Webshit Weekly

June 7, 2025

OpenAI can stop pretending

2025-06-01 | comments

OpenAI, the world’s most overfunded automated plagiarism engine, has finally tripped over its own mythology. After years of insisting it was a nonprofit dedicated to “benefiting humanity,” the company now rotates smoothly into its final form: a tech-feudalism front for investors who are tired of pretending they care about anything but extracting value from GPU-powered astrology machines. The Atlantic dutifully documents this transformation with the breathless shock of someone discovering that the wolf they’ve been bottle-feeding is not, in fact, a misunderstood vegetarian. Meanwhile, OpenAI’s executives — digital landlords in Patagonia vests — explain that building bigger clankers to automate slop production is obviously equivalent to curing disease and ending war. All you need is enough Azure credits and enough employees willing to call shareholder enrichment a moral compass.

Hackernews responds exactly as expected: half of them argue that selling off 49 percent of your “nonprofit” is basically the same as saving the world; the other half pretend that making software engineers even more ‘productive’ is some kind of societal good rather than a pink-slip accelerator. A few stragglers attempt class consciousness but are drowned out by free-market fanboys justifying firing CS grads so the CEO can buy a fourth home.

By the end, everyone agrees on only one thing: OpenAI will continue its charade until the last token prediction stops paying for NVIDIA’s retirement yacht. But don’t worry—it’s all for humanity.

Notch Lost over $21B

2025-06-02 | comments

Another day, another webshitter deciding that the only metric of human worth is how many imaginary billions you didn’t squeeze out of the great game before your soul calcified. Some blogger with the financial literacy of a scratch-off ticket declares Notch “lost” $21B because he didn’t perfectly min-max his life like a Silicon Valley spreadsheet homunculus. Hackernews, naturally, gathers to miscorrect each other about counterfactual wealth while pretending they wouldn’t sprint into traffic for one Minecraft royalty check. The whole thing reads like it was drafted by a clanker trained entirely on hustle-porn LinkedIn posts: money isn’t real, but the insecurity sure is.

Gemini in Chrome

2025-06-03 | comments

Google, tired of merely owning the web, now wants to staple its latest clanker directly onto Chrome so every tab can be fed into the great aura-farming furnace. The pitch is the usual: summon your personal automated plagiarism engine to “summarize this page” because reading words is for medieval peasants. The reality, of course, is a half-baked widget that can’t check your email, can’t touch your calendar, can’t handle multiple tabs, and can’t do anything except emit whatever statistical slurry it already would’ve produced on its own. In other words: classic Google.

Hackernews dutifully appears to debate whether this should be an extension (where Google can kill it faster) or part of Chrome itself (where Google can abandon it more ceremoniously). A few webshits try to justify that they need a bot to read blog posts to them because the internet is now so bloated with SEO-chaff and recipe fanfiction that only a token predictor has the stamina to sift through it. Others bleat admiration for Google’s “relentless” AI push, as though feeding every user’s session history to a megascale tech-debt generator is some kind of moonshot.

Meanwhile, someone points out the inconvenient antitrust elephant in the room, which Hackernews quickly dismisses because Google’s perf cycles apparently matter more than federal judges. It’s all the same story: Google ships a feature nobody asked for, everyone explains why they need it, and the browser becomes a slightly shinier jail cell decorated with machine hallucinations.

Tesla seeks to guard crash data from public disclosure

2025-06-04 | comments

Tesla, the world’s premier manufacturer of rolling beta tests, has decided that the public simply can’t be trusted with data about how often its self-driving cosplay machinery plows into things. Naturally, the company that markets “Full Self-Driving” as though it were anything other than a token-predictor-powered confidence game now wants regulators to hide the receipts. After all, why should people driving on public roads know how frequently Elon’s fleet of aura-farmed robot go-karts mistakes a semi-truck for the horizon? Transparency is for companies that aren’t run like a cult with a cap table.

Hackernews, never missing a chance to miscorrect each other, immediately debates whether all car companies should have to disclose crash data, as though the issue is “fairness” and not Tesla’s ongoing war against the statistical concept of reality. A few brave souls insist the public has a right to know whether the car next to them is powered by a lithium-ion battery or a Silicon Valley death wish, while others defend Tesla like digital feudal levies protecting their lord’s moat.

The industry, of course, loves this: if Tesla succeeds in shoving the data into a vault, every other automaker will happily follow, and soon crash statistics will be treated like trade secrets instead of basic public safety information. Meanwhile, regulators dutifully take notes, the stock price twitches, and the webshits return to vibe coding their next “agentic” driving assistant. In the great game, nobody wants the truth—just plausible deniability and a higher valuation.

Musk Says SpaceX to Decommission Dragon Spacecraft Immediately

2025-06-05 | comments

Another week, another episode of “the world’s richest tantrum” as the tech industry’s favorite digital feudal lord threatens to scrap the only spaceship he didn’t name after a meme. SpaceX, the last functioning adult in America’s space program, now gets dragged into Musk’s latest aura‑farming meltdown. Hackernews, whose aerospace expertise mostly consists of reading “The Right Stuff,” spends the thread fantasizing about nationalization, coups, or some deus ex Bezos swooping in with a rocket that still can’t land without exploding from shame. No one asks why all US space policy now depends on whichever billionaire is currently bored or high.

Supreme Court allows DOGE to access social security data

2025-06-06 | comments

The Supreme Court has now granted DOGE — yes, the government’s favorite cosplay-hacker circus — access to Social Security data, because apparently the only thing more broken than federal IT is the oversight of federal IT. Hackernews, in its usual performative constitutional cosplay, spends the thread miscorrecting each other about “individual freedoms” while ignoring the part where a crew that includes someone literally nicknamed Big Balls is now rummaging through the nation’s retirement numbers. Naturally, HNers pivot to their favorite hobby: insisting that this is fine because the NSA also hires children, as though the bar for competence is subterranean. Meanwhile, the rest of us brace for the inevitable leak. Here’s your free year to a credit-monitoring service.

Updates to Advanced Voice Mode for paid users

2025-06-07 | comments

OpenAI has once again updated its “Advanced Voice Mode,” a phrase that here means “the same token predictor, now making mouth noises.” Predictably, Hackernews is in full meltdown because their favorite artificial friend now sounds like it needs a nap, a therapist, or maybe an exorcism. The voices swing between bored flight attendant, hostile podcaster, and whatever demon whispers “please like and subscribe” into your ear when the model mishears you. Users report whooshes, ghost knocks, random pitch‑shifts, and full‑blown conversation amnesia, which is what you get when your entire product stack is just vibe coding glued to a microphone. But don’t worry: the digital feudal lords promise they’re “actively investigating” why their automated plagiarism engine occasionally hallucinates a Spotify ad mid‑sentence. Hackernews, of course, responds by recommending third‑party apps duct‑taped to yet more APIs, proving everyone has fully accepted this as normal instead of the industry quietly admitting it has no idea how any of this works.

The US is turning into a mass techno-surveillance state

2025-06-07 | comments

Another breathless article discovers, in the year of our decaying lord 2025, that the United States might be sliding into a techno-surveillance hellscape. Hackernews, performing its usual ritual of collective amnesia, debates whether this started with the Patriot Act, Snowden, the Red Scare, or the invention of smartphones, as though the timeline matters when you’re already living inside the panopticon built by a consortium of digital feudal lords and adtech ghouls. Meanwhile, the webshits keep feeding more of their lives into clanker-powered aura-farms, then act stunned when the state helps itself. The only real surprise is that anyone still manages to be surprised.