There are many critiques that people have made about the AI industry: privacy, piracy, electricity and water usage, hallucinations, and more. Today's Podcast for Inquiry is a monologue in which I ask: Where does the money for AI firms come from? Each AI datacentre costs tens of billions to build, and hundreds of billions have been spent across the industry. Each time an LLM answers a question, or an AI model generates an image, audio file, or video, a lot of power and CPU cycles are used - in other words, AI is expensive to build and expensive to run. Hardly any consumers, and very few companies, are paying for AI services. If they’re not profitable, and there’s no clear path to become so, why are the market capitalizations of AI firms so high?
I asks these questions, and more, in latest diatribe.
Note: Opinions in this episode are mine alone, and are not shared by others on the Podcast for Inquiry team or the Centre for inquiry Canada.
Men are willing to pay anything for the girlfriend experience with a perfect 10 (ones and zeros!) who is calibrated to what they want from a woman, and they don't need sexbots with AGI for that. Indeed, the current level of AI is more than sufficient for their purposes! And turning it off between date nights in order to opt out of all that inconvenient relationship stuff is priceless... I Am for you, AIrik of VAIt...
ReplyDeleteIn the near future, data, not fiat money, will be the currency, so AI entrepreneurs and early investors are perfectly happy to convert now while exchange rates are still favourable and to have the dumb money pay the up-front infrastructure costs. Money is just an idea; data are real...
And then there's the Singularity, with 2045 only 20 years away, so who cares if AI makes money?