- 12 hours ago
Category
🗞
NewsTranscript
00:00How's everybody doing?
00:05Rudina, thank you so much for doing this with us.
00:08We're going to try and cover a lot of ground in the next 10, 12 minutes.
00:12But I want to start with the macro, because I feel like it's kind of remiss if we ignore the
00:16elephant in the room.
00:18And obviously, we continue to see policies that impact the investment world, our world at large.
00:25How does any of this stuff factor in as you think about investments?
00:29You've been investing for a long time, seen different cycles, different White Houses.
00:33Does any of this impact kind of your investment decisioning or decision-making right now?
00:38Thank you, Carol, and good morning, everyone.
00:40So I'm an early-stage investor at the very edge of artificial intelligence and frontier tech for enterprise and cybersecurity.
00:49And it's an environment of change.
00:51And change can be good, and change can be disruptive.
00:54As far as AI adoption, we're only in the beginning innings of its adoption.
01:01The regulatory environment is as loose as it has been in a lot of years.
01:04That, in many ways, opens up the door for investments.
01:08QSBS, for example, Qualified Small Business Stock, being an opportunity to sort of loosen its definition
01:16and made it easier to obtain that with a goal of spurring more innovation.
01:21And then, of course, there is overall uncertainty in the political environment.
01:25So it's a lot of change in both directions.
01:28A lot coming at us.
01:29Regina, when it comes to AI, you've been a long-time investor.
01:35In terms of AI-related opportunities, how much stuff is coming across your desk?
01:39And how much of it is actually you think, okay, these are opportunities?
01:45So a lot, because effectively everything that we do going forward will be at a minimum AI-enabled.
01:54I caught a glimpse of the prior panel when they were talking about building on top of the LLMs or
02:00the foundation models or frontier models.
02:02I think I refer to them as such.
02:04If you think about it, any type of tool we use, whether it's in our personal lives or in our
02:10business lives,
02:12we'll have a layer, we'll have a wrapper that's human-like interaction, human-like experience.
02:18That, right there and there, has been impacted by AI.
02:21That's actually not an AI company.
02:24That's an AI wrapper.
02:25That's an AI enablement.
02:26And every interaction from shopping to your personal finances will be impacted by that.
02:32What's coming to my desk is more qualified opportunities, I hope, particularly around vertical AI,
02:39so vertical industries where deep domain expertise is needed,
02:43where you can have unique modes and defensibility around the data
02:48and around the proprietary potentially models or unique models to the vertical or the problem at hand
02:54so that the large incumbents can't respond easily.
02:58So I want to get into more of what you mean in terms of vertical AI,
03:01but in terms of the horizontal players, the large language model players,
03:05whether it's your OpenAI, TouchyBT, whether it's Gemini, whether it's Anthropic.
03:09Mr. Al, let's add to them.
03:11Right.
03:11The companies that we are talking about day in and day out, lots of money going into them.
03:16Is that the opportunity going forward, or is that just providing for the opportunities going forward?
03:23That is providing for the opportunities going forward.
03:26And again, I'm going to try to balance, not to be too techie.
03:28But what we are seeing is a rebuilding of the technology stack.
03:33So, you know, these models are the very foundation.
03:36They're just right above the hardware layer of infrastructure.
03:39They're really the software or the AI infrastructure for this stack.
03:42Above it, we will have, and we do have already, the middle layer,
03:46which is tools that help sort of get you from, okay, I have access to a model.
03:52What do I do now?
03:53To that application layer.
03:54But the real opportunity set is that the application layer,
03:58very analogous to what we saw with the cloud,
04:00where we have AWS and Azure and others, GCP with Google.
04:04But it opened up the sort of the opportunity set for a whole new class of SaaS companies,
04:10which they are challenged.
04:11That's the nature of disruption.
04:13But it created an entire ecosystem from NetSuite to Workday to you name it.
04:19You did say on our planning call that there's a war around AI.
04:23And you talked specifically about these horizontal players.
04:27And your point was kind of they're going to take us so far, but not far enough.
04:31Correct, because it's both a business reason and a performance reason.
04:38The horizontal players, the foundation model players,
04:40have actually ingested all of the web data.
04:42So horizontally, it's extremely difficult to compete with them,
04:46and they seemingly have access to capital that feels like unlimited.
04:50Anytime they want to raise, they've been raising.
04:51And we can have a whole different discussion on how many players can this market support.
04:55But fundamentally, that's been done.
04:57If I am going to disrupt the freight shipment industry,
05:02I have a portfolio company called Ship Angel that's doing just that here in New York.
05:07And, you know, take us from the 1980s where literally if a brand is shipping any good
05:12from any part of the world, what they are doing is looking at a piece of paper with rates,
05:17who knows if they're accurate or not, calling somebody to book,
05:21and then they receive, you know, the invoice, and they don't know how to audit it.
05:25There is no way.
05:27Web industry, it's vertical, it's narrow.
05:30I don't see open AI or an anthropic going after it.
05:33So the mass aggregators of data, the large language models,
05:36that's not the value going forward is basically the point.
05:38At least for investors, that's my point.
05:40All right.
05:41So what about the software companies?
05:43As the AI scare has made its way through the market, through different sectors,
05:46it's been a little crazy.
05:49What about the software space?
05:50But are they the ones who can be opportunistic because of their vertical data sets?
05:55Or are they unique data sets?
05:57Or do they have unique data sets?
05:59AI is software, just saying.
06:02So I think this notion of old software and new AI, there's a little bit of misnomer going on there.
06:08But fundamentally, it will depend on whether incumbents can disrupt themselves.
06:13It comes down to that.
06:15If you are Salesforce and you coined the term AI agents, and yet, you know, you are challenged in all
06:21different directions,
06:22can you disrupt slash innovate slash acquire faster than the two guys in a garage,
06:29vibe coding from scratch, getting to revenue overnight, and who's going to win?
06:34So it becomes some about performance and a lot about execution.
06:39What about something like Anthropic?
06:41They've just recently announced, it was just this week, updates to Claude Cowork.
06:45And it is extending abilities into things like human resources, investment banking.
06:50This is all through plug-ins.
06:52And what's interesting is they've worked with companies like FactSet Research, S&P Global, to make these capabilities possible.
06:59So what's to stop them from using that knowledge and kind of running with it and kind of taking over
07:08and making those companies obsolete?
07:10And these companies feel like they would rather be inside the machine than outside.
07:13So they're saying, let me work with you.
07:16We're asking the same question we should have and would have asked when the Internet and the Web was invented.
07:21So, yes, it's a disruption.
07:23Therefore, it will give a rise to new companies and new opportunities.
07:26But the existing, the Bloomberg terminal didn't go away when the Web, I remember, I was an investment banker in
07:31a past life.
07:32We would go to the terminal.
07:33It didn't go away, right?
07:34Because we're really good.
07:38Truly.
07:39But it became Web-enabled.
07:41So I come back to the notion of, yes, there will be new players.
07:45By the same token, there is an opportunity for these existing players to disrupt themselves with their unique data and
07:51automate your workflows.
07:52Right now, if you want to win an AI, it's about automating workflows.
07:56What's a recent investment?
07:57So explain this vertical AI.
07:59So you have to be thinking, you're early stage.
08:02So you're thinking five to ten years, this is a company that's going to be around and maybe disrupt the
08:06disruptors.
08:07So give me an idea of a recent investment.
08:09Happy to.
08:10So, and I'm going to pick one that's hopefully relatable to all of us as humans.
08:15Pharmacy prescription fulfillment.
08:18How much fun do you have when you have to deal with that and prior authorizations?
08:22I mean, it's a very antiquated process or workflow whereby the doctors, 20% is headwritten, 80% is digital
08:30faxes.
08:30They literally get faxed.
08:32Then it gets into the pharmacy.
08:33Depending on which pharmacy it is, they have many different systems.
08:36They have to input it from scratch.
08:38Then if it's a prior authorization, it has to go through a whole process.
08:41It's messy as messy can be.
08:44Well, with AI, you can actually automate a lot of the workflows from needing to call.
08:49You can now have an AI engine calling to, or, you know, an agent.
08:54All the way to how you replenish 30% of the drugs that actually do not get picked up by
09:00the patients.
09:01So it's a multi, multi-billion dollar workflow filled with waste and difficult experiences for the patients and for the
09:10providers.
09:10Go for that.
09:11What about the financial industry?
09:13What should they be thinking?
09:14I mean, they are obviously super familiar with data, with algorithms, with AI for a long time.
09:20But what does it mean for this space specifically?
09:23That our roles change and change dramatically.
09:26So Claude Cowork came out about a month ago.
09:29I sat down over the weekend last Sunday and played with it, as I like to say, for a couple
09:35of hours
09:36and literally updated the entire mapping of the entire AI techniques, architectures out there.
09:43My job is changing.
09:45Our jobs are changing.
09:46We're no longer going to be number crunchers.
09:48The number crunching is going to practically happen on its own, whether it's a Claude or something else.
09:53We are going to be interpreters.
09:55We're going to be more empowered decision makers.
09:58So what I will say is there's a big fear around jobs.
10:01I'm not worried about jobs going away.
10:03In fact, any research you see indicates that there will be more jobs, especially financial services.
10:09Our jobs, though, will change.
10:11And where will we be in that restructuring of our roles?
10:16An analyst will be a lot more about interpretation.
10:18Will be a lot about this collaboration between human and machine or human and AI to deliver different outputs.
10:25So we need to be retrained.
10:26Just last question in terms of your advice, a takeaway, because I do think about anyone who's here.
10:34I keep hearing about dislocation.
10:36Older folks might have some problems kind of making the transition.
10:39I hate to put an age thing on it.
10:41But you had a great piece of advice, and I know I've done this with my 22-year-old.
10:45What do you suggest?
10:46I suggest hanging out with the 25-year-olds.
10:49They're mature enough to kind of have some level of sense of what's right and wrong, I hope.
10:56But they're young enough to actually adopt new tools and think differently and not be bound by it has been
11:04done this way.
11:05Therefore, it should continue to be done that way.
11:07Their naivete is an advantage in these types of disruption and transformation.
11:12Covered a lot of ground.
11:13Hope you guys enjoyed it.
11:15Regina, thank you so much.
11:16Really appreciate it.
11:17Regina Siseri, everybody.
11:19Thank you so much.
11:19A blast being bent.
Comments