Meeting Title: Uttam <> Chang Date: 2024-05-01 Meeting participants: Chang Ho, Uttam Kumaran
WEBVTT
1 00:00:12.080 ⇒ 00:00:12.800 Uttam Kumaran: Hello!
2 00:00:13.230 ⇒ 00:00:14.359 Chang Ho: Hey? Can you hear me?
3 00:00:14.360 ⇒ 00:00:16.100 Uttam Kumaran: Hey? I can hear you. How’s it going.
4 00:00:16.100 ⇒ 00:00:20.999 Chang Ho: Awesome. I’m just gonna turn the video on. So I got carried away washing the dishes.
5 00:00:21.348 ⇒ 00:00:25.530 Uttam Kumaran: No problem. I’m actually just running to another meeting. So I’ll be
6 00:00:25.950 ⇒ 00:00:30.870 Uttam Kumaran: video off. I’m just leaving my house. But no problem. I know it’s late. Your time.
7 00:00:31.810 ⇒ 00:00:37.140 Chang Ho: We’ll start the video. Let’s just try and get the. So do you want to just stay off video, then.
8 00:00:37.140 ⇒ 00:00:38.870 Uttam Kumaran: Yeah, that’s fine. Cause I’m just gonna get.
9 00:00:38.870 ⇒ 00:00:39.300 Chang Ho: As also.
10 00:00:39.300 ⇒ 00:00:40.810 Uttam Kumaran: Real, quick.
11 00:00:40.810 ⇒ 00:00:42.100 Chang Ho: Sure, sure, sure!
12 00:00:42.530 ⇒ 00:00:43.460 Uttam Kumaran: How’s everything?
13 00:00:45.140 ⇒ 00:00:49.930 Chang Ho: Very good. Yeah, it just. It’s just been a very. It’s been a mental month through time. So I’m glad that we managed to catch up now
14 00:00:50.655 ⇒ 00:01:02.744 Chang Ho: like uncharacteristically mental. I’m like Rebecca. My wife and I were just discussing this. We were in Toronto, as you know. That’s when we chatted and then also ended up being in Norway.
15 00:01:03.090 ⇒ 00:01:03.680 Uttam Kumaran: Wow!
16 00:01:03.680 ⇒ 00:01:04.410 Chang Ho: What’s kind
17 00:01:04.709 ⇒ 00:01:08.859 Chang Ho: through skiing with some with some Norwegian friends there, followed by an
18 00:01:08.870 ⇒ 00:01:20.890 Chang Ho: I to give a presentation on AI applications. Well, at least applications of generative AI in infection diseases and clinical microbiology. Just last weekend, also in infection prevention control.
19 00:01:20.930 ⇒ 00:01:24.359 Chang Ho: They invited me to talk, based the European
20 00:01:24.920 ⇒ 00:01:31.890 Chang Ho: Conference, and it’s held in Barcelona. It was obviously fucking sweet. But yeah, all-consuming
21 00:01:32.060 ⇒ 00:01:36.259 Chang Ho: or consuming 15,000 15,000 people to send upon
22 00:01:36.500 ⇒ 00:01:38.309 Chang Ho: the city all at once.
23 00:01:38.610 ⇒ 00:01:39.530 Uttam Kumaran: Wow!
24 00:01:39.530 ⇒ 00:01:44.629 Chang Ho: Is mad is a cavalcade. It’s all. It must feel like the European version of Bernard
25 00:01:46.230 ⇒ 00:01:49.849 Chang Ho: anywhere near as anywhere near as much class. A drugs.
26 00:01:49.990 ⇒ 00:01:50.510 Uttam Kumaran: Yes.
27 00:01:50.941 ⇒ 00:01:54.819 Chang Ho: Yeah. Anyway, how how are tricks with you?
28 00:01:55.110 ⇒ 00:02:07.410 Uttam Kumaran: Things are good. You know. I think I’m similar kind of crazy mode. I was doing some traveling to India just really nice to go see some family, and you know now that my life and businesses
29 00:02:07.410 ⇒ 00:02:25.090 Uttam Kumaran: sort of one was able to kind of have some nice business chats and, like my cousin, is doing some business there that would love to kind of explain a bit more is something that I message you about. And it was nice getting some time with family. I haven’t taken a break in like about a year, basically.
30 00:02:25.822 ⇒ 00:02:28.055 Uttam Kumaran: So that was a weird feeling.
31 00:02:29.145 ⇒ 00:02:29.490 Chang Ho: This
32 00:02:30.960 ⇒ 00:02:38.935 Uttam Kumaran: It was nice. It was nice, you know. I can’t work during the day, because there’s no work to be done. But it it was really nice.
33 00:02:39.220 ⇒ 00:02:39.810 Chang Ho: Life.
34 00:02:40.140 ⇒ 00:02:48.791 Uttam Kumaran: It’s a good. It’s a good motivation to kind of like reassess work and stuff. And yeah, the business is going well, we had it. We assigned another client.
35 00:02:49.080 ⇒ 00:02:49.650 Chang Ho: Well.
36 00:02:49.650 ⇒ 00:02:55.630 Uttam Kumaran: His clients doing like quote. They’re doing like quoting in like me, metals like metal related
37 00:02:55.650 ⇒ 00:03:03.909 Uttam Kumaran: voting for steel here in the Us. And they’re like a subsidiary of like a larger company. And we’re helping them with some data management. So that’s.
38 00:03:03.910 ⇒ 00:03:04.310 Chang Ho: Very well.
39 00:03:04.310 ⇒ 00:03:05.960 Uttam Kumaran: It’s going well.
40 00:03:06.750 ⇒ 00:03:09.549 Uttam Kumaran: yeah, I mean overall. I think things are going
41 00:03:10.060 ⇒ 00:03:18.449 Uttam Kumaran: as well as they can be. The only other thing that’s interesting we’re doing is we’re hiring a project manager. So this will kind of allow me to get a little bit out of the day to day.
42 00:03:18.670 ⇒ 00:03:19.600 Chang Ho: Yes. Yeah. So.
43 00:03:19.600 ⇒ 00:03:35.637 Uttam Kumaran: So, and I could. I can come in and do engineering work which I’m happy to do. It’s more of like a project management, and the communication like making sure everything’s on time. That’s like, actually takes a lot a lot of contact switching and I’m a much more effective engineer than I am that
44 00:03:36.280 ⇒ 00:03:39.140 Chang Ho: Fair enough you’re telling. You can’t be all up, all everything, you know.
45 00:03:39.140 ⇒ 00:03:42.790 Uttam Kumaran: Exactly. It’s quite a lot of contact switching from like that
46 00:03:42.820 ⇒ 00:03:56.029 Uttam Kumaran: from engineering to the project manager to then, just like broadly, like meeting people for business. It’s quite a lot of hoops to jump through, so I prefer to just do like 2 of those and engineering. There’s no I like and really enjoy doing so.
47 00:03:56.030 ⇒ 00:03:56.536 Chang Ho: Yes, that’s.
48 00:03:56.790 ⇒ 00:03:58.667 Uttam Kumaran: That I would prefer to do
49 00:03:59.290 ⇒ 00:03:59.970 Chang Ho: Nice.
50 00:04:00.190 ⇒ 00:04:02.403 Uttam Kumaran: But otherwise things are going well, I mean,
51 00:04:03.000 ⇒ 00:04:16.088 Uttam Kumaran: yeah, I think we’re we. We were working a little bit on one like AI client for a bit, but they’re like a startup, and they’re like reassessing their entire like existence. So I kind of slowed down.
52 00:04:17.970 ⇒ 00:04:18.565 Uttam Kumaran: I’m
53 00:04:19.240 ⇒ 00:04:27.773 Uttam Kumaran: I’m working with some people here locally in Austin. We’re planning on building an AI email scheduling assistant where basically.
54 00:04:28.180 ⇒ 00:04:29.150 Chang Ho: CC.
55 00:04:29.150 ⇒ 00:04:30.890 Uttam Kumaran: You CC, in like a
56 00:04:30.950 ⇒ 00:04:36.937 Uttam Kumaran: like this email. And then it basically does the back and forth with whoever you’re talking to,
57 00:04:37.310 ⇒ 00:04:49.629 Uttam Kumaran: just like 2 friends of mine that are engineers here. We’re like, Hey, let’s just try to spin this up. And we saw a product launch that was very similar that like, got a lot of traction we’re like, it’s pretty easy to do and like, maybe do it like a couple of weekends. So
58 00:04:49.690 ⇒ 00:04:53.099 Uttam Kumaran: that’s been kind of a fun little side project to work on as well. So.
59 00:04:53.590 ⇒ 00:04:54.660 Chang Ho: Oh, wow!
60 00:04:54.960 ⇒ 00:04:57.110 Chang Ho: That does sound really fun, actually.
61 00:04:57.110 ⇒ 00:05:21.475 Uttam Kumaran: Yeah, so so basically, all it is is like it, though the whole system is like again, it processes the email. It pulls your Google Calendar looks at your events and your availability, and then basically acts as like it’s presented like a like an assistant where it says, Hi, like, I’m so and says, assistant like, Here’s the scheduling options. And then based on the reply, it then, you know, basically it can answer, and then book the meeting.
62 00:05:21.890 ⇒ 00:05:25.160 Uttam Kumaran: But the language model makes it way easier and way cheaper
63 00:05:25.180 ⇒ 00:05:31.324 Uttam Kumaran: to handle, like all sorts of used cases. Where we were like, okay, I wonder if you could book flights through email or book.
64 00:05:31.570 ⇒ 00:05:31.989 Chang Ho: Oh, my God!
65 00:05:31.990 ⇒ 00:05:32.969 Uttam Kumaran: 40,000. Yeah.
66 00:05:32.970 ⇒ 00:05:37.110 Chang Ho: Love to see that in action, or even learn how to do some of that
67 00:05:37.547 ⇒ 00:05:41.220 Chang Ho: engineering part of that, because that sounds really fun. So.
68 00:05:41.520 ⇒ 00:05:48.100 Uttam Kumaran: Yeah. So that’s been really fun. Like, we got a landing page and stuff up, I’ll I’ll shoot it to you. And we’re just working on like
69 00:05:48.605 ⇒ 00:05:52.749 Uttam Kumaran: somewhere, like the basic, like Google passing Google Calendar to an Llm
70 00:05:53.010 ⇒ 00:05:57.019 Uttam Kumaran: crafting the responses and then doing all the email and deployment. But
71 00:05:57.170 ⇒ 00:06:02.195 Uttam Kumaran: that’s like, that’s a little bit more of like fun, and still trying to do something in AI while.
72 00:06:03.076 ⇒ 00:06:06.249 Uttam Kumaran: just like with some with some friends. So.
73 00:06:06.540 ⇒ 00:06:06.925 Chang Ho: Yes.
74 00:06:08.040 ⇒ 00:06:16.490 Chang Ho: the the irony there every time of of one wanting to be like, I guess, wanting to be involved in that kind of thing, and and see that kind of product being
75 00:06:16.640 ⇒ 00:06:23.060 Chang Ho: being built. It is more for me, more just for the interest of it, because I know that if I was deployed myself on my own calendar, which is
76 00:06:23.210 ⇒ 00:06:24.680 Chang Ho: bare bones. Replace
77 00:06:25.300 ⇒ 00:06:32.839 Chang Ho: any data that needs to be replaced with automatically insert meetings everywhere. I do not want it to you.
78 00:06:32.840 ⇒ 00:06:34.099 Uttam Kumaran: Yes, yes, I mean.
79 00:06:35.530 ⇒ 00:06:40.010 Uttam Kumaran: the nice thing is like, you see, see it in. And then maybe like it’s mainly to handle like
80 00:06:40.530 ⇒ 00:06:42.799 Uttam Kumaran: I commonly have this like back and forth
81 00:06:43.050 ⇒ 00:06:48.639 Uttam Kumaran: process where it’s like, What are you free? What are you free and like it would I would love for someone else to handle that but.
82 00:06:49.026 ⇒ 00:06:49.799 Chang Ho: Not everyone.
83 00:06:49.800 ⇒ 00:07:00.659 Uttam Kumaran: I can have assistance, and it’s like I don’t even I’m not. I don’t mind searching for an assistant right now, and that’s like it whatever. And we’re like, okay, I didn’t do this possible. And we saw a demo of it working this other product. We’re like.
84 00:07:00.890 ⇒ 00:07:07.599 Uttam Kumaran: well, I think we could do this, basically in like a couple of weekends. So that’s a little fun, AI thing trying to do.
85 00:07:08.260 ⇒ 00:07:12.149 Chang Ho: That’s really cool. And which which lemme are you using for that.
86 00:07:12.340 ⇒ 00:07:18.349 Uttam Kumaran: I think we’re just gonna use Openai. But then I would, based on the cost I would love to use like llama 3.
87 00:07:18.600 ⇒ 00:07:22.219 Uttam Kumaran: Yeah, we actually don’t really need it to be super fast.
88 00:07:22.280 ⇒ 00:07:24.770 Uttam Kumaran: because we kind of want it to seem like an assistant
89 00:07:24.910 ⇒ 00:07:31.529 Uttam Kumaran: like the other product that is kind of in market. They’re like, really, really rapid. And we’re kind of like, well, what’s the actual
90 00:07:31.720 ⇒ 00:07:37.439 Uttam Kumaran: like? What’s the? Does? It actually need to be like immediately respond. It could be in like a minute or 2. But
91 00:07:37.500 ⇒ 00:07:46.740 Uttam Kumaran: yeah, mainly, we’re gonna look. We’re gonna have some several like test cases and then evaluate against like a bunch of different like rescheduling and like interesting answers, like test cases.
92 00:07:47.683 ⇒ 00:07:51.549 Uttam Kumaran: And then hopefully swap out to get to the cheapest.
93 00:07:51.900 ⇒ 00:07:55.319 Uttam Kumaran: you know version, but not too opinionated right now.
94 00:07:55.450 ⇒ 00:07:56.110 Chang Ho: Yes.
95 00:07:57.430 ⇒ 00:08:08.229 Chang Ho: cool, cool, cool. I yeah, when we can talk hopefully. Yeah, I mean, we can talk about a different time or on via slack. I would love to see the how one goes by implementing that actually
96 00:08:08.500 ⇒ 00:08:09.150 Uttam Kumaran: Yeah.
97 00:08:09.150 ⇒ 00:08:17.609 Chang Ho: Goes about engineering and building that because that’s I’m just curious, just pure, like, honestly, purely. So I can just maybe apply it to my clinical calendar.
98 00:08:18.370 ⇒ 00:08:31.449 Uttam Kumaran: Yeah. And even not even that. It’s even just like, how do you? How do you roll this up like as cheap as you can right like, can you go from like idea to prototype to like something production? And like.
99 00:08:31.500 ⇒ 00:08:35.059 Uttam Kumaran: you know, a few weeks. So, yeah, those are the interesting thing.
100 00:08:36.080 ⇒ 00:08:39.005 Chang Ho: Yeah, totally. I’ll I’d love to see that I’d love to see that.
101 00:08:39.230 ⇒ 00:08:40.870 Uttam Kumaran: Yeah, and there’s talk about what’s happening.
102 00:08:40.870 ⇒ 00:08:42.190 Chang Ho: But India. Yeah.
103 00:08:42.190 ⇒ 00:08:46.999 Uttam Kumaran: Yeah, yeah, yeah. So sorry. I’m just disconnecting from my wi-fi. Give me 1 s.
104 00:08:47.700 ⇒ 00:08:48.540 Chang Ho: No worries.
105 00:08:52.700 ⇒ 00:08:59.087 Uttam Kumaran: Yeah. So I was talking to I don’t know if you had a chance to read those dots. It’s a little bit tough without a lot of context. But
106 00:08:59.370 ⇒ 00:09:00.700 Uttam Kumaran: basically, my.
107 00:09:00.700 ⇒ 00:09:03.605 Chang Ho: I have skimmed it for sure.
108 00:09:04.370 ⇒ 00:09:05.620 Uttam Kumaran: Yeah. So my cousin is working.
109 00:09:05.620 ⇒ 00:09:06.079 Chang Ho: Even though we.
110 00:09:06.080 ⇒ 00:09:06.960 Uttam Kumaran: Out there.
111 00:09:07.510 ⇒ 00:09:23.270 Uttam Kumaran: My cousin’s working with a number of developing countries in India related to projects around like population measurement. In particular, he’s working really closely with Timor, which is a country I actually hadn’t heard of until I started.
112 00:09:23.270 ⇒ 00:09:24.760 Chang Ho: Yeah. Timor, leste.
113 00:09:25.110 ⇒ 00:09:26.040 Chang Ho: yeah.
114 00:09:26.780 ⇒ 00:09:31.539 Uttam Kumaran: I don’t know if you have. You heard about it before, like I I guess, like I didn’t really even know much of the history.
115 00:09:33.120 ⇒ 00:09:36.720 Chang Ho: All I know is that it used to be a I think, a Portuguese colony.
116 00:09:36.980 ⇒ 00:09:39.719 Chang Ho: and it’s desperately poor.
117 00:09:39.850 ⇒ 00:09:40.340 Chang Ho: Yeah.
118 00:09:40.340 ⇒ 00:09:40.990 Uttam Kumaran: And.
119 00:09:41.150 ⇒ 00:09:44.509 Chang Ho: They have. There’s been a massive exodus in
120 00:09:45.330 ⇒ 00:09:49.339 Chang Ho: immigrants from there, some of whom have landed in Oxford actually so.
121 00:09:49.340 ⇒ 00:09:49.850 Uttam Kumaran: Oh!
122 00:09:49.850 ⇒ 00:09:55.270 Chang Ho: We? We end up looking after a few Tim or Sdn patients in our clinics.
123 00:09:56.040 ⇒ 00:09:58.750 Uttam Kumaran: Oh, yeah, it’s.
124 00:10:02.770 ⇒ 00:10:03.590 Chang Ho: Hello!
125 00:10:04.030 ⇒ 00:10:07.119 Uttam Kumaran: Yeah, I’m here. Sorry. I just was connecting to my car.
126 00:10:07.400 ⇒ 00:10:08.530 Chang Ho: It’s not. So. Yeah.
127 00:10:08.530 ⇒ 00:10:09.870 Uttam Kumaran: The the
128 00:10:09.970 ⇒ 00:10:30.277 Uttam Kumaran: the the thing that we’re working on actually is, well, he’s working on is is mainly he’s working with the country to do things around populated survey and creating an app that allows the population to have get access to government services as well as the government to have access to like population. Census
129 00:10:31.480 ⇒ 00:10:36.340 Uttam Kumaran: One avenue that they’re kind of looking into as well is
130 00:10:36.470 ⇒ 00:10:43.320 Uttam Kumaran: 1 one avenue that they’re looking into there, and they’re they’ve been working with the Sri Lankan government. And it’s a more government. Basically.
131 00:10:43.350 ⇒ 00:10:54.259 Uttam Kumaran: their government is gonna get like an Imf grant. And then some of the funds will be earmarked for development in this area. So that’s like one area that they’re actually interested in working on
132 00:10:54.869 ⇒ 00:11:16.499 Uttam Kumaran: there was another area. There’s another area related to like medical and medicine. I think something about the clinician. But I I actually need to really follow up. I haven’t read that document in a little bit. But I I I just thought it would be interesting conversation to kind of get me you and maybe my cousin, and, like W. One of the counterparts on the team is actually
133 00:11:16.500 ⇒ 00:11:29.199 Uttam Kumaran: he was. He’s like a previous doctor, and is really working a lot on the clinical side, and I was the one talking to him a little bit about our conversation about like where you can apply AI and things like that. So.
134 00:11:29.200 ⇒ 00:11:30.379 Chang Ho: Yes, yes.
135 00:11:31.490 ⇒ 00:11:36.700 Chang Ho: yeah, I can see that. There’s I mean, I’ve already flicked through it. There’s quite a lot that is related to
136 00:11:36.820 ⇒ 00:11:44.990 Chang Ho: not just AI usage, but actually just. But maybe it was almost, I guess, AI in very different, very different ways, or layers of this
137 00:11:45.970 ⇒ 00:11:51.870 Chang Ho: very complicated information extracting siphoning pipeline, because that’s how I see this.
138 00:11:52.230 ⇒ 00:11:53.170 Uttam Kumaran: Always, see.
139 00:11:53.170 ⇒ 00:11:56.130 Chang Ho: Most of the value added that we can get out of. You know.
140 00:11:56.500 ⇒ 00:12:05.080 Chang Ho: general, today, our models and so on is all about, I mean, yeah, we’ve always been about, and we will always be about information, personal information.
141 00:12:05.600 ⇒ 00:12:06.270 Chang Ho: extraction.
142 00:12:06.270 ⇒ 00:12:06.810 Uttam Kumaran: Yes.
143 00:12:06.810 ⇒ 00:12:23.159 Chang Ho: How to distill that and then repackage it for a particular solution. Just try and address a certain challenge, you know more problem. And so, you know, with with that in mind, I am not at all surprised that I I end up seeing a document that I’m I swear I’ve seen 15 different iterations of.
144 00:12:23.530 ⇒ 00:12:25.099 Chang Ho: you know, from from various different
145 00:12:25.680 ⇒ 00:12:45.229 Chang Ho: government. So what they’d love to do with this new technology or this suite of tools that we now have at our disposal, including your GPS data, you know, local clinic levels, data, unofficial healthcare provider data and so on and so forth. So it’s it’s it’s definitely on the tips of lots of people’s fingers. And
146 00:12:45.690 ⇒ 00:13:03.600 Chang Ho: and it was cool to see this sort of very high level document that was trying to address the issue of one health which you probably encountered before, but basically just refers to the it’s a who and UN joint sort of collaborative piece and trying to move the whole scientific world forward, on the notion that
147 00:13:04.130 ⇒ 00:13:05.650 Chang Ho: the health of our
148 00:13:05.850 ⇒ 00:13:07.960 Chang Ho: all earth.
149 00:13:08.820 ⇒ 00:13:18.629 Chang Ho: and therefore the help of us, are intricately into linked, and therefore every other activity we undertake must be considered to be impactful to help.
150 00:13:18.640 ⇒ 00:13:22.600 Chang Ho: And as you know, there’s plenty of issues with what climate change can do for
151 00:13:22.640 ⇒ 00:13:25.140 Chang Ho: yeah, for health, you know, in terms of health
152 00:13:25.620 ⇒ 00:13:28.029 Chang Ho: consequences. It’s it’s gonna be disastrous already.
153 00:13:28.030 ⇒ 00:13:28.530 Uttam Kumaran: Has.
154 00:13:28.530 ⇒ 00:13:30.289 Chang Ho: Started affecting us. As you know.
155 00:13:30.370 ⇒ 00:13:32.500 Chang Ho: heat, heat, stroke.
156 00:13:32.500 ⇒ 00:13:33.200 Uttam Kumaran: But yeah.
157 00:13:33.200 ⇒ 00:13:38.620 Chang Ho: All dying deprivation from water and clean water and sanitation. And so.
158 00:13:38.690 ⇒ 00:13:42.930 Chang Ho: anyway, this yeah, this document was interesting for
159 00:13:43.310 ⇒ 00:13:45.230 Chang Ho: being a sort of nice summary piece.
160 00:13:45.660 ⇒ 00:13:48.610 Chang Ho: really, of a lot of the a lot of the issues that are common to
161 00:13:48.810 ⇒ 00:13:56.679 Chang Ho: so very similar proposals. I guess I guess a couple of questions for you, which is that, did they have they already obtained funding to do? To
162 00:13:56.870 ⇒ 00:13:59.060 Chang Ho: to think about how zoom implements
163 00:13:59.510 ⇒ 00:14:03.620 Chang Ho: sort of a grand governance infrastructure which seemed to be the main thing. They’ll.
164 00:14:03.620 ⇒ 00:14:30.109 Uttam Kumaran: Yeah. So they kind of were thinking for this, like, government OS, kind of a government infrastructure. So I I did get like a little bit of like. I mean, when I met them one day they are. They do have some private funding that they’ve secured for kind of like paying for a lot of the engineering development but their main goal is to kind of like begin to build these the government, the way the process, they articulated. And I’ve never worked with like the government they’re they’re they’re honestly meeting with, like
165 00:14:30.350 ⇒ 00:14:50.639 Uttam Kumaran: the Prime Minister. And like all these big big shot guys. And I’m like, I have no idea about this process. But basically, they’re showing like small prototypes and small advancements and gathering requirements and continuing to build to more. And and they’re working with a few other countries are beginning to acquire these like Imf Grants for funding and.
166 00:14:50.640 ⇒ 00:14:51.080 Chang Ho: Yeah.
167 00:14:51.080 ⇒ 00:14:58.137 Uttam Kumaran: Again like parts of that. They’re like hoping to buy for those contracts in order to provide these services, but they they are privately funded.
168 00:14:58.420 ⇒ 00:14:58.930 Chang Ho: Yes.
169 00:14:58.930 ⇒ 00:15:22.170 Uttam Kumaran: And they have, like some private investors in India, that have funded the project in order to just get it off the ground and get resources and things like that. So it’s quite a big effort. And the timeline, you know, of course, like working with these governments they mentioned, it’s it’s quite an arduous process, but I would say they tried a lot in Sri Lanka, and it was kind of a bit slow, and then my cousin, he had some connections into more, and has been traveling, and
170 00:15:22.170 ⇒ 00:15:35.970 Uttam Kumaran: they they like met with the Prime Minister, and met with a lot of people, and they seemed to have really good momentum onto actually like getting something implemented which they may start small. But I mean again, you can see how widespread that document is, and how many things it covers.
171 00:15:37.090 ⇒ 00:15:38.460 Uttam Kumaran: yeah, there’s extreme.
172 00:15:38.460 ⇒ 00:15:39.720 Chang Ho: Broad scope, so it would be.
173 00:15:39.720 ⇒ 00:15:40.560 Uttam Kumaran: Jeremy Broadscom.
174 00:15:40.560 ⇒ 00:15:42.020 Chang Ho: For us to
175 00:15:42.130 ⇒ 00:15:45.269 Chang Ho: to narrow down on. I guess what you what particular
176 00:15:45.700 ⇒ 00:16:05.196 Chang Ho: killer, what particular use case or avenue they would want you and I or you know, and our team, I guess to sort of get involved in if we do get involved in it at all. Yeah, for sure. I think one of the cool things just say off the bat about in this is in Sri Lanka, but also but this is a cousin in India, based in India, right? Because.
177 00:16:06.080 ⇒ 00:16:07.009 Uttam Kumaran: He’s based in China.
178 00:16:07.010 ⇒ 00:16:09.650 Chang Ho: Based in India model. It’s 19
179 00:16:10.160 ⇒ 00:16:20.620 Chang Ho: is that from a governance and regulation angle they are just a few years behind the Us. And Uk, and and the rest of Europe in terms of AI applications and health.
180 00:16:20.930 ⇒ 00:16:25.280 Chang Ho: Well, what I mean by that is that they have yet to, from what I can see just
181 00:16:25.860 ⇒ 00:16:29.619 Chang Ho: distinguish between diagnosis and and simply
182 00:16:29.880 ⇒ 00:16:32.189 Chang Ho: logging like automatically logging.
183 00:16:32.190 ⇒ 00:16:36.930 Uttam Kumaran: Well, you you you’d be surprised like I. I actually talked to them about
184 00:16:37.020 ⇒ 00:16:44.200 Uttam Kumaran: what kind of like a lot of talk it was that we talked about. And he was saying you’d find it. How hard it is to get people need to write things down.
185 00:16:44.310 ⇒ 00:16:47.510 Uttam Kumaran: let alone like record in a computer. He’s like that.
186 00:16:47.510 ⇒ 00:16:47.950 Chang Ho: Oh, yeah.
187 00:16:47.950 ⇒ 00:16:55.109 Uttam Kumaran: Too far behind, and he was like he he was telling me. He’s like, you know, in the Us. Your physician may see, like
188 00:16:55.160 ⇒ 00:17:00.089 Uttam Kumaran: you know, couple of people a day. He’s like these people are. These aren’t racing like 100 people a day.
189 00:17:00.090 ⇒ 00:17:00.500 Chang Ho: Yes.
190 00:17:00.500 ⇒ 00:17:00.860 Uttam Kumaran: Like.
191 00:17:00.860 ⇒ 00:17:01.940 Chang Ho: Bad. It’s yeah.
192 00:17:01.940 ⇒ 00:17:03.180 Uttam Kumaran: Insane, and then he was like.
193 00:17:03.180 ⇒ 00:17:03.640 Chang Ho: Yeah.
194 00:17:03.640 ⇒ 00:17:22.979 Uttam Kumaran: You really have to show so much value before people will adopt because there’s no time. And so it’s a diff completely different set of challenges. That conversation really opened my eyes, and thinking back to what you said I was like, oh, like they’re writing things out like piece of paper, and it’s still like very, very rudimentary.
195 00:17:23.400 ⇒ 00:17:25.329 Chang Ho: Oh, come, absolutely.
196 00:17:25.790 ⇒ 00:17:26.899 Chang Ho: I think one of the
197 00:17:27.190 ⇒ 00:17:32.399 Chang Ho: well, I’m sure the project’s that quite a few different techies over in India are probably dreaming up
198 00:17:32.530 ⇒ 00:17:33.710 Chang Ho: would be
199 00:17:33.980 ⇒ 00:17:40.980 Chang Ho: the application of some of these Chat Gbtgbt tools, but, you know, modified for the purposes of
200 00:17:41.250 ⇒ 00:17:45.340 Chang Ho: conversations and documentation of conversations in in a clinical
201 00:17:45.910 ⇒ 00:17:51.450 Chang Ho: context, in in the very, in the, in the mirror, you know, in the many myriad, you know, the myriad of language does not exist.
202 00:17:51.750 ⇒ 00:18:00.580 Chang Ho: And and obviously, you know, off the off the cuff, it will only really be feasible for Hindi. And and we’ll do another highly.
203 00:18:00.580 ⇒ 00:18:01.200 Uttam Kumaran: Yes.
204 00:18:01.200 ⇒ 00:18:05.370 Chang Ho: Prevalently spoken languages. But it is striking, as you know, it is
205 00:18:05.930 ⇒ 00:18:25.460 Chang Ho: our Anglo centric. All these technologies are so far by by default, because we’re, you know. Ultimately it’s all come out of the Us. Mostly so and then, because research is so Anglo centric as well for people who are doing actual experimentation, or all talking and discussing it in English. So it’s it off.
206 00:18:25.780 ⇒ 00:18:31.150 Chang Ho: There’s an opportunity there as well, I think, for you know, how do we think about Lms like that? And and actually.
207 00:18:31.160 ⇒ 00:18:40.789 Chang Ho: if we were to get to the point, I I’m sure this sort of work is happening with Tom would have to look it up. But if we first get to the point where we could create a a competitive
208 00:18:41.281 ⇒ 00:18:43.390 Chang Ho: Hindi, who are doing other like.
209 00:18:43.390 ⇒ 00:18:43.820 Uttam Kumaran: Yeah.
210 00:18:43.820 ⇒ 00:18:45.409 Chang Ho: In in linguistic
211 00:18:46.030 ⇒ 00:19:05.550 Chang Ho: and even a fine tuned Lm, ultimately that would probably be all it will be, anyway. Then it’s it opens up the opportunity for a whole bunch of shit that would in terms of like the value that that product like that could have. I mean, the moats would be the important bit. How do you determine the moat? It would have to be to do with the data that we have
212 00:19:06.000 ⇒ 00:19:09.939 Chang Ho: to fine tune our models. But there’s there’s another thought. So idea I had was.
213 00:19:10.150 ⇒ 00:19:10.740 Chang Ho: yeah.
214 00:19:10.740 ⇒ 00:19:11.080 Uttam Kumaran: No here.
215 00:19:11.080 ⇒ 00:19:11.790 Chang Ho: We’ve discussed it.
216 00:19:11.790 ⇒ 00:19:12.810 Uttam Kumaran: Totally right.
217 00:19:12.960 ⇒ 00:19:16.279 Chang Ho: Doing stuff like this already in medicine, where
218 00:19:16.410 ⇒ 00:19:25.630 Chang Ho: you have a very general interface with with a very generative like general generative AI interface just to then be guided towards more specifically
219 00:19:25.980 ⇒ 00:19:29.090 Chang Ho: fine-tuned Llm. Or generative model.
220 00:19:29.180 ⇒ 00:19:42.590 Chang Ho: And that’s basically the the way in which things are going. Just it’s a sort of like your initial chat. Your initial chat Bot is ridiculously stupid in comparison to what it can then lead you to do in terms of the very fine tune process like, you know.
221 00:19:42.670 ⇒ 00:19:45.519 Chang Ho: hey? I wanna I wanna write
222 00:19:45.970 ⇒ 00:20:00.289 Chang Ho: discharge summary for this patient with tuberculosis. Meningitis like that’s a very specific use case. A very generous AI model will probably not perform as well as it should, for the information extraction needed, like what’s person and what’s not.
223 00:20:00.390 ⇒ 00:20:05.389 Chang Ho: Then, then, you know, by virtue of, then it leading me as a lay user
224 00:20:06.040 ⇒ 00:20:14.470 Chang Ho: towards a more specific model. The more specific ln that’s been finding tuned to that purpose. We kind of open up the possibility of creating these little
225 00:20:14.970 ⇒ 00:20:25.499 Chang Ho: villages of models basically, that could span something like this one health. I don’t know. I it was just that thought that came immediately to mind
226 00:20:26.020 ⇒ 00:20:30.070 Chang Ho: something that I’ve you know, not much time to ruminate over. But yeah.
227 00:20:30.070 ⇒ 00:20:54.099 Uttam Kumaran: Yeah, the thing that’s you know. I I spoke a lot with even about just doing business in India and things like that. The one thing that’s really difficult is business in India requires like a lot of handshakes and a lot of like knowing who to know. I think that’s the one thing I really took for granted here in the Us. Is that it’s very open for business in India. It’s a completely different process to get started and get trust and get funding. And I just think, like
228 00:20:54.100 ⇒ 00:21:06.399 Uttam Kumaran: it’s commonly it’s commonly like underestimated. And I would really underestimated how complicated it is. And that’s why, even for me, it’s even a bigger opportunity. Because I don’t think there’s many people that are bringing this technology there and have.
229 00:21:06.920 ⇒ 00:21:22.023 Uttam Kumaran: Like the level of understanding and depth that you can get like about these technologies here, and bring it to a place like India, where you ha have a local contact that’s able to get engineers get funding and get like actual like ability to apply this sort of technology
230 00:21:22.370 ⇒ 00:21:39.919 Uttam Kumaran: right? And that was the most interesting thing I talked to. A lot of. My my cousin is like, Oh, we do have a really strong link here is that he’s very ingrained and has a lot of connections in the Indian government and in the people that are making decisions to bring on these technologies. But he’s like, hey? The problem is like, people here don’t have access to a lot of them. They’re not has like
231 00:21:39.920 ⇒ 00:22:02.350 Uttam Kumaran: sophisticated in knowledge of like what’s possible if it’s getting closer. But even I was talking to the CTO. Of his company, and and we were having a really great conversation where I was like making a lot of great suggestions. But I could tell that, like even getting talent to do this sort of AI work, you would have to train, you know you. You definitely have to like, find people to train and bring up to speed on this sort of knowledge. So.
232 00:22:02.350 ⇒ 00:22:04.619 Chang Ho: Oh, definitely, you definitely would.
233 00:22:05.440 ⇒ 00:22:06.080 Uttam Kumaran: Yeah.
234 00:22:06.300 ⇒ 00:22:12.070 Chang Ho: It wouldn’t be, it would be quickly. Yeah, you’re right. It wouldn’t be quickly deployable from either human or technological standpoint. It wouldn’t be okay.
235 00:22:12.510 ⇒ 00:22:12.950 Uttam Kumaran: Have.
236 00:22:12.950 ⇒ 00:22:18.170 Chang Ho: Just off the shelf stuff that you can just quickly tweak on a on a training data set. No, not at all.
237 00:22:18.400 ⇒ 00:22:21.910 Uttam Kumaran: But I would say the the amount of impact you can make
238 00:22:22.652 ⇒ 00:22:40.909 Uttam Kumaran: is is is crazy like the amount of people there, and the amount of like both my, his, his parents, my aunt and my uncle are both doctors. My aunt is still practicing, and sees like tons of patients, and it’s crazy just to see them work and like how much
239 00:22:41.040 ⇒ 00:22:43.209 Uttam Kumaran: they have to go through is like.
240 00:22:43.390 ⇒ 00:22:45.750 Uttam Kumaran: there’s a different scales in the
241 00:22:45.810 ⇒ 00:22:47.590 Uttam Kumaran: like in the clinical world.
242 00:22:48.120 ⇒ 00:22:48.890 Chang Ho: Right?
243 00:22:49.300 ⇒ 00:22:57.650 Chang Ho: Right it absolutely is, and certainly compared to how it is here and in the Us. I mean, we complain about not having enough personnel, but it’s
244 00:22:58.060 ⇒ 00:23:00.250 Chang Ho: absolutely mad. What happens.
245 00:23:00.250 ⇒ 00:23:02.370 Uttam Kumaran: Yeah, absolutely.
246 00:23:02.370 ⇒ 00:23:05.710 Chang Ho: Yeah, it’s absolutely mad. The speed at which people have
247 00:23:06.270 ⇒ 00:23:10.979 Chang Ho: to work. I mean, yeah, you must just go home utterly shattered. And you must also have to just assume that you’ve
248 00:23:11.550 ⇒ 00:23:14.730 Chang Ho: you just doing the best with the bet, you know, with the, with the facilities you had.
249 00:23:14.730 ⇒ 00:23:15.979 Uttam Kumaran: But you get these other one.
250 00:23:15.980 ⇒ 00:23:21.613 Chang Ho: Because if you rue the small mistakes and be able and reflect on them, then you’re scuffled, aren’t you? I mean morally.
251 00:23:21.870 ⇒ 00:23:22.450 Uttam Kumaran: Yeah.
252 00:23:22.670 ⇒ 00:23:23.160 Chang Ho: Yeah.
253 00:23:23.160 ⇒ 00:23:26.020 Uttam Kumaran: No, you’re totally correct. And and but then, even like
254 00:23:26.160 ⇒ 00:23:38.370 Uttam Kumaran: feedback and everything again, you just don’t have any time. And and you’re right. But the one thing is they’re they’re so hard working like it’s I don’t know. They make it work, they manage it. I don’t think they have another option.
255 00:23:38.370 ⇒ 00:23:45.939 Chang Ho: Yeah, I didn’t know about you, mate, but I was. Yeah. The scale of everything is just at just another level. I mean, even
256 00:23:46.180 ⇒ 00:23:52.550 Chang Ho: looking at a number of people who apply, I appreciate they’re not like, for, like they’re not apples and apples for Iit.
257 00:23:52.900 ⇒ 00:23:53.490 Uttam Kumaran: Yes.
258 00:23:54.250 ⇒ 00:23:58.110 Chang Ho: Is just staggering 1,000 times one, you know.
259 00:23:58.240 ⇒ 00:24:00.940 Chang Ho: which is just not a level. So.
260 00:24:00.940 ⇒ 00:24:09.589 Uttam Kumaran: No, it’s absolutely like another level, and everything but everyone there. They described me like they described life as like survival.
261 00:24:09.630 ⇒ 00:24:19.869 Uttam Kumaran: like nobody there described life as like, Oh, you just like live, and you relax and you enjoy. They’re like everything is survival here like you really have to find a way to.
262 00:24:20.100 ⇒ 00:24:23.849 Uttam Kumaran: you know, survive in in whatever you do, you know. So
263 00:24:24.000 ⇒ 00:24:24.680 Uttam Kumaran: yeah.
264 00:24:24.680 ⇒ 00:24:25.450 Chang Ho: Yeah, yeah.
265 00:24:26.090 ⇒ 00:24:31.560 Chang Ho: completely. I think one of the so would you. Do you want to? Would you like to basically
266 00:24:31.850 ⇒ 00:24:36.419 Chang Ho: organize a quick meet with 1 one of the guys that one of the co-founders then.
267 00:24:36.830 ⇒ 00:24:38.549 Uttam Kumaran: Yeah, I think that would just see what.
268 00:24:38.550 ⇒ 00:24:44.689 Chang Ho: Opportunity exists. And I mean, obviously just, we could just keep it very noncommittal and see what what.
269 00:24:44.690 ⇒ 00:24:55.320 Uttam Kumaran: No, that’s 100. That’s actually what I told them is like, I think we should just chat. I mean I had. I was had. We may have maybe an hour or 2, but I also. This is my first time learning about their entire.
270 00:24:55.380 ⇒ 00:25:19.270 Uttam Kumaran: you know, business and all the things that they were interested in doing. And you know, it’s quite a lot. And so I think, having one focus conversation and having, you know, one of their their CTO join and just chatting. And that’s exactly even what I mentioned is like, hey, I think there’s a lot of opportunity to share knowledge and to just discuss. And yeah, I think it’ll be as casual as this, and we can just chat.
271 00:25:22.090 ⇒ 00:25:25.960 Chang Ho: I presume, actually, I mean, you know something. One of the Shocking, facts
272 00:25:26.340 ⇒ 00:25:30.049 Chang Ho: discovered just last weekend as a barcelona for this conference was.
273 00:25:30.070 ⇒ 00:25:39.215 Chang Ho: even in one of the most developed countries in the world. Germany, the hospitals have yet more than 90% of the hospitals do not have electronic health records.
274 00:25:40.060 ⇒ 00:25:40.950 Uttam Kumaran: Wow!
275 00:25:41.160 ⇒ 00:25:46.029 Chang Ho: Which is just mad. So, and that means that they’re they’re not going to be the last.
276 00:25:46.330 ⇒ 00:25:50.899 Chang Ho: And there were plenty of other countries like them, and India must be
277 00:25:50.990 ⇒ 00:25:52.659 Chang Ho: light years away
278 00:25:52.890 ⇒ 00:25:53.440 Chang Ho: from.
279 00:25:53.440 ⇒ 00:25:54.030 Uttam Kumaran: Yeah.
280 00:25:54.030 ⇒ 00:25:57.979 Chang Ho: That happening. And I just have to. I was just thinking about
281 00:25:58.260 ⇒ 00:26:03.050 Chang Ho: what kind of platforms then they need there. What kind of
282 00:26:03.430 ⇒ 00:26:05.499 Chang Ho: you know, I guess.
283 00:26:06.930 ⇒ 00:26:08.880 Chang Ho: system agnostic.
284 00:26:10.470 ⇒ 00:26:26.009 Chang Ho: either web or smartphone, or both platforms need to need to exist, to basically enable and incentivize both official and non official healthcare providers. And and I think there’s something else that this Md. Friend of your
285 00:26:26.570 ⇒ 00:26:27.599 Chang Ho: was going to tell you
286 00:26:27.640 ⇒ 00:26:32.484 Chang Ho: is that actually the great majority of prescriptions for antibiotics, for example, in India,
287 00:26:32.880 ⇒ 00:26:34.010 Chang Ho: because
288 00:26:34.720 ⇒ 00:26:39.690 Chang Ho: via pharmacies and via non official.
289 00:26:39.750 ⇒ 00:26:46.199 Chang Ho: non-trained, untrained healthcare providers, because there’s just such a humongous unmet need.
290 00:26:46.530 ⇒ 00:26:47.570 Uttam Kumaran: Wow. Okay.
291 00:26:47.570 ⇒ 00:26:58.470 Chang Ho: If you just rely on official people like officially trained individuals, and so actually to help combat that. And I think this would be something that maybe the Government, depending on the Government will be more interested in
292 00:26:58.520 ⇒ 00:26:59.870 Chang Ho: in it than not.
293 00:26:59.980 ⇒ 00:27:06.169 Chang Ho: is some process by which, or some platform by which we can incentivize, monetarily or otherwise.
294 00:27:06.510 ⇒ 00:27:15.119 Chang Ho: as some of these non official prescribers to to do the right thing, or to be a bit more consistent with the way they prescribe, or maybe not go straight for the big guns.
295 00:27:15.614 ⇒ 00:27:16.950 Chang Ho: That we end up.
296 00:27:17.370 ⇒ 00:27:23.140 Chang Ho: you know, basically polluting the idea of one health, you know, prescribing antibodies that way, too strong way, too.
297 00:27:23.140 ⇒ 00:27:23.980 Uttam Kumaran: Yeah.
298 00:27:23.980 ⇒ 00:27:26.250 Chang Ho: That kind of thing if some
299 00:27:26.360 ⇒ 00:27:29.759 Chang Ho: but could be compelling. It’s just one thought, but that’s one.
300 00:27:29.760 ⇒ 00:27:36.850 Uttam Kumaran: And that. And that’s why I think they’re trying to go after some of the developing countries, especially, I think they have a ability to implement these wide scale.
301 00:27:37.000 ⇒ 00:27:46.230 Uttam Kumaran: and they’re starting a little bit from scratch. And you’re right about you know how poor to more is. But at the same time. I think they have an ability to kind of start with technology.
302 00:27:46.557 ⇒ 00:27:55.829 Uttam Kumaran: And that was the one thing that was very interesting to me is like even seeing India like they skipped a lot of places, don’t have Wi-fi because they skipped Wi-fi. They never needed to have.
303 00:27:56.938 ⇒ 00:27:59.269 Uttam Kumaran: That sort of technology. Yeah. And.
304 00:27:59.270 ⇒ 00:27:59.609 Chang Ho: Yeah, I mean.
305 00:27:59.610 ⇒ 00:28:01.229 Uttam Kumaran: Like a cell. Everything is cellular.
306 00:28:02.460 ⇒ 00:28:12.119 Uttam Kumaran: you know, which was like that was my first like. Oh, I totally like, did not understand that, like they previously skipped like one whole step function of like
307 00:28:12.390 ⇒ 00:28:15.089 Uttam Kumaran: from how people got connected to the Internet. And
308 00:28:15.310 ⇒ 00:28:32.253 Uttam Kumaran: they’re also very quick to adopt. Like most that country runs on Whatsapp. It’s all it’s all. It’s all upi like further payments. So like I does give me some hope that, like we even we couldn’t do a lot of the stuff that they did across. How many people, you know?
309 00:28:32.870 ⇒ 00:28:33.510 Chang Ho: Yes.
310 00:28:33.970 ⇒ 00:28:34.500 Uttam Kumaran: So.
311 00:28:34.500 ⇒ 00:28:38.849 Chang Ho: Yeah, absolutely good time. So yeah, let’s organize a chat with them
312 00:28:38.930 ⇒ 00:28:46.530 Chang Ho: for one. The 2 I would love to be privy to some of the the email scheduling assistance.
313 00:28:46.530 ⇒ 00:28:50.500 Uttam Kumaran: Yes, yes, that’s a lot smaller. That’s a lot smaller. So.
314 00:28:50.500 ⇒ 00:28:52.360 Chang Ho: That’s a lot smaller, know, but that just has.
315 00:28:52.360 ⇒ 00:28:52.790 Uttam Kumaran: Yeah.
316 00:28:52.790 ⇒ 00:28:54.699 Chang Ho: Hilarious fun. Actually so.
317 00:28:54.700 ⇒ 00:28:56.270 Uttam Kumaran: I will totally share we should.
318 00:28:56.850 ⇒ 00:29:00.449 Uttam Kumaran: Stuff by next week, and I’ll I’ll send you some Demos and stuff. Yeah, it’s interesting.
319 00:29:00.450 ⇒ 00:29:02.470 Chang Ho: I’d love to just see how you built it, actually. And.
320 00:29:02.470 ⇒ 00:29:02.860 Uttam Kumaran: Yeah.
321 00:29:02.860 ⇒ 00:29:06.180 Chang Ho: And be inspired by it, because, as you know, tink around with code
322 00:29:06.720 ⇒ 00:29:14.639 Chang Ho: quite a fair bit myself, and just the idea of being able to engineer that into a nice user interface as well. Something I’d be very curious to learn from.
323 00:29:14.640 ⇒ 00:29:15.720 Uttam Kumaran: Yeah, definitely.
324 00:29:16.260 ⇒ 00:29:20.509 Chang Ho: Third thing, and just so that also I have a better, more profound understanding of
325 00:29:20.580 ⇒ 00:29:23.770 Chang Ho: the full stack development process which says, you know, is a kind of
326 00:29:23.850 ⇒ 00:29:26.120 Chang Ho: still remains a unique preserve of
327 00:29:26.660 ⇒ 00:29:28.739 Chang Ho: computer scientists and software developers.
328 00:29:29.040 ⇒ 00:29:29.750 Uttam Kumaran: Yes.
329 00:29:29.750 ⇒ 00:29:46.159 Chang Ho: I’m I might do a fair amount of this of the the machine learning building and the the regression modeling, and this is the nuances of that of those. But very little time is actually afforded, as you know, to making things look super pretty super agile.
330 00:29:46.730 ⇒ 00:29:52.189 Chang Ho: making things run as efficiently as possible in terms of the background. Yeah, that kind of.
331 00:29:52.190 ⇒ 00:29:52.630 Uttam Kumaran: Yeah.
332 00:29:52.630 ⇒ 00:29:59.995 Chang Ho: Is, still remains a unique preserver can be the scientists, not although I would love to spend time doing it, learning, doing that. I just don’t have enough time.
333 00:30:00.420 ⇒ 00:30:21.429 Uttam Kumaran: No, no, totally. And it’s like, also, how do you take something that runs locally and like, okay, we wanna make this something that we can put on the Internet and like share and have some some sort of scalability, you know. And that’s even the part that I’m learning in this, because for me in this this project, like I, I’m doing all the AI logic, and that for me is fun. But I’m like I can get this running in the Jupiter notebook on my laptop
334 00:30:21.450 ⇒ 00:30:38.689 Uttam Kumaran: like, how does this get deployed like? How do you deploy this? An Api? They can get called by services, and that’s where, like one of the other guys, he works for Snapchat, and he’s like that’s what he does at his job. And he does a bunch of like back end sort of like application development. And I’m like perfect team like, we gotta we gotta run through.
335 00:30:38.690 ⇒ 00:30:39.560 Chang Ho: Bless us!
336 00:30:40.510 ⇒ 00:30:44.269 Chang Ho: Think there’s Uto! Do you know anyone who works in
337 00:30:44.700 ⇒ 00:30:46.360 Chang Ho: like clinical woe.
338 00:30:46.920 ⇒ 00:30:52.089 Chang Ho: pro operations, or even just operations, or double, or the more managerial side of any healthcare.
339 00:30:52.810 ⇒ 00:30:53.460 Uttam Kumaran: So.
340 00:30:53.763 ⇒ 00:30:54.370 Chang Ho: The Us.
341 00:30:54.370 ⇒ 00:30:59.249 Uttam Kumaran: I I know last time I was planning on meeting with some people
342 00:30:59.260 ⇒ 00:31:13.790 Uttam Kumaran: in the medical side here, I just ran out of time. I didn’t even get to ask last time we met. About that. The only person I met is the folks that I’ll be connecting you with on us with in India. But here I
343 00:31:14.010 ⇒ 00:31:19.539 Uttam Kumaran: I don’t know. I I think I’m like one degree from couple of people, but I need to
344 00:31:19.590 ⇒ 00:31:21.660 Uttam Kumaran: kind of like. Ask around again.
345 00:31:21.909 ⇒ 00:31:29.810 Uttam Kumaran: I don’t know. I mean like I don’t know any that we’d be surprised. I don’t know any doctors. Everybody I knew who was going to be a doctor, decided to drop out.
346 00:31:30.873 ⇒ 00:31:31.696 Chang Ho: So.
347 00:31:32.520 ⇒ 00:31:34.160 Uttam Kumaran: All those, all those.
348 00:31:34.160 ⇒ 00:31:34.770 Chang Ho: Good! Move!
349 00:31:34.770 ⇒ 00:31:37.249 Uttam Kumaran: Lawyers. They some lawyers made it through. But
350 00:31:38.420 ⇒ 00:31:39.390 Uttam Kumaran: oh, yeah.
351 00:31:39.390 ⇒ 00:31:40.650 Chang Ho: Good! Move!
352 00:31:40.790 ⇒ 00:31:51.210 Uttam Kumaran: I’m I’m I’m looking forward to getting more involved here in like the ut health and things like that. But it’s just been slow, kind of getting ingrained in like more things locally here. So
353 00:31:51.900 ⇒ 00:31:52.610 Uttam Kumaran: excuse me.
354 00:31:52.610 ⇒ 00:32:06.126 Chang Ho: I just the reason the reason why I thought mentioned is because, one of the most exciting, I think aspects of I don’t know, I think one of the more exciting players I’ve seen come out of the woodwork in the in Europe has been
355 00:32:07.280 ⇒ 00:32:12.009 Chang Ho: basically basically focused on it’s called up the hill. I think it’s called uphill health.
356 00:32:12.860 ⇒ 00:32:17.229 Chang Ho: and they’ve won several 1 million euros from Portugal and Spain.
357 00:32:17.450 ⇒ 00:32:23.610 Chang Ho: and essentially the entire premise is one that I think we could
358 00:32:23.970 ⇒ 00:32:26.440 Chang Ho: do not necessarily emulate
359 00:32:26.960 ⇒ 00:32:35.239 Chang Ho: things. I don’t know how what they know workings on, but I think the overall very broad level idea is a good one, and that is many hospitals operate
360 00:32:35.490 ⇒ 00:32:41.840 Chang Ho: clinical and patient workflows. What does what do they mean? Or clinical, basically patient?
361 00:32:42.488 ⇒ 00:32:44.759 Chang Ho: Management flows. What do they mean.
362 00:32:44.960 ⇒ 00:32:51.659 Chang Ho: So, for example, each time when you, if someone were to be referred by another doctor because they think someone has cancer.
363 00:32:52.530 ⇒ 00:32:56.989 Chang Ho: there’ll be an entire operational workflow oriented around cancer
364 00:32:57.260 ⇒ 00:32:57.810 Chang Ho: and.
365 00:32:57.810 ⇒ 00:32:58.270 Uttam Kumaran: Yielding.
366 00:32:58.270 ⇒ 00:33:06.900 Chang Ho: Becomes, it becomes more and more specific, and that that is as broad as you can imagine it to be. So. That includes scheduling
367 00:33:07.580 ⇒ 00:33:09.310 Chang Ho: that includes.
368 00:33:09.600 ⇒ 00:33:11.810 Chang Ho: you know, triaging for the dock.
369 00:33:12.030 ⇒ 00:33:18.849 Chang Ho: maybe by the dock, maybe Semi automatically, or by the sum by a senior nurse, for the more for the less complex cases.
370 00:33:19.220 ⇒ 00:33:24.249 Chang Ho: That means ensuring that they end up going down the right pathway with respect to it, like
371 00:33:24.850 ⇒ 00:33:26.500 Chang Ho: accessory help.
372 00:33:26.890 ⇒ 00:33:33.849 Chang Ho: So we’re talking adjunctive help from local palliative care services or hospice, or what have you?
373 00:33:34.130 ⇒ 00:33:37.589 Chang Ho: And so it becomes this entire process that
374 00:33:37.680 ⇒ 00:33:42.460 Chang Ho: hospitals and healthcare entities have to abide by
375 00:33:42.600 ⇒ 00:33:44.640 Chang Ho: to ensure that they can compete
376 00:33:44.680 ⇒ 00:33:46.459 Chang Ho: and but and provide
377 00:33:47.280 ⇒ 00:33:58.019 Chang Ho: what is deemed to be the Gold Standard service for their patients. Now, naturally, like any system where there’s a lot of manual labor, and there still is a lot of manual labor in this front. If it’s not almost 100% manual
378 00:33:58.440 ⇒ 00:34:01.960 Chang Ho: scheduling and everything else happens via secretaries.
379 00:34:02.554 ⇒ 00:34:04.880 Chang Ho: Triaging happens manually via docs.
380 00:34:05.180 ⇒ 00:34:06.549 Chang Ho: and people do get met
381 00:34:07.330 ⇒ 00:34:23.900 Chang Ho: end up with bogus referrals and all sorts of things, and life gets in the way. Patients are, you know, have been referred to that several months ago, but still haven’t been seen for one reason or another, and it means that you kind of don’t, even though that is, ends up on the pile for
382 00:34:24.250 ⇒ 00:34:26.200 Chang Ho: secretaries and other people to try and
383 00:34:26.989 ⇒ 00:34:34.189 Chang Ho: follow up. You can imagine there’s a great bit of variability in that. And some of that could be, yeah, you’re thinking about this. With the email, Scheduler
384 00:34:34.676 ⇒ 00:34:46.390 Chang Ho: could be automatable to some extent, and I’m certain there must be hospitals that are interested in thinking about that, particularly if they’ve been struggling to hire on that front like hire.
385 00:34:46.550 ⇒ 00:34:47.170 Uttam Kumaran: Yeah.
386 00:34:47.330 ⇒ 00:34:54.079 Chang Ho: Individuals, and not just that. But I think I wonder as to whether there’s any interest in in in the idea of
387 00:34:54.510 ⇒ 00:34:57.330 Chang Ho: like you know, like a virtual.
388 00:34:58.380 ⇒ 00:35:03.539 Chang Ho: like a virtual healthcare management employee that you want could sort of.
389 00:35:03.540 ⇒ 00:35:04.040 Uttam Kumaran: Yeah.
390 00:35:04.040 ⇒ 00:35:06.170 Chang Ho: Create. Yeah, well, basically, we’re just.
391 00:35:06.170 ⇒ 00:35:10.640 Uttam Kumaran: I kind of go function by function. But I think email and email scheduling is one that’s like
392 00:35:10.750 ⇒ 00:35:25.999 Uttam Kumaran: very agnostic, like everybody deals this problem as soon as you become someone that needs to meet with other people. And you’re meeting online, or you’re booking this sort of back and forth and chat bots and things can handle it right now. But some of these may happen through referral through email and.
393 00:35:26.000 ⇒ 00:35:26.540 Chang Ho: Yes.
394 00:35:26.540 ⇒ 00:35:30.589 Uttam Kumaran: You kind of want to make that possible. So we’re kind of starting with like.
395 00:35:30.610 ⇒ 00:35:40.129 Uttam Kumaran: just anybody. But I think one. The second thing is we want to try to do is up the ability for you to run this on your own domain. So like, first, from security standpoint, you can
396 00:35:40.250 ⇒ 00:36:07.890 Uttam Kumaran: right now, it’s gonna email like us, which is like an external company. But you can almost white label the service. So it it’ll still get processed through your email domain and your email server. And then it’ll still have the same action. And then, of course, expanding it to more use case based things. A couple of things we thought of were more consumer based. But I think even Intra company, there’s probably things that you can do or have specific workflows that go from email to like, oh, it’s this sort of request.
397 00:36:07.890 ⇒ 00:36:22.720 Uttam Kumaran: But the thing is, I always thought the medium of email is actually more important than having someone go through or text is more important than having them go through a website or another flow, like the reason why I really like this product is, there’s no ui, the product.
398 00:36:23.380 ⇒ 00:36:26.890 Uttam Kumaran: the demo, the product is that. And then maybe there’s a ui to set your like
399 00:36:26.900 ⇒ 00:36:29.709 Uttam Kumaran: your any sort of small settings.
400 00:36:30.060 ⇒ 00:36:30.410 Chang Ho: Yeah.
401 00:36:30.410 ⇒ 00:36:40.510 Uttam Kumaran: But otherwise, like you basically out, one of the ideas we have is to have the entire onboarding flow happen through email emailing back and forth with the bot. Right? And I think that’s the thing where
402 00:36:40.630 ⇒ 00:36:51.101 Uttam Kumaran: a lot of stuff I’m seeing in the market continues to resolve back to like Whatsapp or text or email. That’s where everybody continues to be. You don’t want another app. You don’t want another workflow, right? So.
403 00:36:51.370 ⇒ 00:36:52.810 Chang Ho: No.
404 00:36:53.360 ⇒ 00:36:56.159 Chang Ho: not at all. And you kind of have to wonder if
405 00:36:56.650 ⇒ 00:37:06.640 Chang Ho: that’s actually the way, we kind of you know, we should be going as well in terms of opening up scheduling, and modernizing, scheduling, scheduling, for example, with the patient
406 00:37:07.110 ⇒ 00:37:16.249 Chang Ho: and some of their clinicians, and so on. So yeah, they don’t have to then subscribe to yet another platform, or whatever they can simply just send a message via Whatsapp, or whatever it is, an interface with.
407 00:37:16.250 ⇒ 00:37:23.010 Uttam Kumaran: I mean, if you look at India, everything’s happening through Whatsapp, every single thing like you’re able to re up your
408 00:37:23.040 ⇒ 00:37:30.799 Uttam Kumaran: your your cell phone like minutes, you’re able to pay all sorts of different things. I think in the Us. We’re behind in that we
409 00:37:31.320 ⇒ 00:37:36.290 Uttam Kumaran: I don’t know. We haven’t really like adopted that level of like frictionless stuff. But
410 00:37:36.510 ⇒ 00:37:42.953 Uttam Kumaran: I mean, I I’m connected with this guy that runs a company here in Austin that allows you to buy products
411 00:37:43.350 ⇒ 00:37:53.029 Uttam Kumaran: via text messaging. So you can have your, you could have a company. You could have your text message, campaign text, a bunch of people. And then you people can buy your products directly from text.
412 00:37:53.595 ⇒ 00:37:57.169 Uttam Kumaran: And he’s seeing a lot of growth. So I think SMS
413 00:37:57.480 ⇒ 00:38:02.900 Uttam Kumaran: after email is one of the big frontiers for like this sort of stuff. But again, everybody has a phone. Now.
414 00:38:03.100 ⇒ 00:38:13.139 Uttam Kumaran: the one thing is, you can bet on this, people have phones and access to Internet, especially in India. They have, like 5G for like ₹1020, you could get 5G and.
415 00:38:13.140 ⇒ 00:38:13.770 Chang Ho: Wow!
416 00:38:14.150 ⇒ 00:38:16.109 Uttam Kumaran: So I think it’s like
417 00:38:16.190 ⇒ 00:38:18.779 Uttam Kumaran: everybody has access to Internet. But then it’s like
418 00:38:18.820 ⇒ 00:38:21.599 Uttam Kumaran: everything comes preloaded with email and text.
419 00:38:21.610 ⇒ 00:38:27.439 Uttam Kumaran: Rely on those mediums to get your thing done. And the nice thing about those mediums is they’re all really conversation based
420 00:38:27.500 ⇒ 00:38:28.780 Uttam Kumaran: right? So
421 00:38:28.990 ⇒ 00:38:35.230 Uttam Kumaran: you don’t really need to have a ui for some of these, if you can have it come through a conversation. So I think there’s a tons of opportunity there.
422 00:38:36.480 ⇒ 00:38:37.250 Uttam Kumaran: Yeah.
423 00:38:38.830 ⇒ 00:38:41.677 Chang Ho: No, it does sound really good. I think.
424 00:38:42.700 ⇒ 00:38:46.889 Chang Ho: I think it’ll be. It’ll be. It’ll be interesting to explore
425 00:38:47.390 ⇒ 00:38:54.210 Chang Ho: all that space. Yeah, just for more operational languages, because then it then you don’t have to worry so much about all the
426 00:38:54.280 ⇒ 00:38:55.709 Chang Ho: the issues with.
427 00:38:56.187 ⇒ 00:39:06.839 Chang Ho: you know, diagnostics and regulation there, and there are lots of different people who are trying their hand at assisting diagnostics and being clinical health tools, etc. I’m not.
428 00:39:07.250 ⇒ 00:39:15.030 Chang Ho: I’m I’m not so sure there are that many companies are trying to help with the opposite side of things. You’re like with the actual managerial and operational side of things.
429 00:39:15.330 ⇒ 00:39:20.610 Chang Ho: And I wondered as to whether there’s value in that, basically, because if it means that, for example, if a hospital
430 00:39:20.620 ⇒ 00:39:26.820 Chang Ho: doesn’t have to, is struggling to hire 2 or 3 people who could do that kind of work, or they struggle to
431 00:39:27.580 ⇒ 00:39:33.770 Chang Ho: so well, it is costly to train them up to a point where they? They are at the level they need them to be.
432 00:39:34.125 ⇒ 00:39:40.019 Chang Ho: That, I wonder as to where you know, as you know, like a subscription model, which is obviously the classic thing to hook them by.
433 00:39:40.040 ⇒ 00:39:45.000 Chang Ho: where you severely undercut even just the hiring of one person per annum
434 00:39:45.580 ⇒ 00:39:46.649 Chang Ho: to be on it.
435 00:39:46.650 ⇒ 00:39:54.199 Uttam Kumaran: Exactly, but it’s also a lot of people don’t don’t have assistance, and they never and they never hire assistant right? And so.
436 00:39:54.380 ⇒ 00:39:59.589 Uttam Kumaran: allowing for this to be possible for 10 a month.
437 00:39:59.700 ⇒ 00:40:03.419 Uttam Kumaran: Yeah, when, again, our math is like an assistant is maybe like
438 00:40:03.830 ⇒ 00:40:06.019 Uttam Kumaran: like the deep business system, maybe like
439 00:40:06.280 ⇒ 00:40:07.939 Uttam Kumaran: like 10 bucks an hour
440 00:40:08.650 ⇒ 00:40:12.490 Uttam Kumaran: and then whatever that amounts to 8 h per day instead. Now.
441 00:40:13.260 ⇒ 00:40:15.250 Uttam Kumaran: And most assistants do scheduling.
442 00:40:16.160 ⇒ 00:40:28.200 Uttam Kumaran: They do this sort of schedule. So that’s the thing we’re kind of thinking about is like, okay, how do we just take care of that one thing and then think about, okay, what are other email based tasks that happen so often in this sort of assistant
443 00:40:28.280 ⇒ 00:40:41.709 Uttam Kumaran: level function. But I think scaling it. What what would be amazing about this is like not just making it available for us for consumers, but making it available for enterprises to say, like, you can roll this on your own.
444 00:40:41.750 ⇒ 00:40:44.690 Uttam Kumaran: You know, you could have this within your your email domain. So.
445 00:40:45.100 ⇒ 00:40:45.810 Chang Ho: Yes.
446 00:40:46.200 ⇒ 00:40:47.939 Chang Ho: yeah, yeah. Yeah. Totally. Man.
447 00:40:48.800 ⇒ 00:40:50.220 Chang Ho: that some.
448 00:40:50.230 ⇒ 00:40:54.770 Chang Ho: The other thing I was gonna say to actually was one of the other areas I’m interested in.
449 00:40:55.259 ⇒ 00:40:58.550 Chang Ho: I don’t know if you’ve had if you’ve had any
450 00:41:01.759 ⇒ 00:41:03.239 Chang Ho: Financial clients
451 00:41:03.670 ⇒ 00:41:07.069 Chang Ho: come up for you where where you had to deal with financial jargon.
452 00:41:08.060 ⇒ 00:41:08.520 Uttam Kumaran: A.
453 00:41:08.520 ⇒ 00:41:10.409 Chang Ho: Yeah, from an investment.
454 00:41:10.410 ⇒ 00:41:21.460 Uttam Kumaran: Yeah, I I have one client that we’re working on right now that they’re building almost like a linkedin for connecting asset managers and wealth managers together.
455 00:41:22.530 ⇒ 00:41:22.960 Chang Ho: Well.
456 00:41:23.287 ⇒ 00:41:48.169 Uttam Kumaran: And and they’re building almost like a linkedin for connecting them together. And basically, the other part of their business is what I’ve been working on is helping them get sec. And like the securities related data into a warehouse where they can run AI on top of it. And basically what they’re gonna do is they’re both creating first party data. When people sign up to that platform are are telling their investment preferences
457 00:41:48.170 ⇒ 00:42:09.330 Uttam Kumaran: which the wealth manager trying to connect with the asset managers. And they’re trying to like get their money to manage. Asset managers have a lot of the wealth managers have money to manage. Asset managers go deploy that capital. So how does that connection process works? And then the second is actually taking all that data and spinning it out and creating like a data business where you can repackage and sell that data to people who want to buy.
458 00:42:09.330 ⇒ 00:42:26.000 Uttam Kumaran: you know, access to those managers and in the Us. There’s only a there’s only about like 500,000 registered managers. So, and it’s a revolving set of people that come in and out of that. And it’s like a registration process. But every one of them has different specialties, different skill sets. And
459 00:42:26.000 ⇒ 00:42:42.459 Uttam Kumaran: there’s always like matching process that happens. And so they’re building a platform to do that. One part that we’re do. What I’m helping them with is bringing in all this sec data and running rag on top. So basically, with any person you can associate that person with the firm. And then you want to find any document
460 00:42:42.480 ⇒ 00:42:56.719 Uttam Kumaran: associated with that firm and then we’re kind of packaging that data up and selling. But it’s a little bit of like a smaller thing. Not. I’m not working with anybody that’s on a lot larger financial institutions. These guys are just a small startup. So.
461 00:42:58.650 ⇒ 00:43:07.120 Chang Ho: I guess. Yeah, I mean, I don’t know. If you were, I don’t know if that would be something that I could get involved, not this particular thing, but if it’s there’s one where I get to be
462 00:43:07.570 ⇒ 00:43:08.350 Chang Ho: opposed.
463 00:43:08.350 ⇒ 00:43:10.000 Uttam Kumaran: Yeah, I’m even. Gonna I’ll just.
464 00:43:10.000 ⇒ 00:43:11.280 Chang Ho: With financial stuff. Yeah.
465 00:43:11.280 ⇒ 00:43:14.034 Uttam Kumaran: Yeah, I’ll just send you. I’ll just send you
466 00:43:14.550 ⇒ 00:43:16.210 Uttam Kumaran: these guys is link.
467 00:43:16.330 ⇒ 00:43:25.949 Uttam Kumaran: And I mean, we’re continuing to work on AI stuff for them, and again I keep you privy of stuff we’re doing. We’re just now testing the new AI features.
468 00:43:26.560 ⇒ 00:43:27.160 Uttam Kumaran: On
469 00:43:28.390 ⇒ 00:43:37.869 Uttam Kumaran: on Snowflake that just came out where you can. You can run asset link. You can run AI directly in snowflake
470 00:43:38.133 ⇒ 00:43:47.149 Uttam Kumaran: to pull documents from like any sort of data store. So we’ve just started testing all that out. I’d be happy. Yeah, I’d be happy to share that with you. And this is the company.
471 00:43:47.996 ⇒ 00:43:50.504 Uttam Kumaran: Www, dot asset.
472 00:43:51.590 ⇒ 00:43:53.410 Chang Ho: I think they are. Yeah.
473 00:43:55.850 ⇒ 00:43:56.879 Chang Ho: where they are.
474 00:43:58.620 ⇒ 00:44:02.550 Chang Ho: Yeah, this would be really good. It’s just, I’m I’m trying to become more and more versed
475 00:44:02.710 ⇒ 00:44:07.179 Chang Ho: in financial jargon. I feel like it’s sort of one aspect of business that just
476 00:44:07.630 ⇒ 00:44:10.640 Chang Ho: so not have any. So I’ve so knowledge on at the moment.
477 00:44:10.650 ⇒ 00:44:12.260 Chang Ho: I’d love to be able to see
478 00:44:12.350 ⇒ 00:44:17.419 Chang Ho: like how. Yeah, how much I need to basically be able to learn to interface with people like this.
479 00:44:17.840 ⇒ 00:44:18.490 Chang Ho: And.
480 00:44:18.490 ⇒ 00:44:20.290 Uttam Kumaran: Yeah, no, no, it’s Google. It’s
481 00:44:21.217 ⇒ 00:44:30.989 Uttam Kumaran: this is, I have a little bit of background in finance from school, and I’ve been able to kind of work my way through a lot of this. But again, in financial management, too, there’s a ton of data and a ton of.
482 00:44:31.050 ⇒ 00:44:34.095 Uttam Kumaran: you know applications for this sort of stuff, too. So
483 00:44:34.450 ⇒ 00:44:35.630 Uttam Kumaran: and
484 00:44:35.640 ⇒ 00:44:41.339 Uttam Kumaran: I think the nice thing about working in healthcare, though healthcare and government is that those clients are very sticky.
485 00:44:41.390 ⇒ 00:44:54.450 Uttam Kumaran: like the odds of them, like kind of removing you is is tough compared to like. If you work for a startup or someone like, and is in the market for shopping. So that’s what I tried to think a lot about these days, too, is like, are we getting clients that
486 00:44:54.460 ⇒ 00:44:59.519 Uttam Kumaran: are going to really value the partnership and want to leverage us for more and more things over time.
487 00:44:59.800 ⇒ 00:45:02.577 Uttam Kumaran: And so the other, the other person I met with is
488 00:45:03.180 ⇒ 00:45:10.309 Uttam Kumaran: a family friend of mine. He’s the like number 3 at the the company in India that
489 00:45:10.320 ⇒ 00:45:14.950 Uttam Kumaran: did the did. The Indian Coronavirus vaccine.
490 00:45:15.600 ⇒ 00:45:16.680 Chang Ho: Oh, right. Wow!
491 00:45:17.081 ⇒ 00:45:19.088 Uttam Kumaran: Forgot. What are they called?
492 00:45:20.660 ⇒ 00:45:26.220 Uttam Kumaran: Let me look at what the link he sent me. He’s like the head of operations like ahead of hr there.
493 00:45:27.255 ⇒ 00:45:29.790 Uttam Kumaran: It was called.
494 00:45:30.456 ⇒ 00:45:31.369 Uttam Kumaran: Let’s see.
495 00:45:33.750 ⇒ 00:45:35.700 Uttam Kumaran: And I was talking to him about like
496 00:45:37.920 ⇒ 00:45:40.170 Uttam Kumaran: I was talking to him about like everything
497 00:45:40.230 ⇒ 00:45:43.110 Uttam Kumaran: that they were working on, and it was insane.
498 00:45:44.534 ⇒ 00:46:02.515 Uttam Kumaran: And he was like he said, he’s been working there for you. It’s like a family owned company, and they’ve been working there since, like before, Covid. And he’s just talking about how insane it wasn’t. He’s like I met every Prime Minister, and they were the only Indian based vaccine, I think, next to Johnson and Johnson, or
499 00:46:02.820 ⇒ 00:46:03.920 Chang Ho: Yes, wow!
500 00:46:03.920 ⇒ 00:46:08.909 Uttam Kumaran: Yeah, I ha! I’ll have to get the company and send it to you. I forgot what exactly it was called.
501 00:46:09.390 ⇒ 00:46:10.060 Chang Ho: Wow!
502 00:46:10.520 ⇒ 00:46:14.689 Uttam Kumaran: Oh, brought biotech BHAR. A. T. Biotech.
503 00:46:15.240 ⇒ 00:46:17.250 Chang Ho: Right for biotech.
504 00:46:18.710 ⇒ 00:46:19.450 Chang Ho: I’m sorry.
505 00:46:19.450 ⇒ 00:46:20.899 Uttam Kumaran: I’ll send this to you.
506 00:46:22.680 ⇒ 00:46:27.399 Chang Ho: I’ll do that. Do that return. I’ll also try and inquire with other people that
507 00:46:27.620 ⇒ 00:46:34.419 Chang Ho: know over in the Us. Hospital sectors about the operational side of things. I didn’t know if that would be something. You might be interested.
508 00:46:34.420 ⇒ 00:46:35.410 Uttam Kumaran: Yeah. Totally.
509 00:46:35.694 ⇒ 00:46:37.119 Chang Ho: Small project to try out
510 00:46:37.498 ⇒ 00:46:46.430 Chang Ho: but just to get a notion of whether there is any value here, or whether the people who are looking for this sort of solution, but have yet to find a small business like brain forged.
511 00:46:46.430 ⇒ 00:46:46.880 Uttam Kumaran: Yeah.
512 00:46:46.880 ⇒ 00:46:54.079 Chang Ho: Like this, because ultimately, as you know, the the activation energy, financially speaking, for
513 00:46:54.200 ⇒ 00:46:59.729 Chang Ho: bringing in the likes of Bcg or Mckinsey, you know, quantum black is massive.
514 00:47:00.060 ⇒ 00:47:00.450 Uttam Kumaran: Yeah.
515 00:47:00.450 ⇒ 00:47:11.209 Chang Ho: And so for for hospitals to justify it. Yeah, maybe some of the likes of mass Gen. Could, or someone someone like you know, Utc. But most
516 00:47:11.420 ⇒ 00:47:19.819 Chang Ho: I would imagine, most hospital networks small enough. The local ones that you still need. This sort of help will not be running off to a.
517 00:47:20.560 ⇒ 00:47:21.260 Uttam Kumaran: Yes.
518 00:47:21.260 ⇒ 00:47:27.349 Chang Ho: Consultancy firm will not necessarily, you know, and they’ll probably end up relying on so much
519 00:47:27.490 ⇒ 00:47:32.890 Chang Ho: smaller vendors that help to automate stuff, but are probably may well not have the same
520 00:47:33.040 ⇒ 00:47:47.227 Chang Ho: flex as a as a more modern and small, like startup mic in terms of creating solutions that are platform agnostic, you know, we’re talking sort of garage projects that basically come to ferocious, you know, 30 years ago, you know, 20
521 00:47:47.971 ⇒ 00:47:53.089 Chang Ho: clunking away with. So yeah, I think that, you know that’s something else I quite like to
522 00:47:53.140 ⇒ 00:48:02.930 Chang Ho: explore, just based on my experiences how things are here and across Europe. I can’t imagine it to be that much more advanced in the Us. Considering
523 00:48:04.360 ⇒ 00:48:18.709 Chang Ho: that even the Us. In terms of healthcare it healthcare, lies in the bottom. 5 of all industries for how much money the percentage of their revenue they reinvest in healthcare, digital infrastructure.
524 00:48:19.820 ⇒ 00:48:20.600 Uttam Kumaran: Wow!
525 00:48:20.740 ⇒ 00:48:22.499 Chang Ho: I didn’t know if you knew. Know that. Yeah.
526 00:48:22.500 ⇒ 00:48:22.920 Uttam Kumaran: Now.
527 00:48:23.340 ⇒ 00:48:27.499 Chang Ho: They’re the bottom 5. I mean, we’re caught. Yeah across the board. So I
528 00:48:27.720 ⇒ 00:48:38.180 Chang Ho: know that there is opportunity. There. You say, there’s also sticky software and issues. So if you come in as a small, agile, relatively low cost to start entity that can
529 00:48:38.210 ⇒ 00:48:47.400 Chang Ho: that can basically help them with a solution where they don’t have to pay a obsc amount of money for it to work and do a whole lot of digital, you know, like infrastructural gymnastics, I
530 00:48:47.940 ⇒ 00:49:01.059 Chang Ho: the gymnastics to make the thing work. Then, you know, I would imagine there’s quite a wealth of opportunity just like there is here at the moment, and that’s something that we’re, you know. I’m actively looking at here with a local hospital, too, because I.
531 00:49:01.060 ⇒ 00:49:02.050 Uttam Kumaran: Yeah, mode.
532 00:49:02.050 ⇒ 00:49:04.450 Chang Ho: There’s there’s quite a lot that
533 00:49:04.490 ⇒ 00:49:12.810 Chang Ho: that could be done, even with poor, even with hospitals that are poor. I can give you examples. In Oxford there were
534 00:49:13.170 ⇒ 00:49:16.800 Chang Ho: more than 5 vacant positions for
535 00:49:17.817 ⇒ 00:49:23.669 Chang Ho: people working on these clinical pathways, these patient pathways, schedulers, etc.
536 00:49:23.730 ⇒ 00:49:27.570 Chang Ho: But they basically just said, this year hands up. We don’t
537 00:49:27.940 ⇒ 00:49:38.930 Chang Ho: enough money to hire. Like all 5 of these people this year, we’re just not gonna hire, although we need them. So basically, there is a demand, but they can’t afford to pay the amount it takes for human capital to be paid.
538 00:49:40.710 ⇒ 00:49:41.280 Uttam Kumaran: Interesting.
539 00:49:41.280 ⇒ 00:49:47.419 Chang Ho: So if you can cut well below that and say, Look, we can come in with a solution that basically will act like fine.
540 00:49:48.170 ⇒ 00:49:51.860 Chang Ho: But for the price of maybe even less than one for an annual subscription, they.
541 00:49:51.860 ⇒ 00:49:52.670 Uttam Kumaran: Go into it.
542 00:49:52.670 ⇒ 00:49:59.089 Chang Ho: So you want to a model that could basically, yeah, proliferate. So that was just the thought, I, I, yeah, that I’m current experience.
543 00:49:59.090 ⇒ 00:50:06.040 Uttam Kumaran: No, I mean, if there’s people here that would be amazing. And even again, just to have casual conversations. That’s the one thing that a lot of people don’t want to get sold to.
544 00:50:06.070 ⇒ 00:50:15.119 Uttam Kumaran: and the thing that I found a lot easier to do is just like, have conversations with people, explain what we know about what’s possible, and then see how things go. So live.
545 00:50:15.120 ⇒ 00:50:15.660 Chang Ho: Yes.
546 00:50:15.660 ⇒ 00:50:22.439 Uttam Kumaran: Better process of doing things. And I think people are. People are used to just talking to sales people or vendors.
547 00:50:22.787 ⇒ 00:50:27.050 Uttam Kumaran: And they just wanna hear like from the technology side, like what’s possible.
548 00:50:27.160 ⇒ 00:50:34.779 Uttam Kumaran: And definitely, if there’s a champion in those businesses that’s like really keen on getting something done. There’s a lot of opportunity so
549 00:50:36.670 ⇒ 00:50:37.630 Uttam Kumaran: nice one.
550 00:50:38.010 ⇒ 00:50:42.230 Chang Ho: I think so, too. I’ll let. I’ll keep you abreast, I mean, like, we’re obviously we’re on slack, anyway.
551 00:50:42.654 ⇒ 00:50:46.879 Chang Ho: Are you doing this bi weekly meeting as well, which I’ll try and.
552 00:50:47.170 ⇒ 00:50:47.570 Uttam Kumaran: We.
553 00:50:47.570 ⇒ 00:50:48.170 Chang Ho: In, on.
554 00:50:48.170 ⇒ 00:51:08.969 Uttam Kumaran: We we we were, but IAI think Brian is kind of in and out on stuff. I mean, I’d be happy just to chat with you every week or every other week or so, and again, it’s good for me, because if I have at least on the calendar, I’ll attend. Otherwise the weeks get a bit crazy. But it’s also it’s also good motivation. I mean, at minimum. I’m gonna set up a meeting for next week for us to talk with the India folks and.
555 00:51:08.970 ⇒ 00:51:10.369 Chang Ho: Yeah, let’s do that. First. Let’s just.
556 00:51:10.370 ⇒ 00:51:20.300 Uttam Kumaran: Let’s do, let’s do that, and then I do owe an a follow up to my family friend at Barad. He! He’s they’re opening a bunch. So
557 00:51:20.350 ⇒ 00:51:29.210 Uttam Kumaran: I mean, it’s like a very crazy store. But basically there, they had to ramp up all the manufacturing of this vaccine, and they’re opening a bunch of new facilities in India, and
558 00:51:29.310 ⇒ 00:51:35.230 Uttam Kumaran: they’re doing a bunch of stuff in IoT. And he’s like there may be an opportunity to bring you guys in for some data stuff
559 00:51:35.240 ⇒ 00:51:44.160 Uttam Kumaran: he’s like, when you guys are, he’s like, when you guys are stable like, give it a thought and like, think about like where their opportunities might be, and so I owe him. I owe him a little bit of like.
560 00:51:44.160 ⇒ 00:51:46.119 Chang Ho: Oh, yes, please.
561 00:51:46.120 ⇒ 00:51:48.339 Uttam Kumaran: Yeah. And here, look at.
562 00:51:48.340 ⇒ 00:51:49.389 Chang Ho: Let’s do something. There.
563 00:51:49.560 ⇒ 00:51:58.130 Uttam Kumaran: Yeah, they have, like a huge facilities that they’re building where they’re. I think they’re just finishing up those facilities. But again, he’s like these are gonna throw off so much data.
564 00:51:58.140 ⇒ 00:52:03.040 Uttam Kumaran: And he’s like he was interested in, like what sort of opportunities there are for.
565 00:52:03.515 ⇒ 00:52:03.990 Chang Ho: Man.
566 00:52:03.990 ⇒ 00:52:07.549 Uttam Kumaran: And I was like, Oh, that sounds like amazing. I’ll I’ll definitely
567 00:52:07.950 ⇒ 00:52:10.589 Uttam Kumaran: give you a ring at some point. So.
568 00:52:10.590 ⇒ 00:52:21.311 Chang Ho: Oh, man, I mean, if it’s a if it really is as attractive as you’re trying to sell it, and it’s like it’s pons a thousand poppies and let them grow. Then fuck it. Yeah, definitely. We’re following up.
569 00:52:21.580 ⇒ 00:52:32.200 Uttam Kumaran: Yeah, cool. So we have some good stuff. So let me get this meeting scheduled for next week, and then, yeah, I’ll keep slacking you stuff. This is a really good reminder to follow some people. So yeah.
570 00:52:32.200 ⇒ 00:52:36.910 Chang Ho: No, it’s all. It’s all good. But I’ll yeah, I’ll I’ll look. I’ll look forward to seeing
571 00:52:37.520 ⇒ 00:52:38.170 Chang Ho: it’s okay.
572 00:52:38.170 ⇒ 00:52:38.680 Uttam Kumaran: Perfect, see.
573 00:52:38.680 ⇒ 00:52:41.160 Chang Ho: Some of the projects hopefully.
574 00:52:41.160 ⇒ 00:52:42.140 Uttam Kumaran: Yeah. I’m glad.
575 00:52:42.346 ⇒ 00:52:44.209 Chang Ho: We’ll end up chatting next week first, for sure.
576 00:52:44.210 ⇒ 00:52:44.860 Uttam Kumaran: Okay.
577 00:52:45.520 ⇒ 00:52:46.360 Chang Ho: Nice. We’ll do time. Thanks.
578 00:52:46.360 ⇒ 00:52:46.780 Uttam Kumaran: Okay.
579 00:52:47.056 ⇒ 00:52:47.610 Chang Ho: Man, perfect.
580 00:52:47.610 ⇒ 00:52:49.050 Uttam Kumaran: Yeah, I appreciate. Thank you very much.
581 00:52:49.050 ⇒ 00:52:50.399 Chang Ho: Have a good rest of the week.
582 00:52:50.400 ⇒ 00:52:54.290 Uttam Kumaran: Yeah, you, too. And I’ll send you stuff about the email things as soon as we get that kind of cooking.
583 00:52:54.660 ⇒ 00:53:01.190 Chang Ho: Nice, and I’ll I’ll I might just chase up some leads I have over in the Us. For the help some of the healthcare operations stuff. Maybe.
584 00:53:01.448 ⇒ 00:53:13.869 Uttam Kumaran: If you wanna see me on anything, feel free? Yeah, whoever. And again, if if they’re within your shot of me, I’m I’m gonna be doing a lot of traveling this next few months for like weddings and stuff like I’ll be going to Boston. I’ll be going.
585 00:53:13.870 ⇒ 00:53:14.790 Chang Ho: Don’t work.
586 00:53:15.010 ⇒ 00:53:16.440 Uttam Kumaran: Obedience.
587 00:53:16.440 ⇒ 00:53:17.320 Chang Ho: Okay. Okay.
588 00:53:17.320 ⇒ 00:53:20.360 Uttam Kumaran: If there’s people there I can me, I’ll be there so.
589 00:53:20.360 ⇒ 00:53:22.120 Chang Ho: Oh, nice one. Okay.
590 00:53:22.450 ⇒ 00:53:23.639 Chang Ho: I’ll keep that in mind.
591 00:53:24.020 ⇒ 00:53:24.465 Uttam Kumaran: Okay.
592 00:53:25.290 ⇒ 00:53:26.060 Chang Ho: So mine.
593 00:53:26.210 ⇒ 00:53:27.519 Chang Ho: You, too, man, catch.
594 00:53:27.520 ⇒ 00:53:28.330 Uttam Kumaran: Alright!
595 00:53:28.330 ⇒ 00:53:30.019 Chang Ho: Have you seen? Have a good week bye.
596 00:53:30.020 ⇒ 00:53:30.610 Uttam Kumaran: Bye.