01 / SPACE / MOBILITY
Tesla Makes the Roadster Fly
Tesla unveiled its next Roadster on October 15 at SpaceX's test site in Texas. A prototype used compressed-air thrusters and fans beneath the car to drive upside down, launch into the air, roll and land again. 1
The thrusters borrow technology from SpaceX, while the fans pull the car toward the road for more grip. Tesla says the Roadster can reach 60 mph in 0.8 seconds and exceed 230 mph, although those figures have not been tested on a production car. 2
The original concept appeared in 2017 and was once due in 2020. Tesla now says production will begin in 2028. It has not announced a final price, range or how long the most powerful version can stay in the air. 1
Why it matters
The Roadster is the clearest product yet to combine Tesla's electric-car engineering with SpaceX hardware. Its fan system and active control may prove more useful than the short flight itself, especially for acceleration and cornering.
Reality check
This was one carefully prepared prototype, not a production car. The airborne version is not approved for public roads, reservations do not guarantee delivery, and Tesla has already missed several Roadster dates.
02 / QUANTUM
Google Cuts a Quantum Bottleneck by 40x
Google demonstrated a new way to prepare one of the special states that quantum computers need for difficult calculations. On its Willow processor, the method reduced errors by about 40 times and produced a state with 99.99% logical accuracy. 3
These prepared states are used up during a calculation, so a large machine would need to make them constantly. Some estimates suggest that this step could consume as much as 95% of a fault-tolerant quantum computer's capacity. 3
Google's method starts with a noisy state and runs a series of checks that reject faulty attempts. The surviving state can then be moved into a larger part of the processor, where it can support an actual program. 3
Why it matters
A useful quantum computer cannot spend nearly all of its hardware preparing one ingredient for a calculation. Cutting that overhead could reduce the size and cost of future machines.
Reality check
This improves one important step, not an entire quantum computer. Google still has to repeat it at much larger scale, and the headline numbers come from the team that built the system.
03 / AI / MARKETS
Anthropic Grows Fast and Still Disappoints
Anthropic told prospective investors that third-quarter revenue reached $16.8 billion, up 45% from the previous quarter. The company also said it was profitable on an adjusted basis. 4
The figures show that a frontier AI lab can produce both rapid growth and positive cash flow. Anthropic still faces more than $500 billion in planned spending on computing and infrastructure over the next decade. 5
Some investors nevertheless discussed a valuation around $1.5 trillion for the planned November share sale, below the roughly $2 trillion they had previously expected. The reason was not falling revenue, but growth that slowed after sales more than doubled in the second quarter. 4
Why it matters
Anthropic's listing will test how much public investors are willing to pay for frontier AI. The reaction shows how demanding those expectations have become: even 45% growth in three months can be treated as a slowdown.
Reality check
The numbers were shared privately with investors and have not been audited. The $1.5 trillion figure was an investor discussion, not a price set by Anthropic, and adjusted profit does not capture every cost of building the models.
04 / LONGEVITY
One Gene Keeps Old Mice Fitter for Longer
Researchers built a tool that estimates a skin cell's age from the genes it is using. They then screened for genes that could move old human cells toward a younger pattern without making the cells lose their identity. 6
The strongest result came from SRSF1, a gene involved in preparing the instructions cells use to make proteins. Switching it on reduced several signs of cellular aging, stress and inflammation in old human cells grown in the lab. 6
In old mice, the experimental gene treatment slowed the rise in frailty by 44% and was linked to a 14% increase in median lifespan. Each group contained only nine or ten mice. 6
Why it matters
Many rejuvenation experiments try to reset cells so deeply that they risk losing their normal function. This study found one factor that improved several measures of aging while the cells remained skin cells.
Reality check
This is early gene-delivery research in cells and a very small mouse study, not a treatment for people. Several authors are named on a related patent, and two hold equity in a company developing therapies for age-related disease.