zlacker

>>davidb+(OP)
Can't say I saw this coming. This is deeply sad to me. OpenAI did so much life changing work so quickly. It has totally changed my life in terms of enabling an absolutely unprecedented increase in my own personal productivity and how ambitious I can be with my projects. And even though I know Sam didn't personally code the key things, I believe that it never would have happened the way it did without his instincts and talents. And I fear that without him at the helm, all the magic is going to quickly dissipate and OpenAI will be just another lumbering tech company without a rudder. The default state of things is stasis and dysfunction. Just look at how Google can't do anything anymore, and how bad Siri and Alexa are-- all despite having so much manpower, money, and market share at their disposal.

I also find it maddening how boards of directors rush to insulate themselves from any possible issue and are so quick to throw overboard the very people who enabled the success that they get to participate in. I'm thinking particularly of Travis at Uber and how he was thrown out of the thing that he built from scratch, which never would have worked without his extreme efforts. If I were on the OpenAI board, the bar for firing Sam would be so ridiculously high that he would have to have done something so outrageous, so illegal, etc., that I struggle to believe what he actually did could even remotely approach that standard.

>>eigenv+Ur
I cannot even begin to understand what makes you think that this technology arose from Sam altman and not from all the other people working there. By saying you doubt they can do anything without him, you're putting one person on a pedestal and giving them all the credit for this. This is the same phenomenon has happens with Elon musk getting all the credit for his tech companies.

>>window+ru
It's not just raw technology. It's a vision for what the product should be, what overall strategy to take, how to fund it, how to introduce it to the world, how to scale it, what order to do things in, what approach to take with 3rd party developers, when things are good enough to launch, who to hire, who to promote, etc. There are a million little decisions that go into a runaway success like this. And a million opportunities to make the slightly sub-optimal or wrong decision. And it doesn't take many of those to kill everything that made the place special, and that's actually my new base case for OpenAI-- that's the base case for any company/organization. The default state is chaos and entropy, and it's a miracle when you can escape that fate for even a few years of hypergrowth.

>>eigenv+7w
> It's a vision for what the product should be

Have you actually used it?

It's the most basic possible website, API, and app for a language model AI. It's barely functional. For example, the iOS app couldn't even display images until the other day.

> There are a million little decisions that go into a runaway success like this.

I agree that GPT-4 is amazing and probably the best. But there are several other competing implementations of language model AIs right now, some are even developed as open source.