I'm reminded a bit of the first couple of decades of Amazon, when they were continually jeered at for being in the red, then a small profit.
This was all during the dotcom boom. Somebody was going to win, and it was probably the one who spent the most money to be the favorite. Tons of money was set on fire, and Jeff Bezos is pushing being a trillionaire.
In a sane world we'd back off and trust that the tech will get invented even if take two months instead of two days. But they believe that there is one winner and everyone else gets nothing.
Ok, so all this tells me is that there is a point to creating llm models, or at least looking like you might - to take a slice of this crazy money being put on the burning pile. Because that's all I see here.
> Project Stargate is financial fiction masquerading as infrastructure. The pitch: $500 billion total, $100 billion right now, trust us on the rest. OpenAI threw in $19 billion of Monopoly money for 40% ownership, which is like buying Manhattan with Pokémon cards.
I’m going to use this analogy somewhere
The math is modern art. OpenAI gets exclusive access[1] to $100 billion worth of data centers for $19 billion in private stock that's worth precisely nothing until an IPO that may never come.
7 comments
[ 2.2 ms ] story [ 35.4 ms ] threadOr use your browser's reading/focus mode.
This was all during the dotcom boom. Somebody was going to win, and it was probably the one who spent the most money to be the favorite. Tons of money was set on fire, and Jeff Bezos is pushing being a trillionaire.
In a sane world we'd back off and trust that the tech will get invented even if take two months instead of two days. But they believe that there is one winner and everyone else gets nothing.
I’m going to use this analogy somewhere
The math is modern art. OpenAI gets exclusive access[1] to $100 billion worth of data centers for $19 billion in private stock that's worth precisely nothing until an IPO that may never come.
This is incredible if true