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.
Despite many recent rumors and various reports, Elon Musk confirmed today that SpaceX is not developing a phone based on Starlink, not once, but twice.
Today’s report from Reuters cited people familiar with the matter and stated internal discussions have seen SpaceX executives mulling the idea of building a mobile device that would connect directly to the Starlink satellite constellation.
Musk did state in late January that SpaceX developing a phone was “not out of the question at some point.” However, He also said it would have to be a major difference from current phones, and would be optimized “purely for running max performance/watt neural nets.”
Not out of the question at some point. It would be a very different device than current phones. Optimized purely for running max performance/watt neural nets.
— Elon Musk (@elonmusk) January 30, 2026
While Musk said it was not out of the question “at some point,” that does not mean it is currently a project SpaceX is working on. The CEO reaffirmed this point twice on X this afternoon.
Musk said, “Reuters lies relentlessly,” in one post. In the next, he explicitly stated, “We are not developing a phone.”
Reuters lies relentlessly
— Elon Musk (@elonmusk) February 5, 2026
We are not developing a phone
— Elon Musk (@elonmusk) February 5, 2026
Musk has basically always maintained that SpaceX has too many things going on, denying that a phone would be in the realm of upcoming projects. There are too many things in the works for Musk’s space exploration company, most notably the recent merger with xAI.
SpaceX officially acquires xAI, merging rockets with AI expertise
A Starlink phone would be an excellent idea, especially considering that SpaceX operates 9,500 satellites, serving over 9 million users worldwide. 650 of those satellites are dedicated to the company’s direct-to-device initiative, which provides cellular coverage on a global scale.
Nevertheless, there is the potential that the Starlink phone eventually become a project SpaceX works on. However, it is not currently in the scope of what the company needs to develop, so things are more focused on that as of right now.
Tesla has added a notable improvement to its Dashcam feature after complaints from owners have pushed the company to make a drastic change.
Perhaps one of the biggest frustrations that Tesla owners have communicated regarding the Dashcam feature is the lack of ability to retain any more than 60 minutes of driving footage before it is overwritten.
It does not matter what size USB jump drive is plugged into the vehicle. 60 minutes is all it will hold until new footage takes over the old. This can cause some issues, especially if you were saving an impressive clip of Full Self-Driving or an incident on the road, which could be lost if new footage was recorded.
This has now been changed, as Tesla has shown in the Release Notes for an upcoming Software Update in China. It will likely expand to the U.S. market in the coming weeks, and was first noticed by NotaTeslaApp.
The release notes state:
“Dashcam Dynamic Recording Duration – The dashcam dynamically adjusts the recording duration based on the available storage capacity of the connected USB drive. For example, with a 128 GB USB drive, the maximum recording duration is approximately 3 hours; with a 1 TB or larger USB drive, it can reach up to 24 hours. This ensures that as much video as possible is retained for review before it gets overwritten.”
Tesla Adds Dynamic Recording
Instead of having a 60-minute cap, the new system will now go off the memory in the USB drive. This means with:
- 128 GB Jump Drive – Up to Three Hours of Rolling Footage
- 1TB Jump Drive – Up to 24 Hours of Rolling Footage
This is dependent on the amount of storage available on the jump drive, meaning that if there are other things saved on it, it will take away from the amount of footage that can be retained.
While the feature is just now making its way to employees in China, it will likely be at least several weeks before it makes its way to the U.S., but owners should definitely expect it in the coming months.
It will be a welcome feature, especially as there will now be more customization to the number of clips and their duration that can be stored.
With the news of a merger between SpaceX and xAI being confirmed earlier this week by CEO Elon Musk directly, the first moves of an umbrella company that combines all of the serial tech entrepreneur’s companies have been established.
The move aims to combine SpaceX’s prowess in launches with xAI’s expanding vision in artificial intelligence, as Musk has detailed the need for space-based data centers that will require massive amounts of energy to operate.
It has always been in the plans to bring Musk’s companies together under one umbrella.
“My companies are, surprisingly in some ways, trending toward convergence,” Musk said in November. With SpaceX and xAI moving together, many are questioning when Tesla will be next. Analysts believe it is a no-brainer.
SpaceX officially acquires xAI, merging rockets with AI expertise
Dan Ives of Wedbush wrote in a note earlier this week that there is a “growing chance” Tesla could be merged in some form with the new conglomeration over the next 12 to 18 months.
“In our view, there is a growing chance that Tesla will eventually be merged in some form into SpaceX/xAI over time. The viewis this growing AI ecosystem will focus on Space and Earth together… and Musk will look to combine forces,” Ives said.
Let’s take a look at the potential.
The Case for Synergies – Building the Ultimate AI Ecosystem
A triple merger would create a unified “Musk Trinity,” blending Tesla’s physical AI with Robotaxi, Optimus, and Full Self-Driving, SpaceX’s orbital infrastructure through Starlink and potential space-based computer, and xAI’s advanced models, including Grok.
This could accelerate real-world AI applications, more specifically, ones like using satellite networks for global autonomy, or even powering massive training through solar-optimized orbital data centers.
The FCC welcomes and now seeks comment on the SpaceX application for Orbital Data Centers.
The proposed system would serve as a first step towards becoming a Kardashev II-level civilization and serve other purposes, according to the applicant. pic.twitter.com/TDnUPuz9w7
— Brendan Carr (@BrendanCarrFCC) February 4, 2026
This would position the entity, which could ultimately be labeled “X,” as a leader in multiplanetary AI-native tech.
It would impact every level of Musk’s AI-based vision for the future, from passenger use to complex AI training models.
Financial and Structural Incentives — and Risks
xAI’s high cash burn rate is now backed by SpaceX’s massive valuation boost, and Tesla joining the merger would help the company gain access to private funding channels, avoiding dilution in a public-heavy structure.
The deal makes sense from a capital standpoint, as it is an advantage for each company in its own specific way, addressing specific needs.
Because xAI is spending money at an accelerating rate due to its massive compute needs, SpaceX provides a bit of a “lifeline” by redirecting its growing cash flows toward AI ambitions without the need for constant external fundraising.
Additionally, Tesla’s recent $2 billion investment in xAI also ties in, as its own heavy CapEx for Dojo supercomputers, Robotaxis, and Optimus could potentially be streamlined.
Musk’s stake in Tesla and SpaceX, after the xAI merger, is also uneven. His ownership in Tesla equates to about 13 percent, only increasing as he achieves each tranche of his most recent compensation package. Meanwhile, he owns about 43 percent of the private SpaceX.
A triple merger between the three companies could boost his ownership in the combined entity to around 26 percent. This would give Musk what he wants: stronger voting power and alignment across his ventures.
It could also be a potential facilitator in private-to-public transitions, as a reverse merger structure to take SpaceX public indirectly via Tesla could be used. This avoids any IPO scrutiny while accessing the public markets’ liquidity.
Timeline and Triggers for a Public Announcement
As previously mentioned, Ives believes a 12-18 month timeline is realistic, fueled by Musk’s repeated hints at convergence between his three companies. Additionally, the recent xAI investment by Tesla only points toward the increased potential for a conglomeration.
Of course, there is speculation that the merger could happen in the shorter term, before June 30 of this year, which is a legitimate possibility. While this possibility exists but remains at low probability, especially when driven by rapid AI/space momentum, longer horizons, like 2027 or later, allow for key milestones like Tesla’s Robotaxi rollout and Cybercab ramp-up, Optimus scaling, or regulatory clarity under a favorable administration.
Credit: Grok Imagine
The sequencing matters: SpaceX-xAI merger as “step one” toward a unified stack, with a potential SpaceX IPO setting a valuation benchmark before any Tesla tie-up.
Full triple convergence could follow if synergies prove out.
Prediction markets are also a reasonable thing to look at, just to get an idea of where people are putting their money. Polymarket, for example, sits at between a 12 and 24 percent chance that a Tesla-SpaceX merger is officially announced before June 30, 2026.
Looking Ahead
The SpaceX-xAI merger is not your typical corporate shuffle. Instead, it’s the clearest signal yet that Musk is architecting a unified “Muskonomy” where AI, space infrastructure, and real-world robotics converge to solve humanity’s biggest challenges.
Yet the path is fraught with execution risks that could turn this visionary upside into a major value trap. Valuation mismatches remain at the forefront of this skepticism: Tesla’s public multiples are unlike any company ever, with many believing they are “stretched.” On the other hand, SpaceX-xAI’s private “marked-to-muth” pricing hinges on unproven synergies and lofty projects, especially orbital data centers and all of the things Musk and Co. will have to figure out along the way.
Ultimately, the entire thing relies on a high-conviction bet on Musk’s ability to execute at scale. The bullish case is transformative: a vertically integrated AI-space-robotics giant accelerates humanity toward abundance and multi-planetary civilization faster than any siloed company could.
Elon Musk confirms SpaceX is not developing a phone
Tesla adds notable improvement to Dashcam feature
