Connect with us

Tech

Dating apps have an AI photo problem. Hinge users are noticing

A young woman was scrolling through Hinge when she noticed something appeared slightly off with Johnno, a young man she matched with.
When she asked him if his photos were AI, he replied, “Hahaha appreciate it nah no AI here.”
But a reverse image search showed otherwise, revealing Johnno, or whoever was posing as him, had used AI to deepfake his face onto photos of other men from Pinterest.
When the woman shared her experience in a TikTok that racked up 1.2 million views, something even weirder happened in the comment section. Women posted screenshots of profiles from Andy, Roger and Ben, each with different facial features and hair, but sporting the same flannel shirt and linen pants Johnno was wearing, against the same beachfront background.
“Me realizing the guy I was talking to was using AI pics from this post…,” one user commented. “I came across a guy that did this too!! Every single photo came up on a photo search from Pinterest,” one woman commented. “It’s a breeding ground for catfish and it’s just gonna get worse so buckle up,” another user commented.
These increasingly subtle, hard-to-detect modifications are an issue dating experts say is becoming more common as users navigate a barrage of artificial intelligence in the dating world, from AI-aided prompts and conversations to increasingly sophisticated photo edits.
Amy Chan, a dating coach and the author of “UNSINGLE: How to Date Smarter and Create Love That Lasts,” says her clients have encountered an increase of AI-altered images and conversations.
How AI is changing the dating world
AI is rapidly changing the online dating world for young people, as more than half of adults under 30 report using dating apps like Tinder, Match, Bumble and Hinge, according to 2023 data from Pew Research.
2026 data from Match and the Kinsey Institute finds that nearly four in 10 U.S. singles have used AI in their dating lives. The trend is especially pertinent in young users, with 67% of Gen Z singles reporting using AI.
A crop of new apps like Your Move AI, FireTexts, Keys AI and Rizz function as AI wingmen, analyzing conversations and providing options on how to respond. And AI-native dating apps like Amata, Known, Sitch and Ditto are using artificial intelligence as the basis of their matchmaking technology.
In other viral videos, women recounted similar experiences with profiles like Johnno.
“This might be one of the most disturbing things I’ve ever seen on a dating app,” content creator Dana Kuritzkes says at the start of the video about the experience of a friend who matched with a man using AI-generated photos on Hinge in the Bay Area. “Her first thought, and I quote, was, ‘Why do all of these photos look like stock images?'”
In the video, she scrolls through screenshots of the profile, examining staged photos of a hike and one at a winery where people’s fingers are missing from the glasses.
Online dating coach Austin Dunham recounted that a friend sent him screenshots of a Hinge user who cloned a photo of him from Instagram in a white shirt against a palm tree.
“The fact that women are catching on to this and can spot where things are coming from goes to show how many guys are actually starting to use AI photos,” Dunham explained in a Sept. 17 TikTok. “I believe it’s going to cause a really big epidemic in terms of the online dating space,” Dunham said.
Chan says there’s nothing wrong with using AI for some grammatical help editing a bio. But using AI to essentially create a personality, or alter the one you have to be more interesting, is an issue.
While suitors may think they’re coming across as more intelligent or funny by using AI to come up with their bio and prompts, Chan warns it may have the opposite effect.
“Someone can read it, detect the AI copy and not be interested in engaging with you further,” Chan says. “Authenticity is a value people crave more than ever, (and) using washed down, AI answers is the antithesis of that.”
Dating apps know they have an AI photo problem. More are instituting live verification
Brandon Wade, founder and co-CEO of Seeking.com, says the platform has seen a proliferation of AI, with some individuals registering accounts under multiple phone numbers and testing out alternate profiles to compare results.
“People have realized hey it’s a numbers game,” Wade says. “With the use of AI they can now run multiple dating profiles on multiple different sites to be able to get the results that they want.”
Wade warns that these AI profiles can also be tied to scammers who target vulnerable individuals who are looking for companionship or professionals with disposable income.
The platform in January 2025 introduced a Selfie Liveness Check to verify a live selfie against uploaded profile photos. Data from 2025 found the new feature caught 53,000 scammers on the platform, many of who were posing as entirely generated users of different races, ages and genders. Seeking.com found that a majority of these scammers were coming from romance scam hotspots in Africa and Southeast Asia.
“We do see more of that sort of activity specifically from scammers because they are trying to cast a much broader net and target people that way. With actual genuine daters, that is not a behavior that we’re seeing as often,” Wade says.
Tinder and Bumble also started requiring similar biometric face checks for U.S. users.
Hinge offers optional AI-enabled features to help users improve their prompts. They employ a combination of AI-based moderation tools and human moderators to vet profiles. In March 2026, they started rolling out Face Check, a verification process that uses a selfie video to confirm a person’s authenticity and that their face matches their profile.
“The authenticity of our community is important to helping daters find safe and meaningful connections,” Ben Celebicic, Hinge’s chief product and technology officer, told USA TODAY. “If a user decides to include generative AI on their profile, it should not be used to misrepresent themselves or their intentions. Our goal is to ensure that our users only leverage AI to build trustworthy connections with other daters, so they can find love and delete the app.”
The issue with these verification processes, commenters online point out, is that many instances of AI do use someone’s real face, bypassing verification processes, but later edit subtle elements of a photo to make their attire, physique or background more interesting. These apps also face privacy concerns from users about how facial verification videos and photos are collected are stored.
Chan says it’s the digital equivalent to lying, and that the same goes for over-filtering a photo. Men in the comments have pointed out that they’ve also encountered AI-fueled photos on women’s profiles, as well as heavily edited ones.
“I have had clients who show me their profile, and I’m like, ‘that doesn’t look like you.’ They get defensive and I have to tell them matter of factly, the photo has been so altered, or it’s a very old photo that they’ve altered, that they’re setting themselves up for a bad date,” Chan says.
Instead, she advises clients to keep their profiles realistic and true to themselves. She recommends exchanging a few messages and going straight to either a video chat or IRL.
Otherwise, users risk going right back to square one.
“The person will feel shocked and deceived and that is usually impossible to bounce back from on a first date,” Chan says.

Continue Reading

Tech

Good Riddance to Elon’s Musk Cybertruck

In early 2024, I started seeing one particular Tesla Cybertruck driving around Brooklyn. It was the first one I’d seen — a matte black version with a decal on the side that said “To the Moon,” a common phrase associated with trading cryptocurrency. I remember seeing it pop up in Williamsburg, in Greenpoint, and scattered spots throughout Bedford Stuyvesant and Crown Heights. At the time, it was a novelty — the only evidence I had seen in my little New York bubble that anyone was willing to buy these things.
Evidently, though, people did buy Cybertrucks. Just not enough of them. All told, Tesla sold around 70,000 of the vehicles between their release in November 2023 and Friday, when CEO Elon Musk announced that the company would wind down production in favor of a new self-driving taxi, the Cybercab. When the Cybertruck launched, Musk predicted it would sell 250,000 units a year, one of many confident claims the tech billionaire has delivered that wildly overshot the mark. Very few of his products, it turns out, do actually go to the moon.
But how could we ask for a better emblem of this stupid, violent, get-rich-quick era of America? The Cybertruck was a vehicle designed to the specifications of one man, who made himself the richest person in the world through a savvy combination of government kickbacks, stock manipulation, and viral marketing stunts. They were designed to look like a futuristic war-vehicle, manufactured in Texas at an absurdly named building called the Gigafactory, and equipped with bullet-proof side panels and doors. The Cybertruck was a car built for the paranoid fantasies that have gripped so much of a country destabilized by generations of economic theft. Electric to assuage the existential worry of climate change, bullet-proof to match America’s extremist gun culture and pervasive fear of crime, shaped like a video-game toy to please the frontal lobes of a generation doused in the recycled franchises of comic book heroes and supersoldiers.
Editor’s picks
They were also absolutely dogshit cars. The electric systems were prone to software bugs. Some trucks were known to shut down completely when they got wet. Their odd shape and bizarre design made them impractical for hauling, loading, or towing anything. In 2025 the company recalled 46,000 of them because it found they had a specific piece of side panel that fell off in the cold. Almost immediately after they came out, automotive YouTubers and professional mechanics posted videos of Cybertrucks that came in with ill-fitting body work and shoddy production quality, in itself a common hallmark of Tesla’s more recent vehicles. To be fair, the doors were bulletproof, provided you shot them with a small enough caliber of gun and didn’t aim for the windows.
Shortly after announcing the Cybertruck, Elon Musk made an overt political move in backing Donald Trump’s 2024 campaign for president, pouring nearly $300 million into a dark money organization called America PAC and other ventures. Online, he ranted about trans people (he has a trans daughter who hates him), immigrationillegal immigrants (he was born in apartheid South Africa), and “woke” politicians pushing DEI policies (his factories were reported to be alleged hotbeds of racism and sexual harassment that have resulted in multiple lawsuits). When Trump took office in 2025, Musk leapt into a role as head of the Department of Governmental Efficiency, or DOGE, which itself was a reference to a 2010s-era Reddit meme about the Shiba Inu dog breed that subsequently inspired one of the first crypto “shitcoins.” DOGE promptly fired thousands of federal workers, ground multiple government agencies to a halt, and eviscerated USAID, a decision which directly contributed to tens of thousands of preventable deaths overseas of disease and famine in resource-starved regions of the developing world.
Excited dorks who had signed up to buy the Cybertruck found themselves having to slap on bumper stickers that said “I bought this before Elon went crazy” so they wouldn’t get vandalized in public. Others, like the crypto nerds and Trump-friendly bros who made up the enormous middle section of the MAGA-Musk Venn Diagram, often decided to double down, wrapping their Cybertrucks with right-wing decals and tinting their windows with American flags. Simply driving one in public became a political statement, whether the driver intended it or not: I’d be lying if I said I didn’t sneer a little every time the matte-black to-the-moon-mobile drove past me in Brooklyn.
Related Content
I’ll admit, I was predisposed to hate the thing. A decade ago, in what feels like another life, I was a tech reporter at a startup publication that catered specifically to the Cybertruck demographic (though the truck itself didn’t exist yet). Our audience loved Marvel movies, sci-fi nostalgia, and Obama-era techno-optimism. Above all else, they loved Elon Musk. Musk, at the time, was presenting himself as a future-focused benevolent tech pioneer, with cameo roles in Iron Man II and a non-stop stream of quirky gags online, like his cars’ “Ludicrous Speed” mode and an Easter egg on the onboard computers that played the “Nyan Cat” meme. If those words mean nothing to you, that’s probably a good thing.
Trending Stories
What it showed, however, was Musk’s core nature, as a nihilistic salesman who would say anything in order to sell his products, no matter how divested from reality they were. Case in point, in 2017 Musk said that the “Fully Self Driving” feature on his Tesla vehicles could make them capable of “hands-free cross-country road trips” by the end of the year. Autonomous driving technology has come a long way since then, but the road trips never happened. Instead, Musk moved on to other schemes and startups, some successful, some not. The Cybertruck’s successor, the Cybercab, hopes to compete with Google’s driverless Waymo project, though neither vehicle is capable of those fully autonomous road trips yet. Maybe in another 10 years.
By that time, I would be shocked if there are any Cybertrucks left on the road. Perhaps they’ll sit on blocks in the front yards of high-school dropouts who hit it big in the 2020s crypto booms. More likely they’ll end up as scrap, their somewhat bullet-resistant steel panels crumpled down or gathering rust while their massive lithium-ion batteries leech chemicals into the groundwater of some unfortunate industrial dump-adjacent community. Elon Musk, of course will still be rich, his companies subsisting off of an endless spigot of government cash and circular stock buybacks, no matter how many elections the Democrats manage to win. But maybe, if you squint a little, you can see a different future, one in which the occasional real and valuable advances made under mercurial hucksters are used for the greater good, instead of hoarded by monopolies enabled by politicians on the take. Maybe someday we’ll see the Cybertruck as a relic of a dark period in our past, and not a dire portent of the even more broken future down the road.

Continue Reading

Tech

Anthropic IPO hits AI backlash after midterms show rebuke data centers

watch now
Americans were unequivocal at the ballot box this week as they expressed disdain for tech billionaires and mistrust of artificial intelligence. With that backdrop, Anthropic is about to try to go public.
The AI company is gearing up for a potentially historic IPO, likely targeting mid-November for its Nasdaq debut. CNBC reported on Friday that Anthropic investors see the company aiming for a valuation of $1 trillion to $1.5 trillion, down from a target of $2 trillion that had been floated in recent months.
From Wall Street to Main Street, it’s becoming a tougher environment for Anthropic. The Nasdaq on Thursday suffered its steepest drop in 19 months, and it’s down to start November after October marked its worst month since March 2025. Chipmakers and neoclouds, the companies most closely tied to the AI buildout, are getting hit the hardest.
Voters rebuked AI in Tuesday’s midterm elections, while sending Democrats to sweeping congressional victories across the country. Candidates who advocated for more stringent AI regulation won handily in states like Michigan, Ohio, Texas and Georgia, which will all now have Democratic governors, and local communities opposed data center-related ballot measures in 40 out of 41 cases.
More than half of voters said AI data centers were a “crucial” or “major factor” in who they supported for Congress, according to an exit poll from NBC News.
“People don’t want a data center going near them if it’s just going to house AI that’s going to help a big tech company profit,” Maura Austin, a behavioral scientist at the University of Virginia, told CNBC. “What Anthropic is doing when they’re going public with such a high valuation is they’re really leaning into that big tech, billionaire kind of caricature. I expect that’s not going to go well with the public.”
Anthropic has long tried to position itself as a safety-conscious company that takes AI’s potential risks and benefits seriously. CEO Dario Amodei co-founded the company alongside a handful of researchers who defected from OpenAI , now its chief rival. Anthropic is best known for developing a family of AI models called Claude, and it primarily sells its products and services to other businesses.
The company’s enterprise focus has helped it avoid becoming the public face of AI, leaving that distinction to OpenAI and CEO Sam Altman. But with Anthropic poised to become the first pure-play AI company to hit the public market, minting a number of new billionaires and many more millionaires in the process, a heightened level of scrutiny is coming.
“Being a public company, just that by itself pushes you in the limelight,” Reena Aggarwal, a professor of finance at Georgetown University, said in an interview. “And then AI is such a hot-button issue that Anthropic will be in the limelight even more.”
Matt Kennedy, a senior IPO market strategist at Renaissance Capital, said that’s not necessarily a bad thing for Anthropic.
Public markets force greater transparency, he said, and the company could attract new users through its debut. And since AI has become such a mainstream conversation topic, several private companies are already household names.
“I would say that a public company does not deserve more criticisms than a comparable private company,” Kennedy said. “So Anthropic should not be put in a harsher public spotlight than OpenAI just because it’s public.”
Anthropic declined to comment.
Data center blowback
Kent Nishimura | AFP | Getty Images
AI backlash accelerated over the summer after a series of high-profile safety and security incidents rattled the public. As concerns mounted, the debate over data centers intensified, while tech’s hyperscalers simultaneously boosted their forecasts for capital expenditures, each reaching into the hundreds of billions of dollars.
Anthropic and OpenAI depend on state-of-the-art data centers to train models and serve customers, and they’re trying to build out capacity as quickly as possible to meet demand. Anthropic said in an amended IPO prospectus this week that it has entered into purchase commitments and lease agreements with an aggregate committed spend of about $895.4 billion, “substantially all of which is related to compute and infrastructure.”
Americans grappling with inflation and stubbornly high gas prices have grown worried that the facilities could exacerbate environmental harms, raise utility bills and contribute to job loss. For many, a proposed data center in their backyard became the physical manifestation of the widespread antipathy toward AI.
In the risk factors section of its prospectus, Anthropic said opposition to data center development has become more organized across jurisdictions, and that the company’s public profile could draw particular scrutiny to projects associated with it.
Nicole Greene, an analyst at Gartner, said she expects that Anthropic’s executives have long been prepared to contend with AI backlash, especially given the disruptive nature of the technology.
“There’s going to be big emotions,” Greene told CNBC. “With big emotions and feelings — fear and excitement — comes heightened scrutiny.”
Amodei acknowledged the souring sentiment around AI in a post on X in August. He said it’s the result of a “crisis of trust,” and that people believe companies, the government and the tech industry are always “cooking up some new way to screw them over.”
Rebuilding public confidence is going to be an uphill battle as the company begins providing quarterly financial updates.
The University of Southern California published a survey on Monday that found that 71% of respondents don’t trust big tech companies to act in society’s best interest “at all” around data center development.
“Once you’ve lost trust, it’s hard to gain it back,” Wändi Bruine de Bruin, a professor at the University of Southern California’s Schaeffer Institute, said in an interview. “The numbers are not good.”
WATCH: Alex Kantrowitz: Anthropic doesn’t have a lot of flexibility on timing of its IPO

Continue Reading

Tech

These five stock types could rally after U.S. midterms, analysts say

Democrats swept in the U.S. midterm elections this week, and that could be a boon for several groups of stocks, according to Wall Street analysts. On Tuesday, Democrats flipped seven Senate and more than 40 House seats, giving the left-leaning party majority representation in Congress. That means Democrats will have greater say over legislative priorities on Capitol Hill — a power that has big implications for companies and their stocks. Stocks were volatile in the trading sessions following Election Day. The S & P 500 finished the week down 2.7%. Tech stocks got hit particularly hard, with the Nasdaq Composite falling 3.4% during the same period. The 30-stock Dow Jones Industrial Average fared better, only sliding 0.9% during the week. However, Wall Street has identified some potential winners amid the churn. A blue tsunami could bolster U.S. solar, airline and healthcare stocks, among others, several analysts said in notes to clients. Here are a few stocks they highlighted. Drone-related stocks With more Democrats soon to be in power, drone makers and drone battery manufacturers are likely to see their shares trade higher, according to William Blair. The investment firm noted that Democrats have stated their intention to lend more support to Ukraine in its war against Russia, which could lead to more money flowing to companies that make drones and the parts that power them. AMPX 3M mountain Amprius Technologies stock over the past three months One such firm that stands to benefit is drone battery manufacturer, Amprius Technologies , per William Blair. Roughly 50% of the company’s revenue has been directly tied to Ukraine or to allies supporting the effort in 2026, according to the investment firm. “We view Amprius as a midterm winner as the ‘blue wave’ brings more support for Ukraine, a large driver of Amprius’s demand,” analyst Mark Shooter said Friday. “The midterm outcome could reverse the U.S. position from decreasing to increasing Ukraine funding, de-risk demand from this conflict, and offer expansion to Amprius revenues in 2027.” Shooter has an outperform rating on Amprius. The stock has pulled back more than 22% over the past three months, but added 1.5% this week. Clean energy stocks Renewable energy firms, which generate energy via solar, wind or hydro power, could get a boost following the midterm elections, according to several industry analysts. “Some smaller policy tailwinds are possible: Congress is still considering a permitting reform bill … [which has been] flagged as benefiting clean energy and geothermal,” Wolfe Research analyst Tobin Marcus said Thursday. “And some incoming governors are proposing renewable mandates for data center power.” FSLR 3M mountain First Solar performance over the past three months One stock that stands out is First Solar , according to William Blair. “Democrats swept the midterms, winning both chambers of Congress and flipping five governorships in Texas, Georgia, Ohio, Iowa and Nevada, as of the time of writing,” analyst Jed Dorsheimer wrote in a Thursday note to clients. “Ultimately, we believe this will unlock support at a local level for more utility-scale solar procurement and siting.” The Bipartisan American Affordability and Jobs Act , which aims to expedite federal permit issuances to solar projects and expand grid development, is likely to become law with Democrats in power, the analyst added. Dorsheimer upgraded First Solar to outperform from market perform. Shares of the solar panel maker were up 15% over the past week, though the stock is down 23% so far in 2026. Healthcare Healthcare stocks are also poised to benefit from Democrats’ congressional gains. Making medical care accessible is a major aim of the political party’s agenda, which bodes well for some healthcare stocks, according to Wolfe Research. “Presumptive Speaker Jeffries has explicitly indicated that [lawmakers’] ‘first 100 hours’ agenda will include fully repealing the Medicaid cuts … and restoring the expanded [Affordable Care Act] premium tax credits that expired at the end of 2025,” Wolfe’s Marcus wrote Thursday. “So it makes perfect sense for investors to be focused on healthcare after the blue wave.” The analyst noted that the potential tailwinds could be greatest for hospitals, although it could take “quite a while” for those to materialize. Shares of hospital operators HCA Healthcare and Universal Health Services have popped 4.6% and 11.1%, respectively, since the beginning of this week. More broadly, healthcare stocks have surged as of late. The Vanguard Health Care Index Fund ETF (VHT) has climbed more than 10% over the past three months as investors have rotated out of technology stocks and embraced more defensive plays due to credit risks tied to the ongoing artificial intelligence buildout. Airline stocks Airline stocks could rally following a big blue sweep this election season, according to Citigroup. “Investors … may be missing the effect that a dramatic increase in consumer sentiment among Democrats in the wake of the midterms could have on consumer discretionary demand, such as air travel,” Citigroup analyst John Godyn said Wednesday in a note to clients. “As long as Republican consumer sentiment does not collapse 1:1 with any improvement in Democrat consumer sentiment (and implicitly assuming that Democrats and Republicans travel roughly a similar amount per capita), one could argue that the midterms may give airlines a boost in demand.” The U.S. Global Jets ETF (JETS) , which invests in U.S. and international airlines, has risen about 2% since Election Day. Delta Air Lines and United Airlines led gains among airlines this past week, jumping 8.8% and 8.2%, respectively. Specialty REITs A few specialty Real Estate Investment Trusts are likely to rally, according to Barclays. “Voters largely favored candidates and ballot measures that oppose data center development … [which] could slow the pace of industry-wide construction,” analyst Brendan Lynch said Wednesday in a note to clients. “We view our coverage as net beneficiaries from the pricing power that limitations on construction will create.” Lynch pointed to Equinix , Iron Mountain and Digital Realty Trust as three potential beneficiaries of that shift. “Our sense is that EQIX and IRM’s pipelines are essentially fully approved, and DLR faces only limited permitting risk for projects that would commence years from now,” the analyst wrote. Those three REITs have fallen between 4% and 13% over the past three months, largely due to rising interest rates that have diminished the appeal of some dividends and lowered property valuations. However, the pullback offers a good opportunity to scoop up shares, according to Barclays.

Continue Reading

Tech

I tried local AI on HP’s Omnibook Ultra

ZDNET’s key takeaways
PCs built on the Nvidia RTX Spark architecture are superb at AI tasks.
Running a local open-weights model has no per-token charges or usage limits.
The biggest advantage of running a local AI model is privacy.
Back at the dawn of the PC era, PC Magazine’s legendary editor, Bill Machrone, came up with Machrone’s Law: “The machine you want will always cost about $5,000.” Maybe that number is a little higher today thanks to the current RAM crisis, but adjusted for inflation, Machrone’s Law still holds true. The proof? I have a $7,000 notebook PC sitting on my desk right now, and I must confess that I really, really want it.
Also: I put Lenovo’s RTX Spark laptop through a completely bonkers AI coding test – and it passed
The HP OmniBook Ultra is an absolute beast. It’s powered by Nvidia’s RTX Spark, with 128 GB of unified memory, 2 TB of fast Gen5 NVMe storage, and a great-looking 16-inch OLED display. It weighs nearly five pounds, but it isn’t really so much a portable PC as it is a personal data center, offering 1 petaflop of FP4 AI performance.
More from ZDNET
I am not the target market for this PC. But after using it for a few weeks, I am seriously tempted. For AI-related work, its GPU has many times the power of a garden-variety Copilot+ PC, and it’s able to take full advantage of the unified memory.
This PC finally gave me a chance to tackle a project I’ve had on my to-do list for years: creating a fully searchable index of 20 years’ worth of my posts here at ZDNET. With the help of a powerful local AI model, this PC got the job done in one day, without costing me a dime in AI subscription charges or sending my data to someone else’s cloud.
How local AI raced through my project
One reason my project completed so quickly was that I was able to run it uninterrupted. I used Copilot to help me set up the working environment initially, using the open-source llama.cpp inference runtime, the open-weights Qwen3.8-27B large language model, and a coding harness called Cline, running as an extension in Visual Studio Code. With a little extra help from Copilot, I configured those tools, wrote a detailed spec for the app I wanted, fed it to Cline, and then mostly stood back and watched.
I had a good general idea of what all those components were doing, but it would have taken me weeks of reading, studying, and trial-and-error manual installation to get everything set up and running. Using the PowerShell scripts that Copilot wrote, I had the project under way within a half-hour.
As the project ran, I got to watch the AI “reason” in real time. Every so often it asked me to approve something, but mostly it wrote code, ran tests, observed the results, modified the code, ran some more tests, and so on. The whole process was a textbook example of “human-in-the-loop agentic coding,” and all I had to do was check in every so often to see how things were going.
I knew it was busy, because when the LLM was working, I could hear the fan spinning from across the room. Surprisingly, the machine never got uncomfortably hot; still, this is the sort of work that’s best done with the laptop sitting on a desktop, not in a lap.
Also: I can access my home PCs remotely for free – how you can too (no VPN needed)
I gave the LLM a spreadsheet filled with links to my articles, which I had painstakingly built over the past two years. Not only did it find and retrieve nearly every article in that list, it built a separate routine to retrieve pages from the Internet Archive’s Wayback Machine for roughly 90 links that returned 404 errors, and then presented me with a list of another 90 or so pages it couldn’t retrieve at all. It took maybe 30 minutes to review and resolve that last set of issues.
After all was said and done, I had two folders, each containing 1,951 files — one set in raw HTML, the second in Markdown format. And to wrap up, I had my local AI build a simple script that I can run to add new posts to the archive as they’re published.
I cannot begin to tell you how happy this result made me.
How much did it cost?
The complete project used 479 LLM inference calls and processed nearly 50 million logical tokens, at a total net cost of exactly $0.00.
One of the most frustrating aspects of using these modern AI coding harnesses is that there’s no way to estimate how many tokens they’ll use.
I could have used a commercial AI subscription to do this job, but I would have been flying blind. None of the companies that sell those plans can tell you how much a given project will cost. Your actual consumption — and thus the final bill — depends on factors you can’t know in advance. How much code does the agent need to keep in context? How many iterations and debugging cycles will it require? Which models will you use?
Also: Is Claude Code truly faster with message threads? I did the math on my logs
ChatGPT Plus (with Codex), Claude Code Pro, or GitHub Copilot all cost between $10 and $20 per month. That’s more than enough for a simple task, but for this project, I would have quickly slammed into usage limits with those entry-level plans. To keep working with a $20-a-month plan, I would have had to take frequent breaks, or buy extra credits, or upgrade to a $100-per-month plan with higher usage limits, like ChatGPT Pro, Claude Code Max, or GitHub Copilot Max.
After staring at those numbers, I got a sense of where Anthropic, OpenAI, and Microsoft’s business models are going. Even at $100 a month, if I decided to run some even more ambitious projects, I could have easily found myself rate-limited, with work having to wait for hours or days until the next cycle began.
For serious coders, those bills can get very pricey. On a recent episode of Nilay Patel’s Decoder podcast, analyst Joey Brookhart offered a startling statistic: “The top 5 percent of Claude Code users are spending $18,000 a year on average. Non-Claude Code users spend well under $75 a month.”
Now I know why.
The biggest benefit of all
Using a free, open-weights model with an open-source coding harness gives you the freedom to experiment without worrying about the bill. But that’s only part of the story. The much bigger benefit is that you’re not sending potentially sensitive or confidential data to someone else’s cloud; with a fully local AI stack, your prompts, documents, and inference workloads don’t have to leave the PC at all.
That wasn’t a big deal for my project, which used content that had already been published to the web. But I have a couple other apps in mind that I’d rather not share with anyone else until they’re ready. One involves a big stack of financial documents, credit card statements, and invoices. I’m happy to feed that data into an LLM on my own PC, but I’m not comfortable uploading those documents to a cloud-hosted coding service.
Also: This free Windows tool is my secret weapon for setting up new PCs
Will I do enough of this kind of work to justify spending $7,000 on a PC? I’d love to say yes, but my CPA would just stare at me for a full minute and then patiently explain how “return on investment” works. “No, Ed, you can’t justify this expenditure.”
The good news is that Machrone’s Law works hand in hand with Moore’s Law, which describes a reliable technology trend: computing power doubles every 18 months or so. That means in another three years I’ll be able to drool over a PC with four times this much computing power for $7,000. But by that time a configuration like this HP OmniBook Ultra will probably be a mainstream business-class machine, available for a fraction of its current price.
Until then, I’ll wait.

Continue Reading

Tech

Elon Musk said the Cybertruck would be Tesla’s ‘best product ever.’ Now he’s winding down production

In a post on X, Musk said: “Tesla will be winding down production of the Cybertruck at Gigafactory Texas around the middle of next year to make room for increased Cybercab production.”
The changes to Cybertruck production come seven years after Musk introduced the vehicle at a splashy Sci-Fi-themed event at Tesla’s design studio in Hawthorne, Calif., complete with pyrotechnics and a live demonstration of the car’s “armor glass” in which Tesla’s lead designer shattered two Cybertruck windows with a steel ball. Just a month before the launch event, Musk had touted the vehicle as Tesla’s “best product ever” during Tesla’s third quarter earnings call in 2019.
“This could be totally wrong. I could be totally out to lunch here, but I think the Tesla Cyber Truck is our best product ever. That’s my opinion,” he said at the time.
Yet, the Cybertruck, which starts at a whopping $92,490, has been popular, but it has struggled to find the mass appeal of other Tesla vehicles like the Model Y and Model 3.
After production delays pushed the vehicle’s launch date to 2023 from a previously promised 2021 launch, production still fell short of the between 250,000 and 500,000 Cybertrucks Musk predicted Tesla would eventually produce annually.
Tesla sold 38,965 Cybertrucks in the U.S. in 2024 and then just 20,237 in 2025, a decline of 48%, according to estimates by Kelley Blue Book. In the first half of the year, Kelley Blue Book’s parent company Cox Automotive also estimated Tesla sold a total of 7,263 Cybertrucks, which would put them on track for yet another year-over-year sales decline this year. Tesla does not break out its Cybertruck sales in its public filings.
Musk has now pivoted the company’s future prospects to the autonomous Cybercab, which he announced to fanfare at an event in 2024. The first Cybercabs hit roads in Austin, Texas, in September. Tesla also has Robotaxi operations, using the Model Y, in Houston, Dallas, and Fort Worth in Texas as well as Miami, Orlando, and Tampa in Florida, with plans for more U.S. cities to be included later.
Tesla did not immediately respond to Fortune’s request for comment.
New ambitions
The shift away from Cybertruck production comes as Tesla seeks to convince investors autonomous technology can drive growth in the future by surpassing traditional EV sales, which have fallen since President Donald Trump ended the EV tax credit last year. Tesla sells the most EVs of any company in the U.S.
The EV maker has positioned its Cybercab, a two-seater without a steering wheel or pedals, as key to its expansion into autonomous transportation and robotics.
But the new vehicle also faces hurdles. Federal auto safety regulators in September opened an investigation into whether Tesla properly certified that the Cybercab complies with federal vehicle safety standards. At the end of October, Tesla submitted its answers to questions from the National Highway Traffic Safety Administration (NHTSA), but requested its responses remain sealed because they are “confidential business information.” The NHTSA has not yet ruled on Tesla’s confidentiality request, according to Electrek.
In the autonomous robotaxi market, Tesla is entering late. Alphabet’s Waymo already operates in 15 U.S. cities and reportedly provides 500,000 rides per week, it said in an X post earlier this year.
For its part, the Cybertruck has been hit with 11 recalls since customer deliveries began in November 2023, according to Reuters. Price changes that made it significantly more expensive than Musk’s original vision have also stood in the way of mass adoption. Tesla advertised a starting price of $39,900 when it introduced the truck in 2019—but by the time deliveries began in 2023, Tesla was advertising a version of the vehicle expected to cost $60,990.
Tesla raised the price of several models of the Cybertruck this month and discontinued orders for the cheapest model, the Dual Motor AWD. That means the cheapest Cybertruck model still available starts at $92,490, up from the $84,990 price the same model reached in August after previous price hikes, according to Electrek.
Tesla has previously tried introducing less expensive versions of the Cybertruck to attract buyers. The company launched a rear-wheel-drive Cybertruck in 2025 but pulled it from the lineup months later.
Musk’s Friday post also leaves existing Cybertruck owners in limbo. Tesla has not said how long it will continue selling remaining inventory or supplying replacement parts, though winding down the vehicle’s production does not mean it will end support for existing vehicles.
Musk, for his part, still appeared enthusiastic about the truck Friday.

Continue Reading

Latest News

Video1 hour ago

Trump offers UFC CEO Dana White press secretary job

President Donald Trump offered UFC CEO Dana White the role of White House press secretary, but White declined, according to...

Video1 hour ago

Democrats at Crooked Con reflect on a historic week

CNN's Donie O'Sullivan talks to Democrats and progressive activists who have gathered in Washington, DC, this weekend for …

Business2 hours ago

NASA unveils $10B Moon base contract as SpaceX, Blue Origin join lunar gold rush

NASA announced Saturday that Elon Musk’s SpaceX, Jeff Bezos’ Blue Origin and six other aerospace companies are eligible to compete...

Business3 hours ago

Trump’s Recent Warning Is Sending Shockwaves Through Wall Street. Here’s What It Could Mean for Your Retirement Savings.

Love him or hate him, President Donald Trump’s time in the White House has historically correlated with strong stock market...

Entertainment4 hours ago

Stephen Colbert rushed out of Harvard event, accosted by union protesters

Liberal former late-night host Stephen Colbert had to be rushed out of an event at Harvard after someone pulled a...

Entertainment4 hours ago

Josh Brolin renews vows with wife Kathryn in Vegas for 10th anniversary

Josh Brolin renewed his wedding vows with wife Kathryn in a conspicuously unbuttoned shirt. In a joint social media upload...

Sports4 hours ago

Alabama finally puts it all together, dominates LSU and leaves Lane Kiffin staring at $40 million CFP disaster

Alabama spent the past five days taking heat from across the country over its No. 5 ranking in the initial...

Entertainment4 hours ago

Tim McGraw says communication is key to surviving marriage changes

NEWYou can now listen to Fox News articles! Tim McGraw shared his insight into what the “most difficult” part of...

Video4 hours ago

Democratic congresswoman calls Trump's chief of staff move 'desperate'

Democratic Rep. Madeleine Dean joined CNN's Jessica Dean and shared her reaction to President Donald Trump picking …

Entertainment4 hours ago

Bridget Fonda makes rare LA appearance 24 years after her last role

NEWYou can now listen to Fox News articles! Bridget Fonda stepped out for a rare public outing in Los Angeles,...

Trending News

Join Our Newsletter

Stay updated with breaking news and exclusive content.