Daily Positive Signals: a cryptographic receipt for AI answers, a $40 HIV-prevention shot, and a gut-microbe shield for transplant patients
On this page
Listen to this brief · about 4 min, story text only
This edition: a cryptographic receipt that proves which AI model answered, a twice-yearly HIV-prevention shot at $40 a year, and a gut-microbe treatment that cut bloodstream infections after transplant. The thread: progress you can check, afford, and trust.
In this edition · 3 signals
Signal 1 of 3 · AI
A receipt that proves which AI model answered
Anthropic has released FlashProve, an open-source research prototype that attaches a cryptographic proof to an AI reply, showing it came from one exact set of model weights (Anthropic, 2026). The provider first publishes a short fingerprint, called a commitment, for those weights, then produces a proof for each reply that anyone holding the commitment, the prompt, and the reply can check without seeing the weights (Anthropic, 2026). Anthropic demonstrated it on Llama 4 Maverick, a 400-billion-parameter mixture-of-experts model, with proofs generated at about 110 tokens per second and checked in seconds on an ordinary processor (TechCrunch, 2026).
Why it matters. A reply alone cannot show which model produced it, so a cheaper model could be swapped in silently. A checkable receipt gives evaluators and paying customers a way to catch that swap.
Evidence status: Anthropic research post, with independent coverage from TechCrunch. Limits: a research prototype, not a deployed service; proving takes about 70 to 95 times the graphics-card time of ordinary inference; contexts are capped at 1,024 tokens; the proof covers an integer version of the model under greedy decoding.
Signal 2 of 3 · Medicine
A twice-yearly HIV shot at $40 a year
Dr. Reddy’s Laboratories has received tentative approval from the U.S. Food and Drug Administration (FDA) for the first generic injectable lenacapavir, a twice-yearly shot that prevents HIV infection (Clinton Health Access Initiative, 2026). Under a September 2025 pricing agreement, the generic will cost $40 per person per year in the 120 low- and middle-income countries covered by Gilead’s voluntary licence, with first supplies expected before the end of 2026 and wider availability in 2027 (Clinton Health Access Initiative, 2026). The originator was approved in June 2025, and partners describe the shot as almost 100% effective at preventing HIV, while noting that a daily pill is hard to keep up (Contagion, 2026).
Why it matters. Two injections a year remove the daily-pill burden, and a price in line with oral prevention removes the cost barrier. The 16-month path from originator to generic, rather than a decade, is the part other access campaigns can copy.
Evidence status: Joint Unitaid, Wits RHI, and Clinton Health Access Initiative release, with independent trade coverage from Contagion. Limits: tentative approval means the generic meets FDA standards but cannot yet be sold in the United States while the originator patent stands; national approvals, procurement, and delivery systems still stand between approval and access.
Signal 3 of 3 · Medicine
A gut-microbe shield for transplant patients
Results published in Nature Medicine show that SER-155, an oral treatment made of 16 cultivated human gut bacteria, cut bloodstream infections in adults receiving a donor stem-cell transplant (Ponce et al., 2026). In the trial’s randomised, placebo-controlled cohort, 34 people went on to transplant, and bloodstream infections occurred in 2 of 20 people given SER-155 versus 6 of 14 given placebo (Ponce et al., 2026). The treatment was well tolerated, with no treatment-related serious adverse events and no infections from SER-155 strains (Pharmacally, 2026).
Why it matters. Bloodstream infection strikes up to 45% of transplant recipients within 30 days, when treatment has stripped their gut defences. Restoring gut microbes, rather than adding another antibiotic, attacks that risk at its source.
Evidence status: Peer-reviewed phase 1b trial in Nature Medicine, with independent summary coverage from Pharmacally. Limits: a small efficacy sample of 34 transplant recipients, with a wide confidence interval; other secondary outcomes were not statistically significant; the tested regimen combined SER-155 with vancomycin conditioning, so the effect cannot be split between the two; a phase 2 trial is still ahead, pending funding.
Editor’s note
Today’s thread is proof before promise. A receipt that ties an answer to the model that gave it. A price agreed before the generic existed, so the breakthrough reaches the people who need it most. A microbe treatment measured against placebo in the patients at highest risk. None of these is finished: a prototype, a tentative approval, a phase 1b result. But each one replaces trust-me with check-this, and that is what turns a good result into a reliable one.
That’s the edition — three signals, every claim verified against its sources. Browse past briefs · tomorrow’s lands at 07:15.
The Progress Brief, by email
Useful progress, without the hype
The daily Positive Signals brief and the weekly deep dive, by email. Evidence-checked, concise, no doom.
Free. One daily email, one weekly email. Unsubscribe anytime.
References
- Anthropic. (2026, November 2). Progress on provable inference for large language models. https://www.anthropic.com/research/provable-inference
- Clinton Health Access Initiative. (2026, November 2). U.S. FDA approves first generic injectable lenacapavir for HIV prevention. https://www.clintonhealthaccess.org/news/us-fda-approves-first-generic-injectable-lenacapavir-for-hiv-prevention/
- Contagion. (2026, November 2). FDA tentatively approves first generic lenacapavir PrEP for 120 LMICs. https://www.contagionlive.com/view/fda-tentatively-approves-first-generic-lenacapavir-prep-for-120-lmics
- Pharmacally. (2026, November 3). Seres reports 77% reduction in bloodstream infections with SER-155. https://pharmacally.com/ser-155-phase-1b-bloodstream-infection-reduction/
- Ponce, D. M., DeFilipp, Z., Ford, C. B., et al. (2026). A live biotherapeutic product for adult allogeneic hematopoietic cell transplantation recipients: A partially randomized, placebo-controlled phase 1b trial. Nature Medicine. https://link.springer.com/article/10.1038/s41591-026-04686-y
- TechCrunch. (2026, November 3). Anthropic is working on a system to prove AI models’ identity. https://techcrunch.com/2026/11/03/anthropic-is-working-on-a-system-to-prove-ai-models-identity/
