TL;DR

Flying faster than sound, over oceans and maybe one day over land, could come back, and the one company building a supersonic passenger jet just hit a milestone, though not in the sky. Boom Supersonic says it finished the first core of "Superpower," a 42-megawatt power turbine built from the engine it designed for its Overture airliner, on Oct. 22. The turbine is meant to make electricity on site for data centers, a market that is buying now. That could give Boom money and real-world hours on its engine's core while the airplane catches up. The catch: the turbine's first customer cancelled in September, every number here is Boom's own, and Overture has yet to fly a single test aircraft. Meanwhile the rules that could reopen U.S. skies to faster-than-sound flight are moving, and NASA's quiet supersonic jet is flying regularly.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 6 milestones, starting with "A supersonic airliner carries paying passengers over U.S. land"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

The most expensive thing about going deep into space is hauling fuel up from the ground, and one company has staked a legal claim on a way around it: wrap an asteroid in a bag, crack it open with focused sunlight, and catch the water that boils out as rocket fuel. TransAstra says the U.S. Patent Office has granted its founder a patent on that whole system, from bagging the rock to delivering the fuel to refueling stations between Earth and the Moon (company claim). Nothing has flown, and no mission date is set. But if the idea ever works, spacecraft could fill up in space instead of carrying every drop from Earth, which is the kind of change that would make staying out there for good far cheaper.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 6 milestones, starting with "Water from an asteroid is sold as rocket fuel in space"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

Anything people build in space, from a Moon base to a ship bound for the outer planets, eventually runs out of sunlight, and a reactor small enough to launch is the leading answer. DARPA, the defense research agency, has now signed up SpaceNukes to design, build and actually run a small "kilopower"-class fission reactor in a nuclear ground test with Los Alamos National Laboratory, according to a trade listing. It is the second funded ground test of a space reactor this fall and the third government-backed reactor effort on the books, after NASA's Mars ship and its Moon reactor. Nothing has flown, and no cost, power level or test date has been given. But the number of teams that are building real hardware, rather than drawing it, keeps growing, and that is how a power source goes from paper to the launch pad.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 5 milestones, starting with "DARPA's small SpaceNukes reactor runs in its nuclear ground test"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

Alzheimer's starts quietly, with sticky plaques building in the brain long before memory slips, and a simple blood draw may now be able to spot them in people who feel perfectly fine. Blood tests for a protein called p-tau217, approved in the U.S. in September only for people who already have symptoms, are being lined up to find healthy volunteers for prevention trials, NPR reported. That matters because the only version of beating Alzheimer's that arrives in time is stopping it before the damage, and those trials need a cheap, fast way to find the right people and to see whether a treatment is working. The honest catch: nothing is approved yet for a healthy person who tests positive, so for now the realistic payoff of a positive result is a place in a trial.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 4 milestones, starting with "A p-tau217 blood test is cleared to screen people without symptoms"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

A power station has to run all day, every day, and the most studied fusion machine, the doughnut-shaped tokamak, naturally runs in pulses. A team working on the DIII-D machine in San Diego now reports the first experimental evidence that radio waves beamed in from outside can push the electric current that holds a tokamak's plasma together, the job that would let such a machine run steadily instead of in bursts. The push was small: about 20,000 amps, give or take 10,000, out of roughly 1 million, a slice of about 2%. Its own authors call it proof of principle at modest power. Still, it turns a long-proposed idea for future fusion plants into something that has actually been measured, and that is one more piece of the path toward fusion power that never has to stop.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 5 milestones, starting with "DIII-D repeats the helicon current-drive test at higher power"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

A rocket that comes all the way home and flies again within days is how getting to space could finally get cheap, and SpaceX may try to prove it as soon as mid-November. Starship's 15th test flight, the first attempt to catch both halves of the giant rocket in the arms of its launch tower, has turned up on a U.S. air-traffic planning slide for no earlier than Nov. 12. That date is reported, not announced: SpaceX has not confirmed it, and the FAA has not yet licensed the flight. The rockets themselves are ready. Both stages have passed their full-power engine tests on the ground. If the catch works, Starship would be the first orbital rocket ever fully recovered, by Elon Musk's account, and the cheap, frequent launches behind bigger plans, from Moon landings to large stations in orbit, would move a real step closer.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 6 milestones, starting with "SpaceX catches a Starship upper stage with the launch tower"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

For decades, Alzheimer's drugs have been tested on people whose memories were already slipping, which may be the main reason so many have failed. A new federally funded trial flips that around. ARPA-H, the U.S. agency for high-risk health research, is putting up to $41 million into a four-year, 20-site trial that will try five to ten ways of heading off Alzheimer's at once, in people who have few or no symptoms but whose blood already shows the disease's early changes. Instead of waiting years to count who develops dementia, it will read a handful of blood markers, so a regimen that does nothing can be dropped fast and a promising one moved forward. If the approach works, preventing Alzheimer's could someday look less like a heavy infusion schedule and more like a pill, a supplement or a drug already sitting on pharmacy shelves, chosen with the help of a blood test.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 5 milestones, starting with "An Alzheimer's prevention regimen is approved for people without symptoms"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

For people who have lost the ability to speak, the gap between thinking a sentence and hearing it said aloud may be closing. Neuralink says a man with ALS, the disease that slowly takes away movement and speech, reached 75 words a minute through its brain implant in his first week, with about 1 word in 17 wrong. Ordinary conversation runs about 160 words a minute, so that is a bit under half the pace of normal talk, and well ahead of the 62 words a minute, with nearly 1 word in 4 wrong, that a Stanford-led study managed in 2023. If that holds up in more people, an implant could move from slowly spelling out messages to something closer to joining a conversation. For now it is a company claim from one man, with no paper behind it, and no brain implant has U.S. approval. But the speed is climbing and trials are opening to more of the country.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 5 milestones, starting with "A brain implant that restores speech to people with paralysis wins U.S. approval by 2031"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

The batteries inside brain implants could soon pack more energy into the same tiny space. EVE Energy, a Chinese battery maker, has rolled out a coin-sized battery with no liquid inside, built mainly for implants that connect the brain to computers, and says it holds 40% more energy for its size than ordinary cells of the same kind. If that holds up, devices that help people with paralysis or lost speech could run longer between charges or replacements, or shrink. It is the same solid-state chemistry the big battery makers are racing to put in cars, arriving first at the smallest scale. For now the figures are the company's own and untested, and no one has said when the cell will be made in volume.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 5 milestones, starting with "EVE Energy names a production date for its brain-implant button cell by 2027"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

A machine that helps around the house could be coming from the same kind of company that makes the fridge and the washing machine. LG Electronics told investors on Oct. 28 that it will "expand into home humanoids in 2030," and that it is sharply speeding up the training that robots need to get there: a target of 500,000 hours of robot practice by the end of this year, five times the goal it set in August (company claims, as reported by two outlets). Half a million hours is roughly 57 years of round-the-clock robot experience. The open question is whether that much practice can turn into hands that work in a real, messy kitchen. If it does, a household robot could stop being a $20,000 early-adopter gadget and become an appliance-store product within this decade.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 6 milestones, starting with "LG offers a home humanoid for sale by 2030"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

Small nuclear reactors promise power plants built like airplanes: made in a factory, shipped and bolted together, instead of poured in concrete for a decade at each site. That idea just took a physical step. On Oct. 29 NuScale, which holds the only small-reactor design approved by U.S. regulators, signed a contract with South Korea's Doosan Enerbility to build the first production-quality part of its reactor module. It is real hardware, not a study. What is still coming is the part that matters most: a buyer. No customer has signed a binding order for a NuScale plant, and this contract names none. If one signs, a factory line would already be warming up; if not, this part stays a proof that the factory can do it.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 5 milestones, starting with "NuScale signs a binding customer order by 2027"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

Driverless rides may be about to stop being one company's product and become a market. In Austin, Texas, two companies now each log hundreds of thousands of miles a week with nobody behind the wheel, and Tesla's share is climbing fast: an unaudited community tracker counts 438,000 driverless Tesla miles in Austin in the week ending Oct. 18, up 30% in a single week. Competition could make rides cheaper and easier to find, and Tesla says its steering-wheel-free Cybercab reaches Dallas next month (company claim). The gap is still large. Nationally, Waymo drives about seven times as many driverless miles as Tesla, and Tesla's numbers come from its own reports and outside counters, not an audit. But one American city now has a real driverless rivalry, seven weeks after the Cybercab's first public ride.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 5 milestones, starting with "Tesla offers robotaxi rides in about ten Sun Belt metros by 2027"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

Victor Frankenstein had the right parts list and the wrong century. This Halloween, nearly every piece of his lab exists for real, scattered across universities and companies that have never compared notes. Minnesota researchers froze a pig kidney into glass for just over a year, warmed it back up, and it made urine. Gene-edited pig kidneys have kept people off dialysis for months, one man for 271 days, until a human kidney came. A Yale system restarted circulation and some cell function in pig organs an hour after the animals died. Harvard kept lab-grown human mini-brains alive for nearly six years. And a de-extinction company is breeding animals with the traits of lost species. None of it is a monster, and none of it brings anyone back to life. Put together, though, it sketches something far friendlier: a bank of frozen, revivable, made-to-order organs that could make the transplant waiting list a lot shorter.

Below the line, for paid subscribers: all five parts with their odds, our step-by-step speculation on how they snap together into a real-life organ bank, the odds for the whole machine and its bottleneck, what would change those odds, and who's most likely to build it.

Read more

Saturday, Oct. 31, 2026 · about a 5-minute read

Good morning, and happy Halloween. Fitting, since yesterday the site covered a vaccine meant to hunt down aging's "zombie" cells, a gene that let old mice live longer and the first age-reversal therapy tested in people. Today's big one is our Halloween special: the real-life parts list for Frankenstein's lab.

The big one

Frankenstein was just early. How the future is being built by mad scientists. Nearly every piece of Victor Frankenstein's lab now exists for real. Minnesota researchers froze a pig kidney into glass for just over a year, warmed it up, and it made urine. A gene-edited pig kidney kept one man off dialysis for 271 days until a human kidney came. A Yale system restarted circulation and some cell function in pig organs an hour after death. Harvard kept lab-grown human mini-brains alive for nearly six years, and a de-extinction company says its dire wolves just turned two.

Why it matters: nobody is building a monster. Put together, these parts point to something far friendlier: a bank of frozen, revivable organs where organs wait for people instead of people waiting for organs. It's Some Assembly Required #3. Read it →

Yesterday on the site

The End of Aging

• AI Could Unlock a Real Tan Without the Sun, or the Sun Damage: a rub-on molecule darkened human skin in a dish in 2017, and AI drug design could finally get it out of the lab.

• Age-Reversal Therapy Clears Its First Safety Test in Humans: no serious side effects in three patients, and a dose three times higher is cleared to go next.

• AI Found a Gene That Let Old Mice Live Longer: AI is starting to point researchers at aging levers humans hadn't thought to pull.

• Old-Age Weakness May Have Switches Scientists Can Flip: two specific switches behind weak old muscle give drug makers real targets, and one drug is headed for a trial in older adults.

• Immune Cells Could Be Trained to Kill Aging's 'Zombie' Cells: an AI-designed peptide turned the body's own killer cells against worn-out cells in the lab, a third route to clearing them.

• Hormones Steer How Skin Renews, a New Clue to Keeping It Young: in mice, sex hormones set the rules skin stem cells follow, a hint that skin repair could be tailored to each person.

The End of Disease

• A Pill That Could Slow Alzheimer's Gets Its Verdict in Weeks: a pill that lowers the stress hormone inside brain cells reports its pivotal results in mid-November.

• A Nose Spray for Alzheimer's Has Finished Its Trial: it goes after inflamed brain immune cells instead of plaques, with results due early next year.

• A One-Shot Gene Fix for Muscle-Wasting Disease Reached Boys' Muscles: the editor worked in three boys, but a fourth died from a reaction blamed on the delivery virus, and safer delivery is now the race.

Thinking Machines

• Did AI Just Overturn a Rule Hundreds of Computer Science Papers Rely On? Experts Are Checking: if the proof holds, it's a machine-made result experts couldn't produce themselves.

Robots & Self-Driving

• Fully Robotic Cars With No Steering Wheel Are Heading to Dallas: Tesla says its pedal-free Cybercab reaches a second city in November.

• Air Taxis Just Got Their First Federal Permit: an electric air taxi can now fly buyers around, the first step toward selling seats.

Limitless Energy

• Electric Cars Could Double Their Driving Range: the world's biggest battery maker doubled the size of its liquid-free test cells.

• A $500 Million Bet on Batteries That Need No Lithium: one supplier is building for sodium batteries at enormous scale, a cheaper way to store power.

Sci-Fi Tech

• Full-Color Screens Now Fit in Sunglasses: 55-gram glasses with a full-color display put a real screen in an ordinary-looking frame.

To the Stars

• Starship Just Lit All Six Engines on the Ground. Next Stop: Space.: the last engine test before SpaceX tries to catch the top half of Starship with its launch tower.

Already up this morning

• The Speed of Aging May Be Set by the Bugs in Our Gut: if a microbe moves the aging clock, a probiotic or diet could be tested in months, not decades.

• Wrinkle Cream, but Make It Mitochondria: Eye Wrinkles Shrank in 8 Weeks: a cream that triggers cleanup of worn-out cell power plants made eye wrinkles shallower.

• Sunglasses With a Movie Theater Inside Just Went on Sale: the first glasses on Google's Android XR platform are on sale from $1,099.

By the numbers

• 108 days: how much one dose of an AI-picked gene added to the median life of old mice, from 778 days to 886, in a study of 19 mice.

• 55 grams: the weight Rokid claims for its full-color display glasses, a little more than two AA batteries.

• 456: Starship launches a year SpaceX has asked to fly from Texas, against 25 allowed now.

New on the odds board

Yesterday's posts put fresh odds on a paying passenger riding an air taxi in a U.S. city, SpaceX catching Starship's upper stage with its tower, another lab repeating the AI-picked gene's lifespan gain in mice, and a one-time CRISPR treatment for Duchenne muscular dystrophy reaching patients. Paid readers see the numbers inside each post.

On the calendar

• Nov. 4 to 7: the Gerontological Society of America's annual meeting in National Harbor, Md., the next big date on the aging-research calendar.

• Nov. 9: CRISPR Therapeutics presents new human data on its one-dose gene editor for high triglycerides.

• Mid-November: XREAL's Aura glasses start shipping, and results arrive for the Xanamem Alzheimer's pill. On Nov. 16, a second Alzheimer's pill, AriBio's AR1001, reports results from 1,535 people.

• November, no day set: Starship Flight 15 and the first attempt to catch the ship with the tower, pending regulators' approval. Public FAA meetings on SpaceX's request for more launches run Nov. 17 to 19.

The story we're following

Digital patients: no new development since Thursday. We're still watching for the first regulator to let a computer model stand in for real patients in a final trial.

Quote of the day

"70% success is like it doesn't work." — robotics pioneer Marc Raibert, on robots doing household chores, from today's 11 AM post on LG's robot butler

Coming up today

• 9 AM: Tesla's robotaxis drove 438,000 miles in Austin in one week, with no one at the wheel

• 10 AM: The nuclear reactor that rolls off an assembly line just got its first real part

• 11 AM: The fridge company is building a robot butler, due by 2030

• Noon: A coin-sized battery built just for brain implants packs 40% more punch

• 1 PM: A man who lost his voice is talking at 75 words a minute, straight from his brain

• Sunday: Grown-ups growing taller? Science now has the parts list. Some Assembly Required #4 leads tomorrow's email.

Until next time

That's today in the future. If someone you know would want to wake up to this, forward it to them, and if you're reading a forward, subscribe here. It's free every morning.

TL;DR

A pair of glasses that hangs a big, sharp screen in the air in front of the wearer, while the real room stays visible through the lenses, is now something people can order at Best Buy. XREAL's Aura, the first glasses built on Google's Android XR software, went on sale on Oct. 27 from a $1,099 launch price and starts shipping in mid-November. It is not yet the all-day, normal-looking pair science fiction promised: a cable runs to a small computer clipped to a belt. But it offers a view wider than most display glasses and opens a race. Samsung's Android XR glasses are reported for November, Google's are expected by year-end, and lighter full-color pairs are being priced in China.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 5 milestones, starting with "XREAL's Aura glasses reach buyers' hands by 2026"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

A skin cream that works on the tiny power plants inside cells, rather than on the surface, now has a peer-reviewed paper behind it. The cream carries urolithin A, a compound the gut makes from ingredients in pomegranates and walnuts, and in lab-grown human skin it switched on the cells' routine for clearing out worn-out mitochondria, the parts that make a cell's energy. In placebo-controlled studies on people, it made the wrinkles around the eyes shallower after 8 weeks. If that holds up in independent trials, people could one day buy an anti-aging cream that targets one of the reasons skin ages, not just how it looks, and it could arrive as an ordinary consumer product rather than a prescription drug. The caveats are real: the wrinkle results were first shown in 2023, and nobody has reported how many people took part or how big the effect was.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 5 milestones, starting with "A urolithin A skin cream with independently replicated wrinkle results is on the market by 2029"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

Some people age faster than others, and part of the difference may live in the gut. Researchers at the University of Hawaii compared the gut bacteria of 123 adults with a blood test that estimates how fast each person is aging, and found that differences in their bacteria went along with about 15% of the differences in that pace, roughly one-seventh. One common species turned up most often in people aging more slowly. It is an association, not proof that bacteria change how anyone ages. But if a microbe does turn out to nudge the pace of aging, the most hopeful part is how cheap the next step could be: a probiotic or a change of diet, checked against the same blood test in months instead of waiting decades to see who stays healthy longer.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 4 milestones, starting with "A controlled trial shows a gut-bacteria treatment, such as a probiotic or a diet, slows an aging clock by 2032"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

A rocket that comes back whole and flies again within days could make getting to space cheap enough to matter, and SpaceX may try the hardest piece in November. The upper half of its giant Starship rocket, the part that actually reaches space, just passed its last full-power engine test before Flight 15, when SpaceX plans to try catching it in the arms of the launch tower for the first time. SpaceX has caught the lower half three times; it has never caught the ship. If it works, both halves of an orbital rocket would come back for the first time, by Musk's account, and the turnaround between flights could shrink from months to days. That is the pace SpaceX would need for its claimed weekly launches in 2027 and for the string of refueling flights a Moon landing requires.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 6 milestones, starting with "SpaceX catches a Starship upper stage with the launch tower by 2026"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

For the first time, a single infusion carried a complete CRISPR gene editor into the muscles of boys with Duchenne muscular dystrophy, the inherited disease that slowly destroys muscle, and the edit took. In three boys it restored a small amount of the muscle protein they are born without. That is proof that a gene can be rewritten inside human muscle with one dose, which opens the door to one-time treatments for muscle diseases that today have none. The cost was severe: a fourth boy died about three and a half days after his infusion, from an immune reaction blamed on the virus-based carrier rather than the editor, and the trial stopped enrolling. The road ahead is clearer for it. The editing works; the delivery has to get safer, and the protein levels have to climb much higher before this changes the disease.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 4 milestones, starting with "A one-time CRISPR treatment for Duchenne muscular dystrophy is available to patients by 2036"; our labeled speculation on what it adds up to; and every source.

Read more

TL;DR

This is a mouse study, not yet tested in people, but it raises a question every future skin treatment will have to answer: does it work the same for men and women? A Brussels-led team followed single skin stem cells, which keep replacing skin, in living mice for five months. Females and males renewed their skin by different rules, sex hormones set those rules, and wounds on female mice closed substantially faster in the first few days. If human skin works like this, the creams and shots now in testing to rejuvenate skin might help one sex more than the other. The report, from a single source, gives no sample sizes, and the authors say it is unclear how closely mouse skin matches ours.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our Odds board on 5 milestones, starting with "A skin-regeneration treatment tailored by sex reaches patients by 2040"; our labeled speculation on what it adds up to; and every source.

Read more