The whole 'China versus America' angle doesn't really interest me, coming from what is (so far) a completely different country, but the news of Deepseek's arrival adds a twist to the already complex world of large lanmguage models. And here's the twist: "The kicker is that DeepSeek created and released its entirely open source project for about $6 million in training costs ('a joke of a budget,' in one expert's words). OpenAI is spending hundreds of millions of dollars." It ould be cool is we could have open source AI that costs so much less to develop, but it's very much early days here, and there might be another shor to drop. Some more links via Mark Oehlert: Anil Dash on DeepSeek, Don't use DeepSeek-v3!, Conor Grennan on DeekSeek, DeepSeek R1's bold bet.
Today: Total: [Share]
] [