OpenAI could be propped up by MS for a while even if it doesn't make a profit, to keep the brand name alive, for the smaller AI startups I guess it will be the same story as with crypto? VC investors put some money in, hype up retail investors to also invest, cash out/exit, then the startups slowly close because they can't turn a profit leaving the smaller investors with most of the losses
edit: also lol at making public bets on this, got nothing better to do?
The Ed Zitron article is worth reading, it talks a lot about what Microsoft is getting out of the deal with OpenAI in exchange for some Azure credits (fake money)
will do, yeah I heard MS paid OpenAI mostly in Azure credits lol, nothing like paying yourself
I've read somewhere that even after that openAI is paying Microsoft 5B $ for azure
Excellent article.
The stuff about their deal with Microsoft is crazy. MS gets 75% of the profit until they recoup $10B, and free access to any technology OpenAI has or will make until they achieve AGI. So anyone investing in OpenAI is basically only betting that they achieve AGI before they go bankrupt.
That also means Microsoft has little reason to buy OpenAI. Based on the deal, it seems MS is very skeptical of their chances at achieving AGI. It's not a situation where Altman can bluff them into thinking they're making progress, MS has access to see what's really going on.
I think they’d just do a like indirect acquisition or whatever they call it when they hire all the engineers and buy the IP. Wait for OpenAI to burn too much cash, hire their best, and get all the rights to the IP then just laugh
Absolutely. Whoever negotiated the deal, on the Microsoft side was an MVP.
The bit about determining the valuation of $1 in Azure credits is the best part, to my mind. Because how do they calculate that? I have to imagine that it's not $1 of cooling, power, and compute that Microsoft paid for their datacenter. Because why would they provide that at cost to OpenAI?
It's exactly the opposite! Prediction/betting markets are possibly the most/only reliable way to forecast things with otherwise limited information. Prediction Markets are extremely accurate since being wrong makes people poor and they use every trick in the book to not be poor.
Huffing farts over here.
I played with prediction markets for fun for a while. They're completely headline-driven and hype focused. I made a steady profit by betting against every "will X happen by Y deadline" available. Thus, while OpenAI has no business model and is a doomed company, I would not take that guy's bet.
You're telling me that betting markets aren't rational???
I had a lot of trouble tying my shoes this morning
https://www.wheresyoured.at/to-serve-altman/
This was a little annoying. Idk if the article is in the thread linked by OP, but I couldn't find it. The link just goes to Twitter, and if you don't have Twitter (never getting one either), you can't see replies or what people are replying to.
All this to say LINK THE ARTICLE IF ITS RELEVANT GOD DAMMIT IM NOT GETTING A FUCKING TWITTER ACCOUNT
Ouch. Sorry. I subscribe to Zitron's newsletter so I had read that far before I stumbled upon the HN discussion about it.
HNs site layout also sucks ass so it's hard to navigate, to get to the Zitron article. I linked to a particular thread of discussion so it's the "On: How Does OpenAI survive" link
