Len Necefer's investigation into AI energy and water use is fascinating. It did challenge my own beliefs, but what I found most interesting is how people will likely look back on this kind of intensive AI use in future. Len's use of Claude likely lost Anthropic a *massive* amount of money, but because of the way the true cost of token burn is kept opaque to the user, people don't realise this. Setting aside environmental concerns that may or may not be justified, consumer-facing AI is quite simply not a viable business and (as I understand it) there is no way of making it a viable business.
So, while I think Len has done some useful work in setting out facts rather than clickbait hyperbole – and setting aside my moral objections to many of the AI uses he cites – I would be very interested to hear his take on the economic future of an AI-heavy workflow like this.
Additionally, this fills me with abject horror. I don't understand how people can think this kind of thing is OK and not think ahead to the consequences. I'm afraid this post made me lose all respect for Len. https://www.instagram.com/p/DWNjwZpjeTU/?img_index=1
Len Necefer's investigation into AI energy and water use is fascinating. It did challenge my own beliefs, but what I found most interesting is how people will likely look back on this kind of intensive AI use in future. Len's use of Claude likely lost Anthropic a *massive* amount of money, but because of the way the true cost of token burn is kept opaque to the user, people don't realise this. Setting aside environmental concerns that may or may not be justified, consumer-facing AI is quite simply not a viable business and (as I understand it) there is no way of making it a viable business.
I recommend Ed Zitron's blog. His posts are long, rambly and ranty, but the analysis is sound. It's all built on economic quicksand. https://www.wheresyoured.at/why-are-we-still-doing-this/
So, while I think Len has done some useful work in setting out facts rather than clickbait hyperbole – and setting aside my moral objections to many of the AI uses he cites – I would be very interested to hear his take on the economic future of an AI-heavy workflow like this.
Additionally, this fills me with abject horror. I don't understand how people can think this kind of thing is OK and not think ahead to the consequences. I'm afraid this post made me lose all respect for Len. https://www.instagram.com/p/DWNjwZpjeTU/?img_index=1