Thomas Renkert🦞 on Nostr: Before the "we have no moat" memo from #google appeared almost a year ago, it seemed ...
Before the "we have no moat" memo from #google appeared almost a year ago, it seemed like it was only a matter of time until google's #LLM would overtake #OpenAI. The memo made it clear that perhaps that wasn't happening just yet - but it also suggested that #opensource models would make it very hard for any corp to form a platform monopoly on the best #AI.
Interestingly, it seems like #microsoft is doing just that anyway. Not only with their VC money in OpenAI, but also with their partnership with #Mistral which was announced yesterday.
This effectively means that all the most capable models (especially for non-english EU languages) are being offered by #MS.
The next commercial LLMs (often below the threshold of actual usability) are #Anthropic's Claude and Google's #Gemini. And all the hopes for truly open models rest on #Meta and #Llama3
Ironically, #Facebook and Google (Gemma) are now opensourcing AI to at least sabotage Microsoft's market dominance a tiny bit.
#MistralAI
Interestingly, it seems like #microsoft is doing just that anyway. Not only with their VC money in OpenAI, but also with their partnership with #Mistral which was announced yesterday.
This effectively means that all the most capable models (especially for non-english EU languages) are being offered by #MS.
The next commercial LLMs (often below the threshold of actual usability) are #Anthropic's Claude and Google's #Gemini. And all the hopes for truly open models rest on #Meta and #Llama3
Ironically, #Facebook and Google (Gemma) are now opensourcing AI to at least sabotage Microsoft's market dominance a tiny bit.
#MistralAI