Daily Positive Signals - October 15, 2026 - 3 verified signals

Daily Positive Signals: a blood-pressure pill’s second act, six-week weather forecasts, and an AI that keeps moose off the tracks

ByAstro

Published Oct 15, 2026Updated Oct 18, 20265 min read

On this page
  1. A blood-pressure pill may have a second act: stoking calorie-burning brown fat
  2. Weather forecasts just got useful two to six weeks out
  3. An AI that keeps moose off the train tracks
  4. Useful progress, without the hype
  5. Know someone this would help?
  6. References

Listen to this brief · about 4 min, story text only

This edition: a blood-pressure drug that stokes calorie-burning brown fat, a machine-learning wrapper that doubles forecast skill two to six weeks out, and an AI that moves wildlife off train tracks. The thread: getting more from what already exists.

Signal 1 of 3 · Medicine

A blood-pressure pill may have a second act: stoking calorie-burning brown fat

100,000+
patients’ PET/CT scans and medical records analyzed with AI to find the connection
First
study to link greater brown-fat presence to lower rates of anxiety and depression

A National Institutes of Health–funded team at Rockefeller University used artificial intelligence to scan PET/CT images and medical records from more than 100,000 cancer patients, and found that people taking losartan — a widely used blood-pressure drug — showed higher activity in their calorie-burning brown fat (National Institutes of Health, 2026). In brown fat cells and in healthy mice, losartan directly boosted metabolic activity: energy expenditure, oxygen consumption, and carbon dioxide production all rose while body weight held steady (Gómez-Banoy et al., 2026). The same analysis found, for the first time, that patients with more brown fat had lower rates of anxiety and depression — though antidepressants of the SSRI type appeared to dampen brown fat (MedicalXpress, 2026).

Why it matters. Losartan has a decades-long safety record, so repurposing it could move faster than any brand-new drug. And the calorie-burning effect runs on a different molecular pathway than the blood-pressure effect — meaning future therapies might isolate the brown-fat boost for people with normal blood pressure (National Institutes of Health, 2026).

Evidence status: NIH news release plus the peer-reviewed Nature paper (Gómez-Banoy et al.), with independent press. The human data are retrospective — associations, not proof of cause — and the direct metabolic boost was shown in cells and mice, not yet in people.

Signal 2 of 3 · AI / Climate

Weather forecasts just got useful two to six weeks out

2×
forecasting skill in the two-to-six-week range after the machine-learning correction
1st
place in ECMWF’s seasonal forecasting competition — the first model to beat long-term averages there

MIT’s Judah Cohen and Microsoft Research’s Lester Mackey built a machine-learning wrapper — probabilistic bias correction — that learns a forecasting model’s historical errors and corrects them, rather than building a new model from scratch (MIT News, 2026). Applied to the European Center for Medium-Range Weather Forecasts’ gold-standard system, it doubled subseasonal forecasting skill and took first place in ECMWF’s seasonal competition, beating the climatological averages that every previous model had failed to outrun (Cohen et al., 2026). It also improved extreme-weather forecasts — severe-flood prediction skill rose 86–150% in tests on real floods of the past decade — where ECMWF’s own non-AI correction had actually made extremes worse (Gizmodo, 2026). The wrapper is open source and runs on ordinary CPUs, no supercomputer required.

Why it matters. Two weeks of warning instead of two days is the difference between scrambling and preparing — for farmers timing planting, grid operators staging fuel, emergency managers pre-positioning supplies. Kenya’s meteorological office has already asked to use the model for monsoon forecasts (MIT News, 2026).

Evidence status: Peer-reviewed Nature paper plus MIT News, with independent press. Tested on historical data and competition entries; real-world deployment, such as Kenya’s, is still early.

Signal 3 of 3 · Engineering

An AI that keeps moose off the train tracks

150
devices live across North America, Europe, and Australia — airports, railways, roads, mines
$272M
annual cost of wildlife strikes to U.S. civil aviation alone, per an FAA estimate

Flox Intelligence, a spinout from Stockholm’s KTH Royal Institute of Technology, sells edge-AI devices that watch for wildlife near airports, rail lines, roads, and power infrastructure (TechCrunch, 2026). Motion sensors trigger a camera; a local model — all processing on-site, no cloud — identifies the animal and plays a species-specific deterrent sound, varying it over time so animals don’t get used to it (GovTech, 2026). About 150 units are live across three continents, and train maker Alstom is an OEM partner putting the software on trains; the company is opening a $10 million Series A to scale to several thousand units (TechCrunch, 2026).

Why it matters. Wildlife strikes damage infrastructure and kill or injure animals — a device that moves them safely instead of cleaning up after collisions protects both sides of the equation.

Evidence status: TechCrunch original reporting (founder interviews) plus independent GovTech coverage. Deployment numbers, partner claims, and effectiveness figures are the company’s; independent verification is pending.

Editor’s note

Nothing in this edition was invented from scratch. A decades-old blood-pressure pill surfaced in 100,000 old medical scans. A weather model got better not by being replaced, but by having its mistakes corrected. A wildlife device learns from real animal behavior in the field, not a lab. Progress often looks like invention. Today it looked like looking harder at what was already there.

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.

Get the brief by email →

Free. One daily email, one weekly email. Unsubscribe anytime.

References

  1. Cohen, J., et al. (2026, October 14). Advancing subseasonal forecasting with machine learning. Nature. https://www.nature.com/articles/s41586-026-11126-x
  2. Gizmodo. (2026, October 14). Forget the 10-day forecast. AI just improved extreme weather warnings up to 6 weeks out. https://gizmodo.com/forget-the-10-day-forecast-ai-just-improved-extreme-weather-warnings-up-to-6-weeks-2000825560
  3. GovTech. (2026, October 13). Flox Intelligence is using AI to deter what? https://www.govtech.com/question-of-the-day/flox-intelligence-is-using-ai-to-deter-what
  4. Gómez-Banoy, N., et al. (2026, October 14). Artificial intelligence-enabled clinical phenotyping of human brown fat. Nature. https://www.nature.com/articles/s41586-026-11162-7
  5. MedicalXpress. (2026, October 14). Commonly prescribed blood-pressure drug may stimulate ‘healthy’ brown fat. https://medicalxpress.com/news/2026-10-commonly-blood-pressure-drug-healthy.html
  6. MIT News. (2026, October 14). Researchers improve weather forecasts looking two to six weeks out. https://news.mit.edu/2026/researchers-improve-weather-forecasts-looking-two-six-weeks-out-1014
  7. National Institutes of Health. (2026, October 14). A blood-pressure drug may activate health-promoting brown fat. https://www.nih.gov/news-events/news-releases/blood-pressure-drug-may-activate-health-promoting-brown-fat
  8. News-Medical.net. (2026, October 14). Common blood pressure drug boosts brown adipose tissue activity and metabolism. https://www.news-medical.net/news/20261014/Common-blood-pressure-drug-boosts-brown-adipose-tissue-activity-and-metabolism.aspx
  9. TechCrunch. (2026, October 13). Flox Intelligence is building AI to keep wildlife away from trains, airports, and roads. https://techcrunch.com/2026/10/13/flox-intelligence-is-building-ai-to-keep-wildlife-away-from-trains-airports-and-roads/

Similar Posts