So here we are. At a time when every 2nd software startup's core value is owed to and owned by OpenAI.
Sam isn't going back and 70% of the employees at OpenAI including the dude who orchestrated the whole thing is planning to quit. The CEO of MS has basically given them all an open invitation to come right on over. The coffee's fresh and there's apple pie in the oven.
We always talked about how odd it was for YC was to consider, interview, and even invite these wrapper startups to their accelerator. Despite this, we knew that it was likely a safe enough bet to march on. Somehow, someway, this got all fucked up. Founders of these companies should be experiencing an existential crisis right abouuuuuut . . . now.
Something tells me that demo day is going to be a little less exciting when the core of pretty much every product is known to be backed by pissed off investors, incompetent board members, and employees who are all hours or days away from quitting (if they haven't already).
I've never seen such an up and coming company torn to shreds over the course of a weekend. It'd be impressive it wasn't so idiotic.
For those who wish to see the silver lining, realize this. For something like this to happen is further proof that none of us have any clue what we're doing. That includes leadership of one of the most important companies to have ever existed. If there was ever a time for you to doubt yourself a little less, it's today. You'll never manage yourself as poorly as the Board of OpenAI.
Thanks and have a great evening!
I beg to differ. The startups that use GPT 4 API for their software can readily switch to PALM 2, Claude 2 or the other competitive Language models.
The startups add a lot of proprietary code to create value out of base language model. They also learn many things about the user behaviour. These things don't change even if OpenAI shuts down completely.
Furthermore, GPT 4 is going nowhere - before everyone leaves OpenAI, it will surely be merged with Msft in the worst case scenario. So there's that.
Also, even if GPT 4 gets shut down, Msft has such clauses in their contract that they can get the weights and run it on Azure.
So all in all, there is no concern for such startups.
Yep, openAI isn't years ahead.
If they go south, by the time the decide to shutdown, Google, Microsoft or Facebook will already be ahead by quite a margin.
Opensource isn't also years behing either, as soon as OpenAI stops moving they'll lose all of their first movers advantage.
It's not like they can take GPT 4 with them, but if you have the staff that understands the model and training process, you're really not that far away (if you have the money to pay for training, like MSFT easily has).
Not only that, I'm willing to bet MS already has an inhouse unreleased LLM already developed in some quantitative capacity.