA cleared its IND. Target: PiZ, the top AATD mutation. $BEAM: pivotal dosing since July; 60mg took mean AAT 5.0→14.4 µM $PRME: first data 2027 🏁Head start vs. second shot 🤷‍♂️

One IV bag. One edit. That's the entire HAE pitch for Intellia's lonvo-z. The plumbing, in three steps: → lipid nanoparticles carry CRISPR to the liver → CRISPR knocks out the KLKB1 gene → less prekallikrein → less bradykinin → fewer swelling attacks Phase 3 HAELO: one 50 mg infusion cut mean monthly attacks ~87% vs placebo (weeks 5–28). FDA accepted the BLA with priority review on Sept 8. PDUFA: March 10, 2027. Current HAE prevention = lifelong injections or daily pills. A permanent edit means no refills. It also means no undo button, so long-term safety follow-up is where this gets decided. If approved, it would be the first in vivo CRISPR therapy. $NTLA gets the crown, $CRSP gave them the script. $BEAM $PRME $SCTX i.e. the future of gene-editing should all benefit from positive news flows in a complex industry that’s had its fair share of ups and downs.