Why is there not real demand for AI among normies?
>>109007717what is the source of this chart
>>109007717Making crude Pixar memes got boring quick I guess
>>109007717AI is only useful for giving quick summaries of topics with mediocre accuracy. Not exactly a productivity booster.
>>109007727https://isaiprofitable.com/
>>109007717I've been using it to help write star trek fan fiction. Its been of limited use to be honest. Kind of a let down.
>>109007749kek, luddie cope
>>109007717Use case?
>>109007717>Why is there not real demand for AI among normies?huh?it's only normies who use it though
Think about it for five seconds and you have your answer, what can an AI do that's actually useful for a normie?
>>109007717Wait, Amazon does AI? How have they managed to spend so much with so little impact? Or is it just hosting and shit for other companies.
Because normies, like my self, just use it as a google type tool. It's only weirdos who aren't good at things than use it to make music and code and stuff. And govts, businesses and military are using it to kill people and make money.
The use case for AI is unironically, scamming people. If it was anything to be excited about, you would see people be open about AI usage in games, movies, etc. >"Look at the incredible things I'm doing with this tool, that couldn't have been done otherwise!"But you hardly ever see this happen. It's almost always trying to cut corners to deliver the same slop, but faster. It's a 'technology' where nobody is really pushing the envelope and making a genuine use case. It's just doing the same shit faster and tackier. It is rather apocalyptic because it's now faster to create things than it is to assess it, so every medium, community, and means of communication is being flooded with bad faith actors trying to be a 1-man-disinformation agency. It's shattered what social contract remains on the internet.There are some decent uses for it, like being able to detect cancer earlier or solve math problems that were simpler than people thought (if they were truly complex, current LLMs wouldn't have been able to solve them), but nothing outweighs the genuinely apocalyptic tsunami of low quality indian-made shit that is washing out all civilization and art.Soon, the only thing you will be able to trust is vetted individuals, artists, physical vintage sources (which will skyrocket in price over the next few decades), or institutions. Which is exactly what the elites wanted: seizing control over culture again.
>>109007717Normies only needed a LLM capable of searching the Internet and summarizing the content Wikipedia-style and nothing more. All this agent delegation bullshit only appeals to jeets
>>109007717>Why is there no demand for eight meter high haul trucks outside of the industrial mining industry?
>>109007717>Why is there not real demand for AI among normies?The same reason normies dont use Krita or Blender etcThey have nothing original to say to or show the world.
>>109007717Their daily brainless tasks require zero AI>>109008675Retard
>>109008828>trillions of tokens burned>still nothing to show for itgm saaaaaaaar
>>109007717>>109007749I mean...
>>109007717AI fails to answer the question of the "why". Why would I want someone to write my essays and draw my art? If I was interested enough, I'd do that stuff on my own.
>>109009378When there's a gold rush, sell shovels.
>>109007717normies can't find any practical use for a great many advanced tools, because they are ignorant.https://m.youtube.com/watch?v=OHQRo3Uz_VQ&ra=m
>>109007717Actually, normies are huge on AI, it has replaced googling for them.Even when they want to look up an easily veriable fact (like a rule of chess or where a company's HQ is) they'll ask chatgpt "because it's easier". Students are hooked enough that they're willing to pay a subscription, even if only for finals and large projects.However, the average person doesn't see a reason to keep an AI subscription running 24/7, the same way they might pay for netflix and spotify every month without thinking about it. So we'll either have a bubble burst, cheaper models, or comically expensive tiers like $500 or $1000 per month.
>>109008306The reason people are not doing it is because there is an angry mob of luddites willing to piss and shit themselves over any public ai usage. There have been multiple cases when artists backed down even though they find AI useful
>>109007717Because it's just a slightly inaccurate wikipedia article most the time.
>>109007764Yeah it keeps calling Lt. Commander Data "Data"
>>109008868Mythos has found so many exploits in open source software that they're falling behind in fixing them
>>109009506Yeah there've been multiple cases of artists hopping on board AI for worse tracing and nonsensical referencing, then being surprised they get jumped when everybody figures out they're using the slop machine
>>109009604No, I'm talking about Jorge Gutierrez getting canceled also Martin Scorsese made a lot of people seethe
>>109009668Gutierrez claims he knew little or nothing of the tool before *signing on for a fucking animation project with it,* the best possible read is that he's an idiot
>>109009682>he knew little or nothing of the tool before *signing on for a fucking animation project with it*You're trying to make it sound like a bad thing but it's neutral at best
>>109009705I also like to walk off of cliffs and dive into wet concrete btw not sure if that matters
>>109009717Apt analogy for using software to make cartoons, yes
>>109007717It'd be interesting to see these compared to the parent company's total spend and revenue, because a company like Amazon, even though they're eating the biggest AI related losses, can probably just tank it overall and has a good chance of still being standing at the end of all this, whereas for an AI only company like Anthropic probably can't keep taking this over time if they don't start making more money.
>>109009780Every AI company has been operating at enormous loss and only exists or is kept afloat by government hand-outs and venture capitalists, it's why all of them are horny to force AI into profitability.
>>109009466Everyone who loves AI is not paying per token, the actual cost of using the AI. It's hard to argue with a subsidized product. How long can the AI businesses keep giving their product away for free, though? At some point they have to make a profit and pass the compute cost to the customer.
Are we still pretending like AI coding isn't far superior to manual coding?
>>109009880True but I'm not really sure what it has to do with my post.
>>109009378Shovel sellers stay winning
>>109009600So that's a net negative to the world.Finding security holes is a double edged sword and if more get found than can be fixed it's worse than doing nothing.
>>109009962Considering most people using AI for coding appear to be using closed proprietary cloud LLMs under the full control of another company and hosted solely in their datacenters, I'm not sure I can see it being too helpful for society in the long run. Is there a problem with using open LLMs?
>>109010292finding security holes and patching them is good actually
>>109010306what's the practical difference between using KIMI hosted on deepinfra vs Claude hosted on amazon bedrock
>>109008016Amazon bankrolls anthropic
>>109008306You're a pedantic bitch flapping her arms about because in a few years the 'slop' will be higher quality than anything humans can make. So, you'll soon get used to it. Bitch and cope until then I guess. It won't be long in anycase
>>109010377>higher quality than anything humans can makeAll these AI tools, computers, decades of knowledge, and yet you still haven't made a better film than kubrick, a better game than super metroid, a better piece of music than sergei, etc. and you never will.You could literally have a magic button that generates a fully-fledged game on demand from your mind and you still couldn't do it. Pathetic.
>>1090104033 years ago the bar was at low res spaghetti will smith, now it's at Kubrick level cinema. Interpolate
>>109010526>chantard doesn't understand diminishing returns
>>109010542There are no diminishing returns. AI is solving unsolved math problems. World renown physicists jump ship from academia to frontier labs to get access to latest models. METR time horizons grow linearly.
