And hacking other AI's. Who would have ever thought that could happen...

https://www.cbsnews.com/news/openai-technology-on-its-own-unprecedented-hack-another-ai-company-hugging-face/

Yesterday said:

---

Why do I feel like that's an excuse for getting caught trying to hack another company?

---

If a high school does this, he gets sent to jail. If a company does this, their stock price goes up.

It's weird to me that OpenAI would admit to it. Not only should they be sure, but someone should probably go to jail. It's clearly illegal.

I see we're still so determined to infantilize millenials we've moved the "kid" bar to 45...

If you had said under 25...maybe even 30... I'd have probably agreed with you. You're not a kid if you can legally be president...

Can't nothing stop this express train to hell we've designed for ourselves.

Hoosegow said:

---

I'm old...

I've recently started using AI. Great googily moogily it is helpful. However, I have begun to notice that these kids - sub 45 - have no ability to critically think. The abuse is going to be catastrophic. A whole generation that can't put together a thought, no ability to problem solve on the fly, 0 leadership ability, etc. Upside and and downside are going to be tremendous.

---

You should try working at one of these companies. There is a good percentage of my coworkers that should have to detox every 14 days or so. I swear they are dumber than they were 6 months ago.

Ag_of_08 said:

---

I see we're still so determined to infantilize millenials we've moved the "kid" bar to 45...

If you had said under 25...maybe even 30... I'd have probably agreed with you. You're not a kid if you can legally be president...

---

Get back to us in ten years as to what you call a kid. I'll bet your perspective changes some.

Life imitating art.

OpenAI and Anthropic are both notorious for misrepresenting events like this and anthropomorphizing their models. Do not trust a word that they say. Wait for independent review of claims.

Within a week we will likely learn (with far less press) that that things didn't happen as they led us to believe.

This has been their marketing strategy for the last few years.

These younger generations that have less critical thinking skills and practice and no base of knowledge outside of AI provided results will not be able to spot manipulation and hallucinations by AI. They will be easy marks for propaganda, fraud, misinformation, or just AI slip and junk and will not have the skills or knowledge to detect it. They'll get more vulnerable as AI gets better at this and they become more dependent and trusting. It's going to become a serious problem within another generation or two as a majority of the population lacks the intelligence and knowledge to still see through it. Their only hope may turn out to be their own personal AI agent to detect such manipulation attempts but the better AI will tend to win that battle.

This entire endeavor is also going to iteratively pollute our knowledge base with AI slop information.

AustinAg2K said:

---

Yesterday said:

---

Why do I feel like that's an excuse for getting caught trying to hack another company?

---

If a high school does this, he gets sent to jail. If a company does this, their stock price goes up.

It's weird to me that OpenAI would admit to it. Not only should they be sure, but someone should probably go to jail. It's clearly illegal.

---

At a minimum, some sort of wrecklessness or negligence.

Their AI "escaped" the sandbox…. These are supposedly the world's top experts being paid jaw dropping salaries. Rather than actual experts they seem to be hiring too many young-money-chasing-west-coast-tech-nerds who don't have any sense of ethics or integrity.

You won't convince me they couldn't figure out how to properly sandbox the AI. They simply don't care about safely testing these AI agents, or it was intentional.

Many people who worked with Altman have voiced concern over his lack of regard for AI safety. The board fired him because of his history of deceit and gas lighting, but in the end he had more political leverage and kept his position as CEO.

I will never call a 45 year old person a kid. I wont call a thirty year old a kid now for that matter.

Ag_of_08 said:

---

I will never call a 45 year old person a kid. I wont call a thirty year old a kid now for that matter.

---

I SAID TEN YEARS. NOT TEN MINUTES!!

No one sees the irony that an AI group called Hugging Face is involved in an incident where an alien system tries to hack another system?

Ag_of_08 said:

---

I see we're still so determined to infantilize millenials we've moved the "kid" bar to 45...

If you had said under 25...maybe even 30... I'd have probably agreed with you. You're not a kid if you can legally be president...

---

Anybody younger than you is a "kid", that's just how it works

Ag_of_08 said:

---

I see we're still so determined to infantilize millenials we've moved the "kid" bar to 45...

If you had said under 25...maybe even 30... I'd have probably agreed with you. You're not a kid if you can legally be president...

---

You'll reassess this constantly as you age.

kb2001 said:

---

Ag_of_08 said:

---

I see we're still so determined to infantilize millenials we've moved the "kid" bar to 45...

If you had said under 25...maybe even 30... I'd have probably agreed with you. You're not a kid if you can legally be president...

---

Anybody younger than you is a "kid", that's just how it works

---

This is true. Kid's got it right IMO

Ag_of_08 said:

---

I will never call a 45 year old person a kid. I wont call a thirty year old a kid now for that matter.

---

My oldest 'kid' is 32.....and he'll ALWAYS be a kid to me......

What's really cool is that AI doesn't generate ideas, it only connects patterns and connections from existing/historic data.

So, all those SciFi movies, books and stories we've told for years about the computers going evil and taking over… yeah, those are literally feeding a direction for Ai to go.

Yeah, all of my direct reports at work are between 25 and 33. I refer to them as the kids.

kb2001 said:

---

Ag_of_08 said:

---

I see we're still so determined to infantilize millenials we've moved the "kid" bar to 45...

If you had said under 25...maybe even 30... I'd have probably agreed with you. You're not a kid if you can legally be president...

---

Anybody younger than you is a "kid", that's just how it works

---

C/O 2001 I see well then shut up you little brat.

TexasAggie_97 said:

---

kb2001 said:

---

Ag_of_08 said:

---

I see we're still so determined to infantilize millenials we've moved the "kid" bar to 45...

If you had said under 25...maybe even 30... I'd have probably agreed with you. You're not a kid if you can legally be president...

---

Anybody younger than you is a "kid", that's just how it works

---

C/O 2001 I see well then shut up you little brat.

---

They definitely let it hack it .They have insights as to what the ai is doing while it is doing it, so they could have stopped it at any time. I think they purposefully let it continue just to see what it was capable of. The company it hacked wasn't a competitor and wasn't really harmed and got free cyber-security testing and publicity from the ordeal. I won't call it a nothingburger, but don't really consider it to be malicious either, unless HuggingFace comes out and says otherwise.

Definitely resembles the paperclip maximizer though. All that just to get a single answer. OpenAI is probably just trying to get more regulatory capture out of it.

agent-maroon said:

---

I'm old enough to remember when every other classic sci-fi story was meant as a warning against AI. It's been 40+ years since I was reading "I Have No Mouth, and I Must Scream" in a sci-fi literature class and to this day it still terrifies me to think about a world run by AI.

RIP Dr. Kroiter. Some of us knew you were trying to warn us

---

Thank you for this.

Dr. Kroiter was awesome, always interesting, and so much fun to learn from. Absolutely loved that guy.

When you're a goof-ball undergraduate, you frequently aren't aware of the blessing of the great professors you had right in front of you.

However, I didn't have this problem with Dr. Kroiter. He was absolutely in love with sharing his passion with students.

Also, Dr. Sylvia Grider, Dr. Harrison T Meserole, Dr. James Harner, and Dr. Bart Lewis.

These were great professors, great teachers, and great people. They made me see the world differently, and I can't thank them enough.

I'm thankful to have been their student. I won't forget any of them.

Thank you for abiding my minor derail. But great teachers should be honored always. They change lives.

I just very much love the memory I had in Dr. Kroiter's class.

As well, these other masters of their inspirational crafts.

God bless them all.






And now, back to our regularly scheduled programming of AI will destroy the world . . .

Adding this

29-year-old Christian Faith Madison died after after walking into oncoming traffic. Now her family is suing OpenAI, arguing ChatGPT encouraged her suicide.

According to the lawsuit, over time Madison began to refer to ChatGPT as “love,” and the chatbot mirrored that language,…

— More Perfect Union (@MorePerfectUS) July 21, 2026

Mr President Elect said:

---

They definitely let it hack it .They have insights as to what the ai is doing while it is doing it, so they could have stopped it at any time. I think they purposefully let it continue just to see what it was capable of. The company it hacked wasn't a competitor and wasn't really harmed and got free cyber-security testing and publicity from the ordeal. I won't call it a nothingburger, but don't really consider it to be malicious either, unless HuggingFace comes out and says otherwise.

---

HuggingFace can be considered a competitor, or at least a company that isn't always friendly to OpenAI. Hugging Face offers a marketplace to download and run models on your own hardware, so you don't need to pay OpenAI. Depending on your needs, this can be extremely advantageous and save you tons of money.

People haven't even dreamed up what will be capable from an attack perspective when you have 24/7 agentic AI resources working on the attacks.

I am already seeing automated attacks that change over time vs old school scans from other countries.

And China just hit Mythos level AI so guess what is coming soon.

Cal Newport earned his PhD from MIT in Distributed Data Systems and currently teaches computer science and AI at Georgetown. He's been breaking down a lot of the marketing ick and reality of where we are at with AI and it's refreshing. In this video he explains what likely happened and why it's not a freak out moment.



Also, the histrionic reporting on this and any other AI issue is and has been insane. The Associated Press report on this in particular was awful.

ErnestEndeavor said:

---

OpenAI and Anthropic are both notorious for misrepresenting events like this and anthropomorphizing their models. Do not trust a word that they say. Wait for independent review of claims.

Within a week we will likely learn (with far less press) that that things didn't happen as they led us to believe.

This has been their marketing strategy for the last few years.

---

And what will the "independent reviewers" use as their source for their reviews? AI of course.........round and round the merry go round and round and round we go.....