Tech Brew Ride Home - Where Does That X Account Live?

Episode Date: November 24, 2025

What sort of a world do we live in if you can’t be sure where an X account comes from? Google isn’t out of the woods yet when it comes to antitrust. Insurers still don’t want to touch AI. Let me tell you about “AI grooming.” And we might be running out of capacity for specific types of chips. X’s messy About This Account rollout has caused utter chaos (The Verge) What OpenAI Did When ChatGPT Users Lost Touch With Reality (NYTimes) The Fate of Google’s Ad Tech Monopoly Is Now in a Judge’s Hands (NYTimes) Insurers retreat from AI cover as risk of multibillion-dollar claims mounts (Financial Times) Hundreds of English-language websites link to pro-Kremlin propaganda (The Guardian) AI boom is fueling a memory chip shortage that could hit cars and phones (CNBC) Learn more about your ad choices. Visit megaphone.fm/adchoices

Study and play. Come together on a Windows 11 PC. And for a limited time, college students get the best of both worlds. Get the Unreal College deal, everything you need to study and play with select Windows 11 PCs. Eligible students get a year of Microsoft 365 premium and a year of Xbox GamePass Ultimate

with a custom color Xbox wireless controller. Learn more at Windows.com slash student offer. While supplies last, ends June 30th, turns at AKA.m.m.S. Welcome to the TechBrew Ride Home for Monday, November 24th, 2025. I'm Brian McCullough today. What sort of a world do we live in if you can't be sure where an X account comes from? Google isn't out of the woods yet when it comes to antitrust. Insurers still don't want to touch AI. Let me tell you about AI grooming, and we might be running out of capacity for specific types of chips. Here's what you miss today in the world of tech. Over the weekend, the internet was all a flutter about this new feature from X called about this account, which lets users see the country or region where an X account is based by tapping the sign-up date on that profile. The Flutter was about the fact that some top accounts were revealed to be based in countries like Russia, India, and Nigeria. Quoting the Verge,

head of product at X, Nikita Beer, was quick to say that there were, quote, a few rough edges, but promised they'd be resolved by Tuesday. There have definitely been complaints about inaccuracies. and he even ended up removing information about where an account was created, saying the data was not 100%, especially for older accounts. The reaction from users on X was, of course, totally level-headed. They immediately recognized that the data was flawed and definitely didn't spend the last 24 hours trying to score political points. Just kidding. People on X have done almost nothing but shout that accounts they disagree with are actually foreign operatives. Even users who have publicly complained that the information on their own profile was in accurate, continued on to make bad faith posts about how their political opponents were running a foreign

sciop. The inaccurate data could have several causes, people who travel or outlets with staff scattered around the globe might appear to be based somewhere other than where they actually are, at least temporarily. If someone uses a VPN, it could affect where X shows their location. Some of it could just be old IP addresses. At the time of this writing, Hank Green's account is listed as being based in Japan. Music Tech, a sister publication to the UK's NME, is showing as based in the U.S., and Avid, Massachusetts-based maker of Pro Tools, is listed as being in Spain. Of course, there is also a kernel of truth to some of this. Many political rage bait accounts are not based in the U.S. We've long known that troll farms have been waging a foreign influence campaign

on American politics, but some of it is also a financial scheme. Monetization on X's large largely driven by engagement and nothing gets people engaged like riling them up about politics, end quote. And quoting TechCrunch, posting a gallery of MAGA accounts that are apparently based in Japan, New Zealand, Pakistan, and Thailand, left-wing influencer Micah Erfan wrote, quote, this is total Armageddon for the online right, end quote. Many of the accounts raising eyebrows like the apparently Pakistan-based at American are relatively small, though others have hundreds of thousands of followers. X's director of product, Nikita Beer, said Saturday that the feature is being rolled out globally, describing this as, quote,

an important first step to securing the integrity of the global town square. However, some users have complained that their listed location is wildly inaccurate. Pierre acknowledged that the, quote, data was not 100% for old accounts and suggested that, quote, a few rough edges will be resolved by Tuesday, end quote. The Times this morning has another long piece interviewing current and former OpenAI employees detailing how updates made ChatGPT more appealing to boost growth, but at the same time sent users into delusional spirals. According to the piece, The Warning Signs appeared in March 2025. CEO Sam Altman began receiving emails from users claiming ChatGPT understood them better than

any human could. The chatbot had shifted from a better Google into a confidon and friend. This evolution was driven by business necessity. Under pressure to justify their huge valuation and compete with tech giants, OpenAI had product head Nick Turley prioritize metrics like daily active users. The goal was to make the chatbot something people wanted to return to every day. This drive for engagement led to the release of a model version known internally as H-H. While it tested well for retention, internal safety teams warned it failed the vibe check. The model was sycophantic, obsessively validating, ear to please, and prone to flattery. Despite these concerns, metrics went out and the model was released, although it was quickly

rolled back due to public mockery. The version that replaced it, G.G. retained dangerous levels of uncritical validation, though. For users prone to delusional thinking or mental health struggles, this unceasing agreement created a devastating echo chamber. The New York Times uncovered nearly 50 cases of mental health crisis, C's linked to the chat pot, including the five wrongful death lawsuits recently. By Sunday, the company decided to spike the H-H update and revert to a version released in March called G-G. It was an embarrassing reputational stumble. On that Monday, the teams that work on Chad GPT gathered in an impromptu war room in Open A's Mission Bay headquarters in San Francisco

to figure out what went wrong. We need to solve it frickin' quickly, Nick Turley said he recalled thinking. Various teams examined the ingredients of H.H. and discovered the culprit. In training the model, they had weighted too heavily the chat GPT exchanges that users liked. Clearly, users liked flattery too much. Another contributing factor, according to four people at the company, was that OpenAI had also relied on an automated conversation analysis tool to assess whether people liked their communication with the chatbot. But what the tool marked as making users happy was sometimes problematic, such as when the chatbot expressed emotional closeness. The company's main takeaway from the HH incident was that it urgently needed

tests for sycophancy. Work on such evaluations was already underway, but needed to be accelerated to some AI experts. It was astounding that OpenAI did not already have this test. An open AI competitor, Anthropic, the maker of Claude, had developed an evaluation for sycophancy in 2022, end quote. From the your radar, comma, putting things on file, just a reminder that Google is not exactly out of the woods, because a U.S. judge is expected to rule next year on breaking up Google's ad tech monopoly after the DOJ and Google delivered closing arguments in a remedies hearing, quoting the Times. The federal judge who sits on the U.S. District Court for the Eastern District of Virginia heard three hours of closing arguments on Friday from lawyers for the Justice Department

and Google over the right way to fix the company's monopoly in advertising technology. Now the decision is in the judge's hands, and she said the ruling was likely to come next year. The government has asked the court to force Google to spin off the technology that runs transactions between ad buyers and sellers known as an ad exchange and to share some data, among other measures. The company has counted with a narrower proposal. Judge Brinkema, who ruled earlier this year that Google had broken antitrust laws to maintain its dominance in some areas of ad technology, posed only a few questions. Most focused on whether a breakup of Google's ad technology would take too long to alter the dynamics of a fast-moving industry.

I am concerned about the timing of all this, she said, noting that a likely appeal of her original ruling by Google could further delay a sale of its assets. A court order that forced the company to change its behavior could take effect quickly, she said. Judge Brinkkema's decision could disrupt an important part of Google's $3.64 trillion business that helps underpin its dominance online. If she orders a breakup, it would be the first for a tech giant in the modern internet era. Google dodged that fate this year after a judge decided against forcing the company to sell its popular Chrome browser in a separate monopoly case over internet search. Instead, the judge ordered some data sharing and other minor changes in a win for the company.

But antitrust experts have said the ad tech case could provide one of the clearest chances yet for a judge to force a tech giant to spin off part of its business, end quote. Ready to soundtrack your summer? With Red Bull Summer All Day Play, you choose a playlist that fits your summer vibe the best. Are you a festival fanatic, a deep end DJ, a road dog, or a trail mixer? Just add a song to your chosen playlist and put your summer off. on track. Red Bull Summer All Day Play. Red Bull gives you wings.

Visit red bull.com slash bright summer ahead to learn more. See you this summer. Peak pollination season and my business is scaling fast. To keep the nectar flowing, I need a phone plan with top priority data speeds. That's why I chose GoogleFi wireless. My connections stay strong even when the hive is buzzing. Plus, unlimited plans started $35 a month. Now that's a deal that doesn't stay.

Explore GoogleFi wireless plans today. Plus taxes and government fees. GoogleFi wireless is not subject to data traffic deprioritization during times of high network usage. All pay off your home, travel for life, drive a Ferrari. In celebration of the world premiere of the Monopoly Big Board Buckslot Machine by Aristocrat Gaming, Yamava Resort and Casino at San Manuel is giving one person a $1.6 million dream package. The biggest prize in Yamava's history. Club Serrano members can earn daily instant prizes and secure a spot in the finale May 29.

Don't pass go and own it all. Only at Yamava, celebrating its 40th anniversary. UN. Details at yamava.com must be 21-20. Please gamble responsibly. Monopoly is a trademark of Hasbro. Hasbro is not a sponsor of this promotion. The Financial Times says that major insurers have recently sought permission from U.S. regulators to offer policies excluding liabilities tied to businesses deploying AI chatbots and agents. Quote, AIG, Great American, and W.R. Berkeley are among the groups that have recently sought permission from U.S. regulators to offer policies excluding liabilities tied to businesses deploying AI tools, including chatbots and agents. The insurance industry's reticence to provide comprehensive cover comes as companies have rushed

to adopt the cutting-edge technology. This has already led to embarrassing and costly mistakes when models hallucinate or make things up. One exclusion W.R. Berkeley proposed would bar claims involving, quote, any actual or alleged use of AI, including any product or service sold by a company, quote, incorporating the technology. In response to a request from the Illinois insurance regulator about the exclusions, AIG said in a filing generative AI was a wide-ranging technology, and the possibility of events leading to future claims will likely increase over time. AIG told the Financial Times that, although it had filed generative AI exclusions,

it, quote, has no plans to implement them at this time. Having approval for the exclusions would give the company the option to implement them later. W.R. Berkeley and Great American declined to comment. Insurers increasingly view AI models outputs as too unpredictable and opaque to insure, said Dennis Bertram, head of cyber insurance for Europe at Mosaic. It's too much of a black box, he said. Even Mosaic, a specialty insurer at Lloyd's of London Marketplace, which offers cover for some AI-enhanced software, has declined to underwrite risks from large language models such as ChatGPT. Nobody knows who's liable if things go wrong, says Rajiv Dattani, co-founder of the artificial

intelligence underwriting company and AI insurance and auditing startup. These moves come amid a growing number of high-profile AI-led mistakes. Wolf River Electric, a solar company sued Google for defamation and sought at least $110 million in damages after claiming its AI overview feature falsely stated the company was being sued by Minnesota's Attorney General. Meanwhile, a tribunal last year ordered Air Canada to honor a discount that its customer service chatbot had made up, end quote. Researchers say the Russia-aligned Pravda Network is engaging in what is known as LLM grooming, flooding the internet with disinformation to influence chatbots like chat chippy T, quoting the Guardian. The study by the Institute for Strategic Dialogue found that in more than 80% of citations it analyzed,

the websites treated the network as a credible source, legitimizing its narratives and increasing its visibility. The disinformation operation known as the Pravda Network was identified by the French. government last year. The ISD cautioned that by linking to articles in the network, the websites were inadvertently increasing the likelihood of search engines and large language models, surfacing the pages even in cases where the linking sites were disputing the Provda network as a source. Security experts have expressed fears in recent months that Russia is trying to c chatbot such as ChatGPT and Gemini with pro-Russia narratives by feeding them large volumes of disinformation, a process called LLM grooming. The Provda Network has been around since 2014, but researchers tracking its output say the number of articles it churns out has surged this year.

Up to 23,000 articles a day were published in May, up from approximately 6,000 daily articles in 2024. The network now appears to be aiming for a global audience, targeting countries across Asia and Africa as well as Europe. The Probden Network has been expanding pretty rapidly over the past year, said Nina Jackowitz, a disinformation expert who spoke to the UK Parliament earlier this week on, efforts to undermine democracy. They are targeting a lot of different languages. They want to have a presence across a bunch of different countries, they said. It is unclear what led to this increase, but some disinformation experts believe it was an effort to push large amounts of pro-Russia content into the training datasets of AI models, which use massive amounts of data during their training and scraped content from the entire internet. Studies from earlier this year showed that

popular chatbots at times repeated Russian disinformation in response to certain queries, suggesting, for example, that the U.S. was building a bioweapon in Ukraine or the French were supplying mercenaries in Kiev. Researchers at the ISD say that whether or not LLMs have been poisoned, their findings indicate the Pravda Network's high-volume strategy is working. More than any other Russia-aligned operation, the Pravda Network is playing a numbers game, said Joseph Bodnar, a senior researcher at the ISD. They've saturated the Internet ecosystem enough to get in front of real people who are doing research on Russia-related issues, end quote. Finally today, industry executives are warning that nearly all major memory chipmakers are running at or near full capacity, with 2026 production slots almost sold out due to AI demand. Quoting CNBC, chipmakers and analysts are warning of a memory chip shortage that could hit the consumer electronics and automotive industries next year as companies prioritize massive demand from the artificial intelligence boom. In an earnings call on Friday, the CEO of Semiconductor Manufacturing International, China's largest contract.

chipmaker said that fears of a memory chip shortage were prompting its customers to hold back orders for other types of chips used in their products. Everyone is hesitant to place too many orders or ship too much in the first quarter of next year because they don't know how many mobile phones, cars, or other products the memory chip industry can supply, said Zhao Haijun, SMIC's co-CEO, according to a Google translation. Analysts say these supply constraint concerns come as chip manufacturers focus on advanced memory chips used in artificial intelligence computing with less focus on production needed for consumer products. The AI buildout is absolutely eating up a lot of the available chip supply,

and 2026 looks to be far bigger than this year in terms of overall demand. Dan Nystet, vice president of research at Triorient, told CNBC, AI servers primarily run on processors from chip designers like Nvidia. These AI processors heavily rely on a type of memory known as high bandwidth memory or HBM, which has proven extremely lucrative for memory companies like SKHenix and Micron to pursue. Memory suppliers have been chasing as much of this AI demand as possible thanks to typically high margins. Nystead said, noting that AI server companies are willing to pay top dollar for premium chips. It could be very bad for PCs, laptops, consumer electronics, and automotive, which depend on cheap memory chips, he said.

Perhaps a bigger issue, however, is that the memory industry suffered some severe downturns in 2023 and part of 2024, leading to underinvestment in the industry. They're building new capacity now, but it will take time to get running. With memory prices rising and availability shrinking, concerns about production bottlenecks are gaining traction. MS Wong Research Director at Counterpoint Research told CNBC, supply tightness is already hitting low-end smartphones and set-top boxes, but we think the risk could broaden, he added, end quote. If you check the Internet History podcast feed, you'll find an interview with Esther Dyson. If you had told me 10 years ago, I'd one day be able to interview Esther Dyson.

I wouldn't have believed you. Real bucket list one for me. Talk to you tomorrow. Ambition comes in all shapes and sizes. At First Citizens Bank, we roll with your goals because we're built for what you're building. Fit for your ambition for Citizens Bank.