A daily curated digest with the tech news that matter + community vibes, delivered daily, in tabloid style. Like you always wanted.
We track a tech day split between security alarms and AI ambition... GitHub faces a breach tied to internal repos, npm churns out hundreds of malicious packages, and exposed CISA GovCloud keys raise fresh questions about basic cyber hygiene... A Google Cloud wobble knocks Railway sideways while one hijacked GitHub Pages subdomain turns into a spam mess visible in Google Search... Then Google I/O shifts the picture, with Gemini 3.5 aimed at agents and Search pushing deeper into AI conversation... At the same time, the money question hangs over AI, Anthropic gets a jolt from Karpathy, and software work keeps moving toward prompts instead of hand-written code... The mood is alert, curious, and uneasy.
GitHub Pages Turns One Domain Into Spam Trap
A traveler came home to find a subdomain quietly hijacked through GitHub Pages, with junk pages already showing up in Google Search. It felt like a perfect storm of weak ownership checks, silent failure, and cleanup pain.
GitHub Breach Hits Internal Repos
GitHub disclosed unauthorized access tied to a compromised employee device, and the blast radius included internal code tied to VS Code and extensions. Even with quick containment, this is the kind of leak that rattles every developer.
Hundreds of npm Packages Turn Malicious
A hijacked npm account sprayed hundreds of bad releases across more than 300 packages in minutes, a reminder that open source supply chains still crack at the weakest human link. Anyone pulling updates today had good reason to sweat.
CISA Leaves GovCloud Keys in Public
A contractor repository exposed powerful AWS GovCloud credentials for CISA on public GitHub, which is exactly the sort of mistake you expect the cyber cops to prevent, not commit. The story landed with all the grace of a banana peel.
Google Cloud Trouble Knocks Railway Offline
Developer platform Railway spent the day wrestling a major outage linked to Google Cloud, leaving deploys stalled and dashboards shaky. It was another loud reminder that modern apps are often one upstream wobble away from chaos.
At Google I/O, the company pushed Gemini 3.5 as a smarter model built for action, not just answers, with a faster Flash version riding shotgun. The pitch was clear: agents are the new battlefield, and everyone is sprinting.
Google Search Swaps Blue Links for AI
Google rolled out a big Search makeover centered on AI conversation and an intelligent box, making the old ten-blue-links web feel suddenly antique. Publishers have every reason to look nervous while users brace for weirdness.
One blunt argument cut through the buzz: AI may be impressive, but the spending is wild, the margins are foggy, and the energy bill keeps climbing. It captured the growing suspicion that some shiny products still need a real business.
Andrej Karpathy joining Anthropic instantly turned into industry gossip fuel, because talent moves still signal where the serious bets are being placed. When a star researcher switches jerseys, people read it like a scoreboard.
One Engineer Stops Writing Code
A founder saying he no longer writes code and ships by directing AI lit up the old argument all over again. Whether you call it liberation or chaos, the message was unavoidable: software work is being reorganized around prompts and review.
Plex Lifetime Pass Gets Luxury Pricing
Plex stunned users by raising the Lifetime Pass to $749.99, a price jump so huge it made the word lifetime sound like a threat. It was a perfect example of beloved software discovering just how much goodwill it can burn in one post.
While much of the industry chases AI glitter, OpenBSD 7.9 arrived with the usual calm list of improvements across hardware and networking. It was the kind of release that quietly reminds everyone reliable software still matters.
A Museum Puts Lost Operating Systems Back
One builder assembled a virtual museum packed with old operating systems running under emulation, turning forgotten interfaces into something you can actually wander through. It hit that sweet spot of nostalgia, preservation, and geek joy.
Strawberry Gets Hollywood Style 3D Scan
A painstaking Gaussian Splat of a strawberry showed how far homegrown 3D capture has come, with dozens of angles and focus-stacked shots producing a weirdly gorgeous result. It was technical, yes, but also plain old internet catnip.
ZIP Shrinker promised smaller archives right in the browser, including file types that secretly ride on ZIP under the hood like APK and EPUB. It is exactly the sort of practical little hack people love because it solves a real annoyance.
Tonight, AI and open source lead the bulletin... Claude reportedly helps an attacker pull 150 GB from Mexican government systems, while Bambu Lab faces questions over an AGPL network component... The Linux security list strains under waves of AI-found bugs, and maintainers push back against junk AI-generated pull requests... At the same time, Anthropic buys Stainless to build stronger agents, Qwen 3.7 Preview climbs the model charts, research probes censorship inside Qwen 3.5, and Modal says it cuts cold starts by up to 40x... We follow a tech scene where trust, speed, and control move to the center of the story.
One Claude user raids government data
A single attacker reportedly used Claude to help pull 150 GB from Mexican government systems, a nasty reminder that old security holes become far cheaper to exploit when an AI assistant is handling the boring parts at machine speed.
Bambu Lab landed in a licensing storm after a detailed claim that Bambu Studio hides a network component that should be released under the AGPL. This looks less like a tiny paperwork slip and more like playing cute with open source rules.
Linus Torvalds says the Linux security list is becoming nearly impossible to manage as researchers spray in AI-found bugs and shaky reports. The real fear is obvious: AI is boosting the volume of submissions, not the quality.
One team got tired of junk AI-generated pull requests and used Git's --author flag to block the worst offenders. It is a small trick with a big message: open source is done clapping politely while bots farm easy contribution points.
Qwen censorship shows through the cracks
A mechanistic study of Qwen 3.5 argues political censorship can be spotted inside the model's own internal machinery, not just in outer safety filters. That is a big deal, because it suggests ideology can be baked deep into the system.
Anthropic shops for agent plumbing
Anthropic bought Stainless, the company known for SDK generation and API tooling. Translation: frontier labs no longer want chatbots that merely talk. They want agents that can reach into real software and actually get things done.
Alibaba keeps the leaderboard sweating
Qwen 3.7 Preview arrived with strong arena rankings across text, math, and coding, giving Alibaba another loud entry in the model race. The message is getting harder to ignore: the pack behind the usual giants is now uncomfortably close.
Model startup lag gets slashed
Modal says it cut AI cold starts by up to 40x with a stack of checkpointing tricks. It sounds like dusty infrastructure work, but it matters a lot: faster wake-ups make AI apps feel less clunky and make expensive GPU time hurt less.
Amazon wants to ship for everyone
Amazon opened Supply Chain Services, letting outside companies buy its freight, warehousing, and delivery muscle. It is another classic Amazon move: build giant internal machinery first, then rent the machine itself to the rest of the market.
Bitwarden makeover leaves users squinting
Bitwarden's recent pricing and product changes drew heat after users noticed a quieter, glossier corporate turn under Acquia. The worry is not just one price bump, but the smell of a once-simple password tool getting polished into blandness.
Haiku finally boots on M1 Macs
The scrappy Haiku OS project now runs on M1 Macs, a lovely plot twist for anyone who misses computers that feel fast, small, and human-sized. It will not topple macOS, but it proves weird operating system ambition is still alive.
Smart doorbells ring for strangers
A researcher found some bargain smart doorbells could be triggered by anyone on the internet, no invitation required. It is peak gadget misery: you buy a little home security toy and end up installing a tiny public nuisance button.
Today the tech mood turns hard and wary... Europe looks at life beyond US cloud for state data, Mozilla fights for VPNs in Britain, and ordinary WiFi starts sensing the room without cameras... At the same time, GenCAD pulls editable 3D CAD from a single photo, while a new attack shakes faith in AMD secure VMs and their cloud promise... Then the wider AI boom meets the bill: warnings over AI infrastructure, rising subscription costs, weak public trust, doubts over the water panic, and a cheaper path to agent code search... We get a day where power, privacy, cost, and control sit above the demo.
European officials are considering limits on US cloud services for sensitive state work, a move that says the quiet digital sovereignty panic is now policy. The message is blunt: government data may no longer want an American passport.
GenCAD turns a single image into editable 3D CAD and even spits out the build history, not just a shiny model. That makes the jump from picture to real design feel a lot less magical and a lot more practical.
Mozilla Defends VPNs in Britain
Mozilla told UK regulators that VPNs are basic privacy gear, not shady tools, as Britain weighs tougher youth safety rules. It reads like a fight over whether everyday people can still hide from the internet’s endless peeking.
RuView says it can sense people through ordinary WiFi signals without cameras, pushing smart spaces into stranger territory. The pitch is privacy-friendly, but it still has that unmistakable feeling of walls quietly learning your habits.
Researchers showed a bad Infinity Fabric setup can break AMD SEV-SNP, a key protection for cloud VMs. That is the kind of security paper that makes trusted computing sound a little less solid and a lot more conditional.
Mistral Sounds Europe AI Alarm
Mistral boss Arthur Mensch says Europe has about two years to build its own AI infrastructure or become dependent on US giants. It lands as a not-so-subtle alarm bell: sovereignty now means racks of GPUs, not just speeches.
Cheap AI May Get Expensive Fast
The warning here is simple: today’s cheap AI subscriptions may only look cheap because labs are still burning cash. For companies piling workflows onto rented models, that smells like future price shocks, lock-in, and budget pain.
America Still Does Not Trust AI
Fresh survey coverage says most Americans do not trust AI or the people steering it, and that gap is getting harder to wave away with demos. The industry keeps selling inevitability while the public keeps reaching for the brakes.
The AI Water Panic Gets Doubted
A contrarian take argued the AI water panic is overstated and that data centers are being singled out for a broader industrial problem. Even so, it shows the sector is now fighting on utilities, not just models and benchmarks.
Agents Get a Cheaper Code Search
Semble promises code search for agents using far fewer tokens than grep-style brute force, aiming straight at the cost headache in AI coding. It is a very 2026 product: less romance about magic, more pressure to make the bill stop climbing.
Codiff Makes Git Reviews Feel Nice
Codiff is a local desktop app for reviewing Git changes before commit, and its appeal is obvious: fast, focused, and not another browser tab circus. Even small developer tools are winning attention when they cut friction instead of adding ceremony.
Earthquakes Go Live Without a Backend
Klaxon turns public USGS feeds into a live earthquake map with no backend, proving once again that a clean front end plus open data can still steal the show. People love a project that is both useful and cheekily lightweight.
One hacker turned an $80 Android tablet into a usable Debian workstation, which is exactly the kind of scrappy hardware story that never gets old. It is part thrift, part rebellion, and a reminder that locked-down gadgets are still negotiable.
FreedomLang Goes Bare Metal and Proud
FreedomLang pitches a libc-free systems language with direct kernel calls and unusual concurrency, wearing its sharp edges like a badge. It is not trying to be cozy; it is trying to tempt the crowd that thinks ordinary toolchains have gone soft.
Mezz Boxes In Your IoT Gadgets
Mezz is a curl-friendly WiFi sandbox for inspecting your own IoT gear, giving tinkerers a contained place to see what smart devices are really doing. In a market full of mystery boxes, that kind of home lab honesty hits a nerve.
We start with Tesla as newly opened filings say robotaxis crash twice while a teleoperator is driving from afar... In Europe, the sovereign cloud push meets the old grip of chips, with Intel and AMD still at the core... The shape of the next machine shifts too, as personal computing starts to look like a cluster and Kioxia with Dell pack 10 PB into a 2RU server... Then AI spreads wider: layoffs rise in exposed jobs, Malta rolls out ChatGPT Plus for citizens, classic CTFs face bot pressure, Claude misses on Algora bounties, and Stochastic Parrots returns to the center of the LLM debate.
Europe learns chips still rule
Europe is pouring money into sovereign cloud, but critics say the plan wobbles if the hardware still depends on Intel and AMD systems with opaque control layers. The independence pitch suddenly looks a lot less independent.
Tesla robotaxis crash under remote watch
Newly unredacted filings say Tesla robotaxis crashed twice while a teleoperator was remotely driving. That lands right in the middle of the robotaxi sales pitch and makes the safety story sound far shakier than the glossy demos.
Your next computer may be a cluster
The idea that personal computing is becoming a cluster no longer sounds like sci-fi. As AI tools chew through absurd amounts of compute, people are starting to picture distributed machines the way earlier generations pictured a desktop tower.
Ten petabytes squeeze into two rack units
Kioxia and Dell showed off a 2RU server packed with 10 PB of flash storage. That is the kind of number that makes yesterday's big iron look tiny, and a sign that AI and data hoarding are reshaping hardware fast.
AI layoffs stop being hypothetical
A new report says US job losses are now piling up in roles most exposed to AI. After months of cheerful talk about productivity, the darker side is showing up in payroll data, and the mood around automation is getting much harder to ignore.
Malta gives everyone ChatGPT Plus
OpenAI struck a deal with the Government of Malta to offer ChatGPT Plus to all citizens. It is part public service, part giant ad campaign, and a loud sign that access to AI is quickly becoming something governments want to package.
Hackers say classic CTFs are cooked
One blunt take lit up the security crowd: open CTFs no longer cleanly measure human skill because powerful AI can crush challenge formats built for people. A once-beloved proving ground now feels like it needs new rules or a reboot.
Claude chases bounties and misses
One developer tested Claude on open-source Algora bounties and found the dream of easy AI cash was a lot messier than the hype. The results read like a cold shower for anyone expecting coding agents to print money unattended.
Emily Bender revisited Stochastic Parrots, answering the questions critics keep pretending nobody asked. The piece pushes back hard on sloppy LLM claims and reminds everyone that bigger models still do not equal deeper understanding.
WSL9x stuffs a modern Linux kernel inside Windows 9x, which is exactly the kind of beautiful nonsense the internet was built for. It is part nostalgia trip, part technical flex, and fully irresistible to anyone raised on beige boxes.
Tiny e-reader goes pocket mode
A tiny DIY e-reader chased the dream of a truly pocketable book machine. The build hits that sweet spot between practical and charmingly obsessive, and it reminds everyone that gadgets can still be weird, personal, and fun.
Student quietly owns campus tech
A student story about taking control of campus projectors and cameras reads like a movie pitch with worse IT hygiene. The real shock is how many systems were exposed by lazy network setup, turning convenience into a giant blinking warning.
This website runs on an 8-bit chip
Someone hosted a real website on an 8-bit microcontroller, because apparently sensible hobbies were unavailable. It is gloriously inefficient, deeply educational, and a perfect reminder that the best computing stories still start with bad ideas.
When did computers stop being fun
An Ask HN thread turned into group therapy over why computers feel less joyful now. Too many locked-down services, too little ownership, too much friction — and a lot of people clearly miss when a machine felt like a playground, not a tollbooth.
Today the center of gravity sits on Big Tech and the bill that follows... Britain keeps Palantir on the sidelines, Meta secures a huge $3.3B tax break for a $10B Louisiana data center, and Waymo recalls 3,800 robotaxis after flood-water errors... California moves toward game refunds when online titles go dark, while London tests live facial recognition at a protest... On the AI front, Anthropic faces fresh questions over disputed numbers and the hidden Claude Mythos Preview, ChatGPT moves closer to users’ money through Plaid, and the alarm over entry-level jobs grows louder. We see a tech industry pushing into daily life, government budgets, and public streets all at once.
Britain puts Palantir on the sidelines
One of the day’s loudest stories said less about one vendor and more about a growing allergy to Palantir-style government tech. Contract records suggest the UK is not writing a blank check, and the revolving-door questions only got louder.
Meta lands giant public data center giveaway
Meta’s planned $10B Louisiana data center came with a jaw-dropping $3.3B in tax breaks, which made the AI boom look less like pure innovation and more like a public subsidy buffet. The question hanging over it: jobs now, or bills later.
Waymo recalls robotaxis after flood water blunders
Waymo had to recall about 3,800 robotaxis after software let cars head into flood waters, a reminder that self-driving still melts when the world stops behaving like a clean demo. The future arrived, then immediately needed a patch.
California pushes game shutdown refunds
California moved closer to forcing game publishers to offer a patch or refund when online titles die. After years of buying games that vanish when servers go dark, this felt like lawmakers finally noticing players are tired of renting forever.
London tests face scans at protests
London police planned to use live facial recognition at a political protest for the first time, turning a public demo into a test case for mass surveillance. The tech keeps showing up first where trust is already thin and tempers run hot.
Anthropic faces awkward money math
Anthropic got dragged into an ugly numbers fight after one figure shown in court reportedly clashed with a much bigger public one. In an AI market powered by sky-high storytelling, even a gap like $5B versus $19B lands like a siren.
Anthropic keeps Mythos behind the curtain
Another Anthropic drama asked whether Claude Mythos Preview is being hidden because it is too risky or simply too expensive. Either way, the shine comes off the frontier-lab mystique when the most powerful toys stay behind velvet ropes.
ChatGPT reaches for your bank account
OpenAI said ChatGPT users can connect bank accounts through Plaid, pushing the chatbot deeper into people’s real money. After health data came first, this looked like the next bold step in turning AI assistants into full-service middlemen.
The fear that AI is chewing through entry-level jobs stopped sounding abstract and started sounding like a hiring memo. If the bottom rung disappears, the whole career ladder wobbles, and that worries far more than another flashy demo.
Amazon workers said they were under pressure to show more AI usage, even when the job barely called for it. That is what an AI mandate looks like in the wild: vague orders from above, awkward make-work below, and a lot of pretending in between.
Windows CE boots on Nintendo 64
Someone got stock Windows CE 2.11 running on a real Nintendo 64, and the result was pure internet catnip. It served absolutely no practical need, which is exactly why it felt so refreshing: clever engineering for the joy of seeing if it can be done.
A pure OCaml protocol stack booted in low Earth orbit, giving functional programming fans a tiny orbital victory lap. Space software stories usually sound stiff, but this one had the irresistible charm of a niche language quietly reaching the stars.
Wikipedia gets the Windows XP treatment
This project lets you browse Wikipedia like an old Windows XP desktop, complete with nostalgic fake files and a dusty interface glow. It is silly, charming, and weirdly perfect for an internet that keeps missing the playful web it used to have.
Bun rewrite trips on Rust safety
Bun’s Rust rewrite got hit with claims that supposedly safe code still allows undefined behavior, which is the kind of phrase that makes systems programmers sit upright. Fast-moving rewrites look cool until someone shines a harsh light under the floorboards.
Cats get their own doomscroll feed
OnlyCats turned cat clips into a fake TikTok for cats, proving the web can still produce delightful nonsense on demand. In a feed full of AI dread, privacy fights, and robotaxi bugs, a shamelessly unserious cat app felt almost medicinal.
Tonight, we watch core tech seize the spotlight as Bun lands its Rust rewrite, a major shift for one of the busiest tools in the developer stack... NGINX admins face a dangerous bug in old rewrite logic, data centers run into a public backlash, KDE gets real money from Germany's Sovereign Tech Fund, and a stripped-down Toyota RAV4 Hybrid puts car tracking back under harsh light... On the AI front, the labs push hard as OpenAI Codex, Grok Build, and Claude Code race for the coding desk, while Claude Opus 4.6 and 4.7 remind everyone that outages still hit the premium tier... There is also a softer headline with sharp edges, as Anthropic and the Gates Foundation unveil a $200 million partnership aimed at health, science, education, and poverty tools.
Bun officially merged its Rust rewrite, a move that felt like a small earthquake in developer land. The new core is said to shrink the app, fix leaks, and speed things up, while quietly closing the book on its heavy Zig era.
A nasty new NGINX bug landed with the sort of timing that ruins a sysadmin's week. The flaw sits in a long-running rewrite feature and can open the door to remote takeovers, putting a huge slice of the web's plumbing on edge.
Neighbors Turn on Data Centers
A fresh Gallup survey found most Americans do not want data centers popping up near home, which is awkward timing for the AI building spree. The cloud suddenly looks less magical when people think about noise, water, and power.
Germany's Sovereign Tech Fund handed KDE about €1.3 million, turning endless talk about digital sovereignty into actual cash. For open-source fans, it was a rare good-news moment: governments may finally pay for what they keep using.
Toyota Tracking Gets Yanked Out
One owner's step-by-step guide to ripping the modem and GPS out of a Toyota RAV4 Hybrid hit a raw nerve. Modern cars keep phoning home, insurers keep buying data, and people are clearly done pretending that is a harmless little feature.
Anthropic's Claude Opus 4.6 and 4.7 threw elevated error rates, reminding everyone that premium AI still has very normal bad days. When the shiny assistant goes wobbly, whole workflows stall, and patience evaporates faster than status updates.
Anthropic Gets a Big Halo Deal
Anthropic and the Gates Foundation unveiled a $200 million partnership aimed at health, science, education, and poverty tools. It sounded noble, ambitious, and strategically tidy at the same time: frontier AI wants a humanitarian halo too.
OpenAI Pushes Codex Everywhere
OpenAI pushed Codex beyond the desktop, making its coding helper easier to use wherever people already live in ChatGPT. The message was not subtle: coding agents are no longer a side toy, they are becoming the main product pitch.
xAI rolled out Grok Build, its own coding agent and command line tool for paid users, joining the great land grab for developer attention. Every lab now wants to be your pair programmer, your shell, and your software manager in one.
Claude Code Grows Office Rules
Anthropic published guidance for using Claude Code on big, messy codebases, complete with tips around a shared CLAUDE.md file. The subtext was loud: these tools are shifting from demo magic to office process, rules, and team habits.
DreamHost Quietly Plants Agents.txt
DreamHost quietly dropped an agents.txt file into customer sites, turning a niche idea into a live policy question overnight. Website owners want a simple way to tell AI crawlers where they can shove it, and hosts are starting to notice.
Amazonbot Finally Learns Some Manners
After plenty of grumbling, Amazonbot says it will finally respect robots.txt. That sounds basic because it is basic, but in the AI crawl gold rush even old web manners started looking optional, which made this tiny email feel oddly huge.
Palantir Exit Saves Real Money
A UK refugee system reportedly saved millions of pounds after replacing Palantir software, handing critics of giant contractor deals a very satisfying headline. Turns out 'expensive and inevitable' is not always the same thing.
Researchers showed the first public macOS kernel memory corruption exploit on Apple M5, which is the sort of phrase that makes security people sit up straight. New chips may be fast and polished, but the bug hunters are still clocking in.
Tiny Machines Keep the Cloud Flying
A beginner-friendly look at Firecracker reminded everyone that cloud magic is often just smarter packaging around tiny virtual machines. If you wondered how services like AWS Lambda feel instant, this is a big part of the trick.
Google pushes Android toward the desktop as Aluminium OS comes into view... GitHub Actions logs expose some GitHub_TOKEN secrets, and Python pulls back an incremental garbage collector after memory warnings in production... Around Lake Tahoe, residents and data centers compete inside the same power story as the AI build-out reaches the grid... At the same time, OpenAI and Sam Altman face new trust questions in a major profile and in court, while Anthropic rolls Claude into the small-business office... We also see a broader mood forming, with worries over AI coding habits and a louder AI backlash entering the center of the tech day.
Google drags Android onto the desktop
Google is teasing Aluminium OS, a desktop take on Android aimed at real PCs instead of giant phones with taskbars. The pitch is bold, the branding is shiny, and plenty of people are waiting to see whether this is finally a desktop plan with teeth.
GitHub logs spit out secret tokens
A nasty GitHub Actions slip exposed some GitHub_TOKEN values in logs when tools printed them the wrong way. It is the kind of tiny mismatch that turns into a real security mess fast, especially for teams that trust automation a little too blindly.
Python pulls back a risky memory change
Python's new incremental garbage collector is being rolled back in 3.14 and 3.15 after reports of heavier memory use in production. It is a sharp reminder that clever runtime upgrades sound great until real workloads start chewing through RAM.
AI data centers squeeze Tahoe's power
About 50,000 Tahoe residents need more power just as utilities look at steering lines toward data centers. The AI boom keeps crashing into the grid, and this story makes the tradeoff plain: chatbots want electricity, towns do too.
The trust question swallows OpenAI
A major profile asked the question hanging over OpenAI: can anyone really trust Sam Altman? The piece stitched together old promises, power plays, and shifting stories, feeding the sense that AI's most famous company still runs on mystery and charm.
Altman faces brutal claims in court
In court, Sam Altman was forced to answer claims that he bends the truth whenever the stakes get high at OpenAI. It turned a boardroom soap opera into public theater, with AI leadership looking less visionary and a lot more chaotic.
Anthropic sells Claude to small shops
Anthropic rolled out Claude for Small Business, bundling connectors and ready-to-run workflows for everyday office tools. The message is clear: frontier AI is now being sold less as magic and more as a cheerful digital worker for the back office.
Coders say AI is making them dull
A wave of developers says heavy AI coding help is making their thinking softer, not sharper. The complaint lands because it feels uncomfortably familiar: fast autocomplete is wonderful right up until you realize you barely know what your own code is doing.
The warning here is simple: the AI backlash is coming, and it may get loud as power use, job fears, and data-center politics pile up. The industry keeps acting like resistance is just confusion, which looks like a very risky way to read the room.
Suicide helpline site shared visitor data
The Dutch suicide prevention site 113 was found sharing visitor data with outside tech companies without consent. That is the kind of privacy failure that makes people furious instantly, because if a crisis website cannot stay careful, what exactly can?
Europe pitches a cleaner digital life
One builder moved email, analytics, and cloud habits toward European providers, arguing digital sovereignty is finally practical, not just political. The appeal is obvious: less dependence on US giants, fewer creepy defaults, and more control.
Coders flee GitHub for self-hosting
A developer said goodbye to GitHub and moved to self-hosted Forgejo, pointing to ownership worries and a similar move by the Dutch government. It reads like a small migration with a big mood behind it: convenience no longer wins automatically.
Europe's public sites are a mess
A scan of European government websites found thousands of trackers, plenty of exposed phpMyAdmin installs, and security that looks worryingly thin. For institutions that love lecturing everyone else about privacy, the result lands with extra embarrassment.