In Brief

Apple has officially entered the touchscreen MacBook Pro era, and apparently, it wants to drain your wallet while it’s at it.

The new touchscreen MacBook Pro starts at $2,899, which is already a hefty price tag in this economy. But if you want the absolute top-of-the-line model with every possible upgrade, prepare to cough up around $14,549.

As MacRumors first reported, Apple’s configurator prices the fully loaded 14-inch model at $14,249 and the 16-inch version at $14,549. Both come equipped with Apple’s top-tier M5 Max chip, featuring an 18-core CPU and 40-core GPU.

You can blame the ongoing memory shortage and AI’s seemingly endless appetite for hardware for some of the pricing pressure. Moving to a 4TB SSD costs an additional $1,000, while upgrading to 8TB adds $3,000. Want the new 16TB option? That’ll set you back another $7,000.

Another $2,000 gets you 128GB of memory, while nano-texture glass adds $150.

For some perspective, the previous 16-inch MacBook Pro topped out at $10,149. The new model raises that ceiling by a whopping $4,400.

Absurd prices aside, the launch marks a significant shift for Apple, nearly a decade after it first experimented with the Touch Bar. The company eventually abandoned that feature, but now it’s ready to embrace a touchscreen MacBook Pro.

Preorders open October 30, with availability beginning November 6.

Google is bringing ads in its AI-powered Search experiences to international markets, betting that generative AI can open up new revenue opportunities even as growth in its core search advertising business slows.

The announcement came Wednesday during Alphabet’s third-quarter earnings call, as the company reported $127.6 billion in revenue, up 25% year over year. But Search advertising growth continued to decelerate, and Alphabet shares fell 8% following the results.

Google Search revenue grew 12% year over year in the third quarter, down from 17% growth in the second quarter and 19% in the first quarter. The slowdown puts added pressure on Google to turn its expanding AI features into a meaningful source of advertising revenue.

“We are now scaling high-quality ads within AI overviews in AI mode to more users and queries,” Philipp Schindler, Google’s chief business officer, said during the call with investors. “This includes taking our monetization footprint global for the first time. Today, we’re announcing the expansion of ads in AI mode into international markets alongside multi-turn ads and AI mode in the U.S.”

Google says early results suggest the formats are working. Click-through rates for sponsored product listings in AI Mode are already comparable to those for organic listings, according to Schindler. The company’s user research also indicates that people find the ads relevant.

The move builds on Google’s May announcement of new Gemini-powered ad formats, including conversational discovery ads that respond to users’ questions and sponsored recommendations integrated into AI-generated answers.

Google CEO Sundar Pichai also said that Search queries grew at their fastest year-over-year rate in five years, though he did not disclose the exact growth rate. He credited AI features with changing how people use Search, including prompting longer and more complex queries.

Queries in AI Mode are three times longer on average than traditional searches, Pichai said, while one in five AI Mode searches globally involves voice, images, or documents. Google has also surpassed 50 billion visual searches per month.

“[Search] is bringing the benefits of AI to more people than any other product in the world,” Pichai told investors. “People are using our AI features more, searching in entirely new ways. Their queries are longer, personalized, and multimodal.”

Alphabet’s Google Services division, which includes Search, generated $96.3 billion in revenue, up 11% year over year. Google Cloud revenue surged 103% to $30.7 billion.

The results follow a busy quarter for Google, including the release of Gemini 4 Argon, the growth of the Gemini app to 1 billion monthly users, and the launch of its Pixel 11 lineup.

Microsoft generated $10 billion in revenue from commercial arrangements with OpenAI, including revenue-sharing payments, in the quarter ended September 30, the company said in a regulatory filing on Wednesday. That represents nearly 11% of Microsoft’s total quarterly revenue of $92.3 billion, highlighting the enormous financial ties that remain between the two companies even as they increasingly compete for the same customers.

(Microsoft’s financial year begins in July, making the quarter ended September 30, 2026 the first quarter of its fiscal 2027.)

For the full 2026 fiscal year, revenue from OpenAI hit $24.1 billion. At its latest quarterly pace, Microsoft would pull in $40 billion annually from OpenAI-related arrangements.

Microsoft also reported $7.2 billion in outstanding receivables from OpenAI at the end of September, up from $6 billion at the end of June. That money represents amounts OpenAI owes for previous transactions, not additional revenue.

It’s not clear how much of Microsoft’s revenue from OpenAI came from revenue-sharing payments or Azure cloud computing services. TechCrunch has reached out for more information. In November 2025, TechCrunch reported, citing leaked documents, that Microsoft had received $865.8 million in net revenue-sharing payments from OpenAI during the first nine months of that year.

OpenAI reportedly shares 20% of its revenue with Microsoft as part of a previous deal in which the software giant invested more than $13 billion in the AI startup. (Neither the AI lab nor Microsoft have publicly confirmed this percentage.) As part of an April 2026 agreement, OpenAI is on the hook to continue sharing revenue with Microsoft through 2030, subject to a cap, even as OpenAI is allowed to work with competing cloud providers.

The companies began loosening the ties of their once-exclusive partnership in October 2025, when Microsoft relinquished its hold on OpenAI to give the startup more freedom to partner with other cloud providers. In exchange, OpenAI committed to buying an additional $250 billion in Azure services as part of an October 2025 agreement.

Microsoft’s ownership stake in OpenAI has also declined slightly from about 25% in June to 24% in September. That continues a gradual decline from roughly 27% following the October 2025 recapitalization. Microsoft did not specify what caused the dilution.

Whatever the future of AI might be, the current situation is still very good for Microsoft. The tech giant reported yet another quarterly revenue record on Wednesday: $92.3 billion, a 19% increase over the year-ago quarter.

Microsoft has now reported record-breaking revenue for 10 quarters in a row.

Image Credits:TechCrunch

This is profitable revenue, too. Microsoft reported that net income reached $36.6 billion, an increase of 32% on a GAAP basis, over the year-ago quarter.

The record-breaking income was largely driven on the strength of Microsoft Cloud, which reached $63.9 billion in revenue this quarter, an increase of 30% over the year-ago quarter, Microsoft said.

Microsoft has reorganized its business reporting into two basic segments: 1) Agents and Infra and 2) Devices and Consumer. On the flip side, revenue in Devices and Consumer was $15.5 billion, decreased of 3%, with Windows OEM and devices revenue falling 22% from the year-ago quarter.

True, not all of Microsoft’s healthy cloud revenue can be attributed to AI. Microsoft 365 accounted for nearly $28 billion. But Azure brought in the most, at $33.2 billion. And again, not all of that is necessarily AI-related. Enterprises run all sorts of the traditional software and SaaS applications on Azure.

Still, a good chunk of the revenue was from AI, because Microsoft also reported $10 billion in revenue from OpenAI. Despite branching out to AWS, Oracle, and others, OpenAI still runs a lot of its services on Microsoft’s cloud. (Anthropic runs on Azure as well.) Microsoft also said its “commercial remaining performance obligation increased 73% to $677 billion.” That represents future revenue committed under existing contracts of the AI labs.

But, like all the big cloud providers, Microsoft is also spending big to prepare to deliver on those commitments. Capital expenditures topped $45.2 billion, another all-time quarterly record.

Personally identifiable information belonging to thousands of U.S. Department of Homeland Security employees has been published online after a hacker stole reams of data in a previously undisclosed breach at the government department earlier this year.

The breach includes names, job titles, and work contact information, such as email addresses, cell phone numbers, and office locations, of more than 40,000 people who work for the federal department, including border and immigration agents, cybersecurity staffers, and senior department officials.

A hacker who goes by the handle Department of Peace claimed they obtained the data from an Okta single sign-on system used by Homeland Security to allow employees and contractors across its various agencies to log in to its central network. The Okta system also contains IT and log-in records associated with the employees’ user account, including details about which systems and internal apps they are permitted to access.

More than half of the dataset — some 21,000 records — relate to employees with U.S. Customs and Border Protection (CBP), who work at airports and other ports of entry.

The data also contains records associated with another 10,000 employees of Immigration and Customs Enforcement (ICE) and its Homeland Security Investigations (HSI) unit, and just under a 1,000 employees who work at U.S. cybersecurity defense agency CISA.

Thousands of additional records relate to staff who work at the department’s headquarters and its other agencies, including U.S. Citizenship and Immigration Services and the Federal Protective Service.

“They have our information, so it’s only fair that we have theirs,” said Department of Peace in a note left with the data, which TechCrunch has seen. The hacker did not respond to TechCrunch’s email requesting comment and with questions about how the breach was carried out.

This is the latest data breach to hit the U.S. government in recent months, following earlier hacks involving the theft of thousands of FBI agents’ personnel records and the personal information of millions of serving and former military service personnel from the Pentagon’s human resources unit. It’s also the second breach involving Homeland Security this year, after the same hacker claimed to have taken thousands of contracts between the department and the private sector, including defense contractors and surveillance vendors.

The latest batch of breached data, which appears authentic, offers a rare — albeit incomplete — view into the scale and operations of Homeland Security, whose deportations, arbitrary detentions, and killings of U.S. citizens under its leadership have become a key issue for voters ahead of the upcoming U.S. midterm elections.

The stolen Homeland Security data was provided to journalist and hacker maia arson crimew, who then shared a copy with DDoSecrets, a nonprofit collective that hosts leaked datasets in the public interest.

TechCrunch verified a portion of the data as authentic by matching the names of several known Homeland Security employees against an offline copy of the database. The thousands of employee directory records appear to have been exfiltrated around mid-August, according to the timestamps on each person’s record. The number of leaked records is equivalent to about one-fifth of Homeland Security’s workforce, which totals around 220,000 staff.

It’s unclear whether the vulnerability that allowed the breach has been fixed or if Homeland Security is aware of the data theft.

A spokesperson for Homeland Security did not immediately respond to TechCrunch’s request for comment about the breach.

When reached, Okta said it was aware of this report. “Okta is secure, and we have no evidence our data or systems were compromised,” a spokesperson said.

While the leaked data includes identifiable information about agents and officers, the dataset also reveals more about the surveillance technologies and cybersecurity tools that Homeland Security employees may be able to access. Critics have called for investigations into tech companies that fund or provide their technology to ICE.

The leaked data contains information about employees’ membership in user groups within the department’s Active Directory, a system that controls access to the department’s internal tools, services, and third-party offerings used for work.

Most of the Homeland Security employees in the dataset were members of the user group that had access to Microsoft 365, the tech giant’s cloud service that hosts the agency’s emails and files, as well as user groups for accessing Adobe Acrobat, Citrix, and other typical desktop software.

The data also shows that around 4,500 users, including ICE agents and deportation officers, are members of a group called “Headspace_users.” Online therapy provider Headspace previously had an $82,500 federal contract with ICE that ended in 2023 to fund “mental health software required for ERO employees,” a reference to ICE’s deportation unit.

A spokesperson for Headspace said the company has not provided any services to DHS or ICE since 2023.

Dozens of other user groups reference Penlink, a web surveillance company that makes technology it claims can track the location data of hundreds of millions of mobile devices. Penlink has had a contract with ICE for years. The user groups vary in size, from a handful of users in local Homeland Security offices around the country to more than 2,000 users across ICE’s HSI unit nationally. One group specifically references Penlink PLX, which the company says is used to capture and analyze “real-time intelligence” from phone calls, texts, web activity, and social media.

A spokesperson for Penlink did not respond to a request for comment.

One user group also includes more than 200 CBP users associated with Mobile Fortify, a facial-recognition app used by both CBP and ICE agents to identify people and verify their information and immigration status.

The user groups also include several employees associated with ChatGPT access, including senior DHS officials, as well as other groups associated with investigations into the China-backed hacking groups known as Salt Typhoon and Volt Typhoon.

Researchers and public information advocates say that the exposed data could provide valuable leads for investigations into alleged abuses by the Trump administration, such as identifying email addresses of agents that can be used in public records requests under the Freedom of Information Act.

“Email addresses and phone numbers are indispensable tools for FOIA researchers, often making the difference between a vague records request that gets stonewalled and one that gets results,” said Ryan Shapiro, the executive director of transparency nonprofit Property of the People.

“Given the Trump administration’s contempt for even basic transparency, the need for such tools is urgent,” said Shapiro.

DDoSecrets co-founder and journalist Emma Best said that the exposed Homeland Security data will be a “catalyst for investigations, empowering journalists, Congressional inquiries, legal challenges, community action and Freedom of Information Act requests.”

The financials Anthropic shared with investors this week suggest that its AI coding tool, Claude Code, is every bit the business success the frontier lab wanted. The company recorded an operating profit in the second quarter, and 900% net revenue retention — that is, spending from the same customers in Q3 2025 grew nine times in Q3 2026.

Anthropic has focused on getting enterprise users to stick around and Wednesday announced a new suite of reporting tools designed to give those customers a more granular understanding of the value they get from their token spend.

Using internal analytics and employee surveys, managers can view a team’s work over a given date range, what the team produced, how much it cost, and where its workflows broke down. Anthropic says it won’t use this data for model training by default. The reports will also offer recommendations on how to improve the results.

Image Credits:Anthropic Image Credits:Anthropic Image Credits:Anthropic Image Credits:Anthropic Image Credits:Anthropic Image Credits:Anthropic Image Credits:Anthropic

This is a clear attempt to demonstrate the return on investment of Anthropic’s tools and to show customers how to improve it — a major concern among large customers. Anthropic is also creating pooled group budgets to allow teams to better manage token spending.

Businesses have been hunting for that kind of information as they figure out how best to deploy AI tools. While usage among venture-backed firms is driving token spend, most of the Fortune 500 are not seeing major gains from deploying AI.

Jellyfish, a company that makes tools for software developers, released a new report yesterday based on the use of agentic coding tools by 276,000 of its customers. Claude Code and other tools built on Anthropic models remain the most widely used in the survey, responsible for nearly a third of all proposed code changes submitted for review. More importantly, spending on those tools has grown 10x since the beginning of the year, with their cost now equal to about 2% of total engineering cost.

“Even if you’re saying 2% is still small, that 10x growth is growth that most CFOs didn’t plan for,” Nicholas Arcolano, the chief scientist at Jellyfish, told TechCrunch. We’re certainly getting to the point where people are realizing, okay, it’s not enough to just spend a bunch. You now have to optimize it, make it work.”

Anthropic’s new reports speak to that challenge. Anthropic leads among enterprise customers, but what makes an agentic coding tool truly sticky is difficult to identify. The release of OpenAI’s Astra model this summer seemed like a potential threat to Anthropic, but so far it hasn’t slowed Anthropic’s adoption numbers.

“It took a long time for us to see Codex really pop in our data at all, and it has come up with Astra [OpenAI’s latest model],” Nikolas Albarran, the AI product lead at Jellyfish, told TechCrunch. Looking at X, he continued, you’d get the distinct impression that the competition is more heated. “Everyone in the Valley is like, ‘Oh, Codex is better.’” In reality, Albarran continued, “Anthropic was just there with great enterprise motions, and they had an incumbent weight of their own — this stuff just worked and worked well, and a lot of people just put a lot of effort into standardizing workflows around it. And there’s not a big enough reason to switch.”

It’s not just a question of models and workflows, either. The realities of business mean that once a frontier lab has gotten its hooks into a customer, switching is harder than it is compared to an individual programmer or small organization. Albarran points to the investment all the major AI labs are making in forward-deployed engineering (FDE) organizations, teams that embed AI consultants inside customer organizations to help them get off the ground. These new reporting tools can be seen as a kind of FDE-lite, a lighter version of that hands-on help, for Anthropic customers.

“We’re at a point now where it’s mature enough, where once you’re in, it’s going to be harder and harder to switch because you’ve spent the effort,” he said. “Even if they’re all mostly the same, they’re all going to be slightly different, have their own slight quirks of how they handle agent.md files, etc., and so that cost to switch is not zero.”

After years of resisting, Apple today launched a MacBook Pro with a touchscreen starting at $2,899. The release comes nearly 10 years after the company first tried integrating touchscreens on its laptops, with its Touch Bar on a MacBook Pro, which it later went on to remove.

The new MacBook Pro more fully embraces the touchscreen, offering a new PROLED XDR Display in both 14-inch and 16-inch screen sizes. The device is powered by M5 Pro and M5 Max chips as options to choose from, along with up to 128GB of unified memory and up to 16TB of storage. The display has 2,000 nits peak HDR and SDR brightness and a 10,000,000:1 contrast ratio.

To make the new MacBook easy to use, Apple designed a hinge that stabilizes the screen when users touch it. This also helps to stabilize the camera when you are on a video call and trying to access content on the display.

In addition, the screen will include a Dynamic Island-style cutout instead of the notch that’s present on most of the current MacBooks. Plus, the new MacBook features a six-speaker array with a new woofer design for better sound.

The 14-inch Pro weighs 3.0 pounds (1.38 kg), and the 16-inch Pro weighs 4.0 pounds (1.82 kg) — lighter than its non-touch predecessors.

Both feature three Thunderbolt 5 ports, an HDMI port, an SDXC card slot, MagSafe, and a headphone jack. Users can hook up three external displays on the M5 Pro variant and up to four displays on the M5 Max variant.

The Touchscreen MacBook Pro 14-inch starts at $2,899, and the 16-inch version starts at $3,399.

Consumers can place a preorder for the new touchscreen MacBook Pro from October 30, and the device will be available starting November 6.

The new Touchscreen MacBook marks two marquee product launches for new CEO John Ternus, who also introduced the company’s first foldable, the iPhone Duo, in September. Various Apple executives, including Steve Jobs and Tim Cook, have scoffed at the idea of a touchscreen Mac in the past. Jobs famously called the idea “ergonomically terrible” at one point. However, rumors of the company working on such a variant started circulating nearly three years before today’s introduction.

The launch follows yesterday’s unveiling of a new standard MacBook Pro with the M6 chip in 14-inch size starting from $1,999.

At the Android Dev Summit on Wednesday, Google’s smart glasses partner Xreal launched its new Xreal Aura wired smart glasses, powered by the Android XR and Qualcomm’s Snapdragon Reality Elite platforms. The companies had teased the glasses earlier in the year at Google’s I/O developer conference.

As Meta expands its push into smart glasses, competitors like Apple and Google are keen to release their own devices. Much like it did with Android smartphones, Google is relying on an assortment of partners, including Samsung and Xreal, to bring devices powered by its Android XR platform to consumers.

The new Xreal Aura smart glasses are tethered to a smartphone-like device for processing power, rather than handling all the computing themselves. The glasses then display the content in the wearer’s field of view.

The device’s design, meanwhile, is closer to Meta’s smart glasses than Apple’s Vision Pro, which is a larger headset.

Xreal Aura uses Sony’s Micro-OLED display with 1200p resolution at 120 Hz and a 95-degree field of view — screen specs that put the glasses above the Ray-Ban Meta Display on paper. However, Xreal’s smart glasses are heavier at 95 grams compared to Meta’s 69 grams, and you have to carry the puck.

At launch, the Aura supports many popular Google Play apps and games along with specific XR experiences, including the NBA App for Android XR; Nevatars; Fallout: Factions; Project Hail Mary: Journey Among the Stars; the NASCAR XR App; and the San Francisco Ballet Company’s “The Nutcracker.”

Users can opt to plug a computer or handheld gaming console directly into the Xreal Aura via USB-C to make it an augmented display. The device can also allow for various entertainment and productivity use cases by spinning up a multimonitor setup or a big virtual screen.

For audio, the Xreal Aura uses Bose’s tech to power its sound while watching videos or listening to music.

Given that the new smart glasses are powered by Android XR, Gemini is available to users as a Q&A assistant and can help them understand on-screen context.

The Xreal Aura is available to order starting today in the United States, Canada, Japan, and

South Korea. The 12GB + 256GB version costs $1,279 with a limited launch price of $1,099 through December, and the 16GB + 512GB version is priced at $1,499 with a launch price of $1,299 through December. The company said it will start shipping by mid-November.

In May, Xreal CEO Chi Xu told TechCrunch that building a smart glass business is hard and everyone is losing money. He also said that the Aura glasses will have a wide range of use cases, and the industry has reached an inflection point of having the right hardware, software, and design smart glasses to reach scale.

“You need all the key pieces ready — you need the hardware ready, the operating system needs to be ready, and then you need a great user interface,” Xu said at that time.

Microsoft has never disclosed Azure’s quarterly revenue in dollars. It has typically only shared a growth percentage or described revenue dollar milestones in vague ways. But on Wednesday, when it posted another record-breaking quarter, it reported that Azure brought in a healthy $33.2 billion in revenue for the July–September quarter, up 48% from a year earlier.

The disclosure lets investors line up Azure’s performance more directly against Amazon Web Services and Google Cloud, whose parent companies already report cloud revenue in dollars every quarter.

Alphabet, which also reported earnings on Wednesday, said Google Cloud brought in $30.7 billion in the same quarter, putting Azure slightly ahead. But Google Cloud, which also includes Google’s Workspace subscriptions, roughly doubled its revenue from a year earlier, growing about twice as fast as Azure, according to TechCrunch’s calculations.

Azure, where businesses rent computing power to run applications, SaaS, and to build and run AI models, is now Microsoft’s largest and fastest-growing product line by the company’s own figures. It accounted for roughly 36% of Microsoft’s $92.3 billion in quarterly revenue.

Azure’s revenue was also about $850 million more than the $32.3 billion analysts expected. Azure’s 48% growth rate also topped the 44% to 45% that Microsoft CFO Amy Hood forecast on the company’s July earnings call.

Azure sits inside Agents and Infra, one of two new business segments Microsoft introduced this quarter. Microsoft previously reported three segments, Productivity and Business Processes, Intelligent Cloud and More Personal Computing. The new Agents and Infra segment, which groups Azure with Microsoft 365’s cloud business, brought in $76.8 billion, up 25% from a year earlier.

Amazon reports its earnings on Thursday. Analysts also expect its cloud, Amazon Web Services, to grow. It’s been the largest cloud since the beginning of time and no one expects Microsoft to best it right now. But with all of them reporting actual dollars, the race is on.

Reddit is the latest social platform to adopt the “For You” feed popularized by TikTok. The company announced on Wednesday that users will now see two options when opening the platform: “For You” and “Following.”

The “For You” feed combines posts from communities users have joined with recommendations tailored to their interests. Reddit says the feature will help users discover relevant communities and conversations they might not have otherwise come across. The “Following” feed only displays posts from communities they have joined.

The company says the change is designed to foster connection, participation, and belonging.

Reddit is likely hoping that the redesigned feed experience will nudge users to engage with content beyond the communities they already follow. It also brings Reddit more in line with the recommendation-driven experience that’s become standard across social media, where algorithms play a significant role in shaping what users see.

With the launch, Reddit joins a number of popular platforms that have adopted “For You”-style feeds, including Instagram, X, Threads, and Facebook.

Of course, not everyone will welcome the change, as some users may prefer the more focused feed experience of only seeing posts from communities they’ve joined, rather than having recommended content mixed into their home feed. Some users who encountered the new feed during its initial testing phase have already expressed their frustration with the change.

Reddit says users can switch between feeds and return to the one they last viewed the next time they open the platform.

It’s worth noting that although Reddit has long recommended posts from outside users’ subscribed communities, the new layout gives those recommendations their own feed on the homepage.

“This new feed experience is a first step in helping drive discovery that leads to something novel and relevant – an idea, a connection, a unique perspective, or even a new interest,” the company wrote in a blog post. “A more human internet is not just something you sit back and watch, it is something you find value in and want to participate in.”

The new feed experience is rolling out globally over the next few weeks, the company says.

Former Yahoo CEO and early Google engineer Marissa Mayer is catching a lot of flak for how she uses AI. In an interview about her new app Dazzle, an AI assistant that learns from your camera roll, Mayer poses what she believes is an impressive flex for the product: It helped her come up with three bullet points about her 10-year-old daughter for a birthday presentation at school, describing her as “empathetic, creative, and an animal lover.”

It’s a vague description, like a generic horoscope — it could probably describe a lot of 10-year-olds.

Some onlookers, like progressive activist Leah Greenberg, marveled that “the fact that Marissa Mayer would voluntarily disclose this to a reporter is a pretty good demonstration of how deeply the culture of Silicon Valley has diverged from the rest of humanity.”

Perhaps Mayer drew so much backlash because we have repeatedly observed sheltered billionaires suggest that we outsource child-rearing — one of life’s greatest tests of patience, humility, and kindness — to a computer that is incapable of understanding human emotion.

We’ve been primed for this frustration. Last year, OpenAI CEO Sam Altman told Bored Ape owner and late night host Jimmy Fallon on national television, “I cannot imagine … figuring out how to raise a newborn without ChatGPT.” The public didn’t love that comment, either, but Altman wasn’t fazed by the negative feedback. A few months ago, he said on X that a “cool use case” of ChatGPT Work would be creating a podcast that talks about your kids’ upcoming birthdays or soccer games and “explain[s] your kids’ interests.” (He was brutally ratioed with comments like, “What if you just talked to your children?”)

After publication, Mayer defended herself in a long X post. She claims she was not relying on Dazzle to explain her child to her, but rather, she was testing what her prototype could do — it resurfaced photos of “forgotten kindergarten projects,” like a “snake devenomer” or “sandcat playground,” illustrating her daughter’s creativity and care for animals.

Fair enough. But Mayer remarks in the same interview that people could also use her AI to outsource the mental load of “figuring out the theme for your child’s birthday party.” She says she’s used her app to plan a playdate for her daughter, book family activities on vacation, and decide what to be for Halloween. She even proposes that her AI assistant could coordinate plans with her husband’s AI assistant. When the reporter asks her if it’s depressing to imagine this “frictionless future … planned by AI,” Mayer replies, “I haven’t really thought that through.”

These leaders seem to write off the naysayers as part of a small, angry fringe group. Yet a Pew Research Center study reports that 50% of U.S. adults think that AI will reduce people’s ability to form meaningful relationships. Only 5% of respondents said AI would make it easier to build these connections.

Meta CEO Mark Zuckerberg, however, joins his Silicon Valley kin in envisioning a world where our social needs are met via AI.

“The average American I think has, it’s fewer than three friends, three people they’d consider friends, and the average person has demand for meaningfully more, I think it’s like 15 friends,” Zuckerberg said last year on a podcast. “The average person wants more connectivity, connection than they have.”

Zuckerberg predicts that it will become more socially acceptable to outsource our emotional needs to AI therapists or even AI romantic partners. “There are all these things that are better about physical connections when you can have them,” he added. “But the reality is that people … feel more alone a lot of the time than they would like.”

Zuckerberg goes on to explain how virtual-reality embodiments of AI avatars are only growing more realistic. Meta seems to be more keen on improving our relationships with these approximations of humans, rather than between the actual billions of humans who use its platforms each day.

Maybe part of this stems from the culture of the tech industry, which prizes efficiency. People try to biohack their own bodies by optimizing their sleep schedules with health trackers or injecting unregulated “peptide stacks” to maximize their productivity.

If you’re building electric vehicles or computer chips, then sure, efficiency is a clear north star. But our relationships with other people are not a tech product that we’ve been challenged to ship with as little capital as possible. Human relationships are inherently messy, but our connections with one another can be what gives our lives meaning. When we delegate those interactions to computers, we surrender the difficult, yet fulfilling experience of being people.

The Trump administration is giving Constellation Energy $5 billion in loans to expand nuclear generating capacity at eight power plants by 1.1 gigawatts.

Most or all of the additional power from seven of those facilities has already been bought by Google, Meta, or Amazon.

Constellation said it plans to use the financing to “uprate” 11 existing reactors, an industry term for expanding capacity at a reactor, often by using more highly enriched uranium and by upgrading equipment to handle the increased power levels.

The energy company said it had already committed to spending more than $7 billion to add 1,080 megawatts of nuclear capacity to deliver power to Google, Amazon, and Meta. The loan commitment, which has yet to be finalized, “could lower financing costs” for the projects, Constellation said in a press release.

The eight power plants include six in Illinois (Braidwood, Byron, Clinton, Dresden, La Salle, and Quad Cities), one in Maryland (Calvert Cliffs), and one in Pennsylvania (Limerick).

Google is already the confirmed buyer of additional power from the Pennsylvania plant and four Illinois power plants, while Amazon is buying electricity from the Maryland plant and Meta is effectively buying the entire Clinton power plant in Illinois.

On Tuesday, the Department of Energy touted the loan as supporting “affordable, reliable power to American homes and businesses,” though there is no evidence that the new nuclear capacity will lower prices for residential ratepayers. Constellation’s announcement does not say whether the projects will maintain or lower electricity rates, and the company would not confirm the government’s claim to TechCrunch.

Hyperscalers, including Google, Amazon, and Meta, have been scouring the grid for additional capacity to support their AI data centers. Tech companies are paying handsomely for energy, committing tens of billions of dollars over the coming years. Given that power has become a competitive advantage, that estimate is probably on the low end.

By lowering the cost of capital for the upgrades, the Department of Energy’s loans could boost Constellation’s bottom line, that of the hyperscalers — or both.

South Korea’s financial institutions have been hit by another wave of cyberattacks, highlighting the growing risk of AI-driven hacking in Asia. The latest incidents affect brokerages Daishin Securities and Mirae Asset Global Investments, as well as the crypto exchange Bithumb.

The incident comes roughly a month after a series of cyberattacks targeting South Korean financial institutions, raising concerns about vulnerabilities beyond just the country’s banking systems. Security researchers linked the September wave of attacks on South Korea to a suspected China-based hacker who relied on an AI penetration testing tool called Artex, though officials are still investigating the root cause of the latest breaches.

According to local news, Daishin reported a cyberattack that exposed the personal information of 38,490 customers, including names, email addresses, and mobile phone numbers. The company has notified affected customers, is cooperating with authorities in the investigation, and will compensate for any losses as required by law.

Meanwhile, Mirae and Bithumb reported separate cybersecurity incidents on Wednesday, per local media, though neither has confirmed that customer data was taken.

South Korea’s financial regulator has called for stronger cybersecurity measures following a string of attacks on financial firms and government agencies, per officials in Seoul on Wednesday. The country also plans to conduct security checks on dozens of IT systems in the private sector.

Yesterday, Japan reported a cybersecurity incident involving GMO Internet, following a series of attacks on its financial institutions. The wave of cyberattacks across Asia underscore growing cybersecurity challenges in the region, often as a result of years of weak cybersecurity; though, investigators have not established a connection between the incidents.

South Korea’s Financial Services Commission (FSC), Daishin Securities, Bithumb, and Mirae Asset Global Investments did not immediately respond to requests for comment.

In Brief

Nvidia on Wednesday released Nemotron 3.5 Transcribe, a free, open-weight model that transcribes audio in up to 40 languages. It also released version 2 of Granary, its open source speech dataset.

Nvidia boasts that Nemotron 3.5 can translate 31 languages to and from English, handle recordings up to an hour long, and tell up to eight speakers apart. It claims to beat rivals on benchmarks, including Hugging Face’s Open ASR Leaderboard.

These tools target developers building apps for business meetings and medical transcription, two of AI’s most established footholds in the workplace. Nvidia says developers can use the model, the updated Granary dataset, and tools like its NeMo Speech model to customize apps to recognize an industry’s jargon in each language. One drawback: Transcribe works with recorded audio, not livestreams.

Still, the company promises Transcribe, at 30 billion parameters, can run on a single chip, from the Jetson AGX Thor robotics controller and DGX Spark chip powering new laptops, to Blackwell datacenter chips. It joins Nvidia’s open-weight voice models, including the super small yet powerful Parakeet (1.1 billion parameters).

Nvidia’s latest voice entry comes during a busy period of AI voice releases: Mistral shipped Voxtral 3 earlier this month; ElevenLabs launched Realtime for instant speech in voice agents; and Salesforce released Launa-Omni, its voice model.

Amazon unveiled five new Alexa devices today, including two new kinds of hardware: the Alexa Photo Frame, a smart picture frame, and the Echo Time, a smart alarm clock.

The Alexa Photo Frame is Amazon’s foray into the world of digital photo displays, a product category that has transcended from a clunky, last-minute holiday gift to a sleek, attractive way to look through your camera roll.

Aura is responsible for much of the growth in this category, so it’s no coincidence that the Alexa Photo Frame has a similar aesthetic feel: The screen is much easier on the eyes than the old, bright photo frames from years past. The device is said to be specifically calibrated for sharing photos, featuring a semi-matte, anti-glare display that makes it look more like a Samsung Frame TV than a laptop screen.

Also like Aura, the Photo Frame’s 10-inch screen looks like a matte photograph, helping it blend in with other artwork on your wall.

Amazon’s new device, though, adds the value of integration with users’ existing Amazon accounts. Prime members get unlimited storage on Amazon Photos, and Alexa+ is supposed to intelligently choose how and when to display photos.

“It knows to choose photos where you’re smiling and the lighting is right, and it matches the season you’re in, so you see last year’s family Halloween costumes in October instead of the middle of spring. It also knows what to skip, like screenshots and receipts,” Amazon said in a press release. You can also add voice notes to photos, and images sync automatically as you take them.

The Alexa Photo Frame is on sale now, and starts at $199.99 in Black, and $229 in Birch or Walnut wood.

Image Credits:Amazon

Meanwhile, the Echo Time clock seems to be a response to the growing sentiment that people don’t want to rely on their phones for alarms, setting morning reminders, or checking the weather.

“Millions of customers start and end their days with Alexa — checking the weather and turning on the lights from bed in the morning or winding down with white noise at night,” Amazon wrote. “But the bedroom asks for something different from a device than the kitchen or living room does — so we designed an Echo specifically for it.”

Amazon says that the Echo Time, which costs $199.99, or $209.99 for a Kids version, has opt-in sleep sensing tools. You can choose to set a smart alarm feature that wakes you up within a 10 to 60 minute window of your alarm time, and can wake you when it detects you’re in a lighter phase of sleep, so that it’s less jarring to begin your day.

Like the popular Hatch alarm clocks, the clock can also mimic sunrise features to ease you awake.

Image Credits:Amazon

Amazon also showed off updates to its existing lines of devices. We now get the Echo Show 16, a new Echo Dot, and Echo Studio with Home Camera.

The biggest upgrade to the Echo Dot is that it now includes a built-in smart home hub, which also includes an eero Wi-Fi extender that adds up to 1,000 square feet of coverage, helping it connect to lights, locks, and thermostats around your home. The new Echo Dot is also powered by Amazon’s custom AZ3 silicon chip, which is said to make it better at on-device processing to speed up conversations with its AI, Alexa+.

Amazon says its most advanced speaker, however, is the Echo Studio with Home Camera. This speaker doubles as a camera so you can check on your home while you’re out, and also has visual intelligence features.

“Alexa can also see and understand what’s in front of the camera and help you in real time,” Amazon explained. “Pull out a board game box and ask Alexa+ to explain the rules or hold up a cereal box and Alexa can recognize it and help you reorder more flavors.”

The Echo Dot is available at $99.99, and the Echo Studio with Home Camera costs $249.99.

Finally, Amazon unveiled the new Echo Show 16, which includes a 16-inch display. Amazon says the device has a 10% larger viewing area than its predecessor, but with a 40% smaller frame. It has Fire TV built in, plus Smart Home tools like Family Organizer. It can also function like a digital photo frame.

The Echo Show 16 is available for preorder for $499.99.

In Brief

TikTok has long been a go-to place for discovering new spots, with viral videos helping local restaurants attract new customers. Now, TikTok is making it easier for users to find their next favorite restaurant with a new feature called “Must Eat.”

Introduced Wednesday, the community-driven guide highlights restaurants getting attention on the platform, bringing food recommendations from across the U.S. into one place.

The move puts TikTok in more direct competition with platforms like Yelp and Google Maps. But instead of relying mainly on search results, maps, and traditional reviews, TikTok is building on what it already does well: connecting people with restaurants through recommendations shared by other users.

“Must Eat” uses more than 40 million posts and comments to pull restaurant recommendations and curated lists together in one place. Users can browse restaurants by city and filter results by cuisine, price, rating, and location, making it easier to compare options before heading out.

The recommendations are split into two categories: “Everyday Eats,” featuring casual favorites and neighborhood go-tos, and “Elevated Dining,” highlighting restaurants for celebrations, special occasions, or a night out.

Additionally, select restaurants can be booked directly through TikTok via OpenTable.

The feature is launching first in New York City, Los Angeles, and San Francisco. Users can find it through the search bar, the Local feed, restaurant pages, and location links attached to videos featuring participating restaurants.

In Brief

Google’s Chrome browser is getting an update that will improve its autofill capabilities using AI. On Wednesday, the company said that Chrome will get an advanced version of autofill that will help users fill out online forms with suggestions pulled from their Gmail.

That means Chrome could suggest things you might need to know, like a flight confirmation number, membership ID, tracking information, order number, address, phone number, or credit card number by referencing your email.

The update is an expansion of Google’s Personal Intelligence, which already allows users to connect their Google Apps to its AI assistant, Gemini, to retrieve certain details and then use them to answer complex questions.

Though seemingly minor, the update could make Chrome easier to use, which could give users another reason to continue to use Google’s browser and stay in its AI ecosystem. It’s reminiscent of one of Apple’s most appreciated features: the ability to automatically detect verification codes sent via text and display them above the keyboard, allowing users to enter them with a single tap.

As with other Personalized Intelligence features, users can decide whether or not to enable enhanced autofill from their Chrome settings. Sensitive data is encrypted, Google notes.

The feature is rolling out to U.S. users on the desktop, and in the iOS and Android apps for Gemini in Chrome.

Google is making it easier for businesses to connect their apps to its AI experiences across Search and Gemini with the debut of its Connected Apps Platform.

Introduced in Gemini and Google Search earlier this year, connected apps let users securely link and interact with popular third-party services, then use them to take specific actions. For instance, you could connect to Canva to design a flyer, or to YouTube Music to build a playlist.

We’ve seen varying takes on the idea by other consumer-facing AI services like ChatGPT and Meta’s Muse. The premise is that such integrations could ultimately shift consumer behavior away from launching and using individual apps to instead using their favorite services through their preferred AI provider.

However, to successfully capture and retain consumers’ attention, there have to actually be a wide range of connected apps to explore.

Google’s idea is to open its doors to any business or developer that wants to reach users through Search or Gemini by allowing them to submit their own connected apps and monitor their performance.

The platform preview is currently available in beta to developers in the U.S., and will allow developers to build and submit apps that can be used across Google’s products.

To get started, developers would integrate with a cloud-hosted MCP (Model Context Protocol) server or Android MCP using the AppFunctions APIs, then use the Google Connected Console to register, test, and submit their app, Google explains.

Essentially, this process lets devs connect their app to Google’s services so it can be used with the search giant’s AI features. Developers can also choose whether they want their app to be available via Search, the Gemini app, or both.

After the app is live, developers can track traffic, latency, and error rates, and maintain control over their brand identity, such as their verified logos and authenticated accounts.

Similar to how SEO allowed companies and websites to show search engine rankings, this could well be one of the initial steps for businesses that want their services to be more visible to users on the agentic web.

Google already has over 50 Connected Apps available in Gemini, including from partners like Adobe, Experian, Localiza, Peloton, and others, in addition to its own expansive set of services, from Workplace and Photos to YouTube Music and Google Home.

These apps will now come to Google Search over the next few weeks. Search today has fewer integrations, featuring apps like Instacart, Canva, YouTube Music, Yahoo Fantasy Football, and others.

Google said meditation app Calm is another new partner on this effort, and that its app will soon launch on both Gemini and Search.

The new Connected Apps platform is in public preview for U.S. developers who sign up for the waitlist. Google did not share when it would become more broadly available.