AI Megathread
-
@bear_necessities said in AI Megathread:
The world of finance and business is not exactly my small little circle here.
Many of the applications is fin/biz are using specific Machine Learning, rather than Generative AI. Whole different ballgame.
Though there certainly are some crappy GenAI uses (replacing blog post writers with shallow slop, replacing customer service agents with unhelpful bots, and the like), this is not a big profit center for anyone. OpenAI lost 5 billion last year while making only 3.7 billion in revenue. 3.7b is certainly not chump change for us normal folk, but for a megacorp like that it’s pitiful and unsustainable. In contrast, for example, Apple made 95 billion in the fourth quarter alone.
-
@Faraday said in AI Megathread:
OpenAI lost 5 billion last year while making only 3.7 billion in revenue. 3.7b is certainly not chump change for us normal folk, but for a megacorp like that it’s pitiful and unsustainable. In contrast, for example, Apple made 95 billion in the fourth quarter alone.
Not to get super nerdy here but it’s not that unusual for a company of that magnitude in early start-up phases to see significant losses, nor is looking at net loss vs gross revenue a great judge of financial performance for start-up companies. I could go into a lot of detail here but this is definitely not the place.
w/ regards to THIS CONVERSATION, it looks like I might be confusing the different AI models but it doesn’t really change the conclusion that the “Machine Learning” AI is just as harmful and will result in significant job losses. It just doesn’t impact our games. Yet, anyway, IDK one day it might.
-
@Jumpscare said in AI Megathread:
@Trashcan said in AI Megathread:
What most people are expressing is the desire for transparency, to know if AI was used to create the content they are engaging with in the MU* space, and if there was more of that, I expect there would be less false accusations to go along with it.
There’s even a real-world example of this! That one Ares L&L game that used AI and was quickly called out for not disclosing it.
This made me discover it was a hot button issue for me to disclose AI usage. Like @Third-Eye eye said up thread, the major thing for me is disclosing whether or not you have used AI, and in what ways you have used it. I could potentially get around to embracing an AI policy that allows for the possibility that players might use AI–
But I cannot embrace a policy where they don’t fucking flag that they are using AI. If it’s about blah-de-blah trust then why aren’t you fucking telling people that you are using AI. HUH. If it adds so much value to what you do, why do you hide it. HUH. Obviously if it is that good you should be able to show it. HUH!!!
Then again, the last time I REALLY saw someone use AI in a game, they were using AI to disguise their poses while they dualboxed alts and used one alt to help the other. They got caught at it. They got banned. You know who you are.
Maybe there IS something to say about AIs and people who use them and trust.
-
@bear_necessities said in AI Megathread:
Not to get super nerdy here but it’s not that unusual for a company of that magnitude in early start-up phases to see significant losses, nor is looking at net loss vs gross revenue a great judge of financial performance for start-up companies. I could go into a lot of detail here but this is definitely not the place.
That’s true, but usually companies like that have a clear plan. Uber can operate at a loss because eventually it’ll drive all the taxis out of business. Then it jacks up the prices and has a profitable monopoly. Amazon operated at a loss for a long time to cement itself as a hub so that it can make a profit on razor-thin margins at tremendous scale.
The concern that has been raised in many economic circles is that these GenAI companies don’t have a coherent plan for how to become profitable. Now sure it’s possible they’ll stumble into some amazing thing that unlocks a huge profit center. It’s just unlikely to come from replacing wikipedia with a plagiarized information source or providing second-rate chat support. Time will tell.
-
@Faraday I do see chatGPT being used a lot now for marketing and “personal branding”. A majority of the stuff I see on LinkedIn is just chatGPT shit anymore. A few business consultants I’ve worked with have suggested using chatGPT to assist with everything from marketing strategies to codes of conduct for employees. So I don’t think it will be too long before they churn a profit off of that market alone.
-
@bear_necessities Even on their Pro plan, ChatGPT is losing money off every prompt their users send it. And a large chunk of their users are just using the free plan. Could they raise prices? Sure. The question is how much cost the market will bear for polishing up LinkedIn posts.
-
@bear_necessities said in AI Megathread:
@Pavel said in AI Megathread:
In your circles. It’s not part of the general zeitgeist, so it’s easily missed or ignored.
I actually can’t tell if you’re being sarcastic or not. The world of finance and business is not exactly my small little circle here.
Neither is the world of medicine, but the specific uses of technology in that field aren’t commonly known nor understood by the layman. You use tools we, the general public, don’t know about.
-
Major newspapers post a summer section of AI slop that includes, among other things, a summer reading list for kids with imaginary books (complete with imagined summaries).
Apart from the obvious debacle, I think that article has a good take on the second-order effects of GenAI that many don’t consider.
- Most “free” internet sites are free because they have ads.
- As more people get their content from AI-generated slop like ChatGPT, the only people coming to sites are AI bots.
- Nobody’s going to pay to advertise to bots.
- Ad money dries up.
- Site goes behind a paywall, or (if it can’t sustain itself with subscriptions) no longer exists.
We’re already seeing more paywalls, and the problem will only get worse.
-
If LLM chatbots weren’t so chronically wrong, using them to dodge adverts and engagement bombardment might actually be a decent use case.
-
@Juniper said in AI Megathread:
If LLM chatbots weren’t so chronically wrong, using them to dodge adverts and engagement bombardment might actually be a decent use case.
Until there’s no more content for them to gobble up because all the websites they stole from have shut down.
-
Part of the reason we’re now drowning in content slop farms that themselves use generative AI is because advertising revenue made content profitable and incentivised creating content as quickly as possible while eliminating any kind of standards for quality.
I don’t have a lot of sympathy for advertisers or those who are paid by them because they form part of the ecosystem that put us in this mess. Let the model collapse, it stopped serving us long ago.
-
@Juniper The existing ad model sucks, but there are other ways to solve that problem. If someone is doing work to put out professional content, they shouldn’t be expected to give it away for free, and it certainly shouldn’t be stolen from them by a plagiarism machine. They deserve compensation, whether that’s through a subscription or ads. I have no problem with, for instance, YouTube’s model where you get to choose between the two.
But regardless of philosophy, what I’m talking about is simple cause and effect. ChatGPT has to get its information from human content creators. If OpenAI drives them all out of business, they’re just shooting themselves in the foot too. But by the time that happens, the damage to all the other creators will already have been done.
-
@Juniper said in AI Megathread:
revenue made content profitable and incentivised creating content as quickly as possible while eliminating any kind of standards for quality
That’s just late-stage capitalism. Internet advertising standards are a symptom rather than a cause.
