00:00When is this going to pay off? It's not just with CoreWeave. It's not just with Microsoft. It's not just with meta platforms.
00:07This is the existential question that every investor is asking themselves right now. When will it pay off?
00:11It's an existential question. However, there's a lot of long-range planning occurring at present.
00:18The commitments that OpenAI and these ecosystem partners are making are stretched out over a very long-term horizon.
00:26And as Dina mentioned, I thought she summarized it well, nobody can get what they need yet.
00:32So it's going to be an ongoing process of seeing allocations, seeing revenues, seeing costs drop, seeing funding, seeing improvement across the ecosystem, which will reinforce investor confidence in this.
00:46I think there's a strong belief that there is ROI, but investors are paid to evaluate risks as well as reward.
00:53And sometimes that pendulum shifts really far in one direction.
00:57And say, in the case of an Oracle who, similar to CoreWeave, is supplying a lot of this capacity, that stock had gotten back on Thursday or Friday to close to where it had been before the big announcement, the big announcement with OpenAI.
01:10And I think investors are viewing what's good as bad until they get some answers to these questions.
01:15That said, I think there's more good to come and we'll get milestones along the way proving out the thesis of the ecosystem and the returns.
01:25That thesis of the ecosystem is really something that I'm having trouble visualizing right now.
01:30And no matter how much Mark Zuckerberg tries to explain superintelligence to me and to shareholders of the company,
01:37I still don't understand what that ultimate payoff looks like, not just for meta platforms, but for the industry in general.
01:43Is this something that will only pay off when there's mass unemployment so companies don't have to actually pay for people because machines and AI are doing the work?
01:54Is it an increase in productivity?
01:56We all keep our jobs, but we have these little friends, these little helpers who help us do it better.
02:01What is it?
02:02Nobody knows.
02:03I think it's a combination of increased productivity, increased revenue, increased velocity of activity.
02:09At a minimum, AI should really supercharge the individual.
02:14Today, the adoption has been very methodical among typical commercial organizations because they have to worry about security and orchestration and liability and all sorts of things.
02:24And that's going to be the case.
02:25But we are seeing the beginnings of real progress.
02:28We're seeing the technologies in the marketplace start to mature.
02:31We're seeing the very beginnings of commercial deployments and production beyond what Palantir is doing.
02:40And so I think we're seeing the beginnings of it.
02:42But as we know, technology trends are overestimated in the short term and underestimated in the long term.
02:47And things move very slowly until they start to move quickly.
02:50And that's just a phenomenon of tech.
02:52But are you using it at all with your job?
02:53Sure.
02:54How?
02:54Yeah, we're using ChatGPT Enterprise.
02:58Our developers are using Cogen tools.
03:00But those things are easier because the finished product doesn't have to be correct.
03:04I can see what works for me and what doesn't.
03:06But it's a very different phenomenon when you have something customer-facing or broadly employee-facing where it has to work out of the box.
03:13And it can't bring your organization down in terms of liability.
03:16So it's going to be a process for sure.
03:17Yeah, I was talking about this last week with somebody on our program.
03:20And they said they're basically like an intern, these LLMs.
03:25You know, eager to help out.
03:27But you really have to check the work to make sure that it's something that can actually go to print.
03:32That's something that is actually accurate.
03:35Finally, I just want to talk about the overall economic effects of this.
03:38A lot of what we talk about when it comes to AI, the real beneficiaries have been the major companies that we talk about every day on Bloomberg Tech.
03:47But when will we start to see these advancements actually affect the bottom lines and the productivity of companies that are not necessarily tech-adjacent?
03:57So such a great question.
03:59And it's funny because people are attributing layoffs to AI.
04:02And I think in part, some of those layoffs may come from the need to fund AI, but not directly from the productivity benefits from AI yet.
04:10So in other words, companies spending money on AI rather than people, and that's why they're making investments there rather than in human capital?
04:17Well, they're spending on AI, but there are other things behind the layoffs, meaning people that overhired during the pandemic, especially in the realm of technology.
04:26Not all those hires were of the quality that they wanted.
04:29They had to pay a lot for those folks.
04:30And also, companies have been bracing for an economic downturn.
04:35My industry contacts tell me everybody's bracing for a downturn, concern over tariffs, inflation, etc.
04:39So I don't think it's directly attributable.
04:41But to answer your question directly, I think next year we start to see, we're starting to see production workloads going to production today.
04:50I think companies will have better worked out the kinks by the second half of next year.
04:55As I mentioned, the technologies are maturing, the costs are plummeting, the ROIs will rise.
05:00And I think into second half of next year are really going to be the first signposts of broader benefits from AI to the average commercial organization.
Be the first to comment