sure, not confirmed, implied, but the premise that this is a disaster because it’s outsized success is costing them, is just short sighted

https://news.microsoft.com/2019/07/22/openai-forms-exclusive-computing-partnership-with-microsoft-to-build-new-azure-ai-supercomputing-technologies/

https://www.theinformation.com/articles/why-openai-spent-barely-a-dime-on-microsofts-cloud-after-1-billion-deal

https://www.theregister.com/AMP/2022/10/21/microsoft_openai_investment/

They’ve already sold it, it will be integrated into Bing next year, and while this cost may be a lot for a small startup, it’s a drop in the bucket for the company paying for it, Microsoft

Umm, the optimal outcome was a viral hit / free marketing, which would lead to an excited user base who would then pay for their product.

It makes me wonder if we need hooks between each of the LLM’s, wondering if they will need to communicate with each other? Will they be so big that they’re all encompassing, or more specific and finite and need to be Daisey chained?

It feels like LLM’s have been a big deal, but only in certain circles. The image synthesis models opened the general idea of AI as an important tool to a much broader audience, who retroactively found GPT-3.

It’s unclear if OpenAI was always going to release ChatGPT or if it was in some ways built as an easier access point than the playground, for a growing community of people engaging with their products.

Whatever the case may be, the timing is pretty good, because if GPT-4 is a decent leap forward, you have developers who have been building on top of GPT-3 for years now (some who have become sizable businesses in their own right), a bunch of use cases in the world and a growing community that is understanding future use-cases — all which will allow GPT-4 to potentially seriously break into the mainstream, not as a name brand per se but as a tool that impacts a much larger part of society.