zlacker

Because you can't train and operate models at scale with ideology. Costs money and compute. Without funding and compute, OpenAI is a bunch of smart folks with laptops talking AI safety in an office until the cashflow dries up.

People who want to get paid or are loyal to Sam will follow Sam. Microsoft is clearly behind Sam. "Whoever holds the gold makes the rules."

EDIT: gkoberger said it far better, and captures the value beyond the physical and fiat components of the power struggle: >>38337782

"I don't think he's on his own. All those people he helped in the past have come out to support him. Almost his entire company (minus the board) has publicly supported him. His team is currently hosting him at the office with hopes to reinstate him. The CEO who temporarily replaced him has publicly and privately supported his return. OpenAI's investors are backing him. His cofounder quit in solidarity.

Turns out the secret to getting through something like this is to simply have spent the past 15+ years genuinely helping everyone you've ever come across."

>>toomuc+(OP)
OpenAI has the most valuable new technology product in a generation and it’s already monetized. It’s also a charity and could give a fuck about growth beyond the extent needed to meet their goals.

The idea that Microsoft is somehow the one irreplaceable part of this formula is completely fucking bizzare.

>>CPLX+v
OpenAI is not a "thing", the workers are a thing. OpenAI is nothing if the engineers/scientists leave the company to work somewhere else if Altman leaves. According to reports, the head of research and other senior staff have said that they will resign if Altman is not restored as CEO.

>>CPLX+v
Well, now we know that once any putative superintelligence learns how to wave money around, we're doomed. All the preparatory "alignment" isn't going to save us from our own greed.

>>CPLX+v
They have the GPUs and there's a GPU shortage already. What do you think is going to happen when they lose that access in $10bn compute? Given the shortage and other vendors being mostly locked up already, there's not much OpenAI can do. And that's not considering that the board is facing immense pressure internally starting from the very temporary CEO that they put in place two days ago to various members of senior leadership threatening to resign (with many employees being paid in options dependent on valuations that the board just destroyed). The board may have de jure power but that is nothing without the de facto power associated with it.

>>notabe+r2
I mean, obviously? Any distant, hypothetical, unlikely AI apocalypse is going to need a lot of willing human accomplices.

>>linkgo+j2
Yes that’s somewhat correct. If he really has the loyalty of a huge percentage of the staff and they leave to follow him you’re right.

But you might want to question why you’re so sure that’s the case. People talk about that a lot but it’s pretty rare to actually see that happen.

The usual way things like this go is most people go huh how about that and go back to work at their job the next day and think about their family and weekend plans.

In his case it’s especially odd given his oddly thin track record of actual success. He seems incredibly good at ingratiating himself with extremely powerful people, but there’s no evidence at all of him being a popular leader or particularly good at building and executing on anything.

Ask yourself how he got his last two jobs. He wasn’t exactly voted in due to widespread popular support.

In both cases however he departed abruptly amidst what looks like interpersonal conflict.

>>toomuc+(OP)
> Because you can't train and operate models at scale with ideology. Costs money and compute. Without funding and compute, OpenAI is a bunch of smart folks with laptops talking AI safety in an office until the cashflow dries up.

Without its exclusive relationship to Microsoft, there are other people with access to money and compute that would like a leg up in the enterprise AI space that Microsoft would be taking a major, even if (optimistically) temporary, step back from its leading position in by terminating its relationship with OpenAI.

>>trajec+X2
Microsoft is going to walk away from the most strategic deal they’ve made in a decade to save the Loopt guy’s ego?

I feel like everyone should be forced to watch a few Game of Thrones episodes before commenting. Organizations act in their own self interest. Leaders are convinced they’re irreplaceable all the way to the gallows steps.

If Sam can sabotage OpenAI or if the remaining people can’t capably run the organization then sure people will walk away. But if they keep things from falling apart nobody is going to give a fuck.

>>CPLX+y3
> But you might want to question why you’re so sure that’s the case. People talk about that a lot but it’s pretty rare to actually see that happen.

Difference here is OpenAI is not Google or FB with tens of thousands of employees bulk of which are engineers most likely not be able to get current level of compensation in case of mass layoff/resignations. OpenAI is couple of hundred people and are highly rated due to perceived level of success. So it is opposite of risk to follow the leader who has gotten them high level of compensation so far.

> He seems incredibly good at ingratiating himself with extremely powerful people

You seem to think it is negative. For most people including OpenAI employees it is great advantage to be working for/with him.

>>pjc50+p3
An apocalypse needs power. That power can be offering money to humans, but doesn't have to be: other (thankfully not yet existent) possibilities include hacking a single clanking replicator, as would offering rewards to unrelated non-human uplift project subjects.