ai
Yesterday, during training (customer service position), one of my coworkers asked if they should be using #LLMs to answer customer queries, and my blood ran cold.
Before anyone could respond, I jumped in and said, "No, absolutely not. Generative #AI doesn't give correct answers. It gives responses which look superficially correct, but are factually wrong. And when we give the customer the wrong answer, they don't want us to blame the machine. The blame lands squarely on our shoulders. Do not use ChatGPT, don't use Google AI, don't use Bing's whatever, just DON'T."
Thankfully, everyone thinks I know what I'm doing, so they accepted that. But this is such a common question that nobody thought it odd to ask, and that makes me worried in a very different fashion.
Just because I don’t “have a master’s degree in computer science” and haven’t “worked in the field for ten years” or “learned anything about software” or “tried” or “cared” doesn’t mean I shouldn’t get to be a Senior Software Engineer!
#VibeCoding #AI
Boost appreciated! ♻️
Ok I've come up with an idea;
I'm not going to post anything more AI related, I'm going to sell that content to help me with life and be able to afford real art commissions
So far I've generated about 227 images, I'm going to sell them as packages, you can see the pricing in the pictures below
I might also open requests ( featuring my own characters that is ) for about 5€? But I'll have to think about how that works
@Fidda for more info!
我認我縮,這星期暫且穿褲子…
#月曜がんばれ 130
「全力全開！パワー全開！ファイトーッ！！」
#松實優麗 #うちの子
#stablediffusion #AI #AI繪圖 #AIイラスト #AIart #AIイラスト好きさんと繋がりたい
This is obscene.
In analyzing both public and proprietary data about data centers as a whole, as well as the specific needs of #AI, the researchers came to a clear conclusion. Data centers in the US used somewhere around 200 terawatt-hours of electricity in 2024, roughly what it takes to power Thailand for a year. AI-specific servers in these data centers are estimated to have used between 53 and 76 terawatt-hours of electricity. On the high end, this is enough to power more than 7.2 million US homes for a year.
https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/
Ich würde gerne #opensource #AI benutzen um meine medizinischen Daten zu analysieren.
Uhhm. I didn't realize the #AI #video generation had reached this level.
The examples on this page are insanely good quality.
Knowing about these, I wouldn't have spotted most of them as created, certainly not in a casual inspection, and for some of them, even with a focused look.
https://www.thealgorithmicbridge.com/p/google-ai-video-just-changed-everything
Ladies, gentleman, and enbies, we are fucked.
Enjoy the awesome movie special effects, because epistemology and veracity of video as a medium is going away.
We obtained records showing how a DOGE staffer with no medical experience used AI to identify which VA contracts to kill.
His tool concluded that more than a thousand deals were each worth $34 million, when in fact some were for as little as $35,000.
https://www.propublica.org/article/trump-doge-veterans-affairs-ai-contracts-health-care?utm_source=mastodon&utm_medium=social&utm_campaign=mastodon-post
Did you know? #Fedify provides #documentation optimized for LLMs through the llms.txt standard.
Available endpoints:
https://fedify.dev/llms.txt — Core documentation overview
https://fedify.dev/llms-full.txt — Complete documentation dump
Useful for training #AI assistants on #ActivityPub/#fediverse development, building documentation chatbots, or #LLM-powered dev tools.
Last weekend, Elon Musk shared some images on his X platform that he claimed had been created by the Grok AI system. They were actually just edits of original work by Australian artist Charles Brooks, a former concert cellist who is known for photographing the insides of instruments. An X user named Eric Jiang had run some of these through Grok, adding people to the original photos so they look like buildings. Musk retweeted the post with the comment "Generate images with @Grok." @404mediaco's Matthew Gault spoke with Brooks. “I feel like, if he’s plugging his own AI software, he has a duty of care to make sure that what he’s posting is actually attributed correctly and is properly his,” he said. “But ‘duty of care’ and Musk are not words that seem to go together well recently.”
#Art #AIArt #GenerativeAI #Grok #AI #ArtificialIntelligence #CharlesBrooks #ElonMusk
I got an email from my bank with the subject line:
"Protect your business from AI-driven fraud."
So yeah, another win for AI, really improving the lives of... scammers?
I am seeing change after change of #ai generated code that is using hard-coded SQL string formatting rather than binding query parameters.
According to a new analysis, AI isn't even close to coming for everybody's jobs, and it's not exactly making workers more productive, either. Via @Futurism. #AI #Chatbot #ArtificialIntelligence #Tech #Technology https://flip.it/DO7dx2
Browsers are still on the news and this time they're caught in the AI wars: https://world.hey.com/ricardo.tavares/ai-browsers-less-local-more-clouds-e19c5386 How does that help or threaten the web?
“If you can’t use a billion dollar #AI system to solve a problem that Herb Simon … solved with AI in 1957, and that first semester AI students solve routinely, the chances that models like Claude or o3 are going to reach AGI seem truly remote.”
Disillusioned & distanced from #Trump, #Musk says he’s exiting Washington & made clear he is frustrated w/obstacles he encountered as he tried to upend the #FederalGovernment🎻
Musk dissed Trump’s signature domestic policy, saying it would add to the #deficit. He complained about a lucrative deal that went to a rival co to build an #AI #DataCenter in the #MiddleEast. And he has yet to make good on a $100M pledge to Trump’s political operation.
#BooMutherFuckinHoo #USpol
https://www.nytimes.com/2025/05/28/us/politics/elon-musk-trump-doge.html?smid=nytcore-ios-share&referringSource=articleShare&sgrp=p&pvid=94709C91-3556-4635-8095-65DC81177EF3
#月曜がんばれ 136
「できるできるっ！君なら絶対できる！がんばれ♥」
#松實優麗 #うちの子
#stablediffusion #AI #AI繪圖 #AIイラスト #AIart #AIイラスト好きさんと繋がりたい
More, faster, and wrong - why #AI won't solve your software productivity problems. 🤖 https://matthewreinbold.com/2025/06/19/theproductivitytrap
Hallucinating #GoldStandard science
"The #MAHA Report Cites Studies That Don’t Exist."
https://www.notus.org/health-science/make-america-healthy-again-report-citation-errors
"Health Secretary #RFKJr says his “Make America Healthy Again” Commission report harnesses “gold-standard” science, citing more than 500 studies and other sources to back up its claims. Those citations, though, are rife with errors, from broken links to misstated conclusions. Seven of the cited sources don’t appear to exist at all."
#AI #DefendResearch #Medicine #ScholComm #Trump #TrumpVResearch #USPol #USPolitics
Senators Demand Transparency on Canceled Veterans Affairs Contracts
—
Following a ProPublica investigation into how DOGE had developed an error-prone AI tool to determine which VA contracts should be killed, a trio of lawmakers said the Trump administration continues to “stonewall” their requests for details.
#News #DOGE #Veterans #VA #AI #ArtificialIntelligence #LLM #Technology #Government
AI is bad compression. Every time you run training material through it, you get a lossy summary of that material back, along with some noise.
You quickly run out of *quality* training material and start dog-fooding the output back in. Then you end up with lossy summaries of lossy summaries, and eventually all your pizza sauce recipes are dog food.
Pretty sure all these #AI guys have utterly misconstrued what art is. They think it's just information/data, but it's not: it's a process; it's a manifestation of interwoven conscious & unconscious thoughts that have ripened in a human mind, sometimes for decades. And the result is an interaction with another human mind, a dialogue. 1/3
BEWARE of Google AI generated video scams -- eight second segments can be a critical clue!
Fraudsters are now using the new advanced #Google #AI video system to create scam videos (including claiming to be from #YouTube) of humans speaking to camera. These are EXTREMELY convincing. One clue (for now) is that the created videos are often limited to eight second segments. To create longer scam videos, fraudsters are stringing together these segments of no more than eight seconds each, usually with a quick fade between them to mask the discontinuities between the segments.
Google is making life much easier for the fraudsters, but much worse for you.