Chinese company DeepSeek announced its new R1 model on January 20. They released a paper on how R1 was trained on January 22. Over the weekend, the DeepSeek app became the number-one free download …
Aside from stitching together a bigger and bigger trenchcoat, has open AI done anything else? Just goes to show how vacuous LLMs are if this is what it takes to catch up/outpace them.
Working on this take:
Aside from stitching together a bigger and bigger trenchcoat, has open AI done anything else? Just goes to show how vacuous LLMs are if this is what it takes to catch up/outpace them.