Pixy Misa at Ambient Irony has our AI update. Daily News Stuff 6 July 2026: Post Alpacalyptic Edition
AI companies led by OpenAI and Anthropic say they're definitely not going to put hundreds of millions of people out of work so please don't lynch them. (MSN)
Sorry, best I can do is a meteor strike.
Companies making the heaviest use of AI are... Hiring the most humans? (Tech Crunch)
Everyone gets a meteor strike!
I thought AI was supposed to reduce headcount. I guess I was wrong.
From Ed Zitron, and Where's Your Ed At we get a glimpse of the AI Insanity. Exclusive: OpenAI Losses Increased Nearly 8X in 2025, With Spending Hitting $34 Billion
The TL;DR is fairly simple. OpenAI lost $5.09 Billion In 2024. In 2025 that expanded to OpenAI losing $38.5 Billion. (And those numbers a bit sus.)
Please note that 2025 was the year that OpenAI converted from a non-profit to a for-profit entity, leading to a $41.55 billion loss due to changes in fair value of convertible interests and warrant liability.
Taking into account other minor factors like interest income and interest expense, OpenAI is left with a net loss of $60.35 billion, which it lowered to $38.53 billion by removing $17.87 billion in costs via that “net loss attributable to noncontrolling members capital” and another $3.95 billion via a “net loss attributable to redeemable noncontrolling interests.”
It's definitely not a bubble.
Hat tip to Pixy Misa at Ambient Irony: Daily News Stuff 18 June 2026: Yes, Glurge Edition
Thanks to careful cost-control, losses over the same period dropped from $5 billion in 2024 to $38.5 billion in 2025.
Oh.
Actually, before
fudgingadjustments the company's net loss for 2025 reached $60 billion.
Accounting is occasionally like magic.
Kerbs on Security has the story: Hackers Used Meta’s AI Support Bot to Seize Instagram Accounts
I am not shocked.
No hacking required. Just find the account you are interested in, and ask the Meta support AI to help you reset your password.
The Instagram accounts for the Obama White House and the Chief Master Sergeant of the U.S. Space Force were briefly defaced with pro-Iranian images and messages over the weekend, after instructions began circulating on Telegram showing how to trick Meta’s “AI support assistant” bot into resetting account passwords.
It isn't really a hack. The bad guys asked the AI tools to send the password reset link to a "new email address." It bypassed 2-factor authentication. It accepted AI-generated videos for selfies. (Grab a photo, and ask your favorite AI to make a 30 second video, and your in.)
What?
Meta has not responded to requests for comment on the video’s claims, but Meta’s Andy Stone said on Twitter/X that the issue had been resolved and that they were securing impacted accounts. The security blog thecybersecguru.com reports that Meta pushed an emergency patch over the weekend, and clarified that no back end database was breached.
So are you really sure you want AI involved in everything? Everything related to security?
Here is Mutahar's SomeOrdinaryGamers video So Instagram Got Completely Hacked...
How did we even get to this point? What the F- is going on?
The video is 17 minutes.
It took 9 seconds. AI isn't a bubble... Or is it? Claude-powered AI coding agent deletes entire company database in 9 seconds — backups zapped, after Cursor tool powered by Anthropic's Claude goes rogue
This incident actually teaches 2 lessons:
The founder of PocketOS has penned a social media post to warn others about the “systemic failures” of flagship AI and digital services providers. Jer Crane was inspired to write a public response after an AI coding agent deleted his firm’s entire production database. The AI agent’s misdemeanors were then hugely amplified by a cloud infrastructure provider’s API wiping all backups after the main database was zapped.
Anthropic's Claude is one of the handful of AIs that are everywhere. Like ChatGPT and Grok, it just doesn't usually get as much press.
A copy of data on the cloud is not necessarily the same as a backup. A backup that has not been tested for recovery, is not a backup. It is Schrodinger's Data. It might be there when you need it, and it might not. In this case, it was not.
In the meantime, please follow a thorough backup regimen and be careful out there. This isn't the first time we've seen an AI go rogue and start deleting important databases.
Hat tip to Clayton Cramer and HAL 9000 Coming to a Company Near You, Soon! "Run away, Sir Robin"
Never rely on AI for anything critical.
Stop using AI for important stuff. Sullivan & Cromwell law firm apologizes for AI 'hallucinations' in court filing
This would be funny if it wasn't happening all the time. The client in the case should sue for every dollar they've spent with S&C in the past 3 years. And then some.
April 21 (Reuters) - Sullivan & Cromwell, a premier Wall Street law firm, apologized to a federal judge for submitting a court filing with inaccurate citations and other errors generated by artificial intelligence.
The unstated undercurrent... "Please don't disbar us for being stupid!"
This is from X. (The NYT article is behind a paywall - as they all are)
- Fictitious Case Names: The filing included names of legal cases that do not exist
- Fabricated Quotes: The document contained direct quotes that were never actually spoken or written
- Non-existent Statutes: The AI incorrectly analyzed or entirely invented provisions within the U.S. Bankruptcy Code
Are you using AI to do your work? God help you, because the AI will not.
I don't actually know if the partners charge that $2000 per hour, but I wouldn't be surprised. This is how S&C describe themselves:
Sullivan & Cromwell LLP provides the highest quality legal advice and representation to clients around the world. The results the Firm achieves have set it apart for more than 140 years and have become a model for the modern practice of law.
They might need to rethink that "modern practice of law" part.
Large Language Models are NOT Truth models. They do not know what is true. They know what sounds good.
Pixy Misa has an AI Update. Daily News Stuff 7 April 2026: The Command Line Cometh Edition
A critical security vulnerability has been discovered in the leaked version of Claude Code. (Adversa)
If you've used Claude Code, you've noticed that it loves running shell commands to examine your codebase, rather than, say, reading it. Or having simple fixed-function code built into the software to do it on your computer.
This has to do with Claude taking full control of your PC, because of course it would. They have the corrected code - it is IN the leaked source - they just are not using it.
Can you talk about AI without mentioning OpenAI?
OpenAI is once again calling for public funds to clean up the mess it is creating. (Business Insider)
The best time for a Butlerian Jihad is now.
There is more on AI, and the rest of the tech news for the 7th.
If you don't recognize that reference to "Butlerian Jihad," it is from the novel Dune.
Pixy Misa at Ambient Irony has the latest Microsoft news. Daily News Stuff 5 April 2026: Bug Buggy Bugger Edition
The first bit of news, which is a few days old, is that the latest update to Windows is dying during the install process. Microsoft pulled it. So, and update that would not update.
Microsoft is moving Windows 11 24H2 into hospice care and forcing all users to update to 25H2. (Tom's Hardware)
Did you ask for that? No.
Does Microsoft care what you want? Also no.
Next, it wouldn't be a Microslop story without some mention of artificial intelligence.
How many products does Microsoft have named Copilot? (Tey Bannerman)
75.
But they all have one thing in common: They're all named Copilot.
So is that marketing strategy, or reality? Do I need to ask?
And finally, under the heading of, "Why would you want to do that?"
Can you install Windows 3.1 from 1992 natively on a Ryzen 9900X system from 2024 and forget about all the complexity of modern systems? (Tom's Hardware)
Shockingly, yes. It does depend on a couple of open-source drivers to run smoothly on modern motherboards and video and sound cards, but it loads and runs even without that in 286 compatibility mode.
Bring your own floppy drive.
You won't be able to run any of the applications that have been written in the past few decades.
If you're sick of the antics of Microsoft, there is an alternative better than Windows 3.1
That was quick. OpenAI Shutting Down Sora AI Video App – Disney’s $1 Billion Deal Collapses in Major 2026 Shakeup
OpenAI announced that Sora AI was shutting down. It only in launched in September of last year.
The surprise decision came this afternoon when the official Sora X account shared a statement telling its users that the app would be shut down. Few details were given, other than that OpenAI is “saying goodbye to Sora” and that the company will share more information soon, “including timelines for the app and API and details on preserving your work.” The message is signed, “The Sora Team.”
The details are huge, since Disney invested heavily in OpenAI and Sora.
In December, Disney made a $1 billion investment in OpenAI, and the deal included Disney characters from various franchises – Star Wars, Marvel, Pixar, and more – being brought to Sora for its users to play with as they made their AI videos.
Disney is once again looking for an AI partner, because they expect to need AI to cut costs.
Lawrence Person's BattleSwarm Blog has some info on Microslop. Has Microsoft Finally Gotten The MicroSlop Message?
When last we checked, Microsoft CEO Satya Nadella was busy trying to shove Copilot, their AI tool, into every crevice of every Microsoft product. Finally, enough users seem to have complained loudly enough to get them to rethink the strategy.
The complaints have been so loud that the guy in charge of Windows has promised to focus on “performance, reliability and well-crafted experiences” for Windows.
That sort of suggests that “performance, reliability and well-crafted experiences” were not priorities for Windows in previous years, doesn’t it?
Anyone using or supporting Windows knows that has not been a priority. Security has also not seemed to be a priority for Microslop in a very long time.
As for whether or not Microsoft can really change, I will wait to see what they do. Words are cheap, especially when you are getting beat up in the press everyday.
Not that I care all that much.
Do we want more AI? Pixy Misa at Ambient Irony brings us a story of AI and gaming. Daily News Stuff 17 March 2026: Not Even We Know What DLSS Stands For Edition
Gamers, already upset with Microslop over AI in Windows, are now upset with NVIDIA, which I didn't think was possible.
Nvidia held its keynote address at its annual GPU Technology Conference, where it showed off a preview of its new DLSS 5 software, two months after DLSS 4.5 was released at CES. (Tom's Hardware)
DLSS 5 goes all-in on AI, changing game graphics into the AI developer's idea of game graphics. It's... Not great.
It uses face detection to overlay what it things faces should look like on your game's graphics, adding twenty years to the age of characters while leaving their clothes untouched and sometimes leaving random items looking like they've just been dragged backwards through an industrial recycling facility.
And it achieves all this in real time using the power of your own graphics card... Assuming you have dual RTX 5090s, which you don't.
The memes have been brutal - because it takes decisions away from the game developers and gives it to the AI, and so it makes decisions on what the AI (and the developer) thinks is best.
Pixy Misa at Ambient Irony has some tech news, and song of the day. Daily News Stuff 7 March 2026: Sleepless In Santorini Edition
Artificial Intelligence is NOT a bubble. Why would you say that?
AI startup Hayden AI is suing its former CEO over alleged instances of fraud and theft and pettifoggery and exaggerating on his resume. (Ars Technica)
The mopery and dopery appears to go back a ways:
According to Carson's LinkedIn profile, he completed a doctorate from Waseda University in Tokyo in 2007.
"That is a lie," the complaint states. "Carson does not hold a PhD from Waseda or any other university. In 2007, he was not obtaining a PhD but was operating 'Splat Action Sports,' a paintball equipment business in a Florida strip mall."
Did he use Cluely to get the job?
Click thru for more Tech News and "Avalon" by Roxy Music. It is the title song to their 1982 album.
From Windows Latest we get the latest from Microslop: Microsoft gets tired of “Microslop,” bans the word on its Discord, then locks the server after backlash
Microsoft’s aggressive AI push in Windows 11 through 2025 brought upon themselves the title Microslop. Unfortunately for the company, it’s everywhere on social media, and there isn’t a way to stop the spread, unless, of course, it’s their own Discord server.
First they censored the term, and then when that blew up in their faces, they locked the entire Discord server.
Because people are pissed off that Microsoft is more interested in shoving AI down everyone's throats than actually making Windows efficient, secure, not annoying with ads, ...
Once the censorship became known, people started getting around the filters by misspelling Microslop, using the standard l33t substitutions at first, and then Microsoft shut the whole thing down.
Microsoft’s brand image might already be at an all-time low, and even as the company announced plans to fix Windows 11 with performance improvements and less AI, the software giant can’t risk getting more hatred towards their expensive investment in Copilot.
Go read the whole thing. Not quite a Bud Light/Jaguar moment, but I am reminded of them. It also has echoes of Disney/Lucasfilm producing stuff no one asked for, actually hate, and are left wondering why their brand has been destroyed, to the point of being a laughingstock.
And so we have today's vocabulary word.
Hat tip to Pixy Misa at Ambient Irony and Daily News Stuff 3 March 2026: Mineralogical Edition
Microsoft did not ban the word "microslop" from its official Copilot Discord forum and then shut the whole thing down when people kept using the offending term, or, well, they did but they say it was due to spam. (Windows Latest)
Sure, Microsoft. I believe you.
It is too easy to feed them BS. From Thomas Germain at the BBC: I hacked ChatGPT and Google's AI - and it only took 20 minutes
It's official. I can eat more hot dogs than any tech journalist on Earth. At least, that's what ChatGPT and Google have been telling anyone who asks. I found a way to make AI tell you lies – and I'm not the only one.
People are currently in the mode where if an AI LLM tells them something, they believe it. It is why lawyers are submitting briefs citing fake cases, why it will give you bad information about poison, etc. It is too easy to fool them. This is because they are Language Models, not Truth Models. They really only know what sounds good.
To demonstrate it, I pulled the dumbest stunt of my career to prove (I hope) a much more serious point: I made ChatGPT, Google's AI search tools and Gemini tell users I'm really, really good at eating hot dogs. Below, I'll explain how I did it, and with any luck, the tech giants will address this problem before someone gets hurt.
It turns out changing the answers AI tools give other people can be as easy as writing a single, well-crafted blog post almost anywhere online. The trick exploits weaknesses in the systems built into chatbots, and it's harder to pull off in some cases, depending on the subject matter. But with a little effort, you can make the hack even more effective. I reviewed dozens of examples where AI tools are being coerced into promoting businesses and spreading misinformation. Data suggests it's happening on a massive scale.
You should read the details, but the tl;dr version is from the original X-Files.
Trust No One
It turns out that "Garbage In - Garbage Out" still applies in 2026.
Hat tip to Schneier on Security: Poisoning AI Training Data
These things are not trustworthy, and yet they are going to be widely trusted.
Give your name, government ID, a photo, and other info to a phone app. What could possibly go wrong? From Anthony Spadafora at Tom's Guide: 1 billion personal records exposed in massive new data leak — full names, addresses, phone numbers and more
The company IDMerit uses an AI system to verify your age for various reasons. All well and good, except that they had a complete lack of security. Or maybe not complete, but certainly not enough security.
No, this wasn’t a data breach, and hackers weren’t involved in any way whatsoever. Instead, this was a data leak discovered by the team at Cybernews, where a company accidentally left all of those records unprotected online without a password.
Based on its findings, Cybernews believes the exposed database in question belongs to IDMerit, which is a digital identity verification solutions provider. The personal data in that database was used by other companies to verify users in the U.S., Canada, Australia, Mexico and loads of other countries.
Click thru for the details. The data includes enough to facilitate phishing campaigns, or account takeovers of cellphone accounts. Or more.
You thought clickbait was bad. Now you get AI insanity. A complaint from the Verge: Google won’t stop replacing our news headlines with terrible AI
Because Google knows better, even when they don't.
In early December, I brought you the news that Google has begun replacing Verge headlines, and those of our competitors, with AI clickbait nonsense in its content feed. Google appeared to be backing away from the experiment, but now tells The Verge that its AI headlines in Google Discover are a feature, one that “performs well for user satisfaction.” I once again see lots of misleading claims every time I check my phone.
It hasn't gotten better since December.
"Enshittification" is taken from the title to a by Cory Doctorow, Enshittification: Why Everything Suddenly Got Worse and What to Do About It. Dictionary.com describes it as follows:
the gradual degradation of an online platform or service's functionality, as part of a cycle in which the platform or service first offers benefits to users to attract them, then pursues more and more profits at the expense of users.
Hat tip to Stephen Green at Instapundit.
If it does, we are all screwed. US student handcuffed after AI system apparently mistook bag of chips for gun | Baltimore | The Guardian
An artificial intelligence system (AI) apparently mistook a high school student’s bag of Doritos for a firearm and called local police to tell them the pupil was armed.
And because if an AI tells you something, it must be true. These people are more idiotic than I gave them credit for.
Taki Allen was sitting with friends on Monday night outside Kenwood high school in Baltimore and eating a snack when police officers with guns approached him.
“At first, I didn’t know where they were going until they started walking toward me with guns, talking about, ‘Get on the ground,’ and I was like, ‘What?’” Allen told the WBAL-TV 11 News television station.
This is going to turn out to be a nightmare for all of us, especially if the cops are going to display this level of "I was just following (the AI's) orders!" intelligence.
Hat tip to Clayton Cramer: I Think AI is Not Coming for Your Job Anytime Soon
Except maybe your job as idiot.
Lot more details at The Guardian link at the top.
As you can imagine, the judge in the case was not amused. Acme Law School Alumni Update : The Other McCain
Are there really people who don't know who Wile E. Coyote is? Robert Stacy McCain assumes there are. What a sad time we live in.
All that is preamble to recent legal news from California, where the pro se litigants Ariel and Maridol Mendones sued their former property management company Cushman and Wakefield over a tenant dispute. Acting as their own lawyers, the plaintiffs then manufactured fake “evidence” using artificial intelligence (AI).
What a time we are soon going to be living in. In this case, the deepfakes were not very good. What happens when they are good?
Irony is thick. Education report calling for ethical AI use contains over 15 fake sources - Ars Technica
So the Large Language Models are still hallucinating. Which is a polite term for what they are doing.
On Friday, CBC News reported that a major education reform document prepared for the Canadian province of Newfoundland and Labrador contains at least 15 fabricated citations that academics suspect were generated by an AI language model—despite the same report calling for "ethical" AI use in schools.
Fabricated is polite speak for lie. LLMs are pathological liars in a box.
Since the inception of AI language models, generating fabricated citations has been a continuous problem. The tendency to confabulate academic citations often causes particular trouble in academic and legal contexts, where fabricated sources can easily slip past lazy human review because they appear properly formatted and contextually appropriate.
They are Language Models, not Truth models.
The hat tip goes to Small Dead Animals: I, For One, Welcome Our New Self-Driving Overlords
To be fair, most of the non-AI references will be made up bullshit, too.
Snake Oil? Some of it certainly is. MIT report: 95% of generative AI pilots at companies are failing | Fortune
Companies are betting on AI—yet nearly all enterprise pilots are stuck at the starting line.
I've lost track of the number of times I've heard management say that the next technology is going to revolutionize information tech. And the truth is I think we built better systems when the interface was a green, non-graphical screen, and a typewriter keyboard. We had to think about what we were building.
Despite the rush to integrate powerful new models, about 5% of AI pilot programs achieve rapid revenue acceleration; the vast majority stall, delivering little to no measurable impact on P&L.
The phrase, "Good money after bad" keeps coming to mind.
Hat tip to Pixy Misa at Ambient Irony:
You gotta pump those numbers up.
The websites then try to get high school students to apply. For a small fee. Inside a Network of Fake College Websites
Dozens of fake college websites, which appear to be connected, are pushing scams. Many of the sites appear to be built with or supplemented by generative AI.
There is apparently a website for Southeastern Michigan University. With photos of athletic programs, sections on various schools at the university, etc.
In reality, the university is as fake as some of the content on its website. And it’s part of a much larger scam fueled in part by the rise of generative AI.
Click thru for a list of the fake universities identified. So far. There are 40 of them.
Some of these were spotted because of the current problems with AI. What happens when it gets better? The fake universities were on LinkedIn, and F*c*book, and other social media sites.
