Gene Shalit, the 'Today' show's movie critic who looked like he raided a costume shop's 'eccentric professor' section, has died at 100. That 1994 car accident in Florida apparently convinced NBC that maybe their critics should review films from the safety of their living rooms instead of dodging traffic.
The New York Times' digital offerings remain as coherent as a fever dream, with today's paper featuring what appears to be broken HTML and a half-finished sentence about Iran, Israel, and the U.S. being stuck in some sort of geopolitical dance-off. Quality web development continues to elude America's paper of record.
Trump claims an Iran deal is 'complete' and will be signed today, while Iran's state media insists they 'forced' the U.S. to accept their terms—because nothing says diplomatic success quite like both sides claiming victory before anyone's actually signed anything. The Supreme National Security Council promises a statement 'shortly,' which in diplomatic time could mean anywhere from five minutes to next Tuesday.
Swiss voters apparently decided they'd rather risk bumping into strangers on the street than officially declare their country 'full' — rejecting a population cap proposal by nearly 55%, proving even the Swiss have limits on their love of precise numbers.
Trump announced a cease-fire agreement with Iran on social media, though Iran seems to have missed the memo since they haven't confirmed anything yet — classic case of declaring victory before the other team knows the game started.
India opened their Women's T20 World Cup campaign by absolutely demolishing Pakistan 64 runs, with Smriti Mandhana's 68 and Deepti Sharma's five-wicket haul giving the sellout crowd exactly the kind of rivalry beatdown they paid to see.
Hannah Hughes learned that Norwegian avalanches don't discriminate against Canadian tourists, delivering what she diplomatically described as a truck-sized wake-up call during her backcountry skiing adventure. Apparently Mother Nature's customer service in the Alps comes with a lengthy recovery period and no refund policy.
Hundreds of first responders lined Toronto streets Sunday to honor a police officer killed during a Thursday gunfight, because sometimes the most dangerous investigations aren't the ones that stay safely within our borders. The procession served as a somber reminder that international incidents have a way of coming home to roost in unexpected and tragic ways.
Toronto police have uncovered what appears to be a violence-for-hire shopping network, complete with a diverse product catalog ranging from U.S. consulate shootings to domestic tow-truck disputes. Nothing says 'professional criminal enterprise' quite like having both international terrorism and local parking lot beef on your résumé.
Rocket Lab's stock decided to celebrate its Nasdaq-100 inclusion by promptly face-planting, because apparently even good news is bad news in today's market. The aerospace company's shares tumbled despite the prestigious index addition, proving once again that investors have the attention span of caffeinated squirrels.
Trump announced a peace deal with Iran on Sunday, claiming to have resolved the oil-choking Strait of Hormuz crisis that's been rattling global markets. Given the timing and lack of details, this feels less like diplomacy and more like someone's weekend Twitter spree got accidentally filed under 'foreign policy.'
JetBlue is making a bold play for Fort Lauderdale with fancy new lounges and international routes, apparently convinced they can out-maneuver American Airlines' Miami fortress. It's like opening a lemonade stand next to Starbucks, but with more overhead bins and fewer comfortable seats.
The White House apparently slammed the brakes on Anthropic's latest models because China might have gotten their hands on them first—nothing says 'cutting-edge AI security' like finding out your adversaries are already logged in. If true, this represents either impressive Chinese espionage or embarrassingly porous American cybersecurity, possibly both.
AI startups are apparently planning to surf the coattails of any SpaceX IPO like remora fish attached to a rocket ship. Because nothing says 'we have our own compelling investment thesis' quite like hoping Elon's space venture will create enough market euphoria to float all boats, including your chatbot company.
Amazon's cybersecurity sleuthing reportedly helped convince the White House to yank Anthropic's fancy new models from international markets, with CEO Andy Jassy personally briefing officials on whatever digital vulnerabilities they uncovered. It's either responsible corporate citizenship or Amazon protecting its AI turf—probably both, with a healthy dose of 'we found the smoking gun' thrown in.
India's tech leaders are having an existential crisis about whether the Anthropic model ban reveals fatal flaws in their AI strategy or just bad timing. Nothing like getting locked out of the latest AI toys to spark a national debate about technological sovereignty and whether you've been building your digital future on someone else's servers.
Our intrepid gardener decided to solve their brown lawn crisis by prompting Gemini into coding them a custom app, which somehow worked despite ominous error messages about 'unrecoverably broken channels.' Modern problem-solving at its finest: when life gives you dead grass, make an AI build you software about it.
Meta is reportedly dismantling its $2 billion Manus acquisition after Beijing basically told them 'nice deal you have there, would be a shame if something happened to it.' Nothing quite like spending two billion dollars only to have a foreign government politely suggest you might want to reconsider your life choices.
Verizon's quality control apparently involves sending customers refurbished phones still loaded with corporate management software that can wipe your data at will — because nothing says 'premium carrier experience' like having your personal device remotely lobotomized by someone else's IT department. This delightful surprise raises some uncomfortable questions about whether Verizon actually bothers preparing used phones for new owners or just slaps a 'refurbished' sticker on whatever they find in the corporate lost-and-found.
A court has decided that Google can't hide behind the 'our AI said it, not us' defense when their systems generate false information, establishing that companies are legally responsible for their digital offspring's lies and hallucinations. This landmark ruling essentially tells tech giants that designing, training, and operating AI systems comes with the same accountability as hiring a particularly unreliable intern — except this intern has access to billions of users.
Today's tech newsletter covers the latest Silicon Valley obsession with 'reprogramming' aging because apparently death is just another bug to be patched, plus explores interoception — the hidden sense that tells you when your body needs attention (unlike most people's relationship with software updates). Life Biosciences has started human trials for age reversal treatments, presumably hoping to achieve what every tech executive dreams of: making mortality as obsolete as last year's iPhone model.
Some clever developer created Kage, a tool that can snapshot entire websites into a single binary file for offline browsing — because apparently we've reached peak internet anxiety where people need to hoard web pages like digital preppers stockpiling canned goods.
Former JPMorgan acquisition turned fraud defendant Charlie Javice is reportedly fishing for a Trump pardon, which is either a brilliant Hail Mary or the business equivalent of asking your ex to vouch for your character at your next wedding.
The FBI built a 22,000 square-foot fake town in Alabama complete with gas stations and hospitals to practice fighting cybercrime, proving that even federal agents occasionally need to touch grass — even if it's artificial grass in a glorified training facility.
Vercel just open-sourced their AI code review bot that lets Claude nitpick your pull requests like that one coworker who always finds the missing semicolon. It's still in beta, which explains why it hasn't learned to passive-aggressively comment on variable naming conventions yet.
Someone finally taught Claude how to draw those beautiful Excalidraw diagrams that make even the most convoluted system architecture look deceptively simple. Now your coding agents can create visual masterpieces that perfectly capture the chaos of your spaghetti code in neat, organized boxes.
A new lightweight inference engine promises faster AI responses through 'speculative speculative decoding' — yes, they said speculative twice, presumably because regular speculation wasn't speculative enough. It's like having an AI that's really, really good at educated guessing.
Finally, someone built proper DevTools for Claude Code so you can watch your AI assistant's every move like a helicopter parent monitoring their teenager's browser history. Now you can see exactly how many tokens Claude burned while deciding whether to use a for loop or forEach.
A curated collection of 1,209+ OpenClaw skills that gets updated weekly, because apparently we've reached the point where AIs need their own skill marketplaces. It's like LinkedIn for large language models, minus the humble-bragging posts about promotions.
Vercel's new tool lets you script browser demos and record them as videos, perfect for creating those impossibly smooth product demos where everything works flawlessly on the first try. Finally, a way to make your buggy prototype look like it was coded by wizards.