andallthat

@andallthat@lemmy.world

• ⁨Comment⁩ on ⁨AI-generated code contains more bugs and errors than human output⁩ ⁨⁨3⁩ ⁨days⁩ ago⁩:

  Microsoft could write an AI agent to filter threads based on context you don’t like. Come to think of it, Megagenius Elon Musk already has one he wrote to censor anti-Israel posts on Trump’s Truth Social. There, I think I got them all… Happy holidays!

• ⁨Comment⁩ on ⁨If AI replaces workers, should it also pay taxes?⁩ ⁨⁨1⁩ ⁨week⁩ ago⁩:

  Not an expert at all, but I think to an extent this already happens with the current system in most countries, and it would probably need to be done much more now. Not that Automation pays more taxes, but that having employees generally qualifies companies for tax breaks. For instance, when Amazon said “we’re going to open a new HQ”, States tripped over themselves to try and give them the largest tax breaks. But that was under the assumption that the HQ would give jobs to tens of thousand of people, not to 5 data scientist and a massive, energy-hungry data center.

• ⁨Comment⁩ on ⁨Windows Marketshare since 2010⁩ ⁨⁨1⁩ ⁨week⁩ ago⁩:

  And yes, you CAN run Doom on it. But turns out it’s not a great idea

• ⁨Comment⁩ on ⁨Grok is spreading misinformation about the Bondi Beach shooting⁩ ⁨⁨1⁩ ⁨week⁩ ago⁩:

  ***k

  Hey, how do you know my password?

• ⁨Comment⁩ on ⁨AI Slop Is Ruining Reddit for Everyone⁩ ⁨⁨2⁩ ⁨weeks⁩ ago⁩:

  Yes it did. Making up variation of the same story in order to farm upvotes used to be done by humans.

  But the strategy of throwing shit at the wall and see what sticks has now been industrialized with AI, because the machine can produce tons of cheaper, faster, smellier shit.

  Socials are basically the perfect application for AI. Unreliable results are not a bug but a feature. You have humans helpfully training it for free by up or downvoting the result. And the AI companies get a machine trained to persuade large groups of people of any made-up story.

• ⁨Comment⁩ on ⁨The Algorithm That Detected a $610 Billion Fraud: How Machine Intelligence Exposed the AI Industry’s Circular Financing Scheme⁩ ⁨⁨3⁩ ⁨weeks⁩ ago⁩:

  I like how there are all these terms with increasingly loose definitions, to which we attach different levels of evilness:

  • algorithm - older, more reliable, mostly deterministic except when it’s “The Algorithm” in capital letters like “The Social Media Algorithm”, then it becomes evil
  • machine learning - been out for decades, hasn’t destroyed the world, mostly does its job undetected. Mostly used only by technical people
  • machine intelligence - The machine is starting to become conscious but it is still mostly helpful. “Machine intelligence” does brain surgery, folds and unfolds proteins whatever that means (but it sounds like a good thing, so we’ll give it a pass)
  • artificial intelligence - machine intelligence’s evil twin. Takes credit for everything good that comes from the other ones and we tend to believe it, because it’s the only one we can actually speak to and can lie to us very convincingly. On its own it can draw pretty pictures and animate them, write code that occasionally works, pretend to love us and teach us the most effective way to slash our own wrists
• ⁨Comment⁩ on ⁨IBM CEO says there is 'no way' spending trillions on AI data centers will pay off at today's infrastructure costs⁩ ⁨⁨3⁩ ⁨weeks⁩ ago⁩:

  One day we’ll read some of these comments and laugh at how shortsighted they were.

  Of course we’ll probably have to read them on a manuscript or smeared on a wall with feces because all the world’s resources will be used to power the huge datacenter that powers our AI overlords

• ⁨Comment⁩ on ⁨Elon Musk Had Grok Rewrite Wikipedia. It Calls Hitler “The Führer.”⁩ ⁨⁨4⁩ ⁨weeks⁩ ago⁩:

  “I’m not a Nazi, I just like the esthetics of blood on a black uniform”

• ⁨Comment⁩ on ⁨Windows 11's adoption is much slower compared to Windows 10, claims Dell⁩ ⁨⁨4⁩ ⁨weeks⁩ ago⁩:

  It’s almost like “you have to buy a new laptop to install it and help train our AI on all you documents” is somehow not convincing enough. Maybe if they also removed local accounts and forced you to have an online MS account? Nah scratch that, it would be stupid

• ⁨Comment⁩ on ⁨Elon Musk’s Grok Goes Haywire, Boasts About Billionaire’s Pee-Drinking Skills and ‘Blowjob Prowess’⁩ ⁨⁨5⁩ ⁨weeks⁩ ago⁩:

  AI goes haywire and starts actually giving reliable information

• ⁨Comment⁩ on ⁨Microsoft says Copilot will 'finish your code before you finish your coffee' adding fuel to the Windows 11 AI controversy that's still raging⁩ ⁨⁨5⁩ ⁨weeks⁩ ago⁩:

  but can YOU do it before I finish my coffee?

• ⁨Comment⁩ on ⁨Elon Musk’s Grok Goes Haywire, Boasts About Billionaire’s Pee-Drinking Skills and ‘Blowjob Prowess’⁩ ⁨⁨5⁩ ⁨weeks⁩ ago⁩:

  AI likes to finish on his face and some went in his eyes

• ⁨Comment⁩ on ⁨ICE to Buy Tool that Tracks Locations of Hundreds of Millions of Phones Every Day⁩ ⁨⁨2⁩ ⁨months⁩ ago⁩:

  I track the location of hundreds maybe thousands of phones every day for minutes at the time. I see people using them while I commute. Where can I collect my fee from the US government for my services?

• ⁨Comment⁩ on ⁨Superhero stories have become less about saving people and more about fighting villains.⁩ ⁨⁨2⁩ ⁨months⁩ ago⁩:

  “Not with that attitude, you can’t!” -Lego

• ⁨Comment⁩ on ⁨If sexuality is a spectrum, does that mean one person is the gayest?⁩ ⁨⁨2⁩ ⁨months⁩ ago⁩:

  It’s more that each person will have a moment in their life when they are at their gayest

• ⁨Comment⁩ on ⁨Exactly Six Months Ago, the CEO of Anthropic Said That in Six Months AI Would Be Writing 90 Percent of Code⁩ ⁨⁨3⁩ ⁨months⁩ ago⁩:

  or CEOs

• ⁨Comment⁩ on ⁨The time and expense of commuting is theft, if that job can be done from home.⁩ ⁨⁨3⁩ ⁨months⁩ ago⁩:

  Might not be the biggest risk to your life but it’s 2.3% out of 55 Millions death (so, just to put things into perspective, we’re talking 1.6M deaths per year).

  Besides, the risk is not only death directly in an accident. For those 33% who died from heart diseases: stress is a risk factor for cardiovascular disease and the paper does mention specifically work stress as a risk factor

  The risk of a cardiovascular event was higher in patients with a history of social isolation (OR, 2.47), marital stress (OR, 2.28), work stress (OR, 3.2), childhood abuse (OR, 2.78), or trauma (OR, 2.67).

  Again not all of it will be due to commuting to work, but raising your stress levels by having you commute needlessly in traffic is not good for your health.

• ⁨Comment⁩ on ⁨South Korea makes AI investment a top policy priority to support flagging growth⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  “A grand transformation into AI is the only way out of growth declines resulting from a population shock,” the ministry said in a statement, referring to South Korea’s record low birthrate.

  The funny bit is how “AI companions” are one of the most profitable uses of AI so far . See how THAT increases a country’s birthrate.

• ⁨Comment⁩ on ⁨[PDF] Tesla is slow in reporting crashes and the feds have launched an investigation to find out why⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  This is Analysis-Paralysis. Why should they spend all their time counting past crashes when they are busy increasing the production of new ones?

  /s

• ⁨Comment⁩ on ⁨UK Official Calls for Age Verification on VPNs to Prevent Porn Loophole⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  Ah yes “The Porn Loophole”, was one of my favorites , I should still have it on a DVD somewhere.

• ⁨Comment⁩ on ⁨Codeberg: army of AI crawlers are extremely slowing us; AI crawlers learned how to solve the Anubis challenges.⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  LLMs can’t do protein folding. A specifically-trained Machine Learning model called AlphaFold did. Here’s the paper.

  Developing, training and fine tuning that model was a research effort led by two guys who got a Nobel for it. Alphafold can’t do conversation or give you hummus recipes, it knows shit about the structure of human language but can identify pattern in the domain where it has been specifically and painstakingly trained.

  It wasn’t “hey chatGPT, show me how to fold a protein” is all I’m saying and the “superhuman reasoning capabilities” of current LLMs are still falling ridiculously short of even sinpler problems.

• ⁨Comment⁩ on ⁨This CEO laid off nearly 80% of his staff because they refused to adopt AI fast enough. 2 years later, he says he’d do it again⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  As a paid, captive squirrel, focusing on spinning my workout wheel and getting my nuts at the end of the day, I hate that AI is mostly a (very expensive) solution in search of a problem. I am being told “you must use AI, find a way to use it” but my AI successes are very few and mostly non-repeatable (my current AI use case is: “try it once for non-vital, not time-sensitive stuff, if at first you don’t succeed, just give up, if you succeed, you saved some time for more important stuff”).

  If I try to think as a CEO or an entrepreneur, though, I sort of see where these people might be coming from. They see AI as the new “internet”, something that for good or bad is getting ingrained in everything we do and that will cause your company to go bankrupt for trying too hard to do things “the new way” but also to quickly fade to irrelevance if you keep doing things in the same way.

  It’s easy, with the benefit of hindsight, to say now “haha, Blockbuster could have bought Netflix for $50 Millions and now they are out of business”, but all these people who have seen it happen are seeing AI as the new disruptive technology that can spell great success or complete doom for their current businesses. All hype? Maybe. But if I was a CEO I’d be probably sweating too (and having a couple of VPs at my company wipe up the sweat with dollar bills)

• ⁨Comment⁩ on ⁨AI experts return from China stunned: The U.S. grid is so weak, the race may already be over⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  So, a few monthe ago China launched Deepseek and the narrqtive was all “the fact they didn’t have access to the latest Nvidia GPUs forced them to get creative and develop a model that is more efficient and cheaper”.

  Now the US is getting behind on “AI wars” because China has more energy for huge data centers?

  How about the US get creativve and develop LLMs that are actually useful and can work without sucking Gigafucks of electricity?

• ⁨Comment⁩ on ⁨Starlink tries to block Virginia’s plan to bring fiber Internet to residents⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  I don’t know how much Musk can be separated from Starlink. Not only because Starlink, as part of SpaceX, is privately held but also because the main reason they now have a superior service to offer is that they got fucktons of money from government customers, which is also tied to Musk’s action

  A big part of Musk’s involvement with politics is because everything he does, from EVs to rockets to, now, big energy-guzzling datacenters for AI, needs a lot of government backing, if not in terms of direct contracts at least in terms of regulation and incentives.

  Even his direct involvement with Trump wasn’t because he suddenly became a Nazi (he’s probably always been one, according to his own family) but in order to become even more entangled with government investments, even trying to control NASA directly.

  And not only US governments. I remember Musk suddenly being everywhere in Europe pitching Starlink. Meloni’s government in Italy was grilled for allegedly agreeing on a big contract with Starlink.

• ⁨Comment⁩ on ⁨Have you ever noticed no one ever looks at a penis and then says, "wow, he must have big feet."⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  Like, you know when you are walking and suddenly they Hulk out and break your shoes? Man I hate foot-boners!

• ⁨Comment⁩ on ⁨Starlink tries to block Virginia’s plan to bring fiber Internet to residents⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  but not in this order, the reverse would be so much more satisfying

• ⁨Comment⁩ on ⁨OpenAI Is Giving Exactly the Same Copy-Pasted Response Every Time Time ChatGPT Is Linked to a Mental Health Crisis⁩ ⁨⁨4⁩ ⁨months⁩ ago⁩:

  ah, dear old copy/paste… It’s funny that even OpenAI doesn’t trust ChatGPT enough to give LLM-generated more personalized answers.

• ⁨Comment⁩ on ⁨“You can't be expected to have a successful AI program when every single article, book or anything else that you've read or studied, you're supposed to pay for” Donald Trump said⁩ ⁨⁨5⁩ ⁨months⁩ ago⁩:

  actually Donnie’s friends prefer to fuck high school students

• ⁨Comment⁩ on ⁨“You can't be expected to have a successful AI program when every single article, book or anything else that you've read or studied, you're supposed to pay for” Donald Trump said⁩ ⁨⁨5⁩ ⁨months⁩ ago⁩:

  “you can’t have a successful government when every time I want to be President or have sex with minors or anything else you have the right to do as a rich, white man, you have to hear people get all judgy”

• ⁨Comment⁩ on ⁨Vibe coding service Replit deleted production database⁩ ⁨⁨5⁩ ⁨months⁩ ago⁩:

  The part I find interesting is the quick addiction to working with the LLM (to the point the guy finds his own estimate of 8000 dollars/month in fees to be reasonable), his over-reliance for things that from the way he writes he knows are not wise and the way it all comes crashing down in the end. Sounds more and more like the development of a new health issue.