Oh sweet baby Jesus no, burn AI to the ground, humanity can't be trusted with it

[Agema]

OpenAI strikes deal with Pentagon, hours after rival Anthropic was blacklisted by Trump

OpenAI CEO Sam Altman said on X that his company has come to terms with the Defense Department for use of its models.

www.cnbc.com

So I guess this is how they're routing in the govt bailout they'll eventually need.

Yep, I think this might be where the USA is going.

Tech bosses haven't flooded Trump with money for nothing: buying seats at his inauguration, his wife's awful film, his ballroom. If they run out of money, the AI boom and their riches come to a grinding halt. Meanwhile, the US economy depends on AI for most of its growth and that growth maintains the reputation of the government: so if the AI bubble pops, the US economy pops, and the Trump administration collapses with it.

They are natural bedfellows: the tech industry needs infinite funding, and the Trump administration needs the tech industry (until Jan 2029, anyway) to prop up the economy. So the US government finds some way to set up a line of near-infinite credit for the tech industry.

 

[The Rogue Wolf]

and the Trump administration needs the tech industry (until Jan 2029, anyway)

He'll need the tech industry after that, for Palantir to help him round up dissidents stop domestic terrorists after he refuses to leave office.

 

[Gergar12]

Yep, I think this might be where the USA is going.

Tech bosses haven't flooded Trump with money for nothing: buying seats at his inauguration, his wife's awful film, his ballroom. If they run out of money, the AI boom and their riches come to a grinding halt. Meanwhile, the US economy depends on AI for most of its growth and that growth maintains the reputation of the government: so if the AI bubble pops, the US economy pops, and the Trump administration collapses with it.

They are natural bedfellows: the tech industry needs infinite funding, and the Trump administration needs the tech industry (until Jan 2029, anyway) to prop up the economy. So the US government finds some way to set up a line of near-infinite credit for the tech industry.

I literally couldn't use Anthropic, they are less general than Gemini, or GPT-5.2. That said they are the best for programming scaffolding, and I regard them as the best AI company. You use Deepseek for the mandarin translation, Gemini for the student license, GPT-5.2 for general purpose and in theory counseling, and Claude for programming. Grok if you want to do NSFW images.

 

[BrawlMan]

​

 

[Zykon TheLich]

​

Woo!, and indeed Hoo!

 

[Thaluikhain]

While nice, what about starting with AI art and then getting a human to mess around with it a bit? Does putting some human touches on it make it copyrightable?

 

[BrawlMan]

While nice, what about starting with AI art and then getting a human to mess around with it a bit? Does putting some human touches on it make it copyrightable?

No. I know there are some artists that use minor AI assistance for tweening animation key frames or certain shadings, but they still originally drew on a digital pad or paper first. If your art is AI from the start, it's still stealing from something. So it's not protected by copyright.

 

[Thaluikhain]

No. I know there are some artists that use minor AI assistance for tweening animation key frames or certain shadings, but they still originally drew on a digital pad or paper first. If your art is AI from the start, it's still stealing from something. So it's not protected by copyright.

Oh, I agree that it shouldn't be, but I wonder if people could use that as a loophole.

 

[Gergar12]

Meanwhile



... That's it, I change my mind on AI, all hail ChatGPT, Claude, Gemini, and Deepseek. I am still not using Grok.

 

[The Rogue Wolf]

AI is forcing students to write more poorly and to use AI detectors to check their writing, in order to avoid accusations of using AI.

We’re Training Students To Write Worse To Prove They’re Not Robots, And It’s Pushing Them To Use More AI

About a year and a half ago, I wrote about my kid’s experience with an AI checker tool that was pre-installed on a school-issued Chromebook. The assignment had been to write an essay about Ku…

www.techdirt.com

 

[BrawlMan]

Following AI generated reviews, Resident Evil Requiem AI guide books flood Amazon, and they're part of a major problem

Following AI game reviews, AI game guides of major releases like Resident Evil Requiem, Pokopia, and more flood Amazon's storefront.

frvr.com

 

[Zykon TheLich]

This has probably been posted before, but I'm going to post it again, because I love this channel.

 

[Gergar12]

https://www.politico.com/news/2026/02/17/data-centers-public-knowledge-5-charts-00769974

So it seems like the public leans towards AI, but doesn't want it to affect them personally. Interesting...

 

[Phoenixmgs]

https://www.politico.com/news/2026/02/17/data-centers-public-knowledge-5-charts-00769974

View attachment 14368

So it seems like the public leans towards AI, but doesn't want it to affect them personally. Interesting...

NIMBYs

 

[BrawlMan]

​

 

[Xprimentyl]

My company has started implementing AI technology. My first encounter was with a co-worker from whom I assumed a weekly Excel report. It's pretty much plug and play; I'm just responsible for pushing it out every Monday now. She complained that there was something wrong with the data I was publishing, and pointed to a specific formula that was giving erroneous data. I looked into it, and while not trying to stereotype, I knew the formula’s complexity was a bit out of her league; she’s generally a dingbat I wouldn’t trust with a basic IF statement or VLOOKUP. But as a matter of course, I asked if she wrote the formula, and she admitted she’d used AI, and the formula was wrong. I asked her to tell me what she expected from the formula and I was able to fix it. She then said she’d been using AI all over the place in her Excel reporting, and has asked me to go behind her and check if all of it is right or not. Great, AI has made her job easier by adding extra work for me. I can’t wait until they lay me off and let dingbats like her use AI to do all of the work wrong for them.

 

[Thaluikhain]

Ouch. Yeah, that seems to be the common experience with people using AI.

 

[FakeSympathy]

I posted this on the complain train thread in Octoboer of last year, but wanted to share it again here as it involves AI;

I did a MAJOR FUCK-UP at work today:

As a data analyst at a research institute, I was supposed to thoroughly review some study articles with a professor I was working with, extracting numbers from the text.

She was talking half-day to respon/confirm ther findings, and I decided to use AI to extract those numbers faster; I didn't tell her I was doing this and forgot about checking in with her, which led to major inconsistencies in the final calculations.....right before our big meeting with the external collaborators. We reported those final calculations as "confirmed" numbers. This fuck-up was discovered after that big meeting. Yeah got chewed out pretty bad.

The good news is that I owned up to it, apologized, and immediately got to fixing them. It was rather a quick fix.

The bad news was that the collaborators took those numbers in as "facts", and everyone was talking about planning a new strategy for the project because of my mistake.

I really have no one to blame but myself. Just want a random burglar to come in and stab me to death or something

Looking back, I realized just how stupid, brain-dead, idiotic it was for me to do this. But you gotta understand, getting data directly from the pdf article is a super demanding and boring. I just thought the AI would help. But I think the bigger fuck-up was not telling the professor about it.

As a part of the support team, I help out with different projects in the institute, and I recently came back to this story. The professor was cool about it now, and this time while I did use the same AI tool, I double-checked, and had her check for the third-time before approving as the final numbers.

Moral of the story:
DO NOT, under any circumstances, take anything AI says as facts

 

[Agema]

My company has started implementing AI technology. My first encounter was with a co-worker from whom I assumed a weekly Excel report. It's pretty much plug and play; I'm just responsible for pushing it out every Monday now. She complained that there was something wrong with the data I was publishing, and pointed to a specific formula that was giving erroneous data. I looked into it, and while not trying to stereotype, I knew the formula’s complexity was a bit out of her league; she’s generally a dingbat I wouldn’t trust with a basic IF statement or VLOOKUP. But as a matter of course, I asked if she wrote the formula, and she admitted she’d used AI, and the formula was wrong. I asked her to tell me what she expected from the formula and I was able to fix it. She then said she’d been using AI all over the place in her Excel reporting, and has asked me to go behind her and check if all of it is right or not. Great, AI has made her job easier by adding extra work for me. I can’t wait until they lay me off and let dingbats like her use AI to do all of the work wrong for them.

Okay, but make sure that is logged and recorded somewhere. Like, prepare a quick report for your manager. Prep a departmental sharing good practice presentation about ways to use AI, where this is an example of potential pitfalls (make sure no-one can identify the dingbat so it doesn't come across as personal, or even get the dingbat in on it). Everyone wins.

Early in my career, I tried to make sure nothing failed. Later in my career, I realised that sometimes you need to let things fail. Incompetent staff sometimes need to be left to make a mess of things, under-resourced projects need to go unfinished or late, and so on. Do your job sensitively, attentively, responsibly and carefully; when you see a clusterfuck let your superiors know in good time; make sure you can be seen to have done your best to attain the right outcome. But you can't be expected to do someone's work, or work way past your hours to do a task that should have had more support, and so on.

I've got a staff member who's into AI. I told her to go and explore and see what sorts of amazing stuff she can find, and I would back her to do so. And then she provides some sort of output to tell everyone else. She gets kudos, I get a bit of reflected kudos, everyone learns how to do their job better.

 

[Xprimentyl]

Okay, but make sure that is logged and recorded somewhere. Like, prepare a quick report for your manager. Prep a departmental sharing good practice presentation about ways to use AI, where this is an example of potential pitfalls (make sure no-one can identify the dingbat so it doesn't come across as personal, or even get the dingbat in on it). Everyone wins.

Early in my career, I tried to make sure nothing failed. Later in my career, I realised that sometimes you need to let things fail. Incompetent staff sometimes need to be left to make a mess of things, under-resourced projects need to go unfinished or late, and so on. Do your job sensitively, attentively, responsibly and carefully; when you see a clusterfuck let your superiors know in good time; make sure you can be seen to have done your best to attain the right outcome. But you can't be expected to do someone's work, or work way past your hours to do a task that should have had more support, and so on.

I've got a staff member who's into AI. I told her to go and explore and see what sorts of amazing stuff she can find, and I would back her to do so. And then she provides some sort of output to tell everyone else. She gets kudos, I get a bit of reflected kudos, everyone learns how to do their job better.

It's corporate culture, and AI is the new "thing" everyone has been encouraged to adopt; I don't want to come off as the contrarian because we've gone through enough change over the past year that contrarians are effectively targeted. It's much easier just to take the training, and put AI in my back pocket where I'll likely never use it. Insofar as documenting its misuses, go back to my former statement; it's easier to just fix what they break than document every time it's misused. My fix is to simply double check anything handed to me and/or do manually what I would need to do regardless as a fail safe, i.e.: anything anyone hands me to be responsible for, I will check for accuracy before publishing what their AI assistant has produced.; if I have to make mention of that, I can, but I'm not about to die on that hill that we're doing dumb shit because AI has convinced us it can do it for us.

An anecdote, my company adopted the open-air office format right around the time the earlier adopters started regretting it; COVID was good for one thing allowing me to work from home. So I imagine we're quite behind the AI curve, and don't expect it to proliferate much beyond making the most menial of tasks slightly easier.