▲ 1287 ▼ Microsoft CEO warns that we must 'do something useful' with AI or they'll lose 'social permission' to burn electricity on it (www.pcgamer.com) submitted 9 months ago by throws_lemy@reddthat.com to c/technology@lemmy.world 350 comments fedilink hide all child comments Workers should learn AI skills and companies should use it because it's a "cognitive amplifier," claims Satya Nadella. in other words please help us, use our AI
[+] Photonic@lemmy.world 10 points 9 months ago (5 children) [removed by mod] permalink fedilink source hideshow 5 child comments replies: [–] ArmchairAce1944@discuss.online 4 points 9 months ago (1 child) Allrecipes has you covered. permalink fedilink source parent hideshow 1 child comment replies: [–] zarkanian@sh.itjust.works 2 points 9 months ago No. No, it really doesn't. I want a vegan recipe that uses turbinado sugar. I get 3 articles and only one of them is a recipe. If I don't like that recipe...too bad. That's what they have. permalink fedilink source parent [–] SparroHawc@lemmy.zip 4 points 9 months ago (1 child) It's not the query that burns through electricity like crazy, it's training the models. You can run a query yourself at home with a desktop computer, as long as it has enough RAM and compute cells to support the model you're using (think a few high-end GPUs). Training a model requires a huge pile of computer power though, and the AI companies are constantly scraping the internet to stealfind more training material permalink fedilink source parent hideshow 1 child comment replies: [–] sobchak@programming.dev 2 points 9 months ago Dunno if that's true or not. Generally, much more compute is used in inference than training, since you only train once, then use that model for millions of queries or whatever. However, some of these AI companies may be training many models constantly to one-up each-other and pump their stock; dunno. The "thinking" model paradigm is also transferring a lot more compute to inference. IIRC OpenAI spent $300k of compute just for inference to complete a single benchmark a few months ago (and found that, like training, exponentially increasing amounts of compute are needed for small gains in performance). permalink fedilink source parent [–] nickiwest@lemmy.world 2 points 9 months ago The number of adults I know who ask ChatGPT for recipes is non-zero. Teenagers use it like it's a search engine. They don't understand the difference. permalink fedilink source parent
[–] ArmchairAce1944@discuss.online 4 points 9 months ago (1 child) Allrecipes has you covered. permalink fedilink source parent hideshow 1 child comment replies: [–] zarkanian@sh.itjust.works 2 points 9 months ago No. No, it really doesn't. I want a vegan recipe that uses turbinado sugar. I get 3 articles and only one of them is a recipe. If I don't like that recipe...too bad. That's what they have. permalink fedilink source parent
[–] zarkanian@sh.itjust.works 2 points 9 months ago No. No, it really doesn't. I want a vegan recipe that uses turbinado sugar. I get 3 articles and only one of them is a recipe. If I don't like that recipe...too bad. That's what they have. permalink fedilink source parent
[–] SparroHawc@lemmy.zip 4 points 9 months ago (1 child) It's not the query that burns through electricity like crazy, it's training the models. You can run a query yourself at home with a desktop computer, as long as it has enough RAM and compute cells to support the model you're using (think a few high-end GPUs). Training a model requires a huge pile of computer power though, and the AI companies are constantly scraping the internet to stealfind more training material permalink fedilink source parent hideshow 1 child comment replies: [–] sobchak@programming.dev 2 points 9 months ago Dunno if that's true or not. Generally, much more compute is used in inference than training, since you only train once, then use that model for millions of queries or whatever. However, some of these AI companies may be training many models constantly to one-up each-other and pump their stock; dunno. The "thinking" model paradigm is also transferring a lot more compute to inference. IIRC OpenAI spent $300k of compute just for inference to complete a single benchmark a few months ago (and found that, like training, exponentially increasing amounts of compute are needed for small gains in performance). permalink fedilink source parent
[–] sobchak@programming.dev 2 points 9 months ago Dunno if that's true or not. Generally, much more compute is used in inference than training, since you only train once, then use that model for millions of queries or whatever. However, some of these AI companies may be training many models constantly to one-up each-other and pump their stock; dunno. The "thinking" model paradigm is also transferring a lot more compute to inference. IIRC OpenAI spent $300k of compute just for inference to complete a single benchmark a few months ago (and found that, like training, exponentially increasing amounts of compute are needed for small gains in performance). permalink fedilink source parent
[–] nickiwest@lemmy.world 2 points 9 months ago The number of adults I know who ask ChatGPT for recipes is non-zero. Teenagers use it like it's a search engine. They don't understand the difference. permalink fedilink source parent