Tech Brew Ride Home - Fri. 05/09 – Apple’s Coming For The Meta Ray-Bans

Episode Date: May 9, 2025

More signs that Apple is diversifying its smart headgear ambitions. Meta wants back in on the stablecoin game it arguably bailed on too early. The Celsius network CEO gets 12 years in prison. And, of course, the Weekend Longreads Suggestions. Sponsors: Quince.com/ridehome Links: Apple Is Developing Specialized Chips for Glasses, New Macs and AI Servers (Bloomberg) Meta Explores Stablecoin Integration as US Senate Blocks Key Crypto Bill (Coinpaper) Google rolls out AI tools to protect Chrome users against scams (TechCrunch) Celsius Founder Mashinsky Gets 12 Years for Crypto Fraud (Bloomberg) Weekend Longreads Suggestions: Can we make AI less power-hungry? These researchers are working on it. (ArsTechnica) Rejoice! Carmakers Are Embracing Physical Buttons Again (Wired) Learn more about your ad choices. Visit megaphone.fm/adchoices

On April 4th, 2023, around 2 in the morning, a man was found stabbed multiple times on a sidewalk in downtown San Francisco. Hey, who did this to you? What happened next turned the story into a political firestorm. Reports have identified the victim as Bob Lee, the founder of Cash App. From Bloomberg Podcasts, this is Foundering, the Killing of Bob Lee, beginning April 16. Welcome to the TechMean right home for Friday, May 9th, 2025. I'm Brian McCullough today. More signs that Apple is diversifying its smart headgear ambitions. Meta wants back in on the stable coin game. It arguably bailed on too early. The Celsius Network CEO gets 12 years in prison and, of course, the weekend long-read suggestions. Here's what you miss today in the world of tech. Apple aims to begin mass production of a chip for its smart glasses with greater power efficiency and multi-camera control by the end of 2026 or in 2027, quoting Bloomberg. The company has made progress on a chip that it's developing for smart classes according to people with knowledge of the matter. The move indicates

that Apple is ramping up work on such a device which would compete with the popular rayband spectacles offered by meta-platforms. The Silicon team has become a critical piece of Apple's product development engine in recent years, especially after it began replacing Intel processors with homegrown Mac chips in 2020. Other semiconductors in development will enable future Macs as well as AI servers that can power the Apple Intelligence Platform, said the people who ask not to be identified because the plans are private. The Glasses processor is based on chips used in the Apple Watch that require less energy than the components in products like the iPhone, iPad, and Mac. The chip has been customized to remove some parts in order to further improve power efficiency. The processor is also being designed to control the multiple cameras that are planned for the glasses.

The company aims to begin mass production of the processor by the end of next year or in 2027, indicating that the glasses, if, successful are likely to come to market in roughly the next two years. As with Apple's other major chips, partner Taiwan, semiconductor manufacturing will handle production. Apple has spent years trying to develop smart glasses, something lightweight that consumers can wear all day. The original idea was to use augmented reality, which superimposes media notifications and apps over real-world views. But AR remains years away from being practical. In the meantime, meta and others have had success with non-AR smart glasses, which can take pictures, play audio, make phone calls and let users talk to a voice assistant. Apple now looks to jump into that market as well,

even while it continues to pursue the AR concept. The company held user studies with employees on the concept last year. Apple is working on both options under the codename N401, a recent shift from the prior internal nomenclature of N50. Tim Cook, the company's chief executive officer, is determined to beat meta in the glasses market, Bloomberg News has reported, but meta is moving aggressively itself. The social networking giant is rolling out a premium model with a display later this year and plans its first true AR spectacles for 2027. Apple is currently exploring non-AR glasses that use cameras to scan the surrounding environment and rely on AI to assist users. That would make the device similar to the meta product, though Apple is still figuring out the exact approach it wants to take. The iPhone maker also needs its own artificial intelligence technology to vastly improve before the company can roll out a compelling AI-centric device.

The company also is spreading its bets. Apple has been working on adding cameras to its AirPods and smartwatches, aiming to turn those products into AI products as well. Bloomberg News has reported that company is developing a chip called Nevis for the camera-equipped Apple Watch and a component named Glennie for the similarly outfitted AirPods. Apple is aiming to have those chips ready by around 2027, end quote. What this segment presupposes is maybe they were just too early. Remember DM? Remember Libra? Well, sources say meta is in discussions with crypto firms to introduce stable coins as a means to manage payouts and has even gone as far as hiring a VP of product with prior crypto experience, quoting coin paper.

Although meta has not finalized a path forward, one source familiar with the matter revealed that the company is considering a multi-token approach, potentially supporting top stable coins like Tether's USDT and Circles USDA. The move could reestablish meta as a major player in the digital finance space after its high, profile retreat from the Libra, later Diem project in 2022. The timing of Meta's renewed interest in crypto is far from coincidental. Stablecoins, digital tokens peg to fiat currencies like the U.S. dollar, have surged in both market capitalization and institutional relevance over the past year. As of May 2025 $22025, the total market cap of stablecoins has eclips $230 billion, reflecting surging demand from retail and institutional users alike.

Metas pivot toward stablecoin payments also aligns with broader trends among major fintech and payments firms. On May 7th, Visa confirmed its investment in stablecoin infrastructure startup BVNK, with Visa's global head of product and partnerships, Rubel Burwet, Gerst, stating that stablecoins are capturing an increasing share of global payments. On the same day, Stripe rolled out stable coin-based accounts for users in more than 100 countries, enabling them to send, store, and convert digital dollars into Fiat. Stripe's move was seen as a clear bet on the growing utility of stable coins in global commerce, especially in countries with limited access to stable fiat currencies. Meta, with its billions of users worldwide, could become the most powerful

distribution channel for stable coins to date. Meta's original crypto ambitions began with Libra in 2019, a digital currency project backed by a consortium of tech and financial firms. The effort was ultimately shelved in 2022 amid regulatory blowback from global lawmakers, particularly in the United States and Europe. Meta sold the intellectual property for Diem to, to Silvergate Bank, exiting the crypto stage entirely. Now, instead of launching its own currency, META appears to be embracing existing Stable Coin infrastructure, a more politically palatable move in light of the U.S. government's shifting stance on digital assets. If META proceeds with stablecoin integration, the implications would be enormous. Platforms like WhatsApp and Facebook

Messenger are already dominant communication tools in many developing economies where stablecoins are often used to avoid inflation and currency instability. Embedding stablecoin payments directly into Meta's apps could unlock an entirely new frontier in financial services, enabling everything from micropayments and remittances to merchant services. Still, many questions remain unanswered. Will Meta partner exclusively with U.S. regulated stablecoins like USD?C.? How will it navigate international licensing and compliance in jurisdictions with strict crypto rules?

And will the company face fresh scrutiny from regulators wary of another tech-led financial revolution, end quote. Google is rolling out AI-powered defenses to help combat scams on Chrome, including the use of Gemini Nano, its on-device, LLM, in enhanced protection mode on desktop, quoting TechCrunch. According to Google Chrome's enhanced protection mode of safe browsing on Chrome offers the highest level of protection, providing twice the protection against fishing and other online threats, compared to the browser's standard protection mode. Now, Google will use Gemini Nano to provide enhanced protection users,

with an additional layer of defense against online scams. Google says this on-device approach will provide immediate insight into risky websites to protect users against scams, including those that haven't been seen before. Gemini Nanos' LLM is perfect for this use because of its ability to distill the very complex nature of websites, helping us to adapt to new scam tactics more quickly, Google said in a blog post. The company is already using this AI-powered defense to protect users from remote tech support claims. Google plans to expand this defense to Android devices and even more types of scams in the future. As for the new AI-powered warnings, Google notes that the risk from scammy sites can extend beyond the site itself through notifications if you have them enabled.

Malicious websites can use notifications to try to scam you, which is why Chrome will now help you be aware of malicious, spammy, or misleading notifications on Android. Now, when Chrome's on-device machine learning model flags a notification as possibly being a scam, you will receive a warning. You can choose to either unsubscribe or view the content that was blocked. If you think the warning was shown incorrectly, you can allow all future notifications from that site. As part of today's announcement, Google shared that it has been using AI to stop scams in search by detecting and blocking hundreds of millions of scammy results every day. Its AI-powered scam detection systems have helped to catch 20 times more scammy pages, Google says. For example, Google has seen an increase in bad actors impersonating airline customer service agents

and scamming people looking for help. The company says it has reduced these scams by over 80%, decreasing the risk of users coming across a scammy phone number on search, end quote. Celsius Network founder and former CEO Alex Machinsky has been sentenced to 12 years in prison after pleading guilty to commodities fraud and misleading investors. Quoting Bloomberg, Mishinsky, who pleaded guilty in December was sentenced Thursday by U.S. District Judge John Kotel in Manhattan. Federal prosecutors had sought a 20-year prison term for the 59-year-old former chief executive officer calling him unrepent. Celsius's bankruptcy was part of the 2022 crypto winter that wiped out billions of dollars in asset value. Under President Joe Biden, the Justice Department launched a series of prosecutions targeting crypto industry figures.

Sam Bankman-Fried, whose FTCS exchange collapsed four months after Celsius is currently serving a 25-year sentence for fraud, and Terraforms Labs founder Du Kwan is still awaiting trial. But Machinsky's sentence comes as the Trump administration is de-emphasizing crypto-related investigations as part of a friendlier stance toward the industry. The Justice Department last month issued new guidelines stating it would no longer pursue several types of criminal cases it said would be better left to regulators. The administration said it would still prosecute cases in which crypto investors were defrauded, though. The case for tokenization and the use of digital assets is strong, but it is not a license to deceive Manhattan U.S. Attorney Jay Clayton said in a statement on the sentencing. the rules against fraud still apply. Celsius gained popularity by pitching itself as being as safe as a bank for crypto except with higher interest rates and on the strength of Michinsky's highly charismatic

YouTube and social media presentations. At its height, Celsius held around $25 billion in assets, but the firm was unable to meet demand for withdrawals amid the crypto downturn in June 22 and filed for bankruptcy a month later. Prosecutors said Thursday that Celsius customers lost access to more than $5 billion in crypto when the network paused withdrawals in June 2022, though they've since recovered around $3 billion. As part of his plea, Mishinsky admitted making a series of misleadingly rosy statements about Celsius's financial health on his Ask Mishinsky Anything YouTube videos, as well as engaging in manipulative trading to inflate the price of Celsius's cell token. Mishinsky made around $42 million from his cell trading, the government said.

Lawyers from Mishinsky, a father of six, who was born in the Ukraine and grew up in Israel, coming to the U.S. in 1988, had asked the judge to impose a sentence of just over a year in prison. They depicted him as being caught off guard by the crypto downturn and said he never meant to defraud anyone. Fighting back tears on Thursday, Mishinsky apologized for his actions and called his false statements inexcusable. He said the decision to halt withdrawals had been made in haste to try to protect Celsius's top customers from losing everything. All my actions were meant to protect my community and I failed, he said, end quote. Time for the weekend long read suggestions. First up, Ars Technica takes a look at the research race to make AI less power hungry. Quote, one way to go about it is reducing the amount of computation, said J. Wan Chung, a researcher at the University of Michigan, and a member of the ML Energy Initiative.

One of the first things researchers tried was a technique called pruning, which aimed to reduce the number of parameters. Jan Lecun, now the chief AI scientist at Meta, proposed this approach back in 1980. terming it somewhat menacingly the optimal brain damage. You take a trained model and remove some of its parameters, usually targeting the ones with a value of zero, which add nothing to the overall performance. You take a large model and distill it into a smaller model trying to preserve the quality, Chung explained. You can also make those remaining parameters leaner with a trick called quantization. Parameters in neural nets are usually represented as a single precision floating point number, occupying 32 bits of computer memory. But you can change the

format of parameters to a smaller one that reduces the amount of needed memory and makes the computation faster, Chung said. Shrinking an individual parameter has a minor effect, but when there are billions of them, it adds up. It's also possible to do quantization-aware training, which performs quantization at the training stage. According to Nvidia, which implemented quantization training in its AI model optimization toolkit, this should cut the memory requirements by 29 to 51%. Pruning and quantization belong to a category of optimization techniques that rely on tweaking the way AI models work internally, how many parameters they use, and how memory-intensive their storage is. These techniques are like tuning an engine in a car to make it go faster and use less fuel.

But there's another category of techniques that focuses on the processes computers use to run those AI models instead of the models themselves, akin to speeding up a car by timing the traffic lights better, end quote. And then Wired says, another one of my long-standing gripes about technology is getting addressed. Quote, automakers that nest key controls deep in touchscreen menus forcing motorists to drive eyes down rather than concentrate on the road ahead, may have their non-U.S. safety ratings clipped next year. From January, Europe's crash testing organization Euro N-CAP, or New Car Assessment Program, will incentivize automakers to fit physical, easy-to-use, and tactile controls to achieve the highest safety rate. ratings. Manufacturers are on notice, EuroenCAP's director of strategic development, Matthew Avery, tells Wired, they've got to bring back buttons. Driving is one of the most cerebrally challenging things humans manage regularly. Yet in recent years, manufacturers seem almost addicted to switch-free

touchscreen-laden cockpits that, while pleasing to those keen on minimalistic design, are devoid of physical feedback and thus demand visual interaction, sometimes at the precise moment when eyes should be fixed on the road. A smattering of automakers are slow. slowly admitting that some smart screens are dumb. Last month, Volkswagen design chief Andreas Mint said that next-gen models from the German automaker would get physical buttons for volume, seat heating, fan controls, and hazard lighting. This shift will apply in every car that we make from now on, Mint told British Car Magazine Auto Car. Despite the proliferation of advanced driver assistance systems, motor crash fatalities in the United States have increased 21% in the

past 15 years. 40,000 people have died on the roads in each of the past three years, for which complete federal records are available. In-vehicle infatement systems impair reaction times behind the wheel more than alcohol and narcotics use, according to researchers at Independent British Consultancy, TRL. The five-year-old study commissioned by road safety charity I Am Road Smart discovered that the biggest negative impact on driver's reactions to hazards came from using Apple car play by touch. Reaction times were nearly five times worse than when a driver was at the drink drive limit and nearly three times worse than when high on cannabis. A study carried out by Swedish car magazine V. Biligare in 2022 showed that physical buttons

are much less time-consuming to use than touch screens. Using a mix of old and new cars, the magazine found that the most straightforward vehicle to change controls on was the 2005 Volvo V-70 festooned with buttons and no screens. A range of activities such as increasing cabin temperature, tuning the radio, and turning down instrument lighting could be handled within 10 seconds in the old Volvo. and with only a minimum of eyes down. However, the same tasks on an electric MG Marvel R compact SUV took 45 seconds, requiring the precious travel time to look through the nested menus.

The tests were done on an abandoned airfield. Distraction plays a role in up to 25% of crashes in Europe, according to a report from the European Commission published last year. Distraction or inattention while driving leads drivers to have difficulty in lateral control of the vehicle, have longer reaction times, and misinformation from the traffic environment, warn the report, end quote. No weekend bonus episodes for you this weekend, but we will have something for you next week because heads up, there won't be a regular 15-minute show on Thursday. I'm going to be attending the Collective Future Conference on Roosevelt Island that day at Cornell's campus,

and it goes all day, so like no time to produce a show basically at all that day. I will have some stuff for you to tide you over that day, however, and I'll be back with the regular stuff on Friday. So four proper shows next week, Thursday off, but I will have a little something extra in your feed that day. Talk to you on Monday.