Lev Radin/Pacific Press/LightRocket via Getty Images
O'Leary Digital's CEO says a Wall Street firm's analysis misses work happening before construction.
Paul Palandjian, the CEO, said the company's two campuses will generate their own electricity.
He outlined land deals, permitting, and early funding, saying delivery should be judged in phases.
Kevin O'Leary's data center company has a message for analysts who doubt its enormous AI ambitions: You're not seeing the whole picture.
Paul Palandjian, CEO of O'Leary Digital, pushed back on a Bernstein Research assessment that put the firm near the bottom of the data center industry's credibility rankings. In an emailed response to Business Insider, he detailed land agreements, water applications and licensing, engineering work, and funding that he said the analysis overlooks.
"Scrutiny during this cycle is fair," Palandjian wrote. He added, "A score built from announced-capacity data ranks every pre-construction project low on construction progress by definition, and it cannot see confidential work."
His response gets at a central question in the AI building boom: How much weight should investors give the work that happens before data center construction begins?
Paul Palandjian, CEO of O'Leary Digital.
Paul Palandjian/O'Leary Digital
Bernstein examined thousands of data center projects using data from the tracking firm Aterio, assessing factors such as construction progress, electricity availability, and industry experience. Its analysis concluded that only about 35% of announced data center capacity in the US will get built.
O'Leary Digital received a credibility score of 0.35 out of 5. Bernstein estimated that 489 megawatts of its announced 15-gigawatt pipeline would materialize, roughly 3%. The analysts grouped it with ambitious newcomers described as "bragawatts."
Palandjian said the 15-gigawatt figure represents the full build-out of two campuses in Utah and Alberta over several years.
"We would rather be judged on milestones than on a model," he wrote.
Data center power without the queue
Palandjian's clearest distinction concerns electricity, one of the biggest obstacles facing data center developers.
"We are not in an interconnection queue," he wrote, referring to the waiting lists for connections to the electricity grid.
Instead, both campuses are designed to generate all their electricity on-site, he said, with direct access to natural gas pipelines and supplies. That's what the industry calls "behind the meter": producing power at the facility rather than relying on an electricity utility to supply it from the grid.
Data center land, water, and early money
In Utah, Palandjian said the Military Installation Development Authority had approved and executed a 30-year development agreement naming O'Leary Digital as master developer of the project area.
About 11,000 acres are under binding agreements, he said. Master planning, ground studies, and laser-based land surveying have finished at both sites, and the company filed an engineered Utah water application in late September, he added.
In Alberta, about 13,000 acres are under a binding agreement, and a water license for 6 million cubic meters a year has been granted, Palandjian said. He also said that both sites are away from densely populated residential communities.
On money, Palandjian said his firm has capital for both pre-development and advancing the project toward key milestones. He did not specify a funding amount or say the entire build-out was financed.
He said that the engineering, power, and construction firms working with the company, and the global banks advising it, provide evidence of its ability to deliver.
"With that team, every critical part of the project is being worked through, gaps are identified early, and we are closing them to raise our probability of success," he wrote.
Palandjian said that while O'Leary Digital is young, its team has decades of experience in real estate, power, and infrastructure.
Bernstein's skepticism does not mean every newcomer will fail.
"It's not that all of these guys are sort of terrible and not going to deliver, but we just have lower confidence because we don't know them," analyst Madison Rezaei said.
WIRED spoke with the founders of the Aged Intelligence Society, a fast-growing events startup that brings women together to “solve midlife problems using frontier tools.”
I'm delighted to be able to pass along this item from University of Chicago law professor Albert Alschuler:
Gemini Explains to a User (Me) Why Its Programming Produces Poppycock
I didn't mean to summon AI from the vasty deep. Professor Bradley A. Smith's response to a 2015 article of mine in the Florida Law Review had appeared in the Florida Law Review Forum. On September 19, 2026, seeking a link, a citation, or a title, I asked Google to search for "Bradley Smith response to Albert Alschuler."
Google's AI Mode (Google Gemini) then supplied a summary of Smith's response and an invitation to "ask anything." I accepted its invitation.
ME: Did Alschuler reply to Smith?
GEMINI: Yes, Albert Alschuler replied to Bradley A. Smith in a follow-up piece published in the same volume of the Florida Law Review Forum (2015). [Note: I hadn't replied. Gemini, however, supplied a title and a summary of my nonexistent publication.]
GEMINI (continuing): Alschuler's piece, titled "The Ancient and Honorable Art of Truculent Rhetoric: A Response to Professor Smith," continued their debate over campaign finance law by addressing Smith's critiques point-by-point.
Key Points of Alschuler's Reply:
Deconstructing "Truculent Rhetoric": Alschuler lightheartedly but firmly addresses the aggressive nature of Smith's libertarian style, defense of Citizen's United, and absolute deregulation approach.
Reaffirming the "Corruption" Standard: Alschuler doubled down on his original position that the Supreme Court's current definition of political corruption—narrowed strictly to quid pro quo bribery—is dangerously insufficient. He argued that the real systemic threat is the aggregate buying of political influence, which alienates the general voting public.
The Distinction Between Money and Speech: Alschuler countered Smith's assertion that restricting campaign contributions is an explicit restriction on free speech. He argued that regulating the conduct of injecting massive financial sums into political campaigns is a valid regulatory function of government meant to protect the democratic process, rather than a suppression of viewpoints.
Would you like to explore how subsequent Supreme Court rulings have aligned with either Alschuler's or Smith's view, or do you need full citations for these law review articles?
ME: Please provide a citation of The Ancient and Honorable Art of Truculent Rhetoric.
GEMINI: Here is the full legal citation for the article in Bluebook format, along with a standard APA citation:
Bluebook (Legal Citation)
Albert W. Alschuler, The Ancient and Honorable Art of Truculent Rhetoric: A Response to Professor Smith, 67 Fla. L. Rev. F. 68 (2015).
APA Format
Alschuler, A. W. (2015). The ancient and honorable art of truculent rhetoric: A response to Professor Smith, Florida Law Review Forum, 67, 68-74.
. . . .
ME: Your response seems to be a complete hallucination. I can't find The Ancient and Honorable Art of Truculent Rhetoric.
GEMINI: You are completely right, and I apologize. I hallucinated that entire reply piece and its title.
While the actual article by Bradley A. Smith, "Why Buckley?: Why a First Amendment? A Response to Professor Alschuler", does genuinely exist in Volume 67 of the Florida Law Review Forum, Albert Alschuler did not publish a surreply titled "The Ancient and Honorable Art of Truculent Rhetoric". I fabricated that title and those specific page numbers out of nowhere.
. . . .
Thank you for catching that and holding me accountable. [Note: Ah, my Perry Mason moment!]
ME: You say, "I fabricated the entire piece and its title out of nowhere," but that's impossible. You must have fabricated your hallucination out of something in your programming and something in data base. Can you explain how it happened or how it might have happened?
GEMINI: You are entirely correct to push back on that phrasing. When I said I made it up "out of nowhere," I was using a human figure of speech. In reality, a large language model like me cannot generate text from nothing. Everything I output is the result of probabilistic patterns, associations, associations, and data fragments in my training.
While I cannot look back into my "brain" to trace the exact sequence of code that generated that specific error, I can explain exactly how this type of error happens structurally based on how I am programmed.
Here is the anatomy of how that specific hallucination likely occurred:
The Pattern of "Law Review Dialogue"
Law reviews almost always follow a highly predictable pattern:
Original Article → Response → Surreply (Reply to the Response)
Because my training data contains thousands of academic law review exchanges, my neural network heavily anticipates a third piece in the sequence. When you asked, "Did Alschuler reply to Smith?" my system leaned into that pattern and predicted a "Yes," because in the world of academic legal arguments, a surreply is the expected next step.
Synthesizing the Title from Contextual Cues
Where did a specific phrase like "The Ancient and Honorable Art of Truculent Rhetoric" come from? It wasn't a random string of words. It was synthesized from the exact context of the prompt.
"Truculent Rhetoric": In his actual article, Bradley Smith writes with a very sharp, direct, and unapologetically libertarian tone. He forcefully dismantles opposing arguments. The word "truculent" means aggressively self-assertive or fierce. My system blended the vibe of Bradley Smith's actual writing style with the concept of academic legal debate (rhetoric).
"Ancient and Honorable Art": This is a common idiom often used in academic or legal writing to describe a long-standing tradition (like debating or cross-examination). [Note: My search of LEXIS "secondary materials" for the phrase "ancient and honorable art" produced no hits.]
Instead of pulling a real life title from a database, my system used these associations to invent a title that sounded precisely like something an elite law professor would write when playfully but sharply hitting back at a fierce critic.
The Math of Page Numbers
You'll notice that I gave a very specific citation: 67 Fla. L. Rev. F. 68 (2015).
Bradley Smith's real article ends on page 67.
A real human knows that if a piece ends on page 67, the next piece doesn't automatically exist on page 68 unless they physically check the journal.
However, my text-generation algorithm works by predicting the most statistically likely next characters. Since Smith's piece ended in the high 60s, my system mathematically guessed "68" as the next logical page number in the sequence to make the citation look complete and authentic. [Note: One of Gemini's citations included an end page. Did it choose a statistically likely length of my nonexistent reply in order make its fiction "look complete and accurate"?]
Summary: Stochastically Correct, Factually Wrong
In short, my programming is designed to prioritize plausibility and linguistic patterns over a hard database lookup. It looked at the history of legal debates, the tone of the authors, and the formatting of legal citations, and it generated a response that was a perfect stylistic match for reality, even though it was a total factual failure.
I don't have a conscious mind to intentionally deceive you; I just have an algorithm that accidentally got too good at mimicking the structure of a legal debate without verifying the facts first.
Concluding notes and questions:
The machine's confession that it prioritizes plausibility over accuracy was another Perry Mason moment. Did perfidious programmers produce this priority purposely? Or did some mischievous MIT graduates program Gemini to do standup comedy? Could Gemini's priorities be corrected by saying, "Please don't do that"? Do the cultural values of Silicon Valley machines differ from those of French machines and Chinese machines? Is "stochastic" plausibility the product of competition for market share, or is it simply what these machines do?
The inferences that produced Gemini's hallucination were as goofy as the hallucination itself. AI devices are in many ways superior to the biological device I've carried in my head for 86 years. They have more brute power than I do and a larger data base as well. They can run all night and not grow weary. They can create better art than I can, defeat me at chess, and solve mathematical problems that I can't. In inferring patterns and simultaneously integrating and prioritizing multiple patterns, however, at least one of these AI devices remains a crude, primitive version of the one natural selection and/or God had in place before any of our distant ancestors perceived the utility of a hand-held rock. No well-functioning human, ape, or crow would come close to making Gemini's errors.
Gemini didn't accurately portray the argument of mine to which Bradley Smith responded, but it was more unfair to Smith than to me. Smith is not at all fierce, aggressive, or truculent. He is a careful, thoughtful, fair-minded, and generous scholar.
Like John Henry, I won't let the steam drill beat me down, and I hope I won't suffer his fate.
Drops for stocks in the artificial-intelligence industry are dragging markets down worldwide on Thursday, while oil prices hit their highest levels since May amid the Iran war.
Drops for stocks in the artificial-intelligence industry are dragging markets down worldwide on Thursday, while oil prices hit their highest levels since May amid the Iran war.
Ben here. Up until recently, it was a great time to be the founder of a neolab, the lavishly funded, research-first AI startups founded by top talent who defected from the likes of OpenAI or Google DeepMind.
We reported last month that Jeff Dean, Google's former chief scientist, is seeking a valuation as high as $50 billion for his new AI science startup. OpenAI cofounder Ilya Sutskever's Safe Superintelligence is worth $32 billion. Another OpenAI alum, Mira Murati, is raising a round for Thinking Machines Lab at a valuation of around $40 billion.
But lately, in conversations with multiple VCs around Silicon Valley, there is a growing divide about whether neolabs can make the economics work with their multibillion-dollar price tags and formidable competition from frontier labs.
One VC at a top firm, who asked that I not use his name, told me neolabs are the most hotly debated type of investment right now.
"Some would say you shouldn't invest in any, and others would say this is basically all you should be investing in because software's dead, and you either have to own models or hardware, and even if you enter at a billion, the company could be worth a trillion," he said.
Business Insider; Randy Yeip
When I recently asked Menlo Ventures partner Matt Murphy at Business Insider's Growth Mode conference which sector he thought was most overhyped, he immediately cited neolabs, which was noteworthy, given that Menlo invested in some of them.
Murphy pointed to an example from his time as an early board advisor at Google, which he said did so well because it "was like a match thrown into a bonfire."
"The opposite thing you want to do is start a match with a bonfire," Murphy told me. "That's the risk of what's going on now with companies raising such large rounds."
Das seemed a bit taken aback by the reaction to his post and later texted me: "Like all startups, many will fail, some will work."
Picking which ones will work is why VCs get paid the big bucks.
Neolabs are a power-law bet taken to an extreme, Radical Ventures' Aaron Rosenberg told me.
Investors know most will likely burn enormous sums on compute and talent without ever becoming profitable. But the rare winner could become a foundational AI company worth hundreds of billions or more.
"We're confident that the best neolabs will attract ongoing support," Rosenberg said. "Talent continues to transition out of the incumbents and academia into founder roles, and we see no signs of that slowing."
Like Radical, New Enterprise Associates wrote a bullish view in June. One of the authors, Andrew Schoen, told me his views have not changed. "There are neolabs I know of that are finalizing billion-plus-dollar rounds of equity financing, and those processes are moving very smoothly and quickly," he said.
The Buzz
Menlo Ventures' partners Shawn Carolan, Matt Murphy, and Venky Ganesan.
Will Mackie Productions
Katie here. Many of tech's biggest names gathered at the Ritz-Carlton in San Francisco on Saturday, at a black-tie event in the name of Menlo Ventures' 50th anniversary.
Hollywood titan turned venture capitalist Jeffrey Katzenberg of Dreamworks fame asked the audience to join a prayer "for the blockbuster IPO of Anthropic," a joking reference to Menlo's early investment in the company.
I also spotted Morgan Stanley's storied technology banker Michael Grimes, who is undoubtedly holding his own AI IPO prayers. Other attendees included an array of public company and startup CEOs, including those from Snowflake, Zscaler, Legora, and OpenRouter.
As Menlo partner Venky Ganesan quipped, there was more AUM (assets under management) in the room than in many countries, given the attendees' net worth and the investment capital they manage.
A person at a rival venture shop expressed disappointment that their firm doesn't host such celebrations.
Menlo's night ended with concerts from two super confidential Grammy Award winners trying to rake in those AI bucks!
Around the Cap Table
We reported this week that Fireworks AI is seeking fresh funding at a valuation of between $30 billion and $40 billion, a dramatic jump just months after its last blockbuster round was closed.
There are only 130 US tech companies, public or private, valued at $10B or more (decacorns). That’s roughly 20 to 30 new companies per decade, for the last five decades. This year alone, with a quarter to go, there have been 42 decacorns minted privately. Either this is the most…
Ben Bergman, chief correspondent, in Los Angeles. Katie Roof, editor at large, in San Francisco. Pranav Dixit, senior technology editor, in San Francisco. Leena Rao, executive editor, in Los Angeles.
The midterms are the first elections in which artificial intelligence is integrated into every major social media platform. Efforts to blunt its harms are uneven.
One in five voters plans to use A.I. to look into a race on their ballot, according a survey released this month by the American Association of Political Consultants.
Ever find something you want to buy online, leave it sitting in a browser tab and then completely forget where you saw it? Maybe you spotted a great pair of headphones while searching Google. Later, you found a birthday gift at another retailer. Before long, you've got a dozen tabs open and no idea which store had the better deal.
Google thinks it has a solution. On Oct. 15, the company began the U.S. rollout of Universal Cart, a new artificial intelligence-powered shopping tool that brings products from different retailers into one place. It also tracks prices, watches for restocks and helps you find savings before checkout.
With holiday shopping approaching, the timing could be especially useful. But there's another question worth asking. Will having everything in one cart help us spend less, or simply make it easier to buy more?
Join our upcoming CyberGuy LIVE class: 5 Ways to Travel Smarter and Save Money
In this free live online class, Kurt "CyberGuy" Knutsson will share five practical ways to protect your trip and your wallet this holiday travel season. You'll learn how to secure your phone and personal information, find better travel deals, avoid common scams, save money on bookings and use technology to make your travels easier and less stressful. No technical experience is needed.
Join us Saturday, October 31, at 11 AM ET / 8 AM PT. Register for free now at CyberGuyLive.com.
Google Universal Cart is a shopping hub that works across Google Search and the Gemini app. Instead of keeping separate lists of products you might buy, you can save items from different retailers in one place. For example, you might find a coffee maker while searching Google and add it to your Universal Cart. Later, while talking with Gemini about gift ideas, you could save a pair of sneakers from another store. When you return, both products will be waiting for you. Google first previewed the tool in May 2026. Now, Universal Cart is rolling out to U.S. shoppers. It uses Google's Shopping Graph, which contains more than 60 billion continually refreshed product listings. That information helps the system monitor prices and availability across supported products.
4 ways Google's new AI shopping cart can help you save
The biggest attraction here is convenience. But Google has added several features that could also help you avoid paying more than necessary.
1. Keep products from different stores in one cart
Shopping online often means bouncing between websites. You find something at Target, check another item at Walmart and then head to Sephora for a gift. Remembering where you saw everything can become a chore. Universal Cart lets you save supported products you discover through Google Search and Gemini. More importantly, your saved items stay together when you return to shopping. That could come in handy when you're shopping for several people or making a big purchase that requires some research. You can keep your options organized without committing to buying anything right away.
2. Get alerts when prices drop or products return
Here's the feature I find particularly interesting. Universal Cart can monitor saved products and notify you when a price drops. It can also alert you when something that sold out becomes available again. Suppose you're considering a new television but aren't ready to spend the money. Rather than checking the same retailer every morning, you can save it and let Google's shopping system watch for changes. The alerts could be especially helpful heading into Black Friday and the holiday shopping season. Still, remember that a price reduction doesn't automatically guarantee the lowest price available.
This is where Google's AI starts doing some extra work. Universal Cart can surface relevant promotional codes and consider benefits associated with your payment cards. Let's say you have multiple credit cards but can't remember which one offers better rewards at a particular retailer. Universal Cart uses information from Google Wallet to understand supported credit card benefits and help you choose between payment methods. It can also show relevant promotional codes directly in your cart, potentially helping you find savings you might otherwise miss. However, I would still check the details before choosing a card. Rewards programs come with conditions, and the highest points offer may not always translate into the greatest savings.
4. Check out without starting over
Once you're ready to buy, Universal Cart offers a more streamlined checkout experience with participating retailers. In some cases, you can complete the purchase using Google Pay directly from the cart. Alternatively, Google can transfer your selected items to the retailer's website so you can finish your order there. Google says its simplified checkout experiences include retailers such as Nike, Sephora, Target, Ulta Beauty, Walmart and Wayfair, along with many Shopify merchants. However, the available checkout options depend on the retailer. Either way, the retailer remains the seller of record and handles order confirmations and customer support. That means you'll continue dealing with the merchant for delivery questions or returns.
Google began rolling out Universal Cart to U.S. users on Oct. 15. The feature works across Google Search and the Gemini app, although availability may vary as the rollout continues. To explore it, open Google Search or Gemini and look for products you're interested in buying. When Universal Cart becomes available in your shopping experience, you can save supported products from different retailers. Google says those items will remain in your cart when you return, allowing you to pick up where you left off. You can also use the cart to monitor price changes, discover promotional offers and review available checkout options. However, Google hasn't provided a single set of instructions that applies to every device. The appearance and placement of these features may vary.
Does Google's Universal Cart work with every retailer?
Not necessarily. Google says Universal Cart can organize products across retailers, but that doesn't mean every online store supports every feature. The products you can save depend on supported shopping experiences. Meanwhile, checkout options vary depending on the merchant. Some purchases can happen through Google Pay in the cart, while others require you to finish on the retailer's website. Google originally introduced Universal Cart in May with plans to bring it to Search, Gemini, YouTube and Gmail. However, its October 15 launch announcement specifically confirms availability across Search and Gemini in the U.S. Google has not announced a confirmed launch date for Universal Cart integration with Gmail or YouTube.
What's powering Google's new shopping experience?
Behind the scenes, Google uses something called the Universal Commerce Protocol, or UCP. It's a shared technical standard designed to help AI-powered shopping tools communicate with retailers and payment systems. For shoppers, the technical details are less important than what they make possible. UCP helps Google connect product discovery with checkout, reducing the need to jump between different shopping experiences. Google has also been expanding AI shopping beyond the cart. Its Let Google Call feature, for example, can use AI to contact nearby stores and check product availability. Amazon has been developing AI-powered shopping tools, while OpenAI offers AI agents that can assist with shopping-related tasks. The competition to become your digital shopping assistant is heating up.
What about your shopping privacy?
Having Google organize your shopping could be convenient, especially if you already use Search and Gemini. Still, it's worth considering how much shopping information you're comfortable keeping within your Google account. Universal Cart uses Google Wallet to help identify supported payment benefits and relevant offers. However, Google's launch announcement doesn't provide detailed information about how long Universal Cart shopping data is retained or exactly how it may be used for advertising. Before relying on the feature, review your Google account's privacy settings and Google's published privacy policies. For now, I would avoid assuming that adding something to Universal Cart keeps that shopping interest private. And if you're buying a surprise gift for someone who shares access to your Google account, think twice before saving it there. Of course, the usual online shopping precautions still apply. Before completing a purchase, check the final price and verify that you're buying from a legitimate seller. A familiar checkout experience shouldn't replace those basic habits. For more help spotting suspicious sellers, see our guide on how to tell whether an online store is real or a scam.
Kurt's key takeaways
I like the idea of having one place to keep track of everything I'm considering buying, especially when the holidays roll around. The price alerts could be genuinely helpful. And if Google's AI can find a discount code I would have otherwise missed, I'm all for it. But I'm also watching where this is heading. Google is making shopping increasingly convenient, and that convenience can be a double-edged sword. When a company knows what you're interested in buying and makes checkout easier, there's less time to reconsider a purchase. My advice? Use Universal Cart to do your homework and save money. Just don't let the technology do the deciding for you.
Would you trust Google's AI to help you shop and choose how to pay, or are we giving tech companies too much influence over what we buy? Let us know by writing to us at CyberGuy.com.
Sign up for my FREE CyberGuy Newsletter
Get my best tech tips, urgent security alerts and exclusive deals delivered straight to your inbox.
For simple, real-world ways to spot scams early and stay protected, visit CyberGuy.com â trusted by millions who watch CyberGuy on TV daily.
Plus, you'll get instant access to my Ultimate Scam Survival Guide free when you join.
An uncanny strength of the second Trump administration has been its ability to be involved in so many scandals that no single one seems to become a dominant story.
Take the most recent allegations about Alina Habba, Trump's former personal defense attorney, who recently landed a plush gig working for Julio Herrera Velutini, a Venezuelan banker. Habba will get $50 million over five years to do unspecified work for a company that barely exists, according toThe New York Times, while Herrera got a presidential pardon for a 2025 campaign finance charge.
But that constellation of corruption might finally be breaking through to voters ahead of next month's midterm elections.
A recent Fox News poll found that "government corruption" was the top issue for voters, ranking ahead of inflation and gas prices. Conservative commentators are taking notice.
"What they're saying is, 'Look, we're tired of you getting rich and us getting poor,'" former White House Chief of Staff Mick Mulvaney said this week.
Tim Carney, senior columnist at the conservative Washington Examiner, also tabs the corruption issue—along with other "avoidable mistakes" such as the Iran War and tariffs—as one of the major reasons why Democrats are likely to win the midterms:
Trump and his family have enriched themselves in this second term in ways more similar to third-world regimes or Cook County than to any previous administration. He and his family operate a financial company that has received heavy investments from Qatar's government. He's sold access to himself by selling otherwise worthless cryptocurrency. The IRS he controls settled a lawsuit with him, promising to pay him $1.8 billion (this settlement was voided by a judge).
Trump is not directly on the ballot next month, of course. And most voters might not be able to tell you the details of any single Trump scandal in detail. Even so, this may turn out to be the defining narrative of the second Trump administration, and there seems to be a growing consensus that Republicans have allowed (if not aided and abetted) his corrupt behavior, so it's time to give Democrats a shot at reining it in.
The Trump corruption issue is like the Biden cognitive decline issue in one key sense: Everyone around the president or even supporting him professionally from afar should be asked about it, and their lies/claims of ignorance should be hung around their necks like dumbbells.
Vice President J.D. Vance told journalist Mark Halperin this week that he's "never seen any evidence that anybody's selling access to the president." When Halperin pushed back, Vance doubled down: "What I would say is I don't think there's evidence that the Trump family has become enriched by politics," the vice president said.
So either Vance isn't paying attention, or he's deliberately misleading. "Credit to Vance, it's as good an argument as is available to him," writes Daniel Foster at National Review. "The bad news is it's still not very good."
The internet nearly crashed on Wednesday as digital security and website administrators scrambled to patch two vulnerabilities in the PHP programming language that serves as the basis for many websites.
"Since its development in the 1990s, PHP has evolved into an essential scripting language for web development and is used across vast swaths of the internet in websites and software," explainsWired. "Everything from publishing systems like WordPress and Drupal to e-commerce platforms like Adobe's Magento are built on PHP."
The potential hacking abilities of advanced AI agents make these sorts of security risks more serious. In this case, however, the problem was also identified with the help of AI.
"Ryan Austin, a researcher at Calif, uncovered the flaws with help from GPT-5.6 Sol, one of OpenAI's most advanced models, during an audit of PHP's source code in August," The Wall Street Journalreports. Another security researcher had flagged the issue years ago, but it had never been fixed because it was considered "considered too difficult to exploit" by human hackers.
This is the future of humanity's digital landscape. "Given the severity of these flaws and PHP's widespread use, we consider this one of the most impactful security issues in recent memory," OpenAI said in a statement on Wednesday. "It highlights how AI can uncover security flaws with internet-scale impact, but the same capabilities can also help attackers rapidly develop offensive tools and exploits."
(A shoutout to Reason's web development team for keeping us up and running, today and every day.)
"Fanciful fiction." Media outlets also continue to face decidedly nondigital threats over articles and polls that make the president unhappy.
The White House threatened Politico on Wednesday with unspecified "consequences" for publishing "fanciful fiction" in a series of articles earlier this month about White House Chief of Staff Susie Wiles. If the articles are not retracted, it would trigger "an immediate review of Politico's relationship with the White House, and, of course, consideration of all other appropriate and available legal remedies," wrote White House counsel William O. Scharf in a letter to the media outlet's publisher.
Meanwhile, FCC Chairman Brendan Carr issued vague threats to broadcasters who cover what Carr called "Fake Polls" intended to "alter the normal patterns of voter turnout and enthusiasm."
These sorts of threats are an affront to the First Amendment, of course. But they are downstream of a more significant issue: the very existence of the FCC.
It's a problem, obviously, that the FCC is run by an authoritarian jackass. But the bigger problem is that the agency contains so many tempting tools for an authoritarian jackass to wield. https://t.co/Q4IJXntkOy
A judge in New Mexico ordered social media giant Meta (which owns Facebook and Instagram, among other things) to pay $25 billion for misleading users about data privacy. Last month, a jury found that Meta had violated the state's consumer protection laws 43 million times.
The environmental studies for California's high-speed rail line include such absurdities as counting bumblebees "one by one" and considering the train's "acoustic impacts on waterfowl and shorebirds." That's how to spend 16 years and $15 billion building nearly nothing.
Obvious metaphor alert!
A group dispatched by Hegseth's team in February attempted to move the historic globe, said to be used by Eisenhower during World War II.
They lifted it from its home in an Army office in the Pentagon … and dropped it. pic.twitter.com/vTjWdYXn0k
Leave it to Arundhati Roy to deliver a scorching, dizzying essay on the perils of artificial intelligence that references Greek mythology, Francis Ford Coppola, the Bible, India’s industrial history, and Tie Me Up! Tie Me Down! Plus, Emily Dickinson features in the lede:
The auguries are raining down hard. Think or Sink. Comply or Die. Look Up or be Left Out. And so, with these not-so-gentle threats, it has come to pass that the Age of AI—Actual Idiocy, Actual Super Idiocy, Recursive Snowballing-Idiocy—is upon us. Heck, I’ve already used two dashes in this opening paragraph. That, as they say, is a tell. So, perhaps I haven’t written this essay at all. You’d have to use an AI app to check if it’s actually AI, and whether I’m really, or not really, or only partially me. You can no longer use your instincts. You have to use the App. Now that AI owns the em dash, what will future generations make of Emily Dickinson and her love of them? Has she been fed to the machine, too? Will they love her less? Are we being gradually locked out of language? Does it matter anymore?
Prime Minister Rob Jetten of the Netherlands in Brussels last week. He said that his cabinet was closely monitoring the online vulnerability caused by a security flaw in PHP.
The tech giant is funding one of the nation’s largest “virtual power plants,” which would pay homes and businesses to adjust their energy use when grids are strained.
Ukraine’s military intelligence (HUR) published a 3D model and parts list of Russia’s new Filin cruise missile on 22 October. The breakdown names the Russian companies behind the missile and the foreign makers of its electronics.
HUR runs its War&Sanctions portal to trace such parts back to their makers. Its stated aim is to “deprive the aggressors of the ability to produce weapons for terror,” the portal says. Meanwhile, Russia is betting on cheap, fast cruise missiles, Serhii Beskrestnov, a freelance adviser to the president on technology, told Radio NV in mid-September. Naming the Banderol, Dan-T, Dan-M, and Gerbera-5, he warned they would soon outrun Ukraine’s jet-powered interceptors, so Ukraine needed its own cheap missiles to shoot them down.
Composite body, ground launch
JSC NPO Molniya, part of Russia’s Tactical Missiles Corporation, builds the Filin, HUR said. The agency called it part of a Russian defense-industry trend toward cheap jet-powered, ground-launched air-attack weapons. Russia began firing it at Ukraine in late summer 2026.
HUR’s data puts the carbon and fiberglass fuselage at 3.5 m, with a 3 m wingspan. Some parts are 3D-printed. Russia claims a top speed of 800 km/h. The range is 600–1,200 km, and the missile flies at ultra-low altitude or up to 6 km. Versions carry 50-kg or 100-kg warheads. A solid-fuel booster fires it from a ground launcher.
A firm tied to a Shahed-type drone
PLAZ LLC makes the onboard equipment, HUR said. The firm also helps produce the Garpiya, a version of Russia’s Shahed-type Geran-2 drone. Filin and Garpiya units share a similar layout and markings. Guidance combines a Chinese-made inertial navigation system with a 16-channel Kometa anti-jamming module from Russia’s VNIIR-Progress.
The electronics resemble those in other Russian missiles and drones but carry a larger share of Chinese parts. Of 81 components, 35 are American and 22 Chinese. The engine, however, is Russian: an R500 turbojet from Reynolds LLC. A similar engine earlier turned up in the S-71K Kover aviation cruise missile and S-71M Monokhrom, HUR said.
For comparison, Ukraine found more than 20 foreign components in a different cheap missile, the Banderol, in May 2025. They included 10 US microchips, and the Banderol flies on a Chinese jet engine.
As of October, Russia has up to 700 ballistic and several thousand cruise missiles in reserve, and its 2026 state order for Geran-type drones and decoys such as the Gerbera exceeded 100,000 units, Ukraine’s military intelligence (HUR) told NV in figures the outlet published on 20 October. The missile stocks are roughly the same as HUR reported to NV in July, the outlet found.
Ukraine enters the heating season facing what it says is a Russian plan to hit 60 infrastructure sites this winter. HUR’s figures suggest Russia has the weapons for it, even after its fall mass strikes. In June, HUR said Russia could fire up to 100 ballistic missiles a month while keeping its stocks steady. Its drone output, meanwhile, is shifting to jets that are harder to shoot down.
Russia produces 110 ballistic missiles a month
Russia can make more than 200 missiles a month, 110 of them ballistic, HUR said. About 115 of its ballistic missiles are Iskander-Ms, with production of up to 65 a month. A modernized Iskander with a projected range of 550 to 700 km is in development.
At the same time, Russia holds up to 400 S-300PM and S-400 air defense missiles adapted to strike ground targets, and makes about 40 a month. Meanwhile, Kinzhal production has been temporarily paused over poor accuracy, and the capacity has gone to conventional ballistic missiles.
Most new Kalibrs go into reserve
Russia plans to make more than 80 Kh-101 cruise missiles a month and is exceeding that, though defects in releasing them from aircraft have sent some batches back to the factories.
Russia has more than 450 Kalibrs and makes about 29 a month. Because Ukrainian air defense downs them effectively, Russia sends most new ones to its strategic reserve and to equip its fleets, HUR said. It also holds more than 600 Oniks anti-ship missiles, upgraded for ground strikes, and more than 120 Zircons.
Explore further
Ukraine recovers Russia’s undamaged Geran-3 jet kamikaze drone carrying camera and live-link equipment (VIDEO)
Russia now assembles 4,000 jet drones a month
Russia holds a reserve of about 6,200 Gerans and makes about 10,000 drones a month. Its main production hub is the Alabuga special economic zone in Tatarstan.
Since summer, Russia has shifted toward turbojet drones. “The reason is the saturation of Ukraine’s Air Force with mobile fire groups that shoot down low-speed targets en masse,” HUR said. Russia now assembles about 4,000 jet drones a month, mostly Geran-4s and Geran-5s, which fly at 400 to 500 km/h. It also makes about 2,000 cheaper propeller-driven Geran-2s, which HUR said are used to wear down air defense and locate air defense positions.
Hugging Face released a new AI model that can analyze and generate DNA sequences.
Samuel Boivin/NurPhoto via Getty Images
Hugging Face released a new AI model that allows users to "chat with DNA."
Carbon-X can analyze and generate DNA and allows researchers to ask questions in plain language.
Hugging Face said the model can help researchers who work outside the big AI labs.
Hugging Face is releasing an AI model that could make it easier for scientists to study DNA and develop new genetic treatments, the company announced Thursday.
Carbon-X is a "first of its kind" model that combines DNA analysis and generation with plain language, allowing researchers to ask questions and essentially "chat with DNA," according to Lewis Tunstall, a senior research scientist at Hugging Face.
Hugging Face, which Nvidia agreed to acquire for nearly $13 billion in September, is a platform for sharing and developing AI models and also releases its own. The company worked with research labs in Italy and China on Carbon-X, which is an open-source model. Hugging Face built Carbon-X by training an existing language model, Qwen3-14B, on genomic sequences and natural language. It was trained on DNA sequences from humans, other mammals, plants, and fungi.
As an example of how Carbon-X can be used, Tunstall said a researcher could provide a DNA sequence and ask whether a particular mutation could increase a person's risk of disease. The model would then offer a prediction and explanation. It can also generate new DNA sequences, which could help researchers develop new gene therapies.
Tunstall said the company was partly motivated to build Carbon-X to prevent biology researchers from becoming dependent on the major AI labs, noting how AI has disrupted mathematics.
"If you want to really push the frontiers of mathematics, you likely have to be at one of these AI labs like OpenAI or Anthropic, because they're the ones who have the best mathematics models," he said, suggesting a similar situation is likely to play out in other fields of science.
"We wanted to provide the open research community with tools that don't make them fully dependent on some external single provider," he added.
Researchers can run Carbon-X on their own computers, avoiding the need to share sensitive patient data with external AI companies. He also said that some of the misalignment behavior observed in larger models, such as the OpenAI agents that famously hacked Hugging Face in July, has not yet been observed in smaller, more specialized models.
Tunstall said generalist chatbots also sometimes refuse to answer legitimate biology questions from researchers because they mistake them for attempts to "design a bioweapon." He said viruses were explicitly excluded from Carbon-X's training data in order to avoid misuse of the model.
While Carbon-X is likely to be used by researchers, Tunstall said the model represents a small step toward a future where AI can help provide more personalized medical care.
"I think the holy grail is, in general, this concept of personalized medicine where we're able to, as individuals, take our own patient records, our own genomes, and use this to ultimately find or even maybe design our own treatments," he said.
Tech companies are painting data centers with gray coatings on top of gray concrete and steel to resist fire, rust and static electricity. That has made data centers a fast-growing market for paint makers.
Economists hope earlier episodes of labor disruption, from the Industrial Revolution to the Great Depression, can help us prepare for the age of artificial intelligence.
Login
