The U.S. has sanctioned a startup and a venture capital firm from China connected to Linwei Ding, a Chinese national and former Google software engineer sentenced to prison last month for stealing Google’s AI-related trade secrets.
The startup, Shanghai Zhisuan Technology, is an AI company Ding ...
Two of the hardest problems in AI right now are creating models that understand the physical world—so called world models—and making sense of the internal workings of AI models—otherwise known as mechanistic interpretability.
Typically, these are treated as two separate problems: researchers working on world models are often not concerned with their internal workings, and interpretability researchers tend to focus on the language models that power most AI tools we currently use. But Sonia Joseph, the co-founder and CEO of a new startup, World Mechanics, argues that world models and interpretability are two sides of the same coin. Joseph joined the latest episode of AI Deep Dive to explain that connection, which could be key to making AI useful in the physical world, from medicine to manufacturing. (The full episode is available here.)
These days, raising a new venture fund can seem like an uphill battle. First-time managers are competing with much bigger funds that can lure founders with rich valuations—as well as track records that can wow limited partners.
But what if you had unfettered access to some sage advice from a seasoned investor and executive?
That’s the case for Ayesha Arora-Hughes, the daughter of Palo Alto Networks CEO Nikesh Arora. Ayesha is raising her own VC fund, Permanent Partners Collective Fund I, according to a securities filing and people with knowledge of the matter—and she counts Nikesh as an adviser to the firm, one of the people said.
Apple sure takes its own sweet time. On Monday, the company sent out invitations to a product announcement happening next Wednesday, which the company’s marketing chief teased in a message on X: “Can’t wait to pro and tell next week.” That’s likely a reference to a new MacBook Pro from the company, which Bloomberg has reported will be Apple’s first touchscreen laptop.
Apparently pigs really do fly. For over a decade, Apple has treated touchscreen laptops like some kind of mutant offspring of computers and tablets, a category it had zero interest in. On a 2012 earnings call, Apple’s then-CEO Tim Cook had a memorable response to a question about the growing number of touchscreen Windows laptops coming on the market: “You can converge a toaster and a refrigerator, but those things are probably not going to be pleasing to the user.”
“We’re not going to that party,” Cook added.
A number of things have changed since then. Touchscreens have become somewhat normalized as laptop features. About one in every five notebook computers sold has such capabilities, by one estimate. More recently, Cook handed off the Apple CEO baton to John Ternus, a longtime hardware engineering executive who is much more of a product guy than his predecessor.
Clearly, a new touchscreen laptop had to have been in the works long before Apple’s leadership transition. Still, it wouldn’t be a surprise to learn that Ternus’ rising stature within the company had something to do with it. Whoever was behind the decision, the first wave of products from the Ternus administration—the iPhone Duo foldable smartphone, HomeView smart display and now possibly a touchscreen MacBook—seems to be sending the message that Apple is ready to embrace categories it previously ignored. That’s a good thing.
Check out today's episode of TITV in which Akash Pasricha speaks to the CIO of Palo Alto Networks Meera Rajavel about the company’s AI agent security strategy.
Start your day with Applied AI, the newsletter from The Information that uncovers how leading businesses are leveraging AI to automate tasks across the board. Subscribe now for free to get it delivered straight to your inbox twice a week.
The Information's WTF Summit returns to Napa Valley, bringing together the women shaping tech, media and finance. Confirmed speakers include Whitney Wolfe Herd (Bumble), Anne Wojcicki (23andMe Research Institute), Kim Posnett (Goldman Sachs) and comedian & entrepreneur Whitney Cummings, with more to come. Expect candid conversations and a room built for meaningful exchange. Tickets are selling fast—reserve your spot now.
More detailsJessica Lessin, The Information's Founder & Editor-in-Chief, and Whitt Butler, EY Americas Vice Chair – Consulting, will lead an intimate executive discussion on the latest trends in AI. Request an invite.
More detailsThe Information's Dakin Campbell will host a panel discussion on the evolving landscape of public listings and exit opportunities during SF Tech Week, sponsored by NYSE.
More details
OpenAI CEO Sam Altman last week told employees about three proposals the industry has discussed for how to safely develop AI, a stark framing that comes just as the Trump administration sets up its first system for policing “superintelligence” developers like OpenAI.
One path involved developing an international organization, similar to the International Atomic Energy Agency, that would enforce safety standards, a scenario Altman appeared to support, said two people with knowledge of the comments. Another option would involve merging all of the AI labs into a single body to develop new AI safely.
Prosecutors allege that the OpenAI employee accused of bugging OpenAI’s office had ties to the Saudi Arabian royal family, according to a Sunday filing from the U.S. Attorney’s office in San Francisco.
The FBI arrested Saudi Arabian national Abdulrahman Alfozan, 35, on October 9 after OpenAI ...
The Information is pleased to announce new reporters who will carry on the publication’s rich track record of reporting exclusive news on the most important people and companies in business. They’ll report on OpenAI, Google, Asia and finance.
Jemima McEvoy, OpenAI
Jemima McEvoy is changing roles to take on the OpenAI beat.
In just over a year at The Information writing for the Weekend section, Jemima has proven herself to be a resourceful, industrious and sharp-eyed reporter with a flair for colorful features and a knack for breaking news. Her run of illuminating stories includes profiles of Epstein-linked investor Masha Bucher, Midjourney founder David Holz and ultra-secretive investor Zach Frankel. An adept collaborator, she also co-authored great reads on OpenAI President Greg Brockman and his wife Anna, Sequoia’s leadership duo and Anthropic “first lady” Cami Clark—not to mention The Information’s inaugural 50 Most Influential People in Tech list. And she has helped deliver scoops on Valar Atomics and Shield AI and deconstructed trends such as the rise of padel.
Jemima joined The Information after five years at Forbes, where she made her bones writing about billionaires and their businesses. She has also interned for CBS News and was editor-in-chief of Washington Square News, the independent student newspaper at her alma mater, New York University.
Ryan Barwick, Google
Ryan Barwick is joining us to cover Google. Ryan comes to us from Axios, where he spent the past year-and-a-half delivering exclusives as a consumer-industry deal reporter for Axios Pro.
Ryan cut his reporting teeth covering advertising technology, giving him a deep understanding of the economic engine that has long powered Google’s core business. Before Axios, he was the senior reporter for Morning Brew’s advertising vertical, where he landed scoops on companies including Meta, Amazon and Twitter as well as Google, and covered the search giant’s battle with regulators. Prior to that, he wrote about the ad industry at Adweek, was a broadcast associate at CBS News and covered the intersection of policy and technology as a fellow at the Center for Public Integrity.
Claudia Chong, AI and Data Centers (Asia)
Claudia is covering the intensifying AI race from Singapore, where both Chinese and U.S. tech companies are staffing up. She also may report on topics including chip smuggling, crypto and anything else she digs up.
Claudia joins us from S-RM, a UK-based risk advisory firm where she specialized in APAC disputes, investigations, and corporate due diligence. Prior to that, she worked on ByteDance’s LLM training operations in Singapore and spent nearly six years as a correspondent at The Business Times, the premier business publication in Southeast Asia, covering venture capital, tech, and private wealth across the region.
Her journalistic work earned her a Singapore Press Holdings (SPH) Media Award for Excellence in 2024 for the investigation into Singapore’s largest money laundering scandal. Beyond training LLMs and conducting due diligence, her skills also extend to M&A, having conducted research that helped with the SPH’s acquisition of the Tech in Asia publication in 2024.
Alex Eichenstein, Tech Capital Markets
We’re excited to announce that Alex Eichenstein is joining the finance team, covering tech capital markets. Alex comes to us from The Deal, where she spent more than two and a half years covering healthcare deals. She regularly broke PE acquisition news while also writing deeper dives on major investing shifts, including how buyout firms’ struggles to find exits for practice rollups fueled a shift toward health tech investments.
We’re excited for Alex to help deepen our finance coverage in her new role, focusing on the fast-evolving debt and equity markets and deals redefining how tech companies fund themselves. Alex is a graduate of Columbia Journalism School and worked as a social media editor at Protocol prior to that.
Elon Musk pledged on X to give one of his SpaceX shares to every kid in America next year, and will do more such gifts in the future.
According to estimates by the U.S. Census Bureau, about 21.1% of the U.S. population is under 18 years old, which translates to about 73 million children. At ...
DeepSeek is working closely with Huawei Technologies to shape the design of future AI chips, two people familiar with the matter said, in a collaboration that could help China build frontier AI models without Nvidia’s technology.
The relationship goes beyond the kinds of conversations companies have with chip suppliers about their semiconductor road maps. Huawei engineers are incorporating DeepSeek’s requests and suggestions into design decisions, such as how models calculate and move data between processors, for chips that will come out next year and beyond, the people said. To do this work, the Huawei engineers have been stationed at DeepSeek’s offices for an extended period, one of the people said.
In the last month, OpenAI, Anthropic and Google have released more than a dozen new models. If you thought that pace was crazy, don’t expect things to slow down anytime soon.
Anthropic is preparing to release a new model that is better in its reasoning and intuition than the Anthropic models that are currently available commercially, according to a person who has used it. The unreleased model is especially good at discerning a person’s underlying intentions and goals even if they don’t give the model much information, such as explicitly stating those goals, the person said.
For instance, a person might ask the model to fix a single bug, and the model might go off and instead identify the underlying class of problems that caused the bug and ask whether it should fix the whole class of problems, said the person who has used it.
The Anthropic model is also better at thinking outside of the box rather than regurgitating common sense advice or solutions, the person said. For instance, if a user asks for advice on how to invest in stocks, rather than providing generic advice like looking at price-to-earnings ratios, the model might come up with its own creative algorithms to predict future stock performance, the person said.
Emil Michael, a senior Department of Defense official and member of President Trump’s “super intelligence” task force, vowed in an interview to use a range of executive tools including export controls to check the behavior of AI companies if they don’t cooperate with the government, while comparing frontier labs calling for more regulation to drunk drivers who demand someone else pull them over.
Michael drew widespread attention as the Trump administration official leading a public feud with Anthropic, pushing for the Defense Department to label the AI company a supply chain risk after it refused to remove safety guardrails. Now, as one of the four senior administration officials leading the Super Intelligence Force that Trump announced this month, Michael is charged with serving as the government’s liaison to Anthropic and other AI companies to develop protocols for handling cyber breaches, hacks and other critical incidents.
Nvidia has asked the suppliers of its next-generation custom high-bandwidth memory chip to use Taiwan Semiconductor Manufacturing Co’s 3-nanometer manufacturing process, the Seoul Economic Daily reported.
Nvidia asked Samsung Electronics, SK Hynix and Micron to standardize the manufacturing ...
DeepSeek’s latest update to its affordable lightweight model, V4.1-Flash, is making a big splash in the industry because of significant improvements from the previous version that bring it closer to frontier models.
The upgraded version released on Saturday, V4.1-Flash-1017, is delivering highly ...
Tech earnings season is back (yes, I know it feels like it never ends). This time, we’ve got a slight twist: Meta Platforms has broken with the pack and is reporting its September-quarter earnings on Wednesday, a week ahead of its big tech counterparts, Google, Microsoft and Amazon. (Apple is even later, on Nov. 2.) By reporting earlier than the others, Meta ensures it will grab more of the spotlight, for better or worse.
Until last week’s market sell-off, Meta had been riding a wave of enthusiasm on Wall Street, thanks to chatter about its Muse personal agent and recently released Muse Wonder AI models. Things changed last week, particularly on Friday, as renewed jitters about the AI expansion tanked most tech stocks. Given Meta’s gargantuan spending on AI development, this week’s update will provide a new test of investor sentiment toward Meta shares.
Anthropic announced the creation of a new fund that will get at least 5% of its shares and support public-benefit causes such as addressing the employment impact of AI.
The Anthropic Public Benefit Fund could get an additional 2% of the company’s shares on top of the initial 5% commitment, at ...
Microsoft has pulled out of a planned U.K. data center Nscale was building for it, an Nscale spokesperson said, after the cloud provider faced delays accessing power to bring the site online.
The cancellation is likely to weigh on Nscale’s planned public offering this month as investor concerns ...
How many young AI cloud providers can the market, and Nvidia, support?
That’s one of the big questions after a week in which investors worried about the data center buildout and the risks that its unproven participants, known as neoclouds, are taking. Those risks include loading up on debt (CoreWeave) or trying to go public without much of a track record (Firmus and Nscale). Firmus abruptly pulled its public offering last week while shares of CoreWeave and its smaller rivals fell around 20%. The concerns even hit bigger AI cloud providers like Oracle.
Data center and neocloud executives told me there’s a limit to the number of AI cloud providers that can exist, even if Nvidia provides many of them with investment or backstops.
The Trump administration has opened the door for entrepreneurs to start new banks in the U.S., many funded by venture capital. But there’s a catch for fast-growing newcomers.
Upstarts like Palmer Luckey’s Erebor Bank, as well as other new entrants like Openreserve and Augustus, will likely have to keep raising money as they grow to maintain the capital cushion they need to satisfy bank regulatory requirements. That’s a contrast to most startups, which use fresh venture money to pay for operating expenses and expand the business.
Regulators require banks to maintain a capital cushion to absorb potential losses and protect depositors. That cushion is typically measured through a minimum tier 1 leverage ratio, which measures a bank’s capital relative to assets including cash, loans and investments. For instance, a bank subject to a 12% minimum tier 1 leverage ratio needs to maintain at least $12 in capital for every $100 in assets.
As a bank attracts more deposits, its assets, such as loans, typically grow too, increasing the amount of capital it needs to meet that minimum ratio. Established banks can build up their capital through their profits, but newer banks that aren’t yet profitable have to rely instead on raising fresh equity from investors.
Startup banks have already been raising significant funding to get their businesses off the ground or as deposits grow.
Erebor Bank is set to raise $1.5 billion in a fresh fundraising round at an $8 billion valuation before the investment, after launching with $625 million in committed capital in February. The fundraising comes as the bank has seen its deposits surge to $7 billion as of the end of September, up from $4 billion at the end of the prior quarter.
Meanwhile, Openreserve, an a16z crypto-backed bank that raised $25 million in September, said it expects to announce an additional capital raise to meet its regulatory requirements ahead of its planned opening early next year.
Then there’s Augustus, a startup bank focused on serving fintech companies. It raised $180 million in a Tiger Global Management–led round in July at a $1 billion valuation and expects to open later this year.
Regulators give capital ratios to each bank based on their assessment of the institution’s business plans and risks. Newly created banks typically face tougher capital requirements than established banks. For the first three years of operation, Erebor and Openreserve must each maintain a minimum tier 1 leverage ratio of 12%, while Augustus faces a 10% minimum. Established banks, by comparison, typically have ratios of between 5% and 6%.
“There’s no question that the regulators treat de novo banks—new banks—differently than they do well-established banks,” according to Brian Graham, a partner and co-founder at consulting firm Klaros Group who advises clients on bank charter applications. Graham noted that the extra capital requirements typically go away after a bank has been operating for three years.
The upstart banks are generally pursuing relatively new business models or focusing on sectors big banks typically avoid. Erebor Bank, for example, has gained traction in areas prioritized by the Trump administration, such as defense startups and crypto firms, and is expanding in geographies traditional banks typically avoid, including Venezuela and Syria.
In addition to making minimum capital requirements, the Federal Deposit Insurance Corp.—a banking regulator—imposed an unusual condition on its approval of Erebor for the deposit insurance program. Under a capital-call provision, Erebor’s investors must contribute additional capital if the bank doesn’t meet its minimum capital requirements, according to a filing. The filing didn’t disclose which investors are required to provide the capital backstop.
In the meantime, upstart banks that are looking to drive revenue growth from lending rather than fees will likely need to tap investors to build their capital.
“To the extent the bank’s business model is reliant materially on the balance sheet for revenue and net income, you’re going to need to raise more capital to drive growth,” Klaros’ Graham said.
Openreserve, for its part, plans to make loans backed by traditional assets and turn them into digital tokens that can trade on blockchain, potentially providing a new source of loans for investors and lenders in the crypto market. For example, it plans to lend businesses money against their trade receivables, turning those assets into digital tokens to make them easier to trade.
“Our ambition is to be the on-chain GSIB,” referring to a globally systemically important bank, said Dee Choubey, founder and CEO of Openreserve. The bank also plans to issue its own stablecoin and to offer tokenized deposits, digital versions of traditional bank accounts that earn interest. Over time, it aims to provide banking and payment services to other digital banks.
Augustus plans to focus on helping fintech, crypto and remittance companies move money across borders using both traditional banking systems and stablecoins. That will put it in competition with existing banks like United Texas Bank, which has become dominant in processing dollar payments for overseas crypto exchanges.
“Whenever someone outside the U.S. touches the dollar through a local bank or local fintech, that always has to be facilitated by a U.S. nationally chartered clearing bank, so we provide that service,” said Augustus CEO Ferdinand Dabitz. In Europe, the bank already counts crypto exchange Kraken among its clients.
Augustus is less focused on activities like lending, however, and has a lower ratio than the other two. The company has raised significantly more than its initial capital requirement, which could lessen the need for more fundraising in the near term. “We are less focused on the credit side as other market participants. We think the clearing side is really where we shine,” Dabitz said.
Welcome, Weekenders!
A song has been stuck on loop in my mind for days now. Yeah, this one.
C’mon, haven’t you heard the big, brassy James Bond theme playing too? We have an honest-to-goodness spy scandal on our hands! That would be the affair over at OpenAI, where the FBI has arrested a former engineer, Abdulrahman Alfozan, for allegedly placing a toilet paper–wrapped recording device in a conference room. Why’d he do it? The Justice Department wasn’t explicit in its press release announcing Alfozan’s arrest, though the statement does make sure to prominently describe him as a “national of the Kingdom of Saudi Arabia,” implying, obviously, that he was hoping to spy on OpenAI for the Saudis.
I expect a not unsizable part of Silicon Valley experienced some schadenfreude when reading about the turn of events at OpenAI. I’d advise against indulging in that feeling, because when I struck up conversations with a few corporate security consultants and surveillance experts over the last few days, they all echoed the same sentiment: Plenty of companies risk ending up in the same situation—especially as they move quickly in today’s era of frenzied, heightened competition. A discomforting thought, eh?
“Organizations can sometimes be moving so fast that they often don’t understand that there’s an issue associated with somebody until that person’s on board and up to no good,” said Bill Priestap, a former head of FBI counterintelligence who now runs his own security consultancy, Trenchcoat Advisors. “Meaning they’re recruiting, and they’re under such pressure to get the right people and get them on board that in many regards, they end up with a minimal understanding of the people they’re bringing on board.” Well, good thing there hasn’t been an enormous, frantic hiring binge lately. Ha ha ha.
If you’re starting to feel a little paranoid, I strongly, strongly advise against talking to someone like David Abbott, a San Francisco private detective, to find out how easy it is to bug an office. Such conversations will only further unnerve you—you might end up holding your next board meeting in your home basement, Tony Soprano style—because according to Abbott, it’s pretty darn easy to install a recording device.
A suitable one can be bought for less than $100. Some of the popular ones resemble phone chargers, which get tucked into a power socket, hidden in plain sight. He’s also seen surveillance devices in fake coffee cups—you know, like the ones that might lie casually around an office—and quite a few in fake smoke alarms.
More bad news: Should you find such a device, it’s like spotting an ant in your kitchen. One ant suggests the presence of more ants. “If I’m doing a bug sweep, just because I find one does not mean there isn’t a second,” said Abbott. “If you find one, there’s going to be two. If you find two, there’s going to be three. You just keep looking.”
What happened at OpenAI underscores Abbott’s warning. There, Alfozan, the alleged spy, did plant two bugs. Well, he tried to do so. Alfozan, not the sort who seems like he would’ve ever earned double 0 status, had placed another bug in a bag that he then mislaid: He reported it missing and had to go pick it up from the company’s lost and found.
OK, if I’ve spooked the heck out of you, I guess I can offer some solace by pointing you in the direction of Eric and Greg Levesque, whose Salt Lake City–based startup, Strider Technologies, has raised $110 million in the last few years to sell AI-powered software to businesses that monitors email and other internal communications, flagging suspicious behavior and possible espionage attempts. (Yes, the company’s name is a reference to Strider from “The Lord of the Rings,” the nom de cloak of Aragorn, the character Viggo Mortensen plays.)
“What we’ve found is that companies are not always dealing with trained intelligence operatives. Like, they’re dealing with scientists or legitimate engineers who have been approached or coerced in some form or fashion and are now working on behalf of someone else,” said Greg Levesque. “They can leave a lot of digital exhaust.”—Abram Brown (abe@theinformation.com)
Listening: “Reaching Out”
As the pandemic shut down the world in 2020, Sayre Quevedo, then a producer for The New York Times’ “The Daily” podcast, quit his job and headed out to Los Angeles, hoping for a chance at self-reinvention. To cope with those tumultuous times, he decided to pursue a documentary project, recording every phone call. Had nothing else of import happened in his life, Quevedo might’ve wound up with an intriguing time capsule. But something quite unusual did occur: Just months after moving to LA, he and his twin brother, Connor, happened to finally locate their biological father, whom they’d never before met. He worked as a barber—blocks from where Sayre lived. “Reaching Out,” a new podcast co-produced by NPR station KCRW that won a top prize in the Tribeca Festival’s expanded podcast categories, is a starkly intimate account of their journey to meet him and their reconciliation with each other, their lives and what life might’ve been like if their father hadn’t been a total stranger who happened to be just down the street.—Abram Brown
Reading: “Sitcom” by Chuck Lorre and “Emperor of Nothing” by Blake J. Harris
Chuck Lorre, the indefatigable TV showrunner, writer and producer, has kept the masses laughing for years—and years—and years, with a list of hits that includes “Two and a Half Men,” “The Big Bang Theory” and “Dharma and Greg.” Those are sitcoms, a term Lorre acknowledges once made him practically break out in hives, since it was often used to cheapen what he sees as grueling, high-touch art. With spicy good humor, he reclaims the word in his new memoir about his time in Hollywood, “Sitcom.”
“The truth is, making a live studio audience laugh for twenty-odd minutes every week, for twenty-two weeks, is incredibly hard and unbelievably stressful,” Lorre writes. “If you’re one of those people who insist on labeling all sitcoms as being cheap, juvenile and unsophisticated, well…all I can say is, blow me.”
I’d say “Sitcom” functions as a good primer on managing talent, too, and Lorre attests to having learned the importance of keeping out of the way. “Creativity needs air to breathe,” he writes. To do the opposite risks turning a leader into “one of those plastic bags you get at the dry cleaners that can kill small children.” Lorre also takes the opportunity to set down what he sees as good comedy’s basic tenets and decry the medium’s basest impulse: to write only about stupid people doing stupid things: “A better, perhaps deeper means to achieving laughter involves pain and suffering.”
Such a sentiment is most certainly shared by Larry David of “Seinfeld” and “Curb Your Enthusiasm,” the subject of “Emperor of Nothing,” a biography that won his reluctant cooperation. (When author Blake J. Harris finally wore David down, Harris received a two-line email from him: “Oh Jesus. If we must.”) As with “Sitcom,” “Emperor of Nothing” charts a remarkable rise across Hollywood from an inauspicious start: Lorre was a wannabe musician before his TV run, while David was a little-known stand-up comic with an unremarkable series of TV writing jobs before co-creating “Seinfeld.”
Like Lorre, David describes the process of penning comedy as an arduous affair. And by David’s own telling, the process of sitting down to write a new “Seinfeld” season would bring on full panic. “That always seemed like a daunting prospect,” he recalled. “Especially because all the good ideas were used up the year before.” Fortunately, David had his friendship with the show’s eponymous star, Jerry Seinfeld, one of many people within David’s orbit who gave interviews for “Emperor of Nothing.” “I would always say, ‘All we have to do is write today,’” Seinfeld said. Or as David recalled: “Jerry always talked me off the ledge.”—A.B.
Watching: “War”
No one does a charming, self-involved cad better than Dominic West, the British character actor who first washed ashore in America as “The Wire”’s Detective McNulty and then again in Showtime’s “The Affair” and in the final two seasons of “The Crown” as Prince Charles. In “War,” HBO’s new high-tension legal drama, he’s tech billionaire Morgan Henderson, a fellow championing a safer, more responsible social media—while embroiled in an increasingly messy divorce from his Hollywood star wife, Carla (Sienna Miller). Much of that mess is Morgan’s own doing: He looks wonderful in his Savile Row trousers—he just has trouble keeping them on. As Morgan and Carla gird for an extended battle, they assemble a posse of cutthroat attorneys, who find their impulsive clients constantly upending their carefully laid strategies. That flock of legal eagles includes a handsome rogue, Johnny (Archie Renaux), a lad from working-class south London who sees a chance to ascend by, uh, redefining zealous representation. Backstabbing and double crosses abound.—A.B.
Login