Very interesting developments. I was suprised by Micrsofts investment initially. But your analysis makes a lot of sense. Microsoft investing shows they are betting on more than one horse.
Do you think in the next few years this will turn out to be a winner-takes-all or a few-takes-all market?
For LLMs, I am now assuming few-takes-all for general purpose with a multitude of niche players that differentiate based on specialized features, personas, or industries. How about you? Any forecast?
I agree. The foundation model market seems to become crowded, with Inflection now entering (and DeepMind on its way); I wonder if the Antphropics and OpenAIs in this world (newscool) can win it in the long run from the big cloud providers (oldscool). Maybe there's room for both. Below that there's much more room to play with.
I'm assuming OpenAI sustains a market share lead for the general purpose LLM with Google PaLM (then Gemini) playing a runner-up role in the proprietary space. It is likely an open-source option will rise to the top of that category and become the primary OpenAI (i.e. MSFT) and Google alternative. Everything else will specialize either by choice or necessity. AWS will become the cloud provider-of-choice for the "everything else" category. Of course, something else could emerge, whether it is Inflection, Anthropic, Cohere, or something else and rise to the top of the proprietary stack through superior economics or performance. However, this doesn't seem likely.
Very interesting developments. I was suprised by Micrsofts investment initially. But your analysis makes a lot of sense. Microsoft investing shows they are betting on more than one horse.
Do you think in the next few years this will turn out to be a winner-takes-all or a few-takes-all market?
For LLMs, I am now assuming few-takes-all for general purpose with a multitude of niche players that differentiate based on specialized features, personas, or industries. How about you? Any forecast?
I agree. The foundation model market seems to become crowded, with Inflection now entering (and DeepMind on its way); I wonder if the Antphropics and OpenAIs in this world (newscool) can win it in the long run from the big cloud providers (oldscool). Maybe there's room for both. Below that there's much more room to play with.
I'm assuming OpenAI sustains a market share lead for the general purpose LLM with Google PaLM (then Gemini) playing a runner-up role in the proprietary space. It is likely an open-source option will rise to the top of that category and become the primary OpenAI (i.e. MSFT) and Google alternative. Everything else will specialize either by choice or necessity. AWS will become the cloud provider-of-choice for the "everything else" category. Of course, something else could emerge, whether it is Inflection, Anthropic, Cohere, or something else and rise to the top of the proprietary stack through superior economics or performance. However, this doesn't seem likely.