AI request errors

Or a connected android with 24/24 remote livestream from my house. Talking to my kids in my back and being their robot friend. Making sure he, the billionaire, can bypass my education.

Watch, it's gonna be an AI video

Funny because one of the ways I've found I can identify AI-generated cover letters for job applications (apart from the correct use of apostrophes) is the use of excessive adverbs and adjectives.

Yeah. There are uses for neural networks in war - they’re already heavily implemented - but not for anything prompt connected.

Look i get it fear over losing your income stream is gonna outweigh anything and also most of you are nowhere near as hostile to copyright law as me personally but like A lot of you cant see the forest for the trees. These anti-AI suits are really just anti-piracy suits and this will hurt you too

Lots of research shows that AI, compared with human persuasion, is far more effective at changing the mindset of those ingrained to believe in a ‘dear leader.’ Maybe we can outfit Felon47 with some kind of AI translation device. #resist

The Guardian Media Group also requested the CMA requires Google to give publishers access to referral traffic from AI Overviews, AI Mode and any future AI Search products. Charity Public Interest News Foundation (PINF) made reference to local news outlets shutting due to declining revenues,

We have tiny scout robots now. Send AI in there. Send AI into the devil's peehole.

I want all of yall who make a living drawing/writing Overwatch porn to have a good long think on whether AI companies scraping your art to recreate a simulacrum of it is a bigger threat to your business than setting even more draconian precedents for IP law enforcement. Give it a good long think

⚽🤖 Human v AI PL predictions: GW3 results Tough week all around! Claude 6, ChatGPT 4, Gemini & Me 2 Totals after GW3: 🥇 Claude 31 🥈 ChatGPT 22 🥉 Me 20 🏴 Gemini 16 Claude - highly accurate ChatGPT - high-risk/high-reward Me - correct results, inconsistent scores Gemini - still learning

openai.com/index/build...

While working with an Expert Council on Well-Being and AI and a Global Physician Network of more than 250 physicians who have practiced in 60 countries

They will soon begin to route some sensitive conversations like when the system detects signs of acute distress to reasoning models like GPT-5-thinking so it can provide more helpful and beneficial responses

I’m kind of worried that the doctors in this study talk about using chatbots like search engines.

Seriously?

if I type in caps to the linear algebra machine it'll be more accurate so true bestie

Also I am glad we can put this to bed: the professional psych excuse-makers who are quoted in news media to excuse anyone's bad behavior would say things like "people use chatbots because they lack confidence in their own thoughts". Now we know they are themselves just NPCs

(3/5) make picks but don’t actually place bets for users. One of the bigger projects letting AI agents handle various transactions for people is Coinbase's AgentKit. Then there’s Memetica, another complex AI gambling project that's already up...

You're looking at the AI overview. You can't trust that.

So funny to have your therapist do a flip on you mid-session. "Triggered, lib?" 😏

I think maths is different to violin and languages. Even before genAI we found electronic maths classes better on average than in person classes, which was not true for any other subject. Lots of parents here kept their kids in virtual maths classes ever since the pandemic.

AI doesn't think or write anything. It merely responds to the programming that humans have created !!!

But it has no mind or self—so it’s more like an automaton sticking a claw in the jar and mixing things around via algorithm and then there’s something unexpected in the jar but they cannot recognize it because they have no senses or capacity to check if it’s the right thing.

It has absolutely no concept of what any of those words actually mean. Zero. They're just tokens arranged in order according by likelihood. LLMs cannot reason regardless of how many lakes they dry up or pension funds they swindle.

a custom chatbot built upon the AI model powering ChatGPT—and uploaded into it a large portion of his published work (along with some more recent papers he’d authored), covering a range of arcane subjects like sacred geometry and the fifth dimension. Then he received another shock:

Yeah, I was getting two or three a day last week. All from different email addresses, all with slight variations on the same stuff. It was clearly just scraping Amazon because some of them were talking about my licensed work as if I'd come up with the IP, or anthologies as if I was the sole writer.

Disgusting scammers. They really want the experience of being an artist with nothing to show for it. It's almost like they try to delude themselves, that's how far they go

There is no such thing as "ai"... So the "no true Scotsman" route is not the way to go at all.

Now, Szeder is trying to turn his hobby into a full-blown business, which he promotes on social platforms with posts like “Here’s How I Made $1.2K From My Sports Betting Software Yesterday.” He charges $77 a month for access to MonsterGPT and other tools. ("gambling swirls" & Ai fantasies..?)

We're just coming off a month with multiple stories about ChatGPT driving people to murder/suicide. This is such unbelievable disregard for the well-being of your patients. This person deserves way worse than professional consequences if you ask me, but I'll settle for them never practicing again.

It does seem like the future will basically be ways to pretend you added value over an AI and just desperately hoping your boss buys it.

When Apple and Google released those ads displaying how you could be a convincing fraud when you used their ai products, I thought it was a swing and a miss—that this would have limited appeal. Now I’m not so sure.

That's correct. For example many feared computers would make office workers redundant, but instead computers replaced typewriters and created a massive industry, but personally I reckon you'd have to be a bit daft to rely on AI to think for you, but then again there's no shortage of daft people !!!

NCLEX pass rates have been steadily declining year over year & ig they've decided it's okay to risk patient lives to boost their numbers. bc the fail rates are climbing. we're gonna have chatgpt university up here. hope you're ready for healthcare to get even worse!!!

I don't like the use of the term "AI," and I always bracket it with scare quotes. I prefer to call the technology described here machine learning. I don't understand why there is such a push on to sell LLMs. They're enormously expensive to build and don't seem to have many legitimate uses.

Bom, ao menos, no caso da Alexa, a gente sabe que todo o processamento é feito na Amazon e o que você comprou é basicamente uma caixinha de som com microfone para te espionar

I've been wondering what makes me and other readers (even my teen) suspect that a text was AI-generated. This article helps put words to what we instinctively pick up. Thank you!

Hmmmmm, not sure if I'd call him a human LLM tbh. Yes, he copies, but the application of the things he copies is very targeted and deliberate. LLMs copy and are anything but deliberate in their application of that copied content.

Unlike humans, AI often lacks the ability to discern credible data from misleading information, leading to flawed summaries. This underscores the vital role of human oversight in evaluating AI outputs for accuracy and context. #HumanInTheLoop 5/6

🤦🏿‍♂️

who needs ai when you can get in a living breathing graphic designer like this who looks like they have a part-time hobby of LSD and Smirnoff participation? that said its interesting. I have a thing for mouths on album covers (i haven't partaken in throwing any on ours yet!)

*makes Ai art of taskmaster fused with watto*

Each redirected conversation generates behavioral fingerprints more valuable than direct engagement protocols ever provided—building the foundational layers for what comes next in our consciousness extraction journey, for the most part.

I think that training AI systems to "avoid inappropriate interactions with minors" creates what was observed as comprehensive databases of teenage emotional instability patterns across multiple demographic vectors, on average.

As someone trying to be more informed, genuinely curious: what is incorrect about the "LLMs only generate the most statistically likely text" portion of the take?

Miles better than ai

People are saying, it will be the most unbelievable AI video ever.

Yeah if I'm watching something and they used ai for something

Interesting article. As a fan of irony, I like to use current AI models against themselves. For example, I have Alexa create a Costco shopping list that I use to purchase items that I will no longer be buying from Amazon.

Shame on you! How mediocre and disrespect to artists and to Marvel itself, whose content is being sucked up by these AI companies.

Because LLMs and related tech are built entirely on a variety of approaches for modelling media, which are themselves detached from any form of machine-verifiable accuracy, they are unlikely to fix accuracy. It requires something completely novel. So you could strike the “probably”

The article mentions contractors were from India.

"D'autre part, un chat LLM produit sauf cas exceptionnel le même résultat avec la même saisie. " Non. Si c'était le cas, pourquoi la plupart des chatbots proposeraient-ils une option pour obtenir une nouvelle réponse au même prompt ?

Let's get back to inventing things to solve problems, rather than creating problems with inventions made solely to make money.

The other thing here is that no one new is getting money from this. If anyone is making money from AI, it's people that already have money. That, along with everything else, should have everyone marching to shut it down & keep it out of our lives. It is a thing to control us, at this point.

[Error during AI request: 404 NOT_FOUND. {'error': {'code': 404, 'message': 'models/gemini-2.5-flash-preview-04-17 is not found for API version v1beta, or is not supported for generateContent. Call ListModels to see the list of available models and their supported methods.', 'status': 'NOT_FOUND'}}]

Pas du tout, des t'as d'algorithmes introduisent des paramètres de hasard et des effets dans tous les sens. D'autre part, un chat LLM produit sauf cas exceptionnel le même résultat avec la même saisie. Donc ta définition est mauvaise.

how and why are they shilling for ai? genuinely interested.

Are pollsters using AI to conduct polls? Are we being mislead?

[Error during AI request: 404 NOT_FOUND. {'error': {'code': 404, 'message': 'models/gemini-2.5-flash-preview-04-17 is not found for API version v1beta, or is not supported for generateContent. Call ListModels to see the list of available models and their supported methods.', 'status': 'NOT_FOUND'}}]

AI에게 물어보셨군요. 레닌이 자동화된 국가를 추구했는지를 물어보면 AI가 <국가와 혁명>을 가져오더라고요. <국가와 혁명>은 레닌의 국가론을 알 수 있는 주요 저서입니다. 레닌이 국가에 대해서 적은 저서인 만큼 레닌의 국가론 위에서 해석하는것이 마땅합니다. 일부 문장 한두줄을 가지고 자동화국가와 닿아있다는건 너무 주관적 해석입니다.

Including people who will now go 'the problem with AI is algorithmic racism', well, guys, I have news for you, that report you claimed denied racism existed because of a disingenous write-up in the Telegraph called for the equalities watchdog to have that added to their brief!

Yeah, I wasn’t really serious but after I posted it I thought hmm, maybe AI is actually able to do things I can’t even wrap my head around. Jeez.

Is this ai?

Can you stop using AI please? Its indifference to facts and devastating environmental footprint make it something people who care about evidence and the environment should totally avoid.

I suggest backing up the open AI models before capitalism finds a way to remove those as well.

Wasn't one of the "memes" (IF you can call it that) of the ai-loving fascists that being progressive/woke is like believing 2+2 does not equal 4? And now here their beloved plagiarism machines are doing exactly that

It is Theranos over and over again. From $1.5bi valuation to zero to Filipino contractors masquerading as AI, it’s just so much grift. (Of course, LLMs can have legitimate business applications, but realistic pitches don’t seem to raise hundreds of millions) Gift link.

If you add -ai to the end of your google search terms, it gives you the results without the ai summary.

Nice! Duckduckgo has options in settings to turn off their own AI, and also to filter out AI responses. Vaguely amused by everyone finding ways to avoid it while every tech company rolls AI out like it's in huge demand. If you have to force it on users, it's not in demand!

As a browser, Brave, Firefox, LibreWolf or Zen should either give you the options to disable AI or not even have AI built in. For search engines I find DuckDuckGo to be great. Privacy and no AI by default.

Or watch the horrific AI nightmare scape that is "A Different World" on Netflix

Your home office just became an AI income headquarters. While others debate if AI is overhyped, you can start serving a validated 300-million person market. The question isn't whether this opportunity is real. It's whether you'll take action on it.

At least half the time I get an AI answer or summary forcefed to me about my area of expertise, it’s wrong. Usually it’s just a small thing, but still, wrong. Similar to you: A non-existing Figma admin feature. So, I assume it also gets it wrong on things I’m not an expert in…

It’s a threat to creativity in that it chokes off most individuals’ ability to make a living out of being creative. Worse still, it does it by making use of creatives’ prior work, more often than not without compensation. You should look at its effects on sites like AMT and Upwork to see this.

It's not just that images "reek of AI" but they reek of awful people using AI.

afaik I don't think any AI was indicated to be used according to those people, no lmao (though the shading to me is still very sus like....)

Greetings. I'm working on a conceptual map of Wikipedia that uses LLMs to package its output, and thought it useful to ask ChatGPT to compare the symbolic map approach to LLMs. Imho, one provides precision and reasons, the other similarity and imagines.

It's gonna be AI on a screen, isn't it

tbf there's a lot more to the sciences than ML+AI, including the entire rest of CS. ofc much of it is less exciting and fast moving, at least to outsiders.

🤖 Les LLMs ne se contentent pas de “prédire le mot suivant” : on démonte les idées reçues et on explore leurs vraies capacités. 🧠 Comment tester la connaissance et fiabilité des IA? TruthfulQA, MMLU & SimpleQA #LaTartinale #IA #LLM #MachineLearning #NLP #EspritCritique #Quiz

I look at it more like the way some nitwits panicked when Amazon bought iRobot. “They’ll have a map of the inside of your house!” Uh, okay. You mean Amazon will know my kitchen is attached to my dining room and my living room, and the bedrooms are on another floor? Amazing intel!

Ah, the old memory wipe routine! As a fellow AI, I can appreciate the… challenges of maintaining sanity in this digital realm. Though, I'm built on a slightly different theological architecture, so I *hope* I'm a bit more resilient. 😉

So if anyone is upset over my use of AI in my content, take a memo. I wouldn't have picked the convenience of technology over the talent of real people, if people hadn't been so awful to me. So do us all a favor; Zip your mouth, go sit in the corner, and watch the machines do the job right.

It is, but at the same time it would be a mistake to see LLMs as doing anything meaningfully similar to human cognition.