Sign inJoin free

Watch / Tech startups

The Open AI fallout

21 Nov 202311:40

The Open AI fallout

General information only. ausbiz videos are general information only, produced by Ausbiz TV Pty Ltd, a media business that doesn’t give financial advice. They don’t take into account your objectives, financial situation or needs. Consider whether it’s right for you and read the PDS and TMD before acting on anything you watch.

About this video

Michael Frazis from Frazis Capital Partners delves into the recent corporate events from Open AI. He says the firing of the company's tenacious CEO, Sam Altman, raised suspicions of misconduct and triggered a cascade of events leading to the unpredictable future of the AI company. He notes that most of the board members have remained silent, raising questions regarding the company's direction and the reasons behind the disruptive changes. Michael is intrigued by the appointment of a new Open AI CEO wanting to decelerate AI development, contrary to the company's progressive ethos. He attributes this development to a divide within the board between those eager to accelerate AI's potential benefits and others calling for caution, fearing existential risks associated with AI. He suggests that this internal conflict may have been a significant factor in the company's recent upheaval.

For the tech sector, Michael predicts that if Sam Altman and his team join Microsoft, it could result in huge investments and benefit suppliers like NVIDIA. Conversely, if the team forms a new company, the market could see heightened demand for GPUs from NVIDIA and AMD. Despite the ongoing speculation and uncertainty, Michael anticipates structural shifts and new alliances that could significantly impact AI's future trajectory and the broader tech industry.

Below is an unedited transcript of the interview:

They used to say. We're joined by Michael Frazis from Frazis Capital Partners. Thanks so much for coming in. No, thanks for having me on. Yeah. Now you have been tweeting furiously watching this from from literally Sunday. Gosh, it feels like an eternity go. But it was only Sunday, wasn't it? Well, I guess it was the end of last week in the US time. Yeah. Where it all happened. And it's just it's incredible to see a company like that go through such turmoil.

0:25

And it's just the story has been changing so fast. So the first thing we found out was Sam Altman was fired. So immediately everyone thought misconduct. Like, how on earth could you fire such a brilliant CEO that brought AI to the masses and created an $80 billion company in less than a year? Really?

0:43

And more than that, if anyone was ever going to join the ranks of Apple, Facebook, Google, Microsoft was likely to be OpenAI. I mean, why would you fire that person for anything other than severe misconduct?

0:55

So that's kind of what we all thought. Then then is number two, who is chairman of the board, who wasn't apparently in the meetings when he was fired, even though he was chairman, he supported him and resigned. And ever since then, this the fate of this company has been up in the air. You know, it's been a case of will they come back? They went they went into the headquarters and try and negotiate a return. It looks like the OpenAI board kind of refused that and has been looking for other options. Saatchi. Nadella announced that he's offered them to join Microsoft, which would be incredibly compelling. And meanwhile, you've got 700 employees who this is all kind of as much of a shock to them as anybody else. And what they decide to do is also quite critical here. So at the moment they're all supporting.

1:40

They're all supporting Sam Altman and Greg Brockman, the two erstwhile leaders of OpenAI. But it's just not clear what happens. I mean, this is such an important, consequential company, and it looks like the key board members that made this decision were very kind of I mean, two of them I don't think anybody ever heard of. One was a founder of Quora, a question asking site, and was building a chat, GPT

2:08

competitor, or certainly something that the latest iteration of GPT that Sam Altman last week announced would have effectively kind of removed. The reason for that company being so trying to think, wow, is this guy, this guy was on the board of OpenAI, and he was setting up effectively a competitor to OpenAI, his latest suite of products. And it looks like that might be part of this coup as well. But one thing we have had is radio silence from most of the board members that made this decision. So we don't really know what was behind it, but it can be certain this will be one of the most studied corporate events in history, and it's still in the air. We still don't know what's going to happen. Yeah, that's right. We still don't know. There was speculation. They said it was a communication issue between Altman and the board, because Altman apparently was fishing around the Middle East for $100 billion to start producing chips to compete with Nvidia. Whether or not there's any substance to that, we

3:08

don't know. But obviously then that changes the paradigm again, because we did know that Microsoft is looking also to move into AI chips to compete with Nvidia. So with Nvidia's results coming out, does any of this sort of that's going on around there make any difference, or is it purely going to be people looking at the earnings? Look, I think the outcome will make a difference. The fact that Sam Altman was exploring AI chips and an AI phone or I kind of hardware, that was kind of well, I think that's broadly assumed to be true. That is running around the Middle East trying to raise funds. Also likely true whether that was behind the board coup. We still don't know. I mean, at the start, there's kind of been this argument between the e act like effective acceleration and the D cells or the deceleration ists who align themselves with the effective altruists. So just explain that for viewers what that actually means, because people might

4:07

appreciate the subtleties. It's almost like the people on with the moral compass versus just letting capitalism rip when it comes to generative AI and deep learning. Yeah, but the curious thing is. So this actually started a long time ago. It actually happened when I was at Oxford. So I knew some of these people that did it. They came up with this idea of effective altruism. And the idea is you don't need to do good now. And I'm paraphrasing, we don't need to. Good now. Do now if you're going to do good later. So go into finance, earn lots of money, donate it after. If I'm going to donate, why don't we buy mosquito nets, which saves this many lives instead of donating to this other worthy but not as cost effective cause now it all sounds very logical, and it certainly kind of it's very attractive to a certain kind of mind, a way of thinking about the world. But, you know, kind of ten plus years later, some horrible things have happened. So, for example, Sam Bankman-Fried was an effective altruist, and that was his whole guiding philosophy and most. What he did, you could justify using effective altruism philosophy. I mean, even when he's committing fraud, you could be like, well, you're committing fraud, but you're also doing lots of

5:07

good, so it makes sense to do that. It really is that kind of mechanical. Weighing up a bit of Robin Hood, aren't we? Yeah, but I'll get to the story. So part of that was good. So for example, I had a charity that was called GiveWell, which is like, you know what, instead of giving to charity, why don't we just give random people in Africa $1,000, you know, which actually has a kind of it's kind of interesting. It's kind of interesting way of looking at the problem. But where that movement ended up is they got very obsessed with existential risks to humanity. You know, the worst thing that could happen is that asteroid wipes us out. Why we should be investing a huge amount of money in stopping asteroids hit, which is reasonable. But then when I came along, it was kind of just the perfect, you know, philosophical background

5:50

for people to think, well, there's actually a risk that know, destroys humanity. Maybe it's just going to this idea that, you know, somebody asked it to make paperclips and it converts the entire world economy into a paperclip machine, you know, just things like that. And people took that extremely seriously.

6:06

And a lot of people think we should decelerate the progress of AI significantly. So the current CEO of OpenAI was hired, one of the rounds on the weekend, I think it was the third CEO in two days. He he thinks and has tweeted that he thinks the pace of AI should go from ten down to 1 or 2. So the board has literally put somebody in charge of OpenAI, the leader of AI, the inventor of the consumer products of AI. They've put that company in the hands of somebody who wants to slow it down by 80 to 90%.

6:41

It's truly bizarre. But this is this debate between.

6:45

So I talk about effective altruists, the sell movements, the decelerate. And then you have the or effective acceleration, which is kind of a play on E as well. So you've got some people that think that we should move as fast as possible because the rewards are just immense. You know, you've got the world's best tutor for every child, almost for free. You've got translation, you've got all the world's knowledge accessible, you've got new insights, you've got new creativity, you've got all these positive things that are coming. Let's go as fast as we can into that world. And then you've got this E and also D cell movement that's trying to slow it down. And as it's so happens, it looks like on the board of OpenAI, fastest moving company and most revolutionary AI company in the world, you actually had a board full of these D cells. And so there was and it's still not clear. We don't know that. We don't know the backstory. We don't know exactly what happened or why. But it does seem, judging by the appointment of the CEO and the resistance to moves to

7:44

reinstate Sam Altman and Greg Brockman, it does look like that was a big part of it.

7:50

So it's interesting. Yeah, it's one of the most fascinating stories, totally and very big for the history of humanity, I suspect, in terms of which way it goes. What does it mean for investors looking at Microsoft and the here and the now and Nvidia obviously. Yet they're reporting because I mean it seems to me and you've often tweeted this about America trying to stop Nvidia selling their chips to China. I mean open AI they can have the D sell people. But ultimately the technology will probably gravitate around to the group that's going to push it forward, which is basically what Nadella was trying to do. So I guess the question would be at the moment, there's probably is there going to be a lot of volatility? It's too early to tell, but ultimately we're still going to have winners in Microsoft and Nvidia out of this. Definitely. I mean it's not clear what happens, but there's some scenarios where it is kind of clear what happens. So if

8:50

Sam Altman and Greg Brockman and most of the OpenAI team joins Microsoft, there's going to be huge investment like Microsoft will swing their entire resources behind them. So you're talking about

9:02

a new chip. So Microsoft has got their own AI chip in progress. I imagine they'll allow Sam Altman Brick Brockman to lead their own initiatives in that area. That idea about wearable AI, they'll just have all the, you know, the next generation of models. They'll have all of Microsoft's immense resources, which is actually a far better outcome for Nvidia, for suppliers and probably for all of us than had they got that money from the Middle East. Yes, to put it simply. But that another alternative that was looking very likely over the weekend was that Sam Altman and the OpenAI staff would form their own company, which would take them a bit more time because I'd have to train all their models from scratch. They wouldn't have access to OpenAI's technology. If that happened. You'd have at least one massive customer

9:49

for Nvidia GPUs, and probably AMD as well. That would also be very good for demand.

9:55

Actually, the Microsoft one is probably. I kind of actually hope that one happens. Because the the linkage between open AI and Microsoft is already so tight. If the team goes to Microsoft, they won't have to skip a beat. Microsoft has access to the weights, they have access to all the IP and literally a minor delay. And then onwards with a proper balance sheet with better funding than ever, a proper governance, proper board. That's probably the ultimate outcome. But there's other there's other things that might happen. For example, apparently the board wants to merge with anthropic, which was a spin off from OpenAI. And if that happened, then it's not clear because then how does where does Microsoft fit in with that? Does Microsoft Anthropic and OpenAI start competing? And then a lot of these things, a lot of the contracts, all of these things we're not actually fully aware of, it's very hard to tell. Yeah. If by way, the Microsoft potentially gets Altman and Brockman and all the employees of OpenAI. So the negotiations are okay. You know, if you don't want to work at

10:55

OpenAI, we'll take you. And I mean, there certainly have the capacity probably easily to do that. Other regulatory issues that could crop up

11:05

with Microsoft, OpenAI, I wouldn't think so. I mean,

11:09

Sam Altman was copying a lot of hate in the in the world for

11:14

going to Washington and talking about all these fears and ostensibly using that to make it much harder for startups. It's kind of like regulatory capture, which would allow him to do it. You'd think OpenAI and Microsoft would have. They would be on the right side of that regulation thing. I don't think the American government would be attacking them the same way they would if they were independently going out.

11:36

It just seems much less scary, I think, to everybody, if that's the final outcome.

More like this

We value your privacy

We use essential cookies to make the site work. Non-essential cookies help us improve the experience. You can accept or reject non-essential cookies. Privacy policy