Tech Brew Ride Home - The “Covid Moment” For AI?

Episode Date: February 11, 2026

Forget the ChatGPT moment, is this the “Covid Moment” for AI, the moment when everything has changed its just that not everyone knows it yet? More weird musical chairs at the AI companies. TikTok but for around the block. And do we finally have a universal translator but for phone calls? ⁠Something Big Is Happening⁠ (Matt Shumer) ⁠OpenAI Executive Who Opposed ‘Adult Mode’ Fired for Sexual Discrimination⁠ (WSJ) ⁠TikTok launches an opt-in Local Feed in the US leveraging users’ precise location⁠ (TechCrunch) ⁠T-Mobile will live translate regular phone calls without an app⁠ (The Verge) Learn more about your ad choices. Visit megaphone.fm/adchoices

It's peak pollination season, and my business is scaling fast. To keep the nectar flowing, I need a phone plan with top priority data speed. That's why I chose GoogleFi Wireless. My connections stay strong even when the hive is buzzing. Plus, unlimited plans start at $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. Welcome to the Techbrewer right home for Wednesday, February 11th, 2026. I'm Brian McCullough today. Forget the chat GPT moment. Is this the COVID moment for AI? The moment when everything has changed, it's just not that everyone knows it yet. More weird musical chairs at the AI startups, TikTok, but for around the block. And do we finally have a universal translator but for phone calls? Here's what you miss today in the world of tech. I'm going to lead with something unusual today. It's an essay from A.A. startup founder Matt Schumer, where he makes the argument that GPT 5.3 Codex and Claude Opus 4.6 can meaningfully contribute to the improvement of AI models, which he says would be a sign of what's coming for most knowledge work within five years. It's been a while since I've seen a piece get this much chatter online, so I'm going to link to it so you can read the whole thing in full, but also I'm going to summarize it right now. Schumer's essay is essentially a warning memo to non-tech friends and family. He says we are at a February 2020 moment for AI. We're still in the

phase where normal life looks intact, and most people think the alarm is overblown. He's, of course, referencing COVID. This is right before the implications become unavoidable, though. He argues this wave will be much, much bigger than COVID, and that the gap between what insiders are seeing and what the public believes is now dangerously wide. He starts by explaining why he's, he starts by explaining why he he's breaking social norms to sound dramatic. He says he has spent years building and investing in AI, but he says he's been giving loved ones a cocktail party version because the honest version makes him sound unhinged. Now he thinks that's no longer defensible. He also emphasizes that most people in AI, including him, don't meaningfully control what's about to happen. The future

is being steered by a small set of teams at a few frontier labs where a single training run can shift the entire trajectory of the technology. The first major claim he makes is the reason tech people are suddenly sounding the alarm is not because they're making distant predictions. It's because the disruption has already arrived for them. In his account, AI Progress used to feel like big jumps spaced far enough apart to digest. Then in 2025, the cadence tightened and the margins got larger. Each release not only was better than the last, but better by more and also arriving sooner.

That culminated for him on February 5th of this year when Open AI and Anthropic release new coding-centric models the same day, and he says the waterline suddenly felt, quote, at your chest. He illustrates that turning point with an anecdote meant to make AI can do my job concrete rather than abstract. He claims he can now describe an app in plain English, what it should do and roughly how it should look, and the system produces the finished product. not a draft, not code you must clean up, but working software. More strikingly, he says, the model doesn't just write code, it runs the app, clicks through features like a user, tests flows, and iterates on design and functionality until it meets its own standards. Then it

hands it back as ready. This experience leads him to a psychological conclusion. What used to feel like, assistant that needs guidance, has crossed into autonomous builder in a way that changes the economics of knowledge work. Schumer argues that coding was not a random early target. It was a strategic prerequisite. The labs focused on making AI great at programming because AI development itself requires enormous amounts of software. If models can write that code, they can help build the next generation of models, which can write better code, which accelerates the next generation, and so on. It's recursive, the definition of, this is his bridge from my job changed to your job is next. Software was the first domino largely because it unlocks a faster path to general capability across the board.

A second claim is that popular skepticism is often based on stale reference points. He responds directly to the common reaction of, I tried AI, and it wasn't that good, by conceding that older consumer experiences, especially from 2023 and early 2024, were legitimately limited and prone to hallucinations. But he insists that in AI time, that's ancient history. In his view, today's leading models are unrecognizable, even relative to six months ago, and people judging from free tiers are often a year behind what paid users see, so society is underestimating what is already available. To justify the this is moving faster than you think framing, he offers a compressed timeline of capability leaps, e.g. arithmetic unreliability in 2022, exam level competence by 2023,

working software and graduate level explanations by 2024, then a step change in early 2026. He points to Meadors measurement approach, tracking how long a model can complete real-world tasks end-to-end without human help, and noting that the autonomous task length has been rising from minutes to hours with a claimed doubling time measured in months and potentially accelerating. He extrapolates that if the trend holds, we should expect systems that can work independently for days within about a year, weeks within two years, and month-long projects within three years, turning AI from a tool into something more like an employee. The piece's most consequential argument is probably about feedback loops. Quote, AI is now building the next AI. He quotes OpenAI's

documentation stating that GPD 5.3 codex was, quote, instrumental in creating itself, used for debugging training, managing deployment, and diagnosing evaluations. He pairs that with Anthropic CEO Dario Amo Dai's comments that AI is writing much of the code at Anthropic at this point, and that we may be one to two years from systems that autonomously build the next generation. Schumer frames this as the early stage of an intelligence explosion, each generation contributing to the next, which increases the pace again. From there, he turns to what this means for work. He cites Amo Dai's public prediction that AI could eliminate 50% of entry-level white-collar jobs within one to five years,

and he argues that many insiders think that's conservative. The key distinction, he says, is that this isn't, like past automation that replace narrow tasks, it's a general substitute for cognitive labor that improves across domains simultaneously. That makes just retrain yourself harder because the thing you retrain into is also being eaten by the same curve. He runs through some examples, legal research and drafting, financial modeling and memos, content writing, software engineering, medical analysis, and upgraded customer service agents, while stressing these are illustrative, not exhaustive. His rule of thumb becomes, if your job happens on a screen, reading, writing, analyzing, deciding, communicating, AI is coming for significant parts of it, and the timeline is already underway.

He challenges a popular comfort belief that humans will remain uniquely valuable because of judgment, taste, creativity, empathy, and strategy. His personal shock is precisely that the newest models sometimes feel like they exhibit judgment, making qualitative choices, refining UX, selecting the right call rather than merely technically. correct calls. Even if some deeply human qualities remain hard to replace, he notes people are already using AI for emotional support and companionship, and he expects that reliance to grow. Meanwhile, he adds, physical labor is not immune either. Robots aren't there yet, but not yet, can flip quickly in AI timelines. The final third of the essay is prescriptive. He says, don't freeze with this knowledge. Get early. His central practical advice is to start using AI seriously and in your actual workflow, not as a novelty search box. He urges paying for the most

capable model tiers, because defaults and free versions lag, selecting the strongest model explicitly, and then pushing it into high-level tasks in your field, contracts, spreadsheets, team performance narratives, etc. He also argues there's a short window where most organizations are still ignoring this, so the person who can produce three days of work in an hour will become immediately valuable. He recommends swallowing ego, building financial resilience with, you know, getting some savings under your belt, caution with taking on new debt in your household, and, quote, leaning into areas that are slower to replace, relationships and trust, physical presence, licensed accountability, and heavily regulated settings, while also stressing these are time buyers,

not permanent moats. He also tells readers to rethink the default guidance for their kids, optimize less for a stable professional pipeline that may be exposed and more for curiosity, adaptability, and becoming a builder who can wield these new tools. Importantly, Schumer doesn't frame this as only Doom. He claims the upside is real and newly accessible. The barrier to building products, learning skills, and creating work is collapsing. Describe an app, he says, and you can get a prototype quickly. Write a book with help, learn from near always available tutors.

His meta advice is to build the habit of adapting. becoming comfortable, constantly relearning tools because models and workflows will churn quickly. He even proposes a concrete practice. An hour a day of hands-on experimentation to stay ahead of the curve. He closes by widening from jobs to geopolitics and existential risk. He repeats Amadai's thought experiment of a new country of millions of superhuman digital workers appearing overnight. Fast, tireless, networked, able to operate anything digital, arguing that would be a historic national security threat. On the upside, he says AI could compress decades of medical progress. On the downside, he points to risks like loss of control, deception and testing, lowered barriers to

bio-weapons, and surveillance state enablement. His bottom line is a set of convictions. This is not a fad. Enormous capital is committing to it. The next two to five years will be disorienting, and the people who do the best in this period will be those who engage now with urgency and curiosity before the future arrives as a headline when the easy advantages are gone. 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

student offer. While supplies last, ends June 30th, terms at AKA.m.m.m.a. Resort and Casino at San Manuel is California's number one entertainment destination for today's superstars. Catch the Jonas Brothers return to the Yamava Theater stage on April 30th, the powerful vocals of Demi Lovato on May 17th, and the signature Southern Country Rock of Eric Church on July 19th. Tickets on sale now at Yamavatheater.com. Only at Yamava resort and casino, celebrating its 40th anniversary. You in? Must be 21 to enter. 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 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. 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. Weird stuff going on again at the various AI startups again. First, the journal says that OpenAI fired VP Ryan Biermeister in January for alleged sexual discrimination. She had earlier raised concerns about the upcoming launch of a adult mode on chat GPT. Quote,

the fast-growing artificial intelligence company fired the executive Ryan Biermeister in early January following a leave of absence, according to people familiar with the matter. Open AI told her the termination was related to her sexual discrimination against a male colleague. The allegation that I discriminated against anyone is absolutely false. Beermeister said in a statement in response to a request for comment,

Beirmeister served as the vice president leading OpenAI's product policy team, which develops rules for how people can use. the company's products and helps design the enforcement mechanisms for those policies. Her ousting came ahead of Open AIs planned launch early this year of a mode that will allow users to create AI erotica in chatchipT. The planned feature, which would permit adult-themed conversation, including sexual topics for adult users, has drawn criticism from researchers at the company who have studied the ways some people develop unhealthy attachments to chatbots, according to some

of the people. They have raised the prospect that sexual content could intensify the feeling, some people have for the AI personas they view as companions. Members of an advisory council on well-being and AI that OpenAI convenes regularly have also expressed opposition to adult mode and urged the company to reconsider plans to launch at people with knowledge of those discussions said before her firing. Beirmeister told colleagues that she opposed adult mode and worried it would have harmful effects for users, people familiar with her remarks said. She also told colleagues that she believed Open AI's mechanisms to stop child exploitation content weren't effective enough and that the company couldn't sufficiently wall off adult content from teens, the people said, end quote.

Meanwhile, another Open AI researcher, Zoe Hitzig, who helped shape how models were built and priced, says in a New York Times op-ed that she has quit after two years at Open AI due to deep reservations about ads and Open AI's overall strategy. I once believed I could help the people building AI get ahead of the problems they would create. This week confirmed my slow realization that Open AI seems to have stopped, the questions I joined to help answer, end quote. Meanwhile, X-A-I co-founder Jimmy Baugh says that he is leaving that company, making him the sixth co-founder to depart XAI in recent months.

Sources and social media posts also show that at least six researchers have left XAI in recent weeks. TikTok has launched an opt-in local feed in the app's U.S. version, including content related to travel, news, events, shopping, and dining near a given user's location. Quoting TechCrunch. The company explains in its announcement that the feed is meant to help users stay connected to their local community and its posts are shown to people based on their location, the content's topic, and when the content was posted. This makes it a more current feed of local information like suggestions of new restaurants to try local events, shopping suggestions, and more.

The new feature also ties into TikTok's push to attract small businesses to its app, not only as content producers, but as advertisers. This could help insulate it against further regulation and help it to claim, as Meta does, that it should not be reined in because so many small businesses rely on its services to reach their customers. TikTok notes that 7.5 million businesses currently use the app to reach global customers, and these businesses support more than 28 million workers per a 2025 Oxford Economics Report. The company also highlighted figures from the Small Business and Entrepreneurship Council, which found that 84% of TikTok small business users said the platform helped grow their business,

and 75% said TikTok helped them reach customers beyond their local area. In addition, another 74% said TikTok helps them connect with their local community. As TikTok sees it, the local feed will help generate real-world traffic and sales for these brick-and-mortar stores across the U.S., end quote. And finally today, this sounds incredibly useful. Quoting the verge, T-Mobile is preparing to test a new AI feature that, translates live phone calls into more than 50 languages. Live translation is launching in beta this spring, according to a press release with registration for eligible T-Mobile customers available starting today. Some of the biggest barriers wireless customers face are the simplest ones,

like being able to understand each other, said T-Mobile CEO Sreni-Gopalan. By bringing real-time AI directly into our network, we're delivering more than connectivity, turning conversations into community, starting with live translation. Enabling it at the network level means users don't need specific apps or devices to use this service. The only requirement is that translation must be initiated by a T-Mobile network user. The feature will be available on the carrier's 5G Advanced Network, but T-Mobile president of technology and CTO John Saw told the verge that live translation works over both 4G LTE and 5G. The key requirement is a voiceover LTE connection, which allows the service to operate reliably across a wide range of devices and network

conditions, said Saw. That flexibility is important because it ensures live translation works for customers wherever they are, not just when they're on the latest network technology. T-Mobile says that beta participants can activate live translation for as long as they need by dialing 87 at no additional cost during the beta period. T-Mobile has not mentioned if live translation will be paywalled when it launches to the general public this year. Beta participants will also be able to activate live translation by saying, hey, T-Mobile later this spring. It's worth flagging that the feature only works during an active call and that T-Mobile won't save call recordings or transcripts. The service is designed to translate conversations in real time and then move on without storing the content of those calls,

Saw said, end quote. This is the latest I've been getting a show out the door in a long, long time. It's not worth going into, but I just had a nightmare of a day. My apologies for being late. Talk to you tomorrow. Did you know if your windows are bare, indoor temperatures can go up 20 degrees? Turn the temperature down with blinds.com and get up to 50% off custom window treatments like solar roller shades and more during the Memorial Day mega sale. Whether you want to DIY it or have a pro handle everything, we've got you. Free samples, real design experts, and zero pressure, just help when you need it.

Shop up to 50% off site wide and huge savings on Doorbusters right now during the Memorial Day mega sale at blinds.com. Rules and restrictions apply.