News
Scientists use AI neural network to translate speech from brain activity
Three recently published studies focused on using artificial intelligence (AI) neural networks to generate audio output from brain signals have shown promising results, namely by producing identifiable sounds up to 80% of the time. Participants in the studies first had their brain signals measured while they were either reading aloud or listening to specific words. All the data was then given to a neural network to “learn” how to interpret brain signals after which the final sounds were reconstructed for listeners to identify. These results represent hopeful prospects for the field of brain-computer interfaces (BCIs), where thought-based communication is quickly moving from the realm of science fiction to reality.
The idea of connecting human brains to computers is far from new. In fact, several relevant milestones have been made in recent years including enabling paralyzed individuals to operate tablet computers with their brain waves. Elon Musk has also famously brought attention to the field with Neuralink, his BCI company that essentially hopes to merge human consciousness with the power of the Internet. As brain-computer interface technology expands and develops new ways to foster communication between brains and machines, studies like these, originally highlighted by Science Magazine, will continue demonstrating the steady march of progress.
In the first study conducted by researchers from Columbia University and Hofstra Northwell School of Medicine, both in New York, five epileptic participants had the brain signals from their auditory cortexes recorded as they listened to stories and numbers being read to them. The signal data was provided to a neural network for analysis which then reconstructed audio files that were accurately identified by participating listeners 75% of the time.
In the second study conducted by a team from the University of Bremen (Germany), Maastricht University (Netherlands), Northwestern University (Illinois), and Virginia Commonwealth University (Virginia), brain signal data was gathered from six patients’ speech planning and motor areas while undergoing tumor surgeries. Each patient read specific words aloud to target the data collected. After the brain data and audio data were given to their neural network for training, the program was given brain signals not included in the training set to recreate audio, the result producing words that were recognizable 40% of the time.
Finally, in a third study by a team at the University of California, San Francisco, three participants with epilepsy read text aloud while brain activity was captured from the speech and motor areas of their brains. The audio generated from their neural network’s analysis of the signal readings was presented to a group of 166 people who were asked to identify the sentences from a multiple choice test – some sentences were identified with 80% accuracy.
While the research presented in these studies shows serious progress towards connecting human brains to computers, there are still a few significant hurdles. For one, the way neuron signal patterns in the brain translate into sounds varies from person to person, so neural networks must be trained on each individual person. The best results require the best data possible, i.e., the most precise neuron signals possible, meaning this is something that can only be obtained by placing electrodes in the brain itself. The opportunities to collect data at this invasive level for research are limited, relying on voluntary participation and approval of experiments.
All three of the studies highlighted demonstrated an ability to reconstruct speech based on neural data in some significant capacity; however, also in all cases, the study participants were able to create audible sounds to use with the computer training set. In the case of patients unable to speak, the level of difficultly in interpreting the brain’s speech signals from other signals will be the biggest challenge. Also, the differences between brain signals during actual speech vs. thinking about speech will complicate matters further.
Elon Musk
Tesla’s Roadster appearance at SpaceX has merger watchers talking
Tesla’s next generation Roadster was spotted at SpaceX’s Hawthorne headquarters, two weeks after its unveiling.
Tesla’s next generation Roadster turned up at SpaceX’s headquarters in Hawthorne, California on Wednesday, less than two weeks after the car flew off a ramp at its unveiling in Texas.
Photos shared on X by SpaceX Engineer (@DeannaRiceSPCX) show the new Roadster parked outdoors beside the famed Falcon 9 that stands outside company headquarters at the corner of Crenshaw Boulevard and Jack Northrop Avenue. Rocket B1019 was the first orbital-class booster ever to land after a launch on December 21, 2015 after it sent 11 Orbcomm satellites to orbit and then flew back to Landing Zone 1 at Cape Canaveral, which would be the first landing in what became SpaceX’s reusable rocket program. SpaceX chose not to fly it again. After a test fire in Florida, the booster was shipped to Hawthorne and raised upright as a permanent monument in August 2016, as Teslarati reported at the time. Putting the Roadster, which flies on SpaceX cold gas thrusters, next to the rocket that started SpaceX’s landings is symbolic in itself.
Small car, big rocket ambitions 🚀🚗 #roadster #falcon9 #spacex #tesla https://t.co/H32xUSXoCl
— Deanna Rice (@DeannaRiceSPCX) October 29, 2026
The Thrust and Rocket versions of the Roadster use cold gas thrusters fed by composite overwrapped pressure vessels, the same type of tank SpaceX flies on Falcon 9, and Elon Musk described the car as the work of Tesla and SpaceX engineers working together. Tesla lists 0 to 60 mph in under 0.9 seconds for Thrust and under 0.59 seconds for Rocket, while the base Performance model relies on a fan and ground sealing system for grip, as Teslarati detailed in its breakdown of the three Roadster configurations.
Tesla has kept the car in the spotlight since the October 15 event near Waco. Last week, the company posted a video of the Roadster’s fighter jet inspired cockpit, including a Track Mode screen with a “Thruster · 5,500 PSI” readout, which we covered in our look at the Roadster’s center display. Musk also said owners will be able to ask the car to perform the stunts autonomously, and promised a looping track at Tesla’s Texas headquarters.
The visit comes as special committees at Tesla and SpaceX weigh a possible merger, a step Musk confirmed on SpaceX’s third quarter earnings call. The Roadster, built on SpaceX hardware, has become one of the clearest examples of the two companies already building products together.
Musk said at the unveiling that Roadster production will begin in 2028. Reservations are open with a $50,000 payment, Thrust is listed as a limited release and Rocket is invite only. Tesla says the car is not approved for public road use.
News
Verizon, AT&T and T-Mobile are trying to shut Starlink out before FCC vote
AT&T, Verizon and T-Mobile want 290 MHz cut from an FCC satellite plan before Thursday.
AT&T, Verizon and T-Mobile are asking the Federal Communications Commission to pull nearly 300 MHz of spectrum out of a proposal that would let satellite companies like SpaceX offer cell service directly to ordinary phones. The request lands one day before the agency votes on it.
The FCC will vote Thursday, October 29, on a notice of proposed rulemaking that would make an additional 482 MHz available for direct to device service, expanding the number of eligible bands from five to 13, according to the FCC fact sheet. It would also scrap the rule requiring a satellite operator to lease spectrum from carriers, replacing it with a certification that the operator has coexistence agreements in place or holds the terrestrial license itself.
That second path matters for SpaceX. As Teslarati reported earlier this month, SpaceX agreed to buy Grain Management’s nationwide 800 MHz portfolio to turn Starlink Mobile into a full US carrier, adding to the spectrum it already bought from EchoStar.
The carriers want two bands removed, 2.5 GHz and 3.45 GHz, which make up 290 MHz of the 482 MHz on the table. Through industry group CTIA, they argued that about 300 million users depend on those bands and that satellite operations in them are not technically feasible. T-Mobile, which calls 2.5 GHz the foundation of its Ultra Capacity 5G network, said in its own filing that satellite interference could wipe out up to 96% of a cell site’s uplink coverage area and cut uplink speeds by 88% to 99%.
Executives from all three carriers met with FCC staff on October 20, and T-Mobile CEO Srini Gopalan met separately with FCC Chairman Brendan Carr. T-Mobile’s role stands out because it is still SpaceX’s US launch partner for Starlink Mobile. The three carriers have also teamed up on their own satellite joint venture to pool spectrum, a move widely seen as a response to Starlink.
SpaceX is asking for the opposite. In an October 23 filing after meetings with FCC staff and commissioners’ offices, the company asked regulators to add the 902 to 928 MHz band to the proposal. “Doing so now saves time later,” SpaceX wrote. The company also asked the FCC to modernize interference rules and align satellite deployment requirements with those for terrestrial cell networks, saying the goal is to let satellite and ground networks work together so Americans stay connected wherever they are.
The carriers have not been shy about how they view SpaceX as a competitor. On Verizon’s third quarter earnings call Monday, its chief executive said Starlink Mobile “just can’t compete with a terrestrial network” and estimated SpaceX would need four to six years and $300 billion to $500 billion to build a full US carrier network.
Thursday’s vote will not settle which bands survive. Adopting the notice opens a public comment period, with final rules to come later in the proceeding. The FCC already considered and rejected 2.5 GHz for satellite coverage once, in 2024. SpaceX also is not the only satellite player watching: the FCC approved Amazon’s $11.6 billion acquisition of Globalstar on October 27, giving its planned direct to device fleet its own mobile satellite spectrum.
A Tesla Semi caught fire in Hayward, California, which caused both directions of traffic to shut down for hours before being completely reopened early on Wednesday, the California Highway Patrol said.
The fire occurred just after 5:30 local time in Hayward on Tuesday, as the Semi caught fire in the southbound lanes of I-880 before CHP officials made the decision to shut down both lanes of traffic due to potential risks from hazardous chemical smoke.
There were no injuries reported.
🚨 A Tesla Semi reportedly caught fire, with California Highway Patrol confirming. The Semi was on I-880 SB
An estimated 5,000 to 7,000 semi-truck and large commercial vehicle fires occur on U.S. roadways every year
— TESLARATI (@Teslarati) October 28, 2026
The cause of the fire is currently unknown, as neither Tesla nor the CHP have not determined a cause.
However, some images taken before the Semi was completely destroyed by the fire seem to indicate this might have been an older unit, as it appears to be one of the designs that Tesla shipped as part of its pilot program, before revising it and starting mass production just a month ago.
@Tslachan From on of the videos looks like it’s an old Semi (not the refresh version that recently started production) https://t.co/KzwmR2kXow
— Area Under The Curve 🎗️ (@AreaUnderCurve) October 28, 2026
At the time of publication, CHP had not returned our request for a comment for additional details, including the Semi involved and any potential collaboration on an investigation between the CHP and Tesla. We will update this story when we hear back from them.
The southbound lanes were reopened at 3:30 a.m. on Wednesday, with the northbound lanes opening last night a few hours after the fire began. The CHP did state that the lithium batteries were “incredibly difficult to extinguish,” according to NBC Bay Area.
The truck’s driver said that the Semi caught fire, but he did not have an explanation.
An estimated 5,000 to 7,000 semi-truck and large commercial vehicle fires occur on U.S. roadways every year.
Tesla’s Roadster appearance at SpaceX has merger watchers talking
Verizon, AT&T and T-Mobile are trying to shut Starlink out before FCC vote
