00:00Explain RPI, besides they minted very good hockey players, up to Hudson, Troy, 10 below zero, and you minted engineers
00:11at Rensselaer Polytechnic.
00:12Yeah, I think it's one of our oldest engineering schools in the country.
00:16The oldest.
00:17It was a great place for me to go to school because it allowed me to do a blend of
00:23strong science and engineering learning, but then application.
00:26I was able to get into the lab, play with things, build.
00:29I was a radio nerd myself at our campus radio station, so it was a really fantastic place to learn
00:36and to do applied work as well.
00:38Bring that over to the silliness, the hype you see in AI now.
00:44What's the thing in the zeitgeist that drives you nuts, Jonathan Curtis?
00:48Well, yeah, so I started my career as an engineer.
00:50The first 12 years of my career, I spent building things, leading software development teams, getting my hands dirty with
00:56software.
00:57And then in 2000, I moved out to the Bay Area to get rich, building internet routers.
01:02That didn't quite work out, and I transitioned into finance.
01:05But I've always stayed close to technology as either a technology sell-side analyst,
01:11and then eventually made my way over to the buy-side at Franklin.
01:14And what is so profound here, and an advantage that I think I have in this AI journey that we're
01:21on,
01:22is that I am tracking right along with what I think the leading lab companies are seeing and doing with
01:29the tools.
01:29I stopped coding in 1995, and I hadn't returned to it until January.
01:36I came back to it because I found that these tools enabled me to build again.
01:43I am able to literally replicate and augment huge portions of our investment research process with these models.
01:52I literally talk to them like I talk to an analyst or an associate.
01:56It builds me the processes and the tools that I want and ultimately allows me to be a developer again
02:02and has dramatically improved my productivity.
02:04So I think what investors are missing is that the productivity that we're seeing in software development
02:12is going to start finding its way into the financial services sector, into the healthcare sector,
02:18because every knowledge worker is empowered to build again.
02:22And I think that's the insight that I bring to the table at this time.
02:26At Franklin Templeton, I don't know if there's a house view,
02:29but how do you guys view this AI from an investment perspective?
02:33I think the very early stages, when we all became aware of it,
02:36when NVIDIA put up that great print four years ago, whenever it was,
02:40we're like, oh, AI is a real thing.
02:42Now, how do I play it?
02:44How do I play it now?
02:46Well, there's definitely no house view.
02:47Franklin's a big place.
02:49We offer our clients all sorts of products and offerings.
02:53I come from more of an innovation and technology lens.
02:57But certainly our view on the technology side of things is that this is just getting going.
03:03If you look at the fund that we manage, we sort of sector it, our bucket into three big buckets.
03:08Build AI, run AI, use AI.
03:11The tech fund that we operate is very oriented currently to the build AI and the run AI companies,
03:18the semiconductor companies, the compute infrastructure companies, the hyperscalers.
03:22But increasingly, we are seeing companies getting real productivity gains on the run AI side of things.
03:29Franklin's a great example of that.
03:31We're seeing it in the tech companies.
03:32They're able to get dramatically more output per engineer because they're using these tools.
03:38So we are seeing real leverage coming out of these models.
03:42And output for engineer.
03:42I had ducky bumps when Reid Weissman of RPI touched down with Artemis.
03:48I mean, out of RPI, the lead astronaut on Artemis 2.
03:52Looking now at the space station and NASA here, it never goes boring back to Ed White of the few
03:58years ago.
03:59It is a spacewalk of two of our astronauts.
04:04And what's interesting is their global sense as well.
04:07Bringing it up here right now as I can.
04:10I'm sorry.
04:11There it is.
04:12She's from Caribou, Maine.
04:14Jessica Meir.
04:15Any work patient for some extra maneuverability?
04:21Spacewalk.
04:22Never gets old.
04:23Nope.
04:24For those of you on Bloomberg Radio, it's a bit of a convoluted.
04:27The lens is right up on their equipment.
04:28But two of our astronauts doing a spacewalk right now.
04:32Jessica Meir is experienced at this out of Sweden in the United States.
04:37And Anil Menon, also a colonel in the United States Space Force,
04:41is the new division that we have.
04:44And they are out there doing it right now.
04:47AI is a big spacewalk.
04:48I noticed, Jonathan Curtis, that Google had four guys walk out the door yesterday.
04:54Yeah.
04:54You've walked out the door, haven't you?
04:57Yes, I've walked out the door before.
04:59These guys walking out is a big deal because Jeff Dean has been such a critical contributor
05:07to that firm over the years.
05:09But they didn't really walk out the door.
05:12I know.
05:12They're doing a sidecar deal.
05:13I read it.
05:14That's exactly right.
05:15And so I think what Google would, I don't know the particulars here,
05:18but Google's keeping one of their best folks very close by.
05:23Right.
05:23I think what they're really doing is opening up another option around scientific discovery and the like.
05:27Just because of time, I've got to get this in.
05:29I mean, Jonathan Curtis can go to logs.
05:31Did you use a slide rule?
05:32Or are you young enough where it's like an HP?
05:36My father taught me how to use a slide rule.
05:37I'm a 12C guy.
05:39There you go.
05:39You're a 12C guy, but you can stay.
05:41I just, I want to know what the short-termism of your world,
05:48what they get wrong when Jonathan Curtis is looking out five years or 10 years
05:53with this American companies.
05:55Ultimately, we think the knowledge worker market is somewhere between 50 and 60 trillion dollars
06:01of spend.
06:03AI is going to augment and or replace a significant portion of that.
06:08I lean into the word augment, but right now we are somewhere around $200 billion of usage
06:15spend on these models.
06:17A small fraction of that, let's say 10% of that $50 trillion knowledge worker TAM is what
06:24is being spent right now.
06:25So we feel like we're in the very early days of this.
06:29And so that is what I think people are missing.
06:32And the tinkerers, the people in your organization who are using these tools every day, get it.
06:39If you are building with them, you understand the productivity.
06:41Do you know the guys at Berkeley that did perplexity?
06:44Did you go to school with them?
06:45I did not go to school with those guys, but we do know the guys.
06:48They're killing.
06:48Should they have sold out?
06:50Yeah.
06:51The real challenge in this market is how do you get enough distribution so your products
06:56get seen?
06:57Right.
06:58And the companies that have the.
Comments