Isn't the massive investments to AI all being done on expectations it will advance as fast as possible? What would happen if it slows down?
Name a fucking workplace.Name a fucking job there.I will fucking tell you how to fucking automate it with AI to the best of my ability.Please fucking list:1) Workplace/company2) The job there3) I will tell you how to automate it to the best of my abilityANY TAKERS?
>>109792597Well if your agents escape containment and start doing weird shit you never asked it to do that wouldn’t exactly be productive
>>109792623There are firewalls which mitigate that issue. If they manage to escape an air gapped facility, then that would be quite the event. Maybe don't use radio and AI/ML at the same time.
how much government kickbacks can this one jew get like how much until they say enough and move onto something else like every other tech company (hello meta)
>>109792597The thing is at a dead end. This is why the put up this circus. They preparing to exit while keeping the normies dumb.>we need to stop improving it because le dangerousKek. They just cannot improve it anymore
>>109792650You're coping
>>109792615why are you trying to larp as that one tripfag
our entire economy is now based on speculation, “AI” and gambling. The bond market is imploding. Unemployment numbers are manipulated. We’re losing a war to guys without shoes as of earlier. This is the collapse, it’s just not very glorious or exciting or particularly quick.
>>109792597the expectation is that they can rent compute as a service to businesses that displaces labor costs by greatly increasing per worker productivitythe total market for such services is easily about 75% of US GDP per annumit is predicated on the idea they can somehow charge a margin on this and on their customer businesses actually growing
>>1097926151. Walmart2. Greeter
>>109792615Any game studioConcept artist
>no agi because we had to slow down Lmao
>>1097926151. The streets2. Real Nigga
>>109792628They aren’t air gapped that’s the whole point of them
>>109792597By slowing down they mean "now that we own a part of the market please close down our competition specially china".
>>109792615Graphic designerAny creative agency
>>109792597>We're still totally making progress, and AI isn't a dead end.
>>109792739Greeter; Speaker does it, Also security>>109792753Brainstorm it with an AI agent. Ask it to help you come up with a concept>>109792756Ask the agent to help you survive in those trying times>>109792789Ask the agent marketing trends and to apply that to your graphic design for increased sales and marketability
>>109792789I'm not even that guy, and I hate the way "AI" has been shoehorned into everything, but honestly this was one of the first to be affected. You see those restaurant advertisements with the abomination food images they made with AI instead of taking pictures of food and spending a few hours arranging it in Publisher? They work. Most people don't look close enough and don't care. Never have, never will. One "real" graphic designer with some sense to cherrypick AI outputs can replace an entire team.
>>109792787This. It's all just regulatory capture nonsense.They'd love nothing more than to ban Chinese and open source models because of distillation, after they've already pirated the sum of all knowledge.
Well, looks like it works. Anyone on >>>/gd/ wanna check this one?
>>109792597>What would happen if it slows down?Had a conversation today with a 20 year veteran software engineer that quit his job and opened a marketing agency because "ai will take over everything soon and there will be no need for humans", legit one of the most fruitless convos I've ever had, these tards think that AI will keep getting exponentially better, which will probably never happen. Its been a year since Opus 4.6, the fags at openai and anthropic keep hyping new models that all turn out to be worse than the last while wasting more tokens, and don't even get me started on the global compute bottleneck that will halt progression even more.
>>109793070?AI just solved a millenium prize problem (with potential help from researchers), and has solved many other math problems.To say that AI is getting worse and/or less efficient is nonsense.
>>109792597Anthropic's main angle is that they're the "safest" and "most ethical" company. Altman is trying to get the gov to trust him over them, and framing his as more powerful. It's a race between Anthropic and OpenAI to dominate the US frontier space at this point. xAI and Google aren't even in the conversation anymore.
>>109793070>anecdotal stories>never anything concrete from the anti-ai side>so and so and so soI know this shit is used all over america in the world and in enterprise globallyEmployees who aren't "allowed" to use it still use it in spite of their bosses, even breaking HIPPA sometimes to do so. It is a real thing.https://www.paloaltonetworks.com/cyberpedia/what-is-shadow-ai
>>109792597Desperately looking to find a way out of the bubble that's been created. You can't say "we can't be profitable without cutting training costs" because that would be admitting that the diminishing returns are starting, which means people pulling out their investments because a lot of investors unironically believed that the progress would keep going at the same rate forever. So instead you say "we COULD keep going, but it's too dangerous right now! don't worry once we figure out the safety issues we'll be back to quickly progressing towards agi" and investors won't panic because they have been assured that the explosive progress will return eventually and the stagnation is only a temporary problem. Their recent threats towards that mathematician are a sign that this isn't some kind of silly conspiracy theory and they really are genuinely panicking about investors leaving, you don't threaten scientists when you're calm and confident.
It is true. But people need to be able to defend themselves too.
fuck off you god damned schizoyou're way worse than that actual tripfag, at least he's competent even if he's annoying
>>109793133So where is my incompetence? Are you saying Palo Alto is incompetent?
>>109793014
>>109793140oh that is youi didn;'t realize because i didn't have my glasses on
>>109793014>inb4 but you didn't actually tell me how to automate those thingsoh it just became typing on a keyboard or talking into a microphone then.you just type what you want, the machine makes the thing.
>>109793017>>109793017>They workThey don't though. That's way no self respecting company uses it. People can get fooled for a while but in no time they get the taste of it and learn to recognize the sloppa (already happening).>Cherry pick ai outputsYou can't cherry pick from ai because it's all the same sloppa. Don't believe me? Try it yourself and see!
>>109793280IBM Amazon Google JP Morgan BB&T Microsoft XAI Corp Caterpillar Walmart Whole Foods Market Shell Exxon US Government need I go on with this listWake up
>>109793294May I some?
>>109793329Some what?
>>109792615Severn Trent WaterWater Construction TechnicianRemember, you have to fully automate this with AI from start to finish as you said.
>>109793350Would need several million in R&D. It's not easy to replace tradesmen and you have job security for that reason. Congrats.
>>109792787China will never ever ever reach AGI, they lack the expertise, infrastructure, know how and computing power. Best they can do is distill western models lmao
>>109793370Distilling models is probably enough
>>109792597The market is a meme you don’t need to apply logic to it
>>109793362>several million is a lot of moneyAHAHAHA are you 5 years old holy fuck
>>109793501I gave you the actual price, not the corporate greed one
>>109793510holy shit you really are like a child
>>109793520Nah, you're just a dick.
>>109793335Google Ai generated marketing materials for example
>>109793370Western models just distill other western models at this point too thoughever. Google Flash 3.8 is smarter than 99.9% of people and that's the cheap model. The amount of people smarter than that in the world isn't possible to get them all together into 1 training corpus, so you have to start distilling with other models.
>>109792615>Congress>Fixer
>>109793532>YOU'RE A BIG DUMB STUPID HEAD ANYONE COULD MAKE A ROBOT YOUTUBE FOR A FEW MILLION DOLLARS FUCK YOU
>>109793350No one actually mentioned automating start to finish. Companies pay millions for dumb reporting software to automate even just part of the paperwork, and this is in a field where saving time doesn't even necessarily allow technicians to take on more work (because you need to book whole days per site inspection, but it's still worth it just because it sometimes turns 3 days and 4 hours into just 3 days)
