Meeting Title: HPI-Real-Estate-Kickoff Date: 2024-09-18 Meeting participants: Joshuadeveyra, Nicolas Sucari, Uttam Kumaran
WEBVTT
1 00:02:05.980 ⇒ 00:02:06.760 Nicolas Sucari: How you doing.
2 00:02:07.470 ⇒ 00:02:08.300 Uttam Kumaran: Hey? I’m Ryan.
3 00:02:10.460 ⇒ 00:02:12.239 Nicolas Sucari: Doing good. How are you.
4 00:02:12.890 ⇒ 00:02:13.580 Uttam Kumaran: God.
5 00:02:14.320 ⇒ 00:02:14.930 Uttam Kumaran: Just
6 00:02:15.920 ⇒ 00:02:19.040 Uttam Kumaran: yeah. I tried not to book any meetings today. So
7 00:02:19.720 ⇒ 00:02:21.890 Uttam Kumaran: library to go do some writing and
8 00:02:22.080 ⇒ 00:02:23.920 Uttam Kumaran: some planning and stuff. So.
9 00:02:25.410 ⇒ 00:02:26.150 Nicolas Sucari: Okay.
10 00:02:26.910 ⇒ 00:02:27.730 Nicolas Sucari: perfect.
11 00:02:28.730 ⇒ 00:02:31.599 Nicolas Sucari: I think the by the coco stuff.
12 00:02:31.640 ⇒ 00:02:35.029 Nicolas Sucari: I think Eddie will review it today. Right? He said.
13 00:02:35.070 ⇒ 00:02:37.059 Nicolas Sucari: It was looking nice, so.
14 00:02:37.780 ⇒ 00:02:38.440 Uttam Kumaran: Can he respond?
15 00:02:38.440 ⇒ 00:02:39.240 Nicolas Sucari: We won’t.
16 00:02:39.460 ⇒ 00:02:41.310 Nicolas Sucari: he responded. Teams, he added, like.
17 00:02:41.310 ⇒ 00:02:41.690 Uttam Kumaran: Okay.
18 00:02:42.593 ⇒ 00:02:44.400 Nicolas Sucari: Bad. Yeah.
19 00:02:44.500 ⇒ 00:02:47.979 Nicolas Sucari: he he will share. He. He says he will share feedback today.
20 00:02:48.290 ⇒ 00:02:49.579 Uttam Kumaran: Okay, cool. Okay. Great.
21 00:02:50.490 ⇒ 00:02:54.439 Uttam Kumaran: Yeah. I just don’t want to get too far ahead if they want to change something or
22 00:02:54.901 ⇒ 00:02:59.419 Uttam Kumaran: but I think we made a good impression that you know so far. So I think that’s great.
23 00:03:01.700 ⇒ 00:03:02.400 Nicolas Sucari: Yeah.
24 00:03:03.240 ⇒ 00:03:07.129 Nicolas Sucari: we need just to fix. How we are
25 00:03:07.350 ⇒ 00:03:18.999 Nicolas Sucari: adding the date and time the bull sheets because we changed something we we get all yesterday, so we can add the time there, and it’s not recognizing it as a date field.
26 00:03:19.000 ⇒ 00:03:19.910 Uttam Kumaran: Oh, okay. Okay.
27 00:03:19.910 ⇒ 00:03:23.190 Nicolas Sucari: I’m gonna but yeah, it’s should be something easy.
28 00:03:25.870 ⇒ 00:03:28.430 Uttam Kumaran: Can you? Can you, ping Miguel? See if he could join?
29 00:03:29.450 ⇒ 00:03:30.150 Uttam Kumaran: Yep.
30 00:04:54.590 ⇒ 00:04:56.379 Uttam Kumaran: I don’t know. For some reason I got kicked off.
31 00:04:57.308 ⇒ 00:05:00.960 Uttam Kumaran: Did I? Did I tell you about this? This real estate thing? By the way.
32 00:05:02.653 ⇒ 00:05:11.840 Nicolas Sucari: No, I think like this. There is like this channel this like channel, like a while ago. But yeah, we never talked about this project.
33 00:05:12.110 ⇒ 00:05:17.309 Uttam Kumaran: Okay, cool. Yeah, this project has kind of evolved but we’re finally getting
34 00:05:17.720 ⇒ 00:05:20.160 Uttam Kumaran: somewhere close.
35 00:05:23.700 ⇒ 00:05:28.780 Uttam Kumaran: so I kind of want to get Miguel to start thinking about some stuff and start planning.
36 00:05:31.900 ⇒ 00:05:33.000 Nicolas Sucari: Yeah, okay.
37 00:05:34.400 ⇒ 00:05:36.200 Nicolas Sucari: I’ve been reading the notion.
38 00:05:36.510 ⇒ 00:05:37.180 Uttam Kumaran: Okay. Cool.
39 00:05:38.640 ⇒ 00:05:41.030 Nicolas Sucari: to understand a little bit more. Yeah.
40 00:06:13.070 ⇒ 00:06:21.259 Nicolas Sucari: what we. What I understood of this project is that they want to reduce, like the negotiation, the negotiation times right.
41 00:06:21.270 ⇒ 00:06:22.510 Nicolas Sucari: or the leases
42 00:06:22.910 ⇒ 00:06:24.200 Nicolas Sucari: that.
43 00:06:24.370 ⇒ 00:06:25.000 Uttam Kumaran: Yeah.
44 00:06:25.640 ⇒ 00:06:27.731 Nicolas Sucari: Yeah, I don’t know what is
45 00:06:28.620 ⇒ 00:06:31.039 Nicolas Sucari: like the actual process that they need.
46 00:06:31.980 ⇒ 00:06:32.300 Uttam Kumaran: Yes.
47 00:06:32.300 ⇒ 00:06:32.680 Nicolas Sucari: That’s gonna.
48 00:06:32.680 ⇒ 00:06:37.500 Uttam Kumaran: It’s gonna be heavily on the yeah. The actual solution is gonna be what I kinda wanna
49 00:06:37.590 ⇒ 00:06:39.860 Uttam Kumaran: talk a little bit about today. And
50 00:06:40.060 ⇒ 00:06:41.370 Uttam Kumaran: I’m sort of
51 00:06:41.880 ⇒ 00:06:44.229 Uttam Kumaran: think with Miguel. I’m like what to do.
52 00:06:44.840 ⇒ 00:06:47.119 Uttam Kumaran: Denise. Anything before we get started.
53 00:06:47.120 ⇒ 00:06:47.920 Nicolas Sucari: No.
54 00:06:48.570 ⇒ 00:06:53.270 Nicolas Sucari: we I I just think him, but he didn’t. Oh, here he is.
55 00:06:56.910 ⇒ 00:06:57.949 joshuadeveyra: And again. Sorry.
56 00:06:58.470 ⇒ 00:06:59.070 Uttam Kumaran: Offline.
57 00:06:59.070 ⇒ 00:06:59.920 Nicolas Sucari: Hi, Real.
58 00:07:00.320 ⇒ 00:07:02.350 joshuadeveyra: I thought my shift was earlier. Sorry.
59 00:07:02.970 ⇒ 00:07:03.900 Uttam Kumaran: You’re good.
60 00:07:04.300 ⇒ 00:07:05.780 joshuadeveyra: Yeah, my bad, my bad.
61 00:07:08.450 ⇒ 00:07:09.890 Uttam Kumaran: cool. I
62 00:07:11.140 ⇒ 00:07:17.279 Uttam Kumaran: I wanted to talk a little bit about this project. I don’t know, Miguel. Did you have a chance to look at a notion, or I can go kind of go through
63 00:07:17.770 ⇒ 00:07:18.539 Uttam Kumaran: the whole thing.
64 00:07:18.540 ⇒ 00:07:20.600 joshuadeveyra: It would be yeah nice to
65 00:07:20.650 ⇒ 00:07:25.000 joshuadeveyra: kind of go through. I did have like a little bit of read earlier. But yeah.
66 00:07:26.640 ⇒ 00:07:28.179 Uttam Kumaran: Okay, great. So the
67 00:07:28.810 ⇒ 00:07:35.184 Uttam Kumaran: Hpi is. Actually, I’ll even go even further back. So I’m working with
68 00:07:35.700 ⇒ 00:07:41.578 Uttam Kumaran: A guy here in Austin who is a he’s a former like public company. CEO,
69 00:07:42.120 ⇒ 00:07:44.560 Uttam Kumaran: And he ran this company
70 00:07:45.950 ⇒ 00:07:53.169 Uttam Kumaran: And he. They had a big office here in Austin, and he ended up becoming like a little bit of a venture capitalist sort of like
71 00:07:53.260 ⇒ 00:07:59.389 Uttam Kumaran: angel investor. And he I’ve been working with him, and he’s been interested in starting a product in like the real estate
72 00:07:59.870 ⇒ 00:08:06.060 Uttam Kumaran: sector. So we’ve been kind of chatting for over a year about several different ideas, all kind of around AI,
73 00:08:06.090 ⇒ 00:08:11.149 Uttam Kumaran: and we kind of landed on one idea in particular, which is this sort of like contract
74 00:08:11.540 ⇒ 00:08:13.250 Uttam Kumaran: negotiation
75 00:08:13.330 ⇒ 00:08:14.700 Uttam Kumaran: spurtive agent?
76 00:08:15.556 ⇒ 00:08:16.670 Uttam Kumaran: The biggest.
77 00:08:17.870 ⇒ 00:08:20.360 Uttam Kumaran: thing I’ll kind of explain is
78 00:08:20.750 ⇒ 00:08:31.419 Uttam Kumaran: and I don’t really know yet what the solution is going to be. But to talk about the problem. If you read that notion there’s a lot of issues that happen when a contract for a commercial real estate building
79 00:08:31.430 ⇒ 00:08:33.240 Uttam Kumaran: is getting signed.
80 00:08:33.639 ⇒ 00:08:42.599 Uttam Kumaran: There’s a lot of the back and forth with lawyers, and a lot of the times, the negotiation with the lawyers. They end up negotiating the same policies
81 00:08:42.610 ⇒ 00:08:44.249 Uttam Kumaran: over and over and over again.
82 00:08:44.260 ⇒ 00:08:48.921 Uttam Kumaran: So one of the things that I was talking to them about was
83 00:08:50.980 ⇒ 00:08:55.899 Uttam Kumaran: the fact that if we were able to provide an AI agent with like, hey, they typically
84 00:08:56.365 ⇒ 00:09:01.119 Uttam Kumaran: for example, there’s certain clauses in the contract that if they were to
85 00:09:01.618 ⇒ 00:09:13.740 Uttam Kumaran: if the if the person who’s wants to lease the space proposes they’re more okay with, there’s certain things that they’re not okay with taking out. So there’s all this kind of negotiation that happens. But these. So Hpi is the biggest
86 00:09:14.221 ⇒ 00:09:17.099 Uttam Kumaran: owner of commercial real estate in Austin.
87 00:09:18.610 ⇒ 00:09:24.170 Uttam Kumaran: and so they’re like a huge huge company. The guy who runs the company is actually like, kind of a big
88 00:09:24.190 ⇒ 00:09:26.500 Uttam Kumaran: guy here in Texas.
89 00:09:28.070 ⇒ 00:09:31.750 Uttam Kumaran: that being said, we have a good opportunity to.
90 00:09:31.940 ⇒ 00:09:36.200 Uttam Kumaran: you know. Try and work on a solution and a prototype for them.
91 00:09:36.488 ⇒ 00:09:45.339 Uttam Kumaran: To kind of like. Show them a little bit about what we can do. And I think there’s, you know, I think this out of all the things that we’re working on
92 00:09:45.867 ⇒ 00:09:50.590 Uttam Kumaran: in the company. This is probably has the potential to be like the biggest
93 00:09:51.230 ⇒ 00:09:52.859 Uttam Kumaran: project out of all of them
94 00:09:52.910 ⇒ 00:09:55.349 Uttam Kumaran: in terms of like revenue size.
95 00:09:55.810 ⇒ 00:09:57.030 Uttam Kumaran: I would say that
96 00:10:00.340 ⇒ 00:10:03.485 Uttam Kumaran: problem that I’ll share in that.
97 00:10:04.590 ⇒ 00:10:21.489 Uttam Kumaran: This company is like a little bit skeptical on like what’s possible. So we’re not actually like contracted with them. What we have, though, is like we have an in and a relationship with, like a pretty high, level real estate firm that not many other firms can get.
98 00:10:21.590 ⇒ 00:10:24.879 Uttam Kumaran: And we’ve received leases and a bunch of contracts from them.
99 00:10:24.940 ⇒ 00:10:30.239 Uttam Kumaran: I want to begin prototyping some stuff around this idea of this like
100 00:10:30.300 ⇒ 00:10:32.730 Uttam Kumaran: contract negotiation agent.
101 00:10:33.104 ⇒ 00:10:44.465 Uttam Kumaran: Almost. It’s gonna start hopefully, like a co copilot for their head of leasing and their like staff, so that when they get feedback back from clients they can sort of ask questions.
102 00:10:44.900 ⇒ 00:10:50.890 Uttam Kumaran: But I do think that there’s a big opportunity to help them streamline stuff and automate things internally.
103 00:10:51.110 ⇒ 00:10:59.779 Uttam Kumaran: So that’s kind of like where I want to start. There’s a lot more details in the document than I’ll go through today. So if you guys can read that and share questions, that would be
104 00:11:00.532 ⇒ 00:11:02.820 Uttam Kumaran: really nice. But any
105 00:11:02.900 ⇒ 00:11:09.850 Uttam Kumaran: any thoughts or questions so far anything I can answer about the client about the project. And we’ll have another meeting.
106 00:11:09.910 ⇒ 00:11:13.349 Uttam Kumaran: you know, later this week or or next week as well. So.
107 00:11:21.288 ⇒ 00:11:25.949 joshuadeveyra: So it’s basically something like, what was the company we met? Uta.
108 00:11:27.760 ⇒ 00:11:31.060 joshuadeveyra: Flowmatic, right? Because I believe it’s something like that.
109 00:11:31.580 ⇒ 00:11:34.170 joshuadeveyra: Remember, it’s like they’re more on documents type.
110 00:11:35.250 ⇒ 00:11:37.149 Uttam Kumaran: Yeah, but I guess it’s like.
111 00:11:37.520 ⇒ 00:11:40.035 Uttam Kumaran: I mean, we. In this case, we’re flomatic.
112 00:11:40.350 ⇒ 00:11:42.200 joshuadeveyra: Yeah, yeah, that’s a, yeah.
113 00:11:42.200 ⇒ 00:11:51.240 Uttam Kumaran: Yeah, but yeah, I I think this is definitely more document heavy. So this is like all word, document and Pdf, heavy stuff.
114 00:11:52.640 ⇒ 00:11:56.370 Uttam Kumaran: Additionally, you know, in word documents you can keep track of like revisions.
115 00:11:59.490 ⇒ 00:12:00.830 joshuadeveyra: Yeah, yeah, yeah, in.
116 00:12:01.166 ⇒ 00:12:04.873 Uttam Kumaran: Yeah. So one of the things that when they work with
117 00:12:05.410 ⇒ 00:12:26.090 Uttam Kumaran: like lawyers is that they have like this revision process and the revision process. We can actually keep track of changes. So, for example, like when a when a clock, when a when a tenant in a building, and again, commercial real estate. These are like massive office based buildings. So hundreds of thousands of square feet of office space
118 00:12:26.110 ⇒ 00:12:39.769 Uttam Kumaran: couple 100 million dollar buildings, they lease out each individual portion. And so when the tenant is is negotiating, they’re negotiating all sorts of stuff like the lease term the payback insurance all sorts of stuff.
119 00:12:39.890 ⇒ 00:12:46.320 Uttam Kumaran: So Hpi has a certain criteria for things they will or won’t negotiate, but they never wrote this stuff down
120 00:12:46.560 ⇒ 00:12:48.620 Uttam Kumaran: right? So every deal is different.
121 00:12:48.780 ⇒ 00:12:51.729 Uttam Kumaran: but every deal is also kind of similar
122 00:12:51.760 ⇒ 00:13:12.040 Uttam Kumaran: right. And so that’s the thing that I propose is that we almost need like sort of like a playbook for how they negotiate deals, and then we can pass that requirements file to an AI agent pretty easily to be able to say, Hey, we just got this email from a tenant. They’re requesting that we change this clause in the contract.
123 00:13:12.470 ⇒ 00:13:21.520 Uttam Kumaran: What do we respond with is this something that we’ve done before? Right? So I would. I think this is going to be like a rag heavy.
124 00:13:21.920 ⇒ 00:13:22.380 joshuadeveyra: Yeah.
125 00:13:22.380 ⇒ 00:13:23.330 Uttam Kumaran: Project.
126 00:13:25.640 ⇒ 00:13:27.449 joshuadeveyra: Yeah, definitely. Yes. Yeah.
127 00:13:27.450 ⇒ 00:13:30.989 Uttam Kumaran: So what do you think reading it, Miguel? Like, what do you think.
128 00:13:33.490 ⇒ 00:13:37.040 joshuadeveyra: Yeah, this is good. Yeah, I agree with you. It’s gonna be like a lot of rag.
129 00:13:38.780 ⇒ 00:13:39.670 joshuadeveyra: But
130 00:13:39.970 ⇒ 00:13:46.309 joshuadeveyra: but I remember I told you when we were on the call that I built something similar before. But that’s just like for an accounting firm
131 00:13:47.310 ⇒ 00:13:48.759 joshuadeveyra: where we basically just check.
132 00:13:48.760 ⇒ 00:13:50.069 Uttam Kumaran: Cancel. What was that?
133 00:13:52.687 ⇒ 00:14:04.070 joshuadeveyra: It was more of like compliance thing like, for example, because, like in the Uk, there’s this compliance list. I think it’s under Nda, so I can’t really tell the details.
134 00:14:06.870 ⇒ 00:14:12.810 joshuadeveyra: But, for example, in this page, for, for example, page 5, it has to contain this type of you know.
135 00:14:14.069 ⇒ 00:14:14.429 Uttam Kumaran: This. Yeah.
136 00:14:14.430 ⇒ 00:14:18.370 joshuadeveyra: Thanks. Yeah, yeah. So it’s, I think, something pretty similar.
137 00:14:19.000 ⇒ 00:14:19.400 Uttam Kumaran: So.
138 00:14:20.010 ⇒ 00:14:20.620 joshuadeveyra: Yeah.
139 00:14:22.710 ⇒ 00:14:23.370 joshuadeveyra: Just.
140 00:14:23.810 ⇒ 00:14:25.210 Nicolas Sucari: You’re you’re cutting off a little bit.
141 00:14:25.210 ⇒ 00:14:27.440 joshuadeveyra: You’re cutting off yeah, like, heavily.
142 00:14:29.759 ⇒ 00:14:32.509 Uttam Kumaran: Sorry. Well, I said, what was the technology that you used there?
143 00:14:33.030 ⇒ 00:14:36.720 joshuadeveyra: I just used express, and then just passed it to Openai. To be honest.
144 00:14:38.890 ⇒ 00:14:39.490 Uttam Kumaran: Okay.
145 00:14:39.870 ⇒ 00:14:42.879 joshuadeveyra: And then a lot of ui, yeah, the the ui was honestly the hard part.
146 00:14:44.600 ⇒ 00:14:50.880 Uttam Kumaran: So I have. So I’ll give you a little bit more context of the project. So we’re working with this guy. He’s actually in the Channel, but
147 00:14:51.330 ⇒ 00:14:53.230 Uttam Kumaran: I don’t know how active he is.
148 00:14:53.230 ⇒ 00:14:54.939 Nicolas Sucari: What was the last 2.
149 00:14:54.940 ⇒ 00:14:55.520 Uttam Kumaran: Oh.
150 00:14:56.100 ⇒ 00:14:56.949 Uttam Kumaran: how about now?
151 00:14:57.950 ⇒ 00:15:01.315 Nicolas Sucari: You. You started saying, we are working with, and we lost you so you can.
152 00:15:01.540 ⇒ 00:15:03.655 Uttam Kumaran: Okay. Sorry. I’m just driving to the library.
153 00:15:04.520 ⇒ 00:15:21.779 Uttam Kumaran: we’re we’re working with this guy. Scott Harman, who’s actually in the Channel. But I don’t think he’s so active. Basically, he’s kind of like our sponsor for the project. So my goal for this, and probably for the next 2. I’ll tell you. We have a demo on the 1st week of October.
154 00:15:21.830 ⇒ 00:15:27.099 Uttam Kumaran: so I want to put together a bit of a demo of something around this project working
155 00:15:28.860 ⇒ 00:15:36.090 Uttam Kumaran: either. And like, let maybe we can. We need to have probably another meeting deciding on like what exactly we want to try to prototype.
156 00:15:36.450 ⇒ 00:15:44.879 Uttam Kumaran: But I want to consider a deadline of like, Yeah, October second, I think, is our meeting, and I want to think a little bit about what a
157 00:15:45.000 ⇒ 00:15:47.550 Uttam Kumaran: demo could look like of this process work.
158 00:15:47.550 ⇒ 00:15:48.210 Nicolas Sucari: So
159 00:15:49.000 ⇒ 00:16:12.329 Nicolas Sucari: th. The process is like they receive like a question or a request from the tenant, and they need to review the contract, for of the leasing. So maybe we can have, like a couple of samples of those leasing documents make them go through an AI agent or review, and then give some like options on how to negotiate right.
160 00:16:12.730 ⇒ 00:16:16.919 Uttam Kumaran: Exactly so. I actually have, like a ton of documents that I got from them.
161 00:16:17.500 ⇒ 00:16:20.099 Uttam Kumaran: A bunch of executed leases
162 00:16:20.160 ⇒ 00:16:21.667 Uttam Kumaran: as well as
163 00:16:22.830 ⇒ 00:16:24.698 Uttam Kumaran: as well as a bunch of
164 00:16:25.260 ⇒ 00:16:27.450 Uttam Kumaran: a bunch of documents that are like
165 00:16:29.020 ⇒ 00:16:31.799 Uttam Kumaran: the back and forth process. So with the revisions.
166 00:16:31.800 ⇒ 00:16:32.500 Nicolas Sucari: Sadly.
167 00:16:33.140 ⇒ 00:16:54.029 Nicolas Sucari: So maybe we need to like the the process. I think it’s pretty clear, is getting all of those documents and creating kind of our context for that. AI agent to go check. What we like negotiated before. What are the options? And then like, give some? Yeah.
168 00:16:54.030 ⇒ 00:17:09.699 Nicolas Sucari: just go ahead and give some ideas for the actual agents to go and and do that documents. But what we need to focus also is on the output on how we’re gonna communicate all of these options, and how we’re gonna
169 00:17:09.940 ⇒ 00:17:10.710 Nicolas Sucari: Ashley.
170 00:17:10.710 ⇒ 00:17:11.089 Uttam Kumaran: Yeah.
171 00:17:11.099 ⇒ 00:17:17.059 Nicolas Sucari: That we are changing something on the documents you are receiving. And what is like those changes right?
172 00:17:17.900 ⇒ 00:17:21.300 Uttam Kumaran: Yeah. So what I want. So if you read the document, there’s a concept.
173 00:17:22.810 ⇒ 00:17:44.849 Uttam Kumaran: What you see is called a doc. So the kind of the idea there came from golf, which is like, you know, you have, like a fair way, right? And and all these lease negotiation processes, there is things that they’re more strict on, and things that they’re less strict on. But it’s a it’s a fluid. It’s not very like we always do this. We do that. If the client is Facebook, they may make an exception.
174 00:17:44.850 ⇒ 00:17:56.130 Uttam Kumaran: If the client is brain force, they may say, like, fuck off like you guys are nobody. Right? So that’s the that’s the things that I want to create. Basically like this, almost like a
175 00:17:56.830 ⇒ 00:17:59.159 Uttam Kumaran: like a sop for
176 00:17:59.630 ⇒ 00:18:25.239 Uttam Kumaran: how they do negotiations. But I want that to come from their negotiate, from the history of their leases. But also I want that to be the thing that we can get feedback from them on. For example, this could be things like, Hey, on on a on this sort of contract term. You guys are very inflexible, and maybe there’s like a score which is like you have a flexibility score of 2. Here. You’re more flexible on this
177 00:18:25.240 ⇒ 00:18:42.890 Uttam Kumaran: right that way. When we pass new comments or new contracts to agents. The agent can go ahead and almost score like each of the terms. And then, when a new comment comes, because invariably everyone renegotiates these leases right so.
178 00:18:43.480 ⇒ 00:18:50.280 Uttam Kumaran: but always will be like, I don’t like this term. I want this term, and it’s a constant back and forth process between
179 00:18:50.550 ⇒ 00:19:10.129 Uttam Kumaran: the head of their contracts and the lawyers. And yeah, like you could tell, I wrote down a little bit about what the process is. And I also have the whole transcript. And this is something we’ve been talking about for a bit. But it’s a little bit real estate heavy. So that’s gonna be where I think for you guys
180 00:19:10.720 ⇒ 00:19:14.769 Uttam Kumaran: to like, I’m gonna try and get some materials together
181 00:19:14.780 ⇒ 00:19:18.630 Uttam Kumaran: about how to learn a little bit about this like real estate.
182 00:19:18.800 ⇒ 00:19:28.219 Uttam Kumaran: like a contract negotiation process. But for Mika I work. I worked, you know, we work for a bit, so I have a little bit of familiarity, but there’s a lot of jargon
183 00:19:28.290 ⇒ 00:19:33.160 Uttam Kumaran: and a lot of like very industry, specific things here that aren’t that
184 00:19:33.280 ⇒ 00:19:34.949 Uttam Kumaran: like it’s not
185 00:19:34.990 ⇒ 00:19:44.700 Uttam Kumaran: that common. So that’s the thing I think, is going to be the one of the bigger hurdles. But again, we have like about 2 weeks that I want to put this like demo together.
186 00:19:45.940 ⇒ 00:19:48.670 Uttam Kumaran: so I think that’s more than enough time to try to do something.
187 00:19:51.130 ⇒ 00:19:54.549 Uttam Kumaran: I guess, Miguel, like, what do you think we can try
188 00:19:54.840 ⇒ 00:19:57.680 Uttam Kumaran: first, st in terms of like
189 00:19:57.870 ⇒ 00:19:59.500 Uttam Kumaran: prototypes, like.
190 00:19:59.550 ⇒ 00:20:01.450 Uttam Kumaran: I have all these documents.
191 00:20:01.933 ⇒ 00:20:06.420 Uttam Kumaran: We. The goal is to build this sort of like fairway document.
192 00:20:06.550 ⇒ 00:20:12.650 Uttam Kumaran: which which basically outlines how they do negotiations. That fairway document then becomes the prompt
193 00:20:12.770 ⇒ 00:20:17.630 Uttam Kumaran: for the agent. Actually, right? So we kind of have these like different phases. But
194 00:20:18.180 ⇒ 00:20:23.039 Uttam Kumaran: like, what can we? I want to try to test like, I want to try to test like anything we can.
195 00:20:23.060 ⇒ 00:20:26.010 Uttam Kumaran: Right? Like, I wanna, I wanna try Openai assistance. But
196 00:20:26.190 ⇒ 00:20:27.640 Uttam Kumaran: like, I wanna
197 00:20:27.970 ⇒ 00:20:36.419 Uttam Kumaran: given like, how important these guys could be if we close them, I want to try anything that we have in our arsenal to try to see whether we can get results.
198 00:20:37.310 ⇒ 00:20:41.170 joshuadeveyra: Yeah, Utama have like a quick question, because it will, you know. But
199 00:20:41.350 ⇒ 00:20:43.589 joshuadeveyra: it’s gonna affect the project really big?
200 00:20:43.939 ⇒ 00:20:50.429 joshuadeveyra: How long is like I I assume it’s like a Pdf. Right? How many pages is like one contract on average.
201 00:20:50.620 ⇒ 00:20:53.349 Uttam Kumaran: Let me find out I didn’t open it.
202 00:20:53.580 ⇒ 00:20:57.259 Uttam Kumaran: If I just got them. I just loaded them into Google drive, actually, yesterday.
203 00:20:57.563 ⇒ 00:20:59.669 Uttam Kumaran: But let me find that out for you.
204 00:21:00.180 ⇒ 00:21:01.070 joshuadeveyra: Okay, yeah, I.
205 00:21:01.070 ⇒ 00:21:02.450 Uttam Kumaran: Assume it’s going to be
206 00:21:02.540 ⇒ 00:21:03.810 Uttam Kumaran: minimum.
207 00:21:04.720 ⇒ 00:21:05.280 Uttam Kumaran: I mean.
208 00:21:05.280 ⇒ 00:21:06.319 joshuadeveyra: 100 right?
209 00:21:06.640 ⇒ 00:21:09.979 Uttam Kumaran: I don’t know if it’s gonna be a hundred, but I think it may be
210 00:21:10.070 ⇒ 00:21:11.639 Uttam Kumaran: 2030 plus.
211 00:21:11.870 ⇒ 00:21:12.880 joshuadeveyra: Okay, yeah, that’s fine.
212 00:21:12.880 ⇒ 00:21:15.590 Uttam Kumaran: Cause. These are these are 10 to 20 year.
213 00:21:15.860 ⇒ 00:21:20.420 Uttam Kumaran: Well, these are anywhere from 3, 5, 10 year office space leases
214 00:21:20.780 ⇒ 00:21:22.840 Uttam Kumaran: like minimum
215 00:21:23.120 ⇒ 00:21:29.539 Uttam Kumaran: 1 million dollars a year type stuff like I’m driving in Austin right now. I’m staring at the building they own.
216 00:21:29.720 ⇒ 00:21:30.300 Uttam Kumaran: like
217 00:21:31.865 ⇒ 00:21:34.900 Uttam Kumaran: they own, like most of the commercial real estate in Austin.
218 00:21:36.700 ⇒ 00:21:40.180 Uttam Kumaran: so these are kind of like, it’s kind of a they’re pretty technical leases.
219 00:21:41.350 ⇒ 00:21:45.100 joshuadeveyra: Is Dallas? In? No, no, my village is in Austin or no.
220 00:21:45.100 ⇒ 00:21:46.739 Uttam Kumaran: Mavericks are are in Dallas.
221 00:21:47.449 ⇒ 00:21:50.690 joshuadeveyra: Different city. I think I’m not familiar. Okay? Okay?
222 00:21:54.920 ⇒ 00:21:56.280 joshuadeveyra: Okay. Yeah.
223 00:21:58.190 ⇒ 00:22:02.849 joshuadeveyra: I’d like to see, like one of the documents like a sample lease, and then
224 00:22:03.790 ⇒ 00:22:05.470 joshuadeveyra: analysis, because.
225 00:22:05.470 ⇒ 00:22:05.800 Uttam Kumaran: Okay.
226 00:22:05.800 ⇒ 00:22:10.969 joshuadeveyra: What we might have to do, cause, I assume, like Lisas, have like chapters in them, right?
227 00:22:11.500 ⇒ 00:22:12.979 joshuadeveyra: I’m not sure what it’s called, but.
228 00:22:12.980 ⇒ 00:22:14.660 Uttam Kumaran: And be, there’s gonna be sections. Yeah.
229 00:22:14.660 ⇒ 00:22:19.650 joshuadeveyra: Sections. Yeah, yeah. Sections. So that’s probably like the best way we could, you know.
230 00:22:21.018 ⇒ 00:22:25.809 joshuadeveyra: we could analyze it instead of putting it one whole thing cause that
231 00:22:25.940 ⇒ 00:22:32.390 joshuadeveyra: all models will hallucinate if you throw like 50 pages there. So we might have to divide that
232 00:22:32.490 ⇒ 00:22:34.619 joshuadeveyra: and then feed that to the AI one by one.
233 00:22:35.600 ⇒ 00:22:38.308 Uttam Kumaran: The other thing is, we may have to like
234 00:22:39.080 ⇒ 00:22:40.840 Uttam Kumaran: We may have to.
235 00:22:41.665 ⇒ 00:22:43.119 Uttam Kumaran: What’s it called?
236 00:22:44.250 ⇒ 00:22:45.890 joshuadeveyra: In our own model.
237 00:22:46.010 ⇒ 00:22:51.509 Uttam Kumaran: Yeah, train our own model or like, vector embed certain things.
238 00:22:51.930 ⇒ 00:22:54.919 Uttam Kumaran: But let’s let’s see, I want to try something this week.
239 00:22:54.920 ⇒ 00:22:55.330 joshuadeveyra: Yeah, yeah.
240 00:22:55.330 ⇒ 00:22:56.810 Uttam Kumaran: And then let’s see where we’re
241 00:22:56.880 ⇒ 00:23:02.429 Uttam Kumaran: failing. And then we can call the flowmatic guys. We can call a bunch of friends. And and they do some research.
242 00:23:02.950 ⇒ 00:23:03.715 Uttam Kumaran: Yep,
243 00:23:05.390 ⇒ 00:23:12.729 Uttam Kumaran: so let’s do that. So right now let me send you I’m just parking, so let me send you
244 00:23:12.810 ⇒ 00:23:15.240 Uttam Kumaran: 2 things. One
245 00:23:17.040 ⇒ 00:23:18.819 Uttam Kumaran: I’ll send you the
246 00:23:18.870 ⇒ 00:23:21.639 Uttam Kumaran: a couple of examples of leases.
247 00:23:23.263 ⇒ 00:23:27.659 Uttam Kumaran: I’ll also send you a couple. I’ll we’ll kind of. I’ll kind of note down
248 00:23:27.800 ⇒ 00:23:30.050 Uttam Kumaran: the things that we’re kind of driving towards
249 00:23:30.717 ⇒ 00:23:32.420 Uttam Kumaran: on our side.
250 00:23:32.560 ⇒ 00:23:33.780 Uttam Kumaran: and then
251 00:23:35.290 ⇒ 00:23:37.890 Uttam Kumaran: let’s kind of see? Maybe by
252 00:23:37.950 ⇒ 00:23:41.189 Uttam Kumaran: maybe I’ll put some time on for Friday. We can discuss again
253 00:23:41.640 ⇒ 00:23:43.190 Uttam Kumaran: the vitacoco stuff.
254 00:23:43.280 ⇒ 00:23:47.070 Uttam Kumaran: Let’s wait until they give us feedback. I don’t want to go too far without them
255 00:23:47.130 ⇒ 00:23:52.589 Uttam Kumaran: telling us, and we? We did like work like super super quickly, anyway. So we have some time.
256 00:23:54.450 ⇒ 00:23:58.940 Uttam Kumaran: so let’s do that. I’ll send you the leases. Let’s just chat in that channel.
257 00:24:00.120 ⇒ 00:24:00.529 joshuadeveyra: Okay. Sure.
258 00:24:00.530 ⇒ 00:24:03.680 Uttam Kumaran: And then, in case we need, in case we have any questions.
259 00:24:03.920 ⇒ 00:24:07.929 Uttam Kumaran: I can ask Scott. Who’s there to answer anything?
260 00:24:08.140 ⇒ 00:24:10.610 Uttam Kumaran: He’s gonna be a huge resource.
261 00:24:10.890 ⇒ 00:24:13.530 Uttam Kumaran: I just want to be careful because
262 00:24:13.740 ⇒ 00:24:15.950 Uttam Kumaran: he’s not as technical.
263 00:24:16.010 ⇒ 00:24:17.180 Uttam Kumaran: and
264 00:24:17.310 ⇒ 00:24:22.659 Uttam Kumaran: I want to show him a demo to get him feedback. I don’t want to ask him about how to actually do the thing.
265 00:24:24.500 ⇒ 00:24:25.780 Uttam Kumaran: so okay, great.
266 00:24:25.780 ⇒ 00:24:28.560 joshuadeveyra: So he’s on like the opportunity. Hpi. Right should we.
267 00:24:28.560 ⇒ 00:24:29.000 Uttam Kumaran: Maria.
268 00:24:29.000 ⇒ 00:24:29.730 joshuadeveyra: Or.
269 00:24:29.730 ⇒ 00:24:33.000 Uttam Kumaran: Yeah, we should chat there. Yeah, I mean, so Scott Scott isn’t
270 00:24:33.420 ⇒ 00:24:37.130 Uttam Kumaran: at Hpi. Scott is like the guy who who
271 00:24:37.220 ⇒ 00:24:40.589 Uttam Kumaran: Spa is like a venture capitalist like an angel investor. Guy.
272 00:24:41.060 ⇒ 00:24:41.400 joshuadeveyra: He is.
273 00:24:41.400 ⇒ 00:24:45.569 Uttam Kumaran: Like, hey? He wants to work with us to kind of develop the solution.
274 00:24:45.710 ⇒ 00:24:55.099 Uttam Kumaran: And then think about how we can sell this to other real estate firms. Hpi is like a separate thing, but Scott has a relationship with the guy that runs Hpi
275 00:24:55.220 ⇒ 00:25:01.379 Uttam Kumaran: cause Scott is like well known around here so like, feel free to message in there. Let’s keep all communications in there.
276 00:25:02.660 ⇒ 00:25:03.670 Uttam Kumaran: and then.
277 00:25:04.320 ⇒ 00:25:14.319 Uttam Kumaran: yeah, I think we can leave the notion kind of as is. I just want to be really scrappy and just drive towards a demo. Of course they haven’t signed anything, so we’re not getting paid yet. But I,
278 00:25:14.850 ⇒ 00:25:26.999 Uttam Kumaran: this is kind of an avenue where these guys are very, very unsophisticated in this sort of technology. So I want to show them something that seems like, really magical that they can be like, Okay, cool, like, we want to give you guys some budget to go do this.
279 00:25:28.250 ⇒ 00:25:32.929 Nicolas Sucari: I mean, we, we can create something visual so that they can like, buy
280 00:25:33.313 ⇒ 00:25:41.209 Nicolas Sucari: and see like, what? What’s the output. But the ultimate goal here is to create, like the right model. And how to
281 00:25:41.400 ⇒ 00:25:52.610 Nicolas Sucari: run those scripts right? Like, we need to give context to an an agent and see what are the the output, what’s what? What is the output, so that we can then leverage that into the into the process.
282 00:25:53.070 ⇒ 00:25:57.823 Uttam Kumaran: Yeah, we just wanna get to again. We want to get to a point where we can start asking questions about feedback.
283 00:25:58.070 ⇒ 00:26:01.960 Uttam Kumaran: So some like signal that’ll start to draw them in, you know.
284 00:26:03.220 ⇒ 00:26:03.750 Nicolas Sucari: Right.
285 00:26:03.750 ⇒ 00:26:04.550 joshuadeveyra: It should be.
286 00:26:07.810 ⇒ 00:26:12.420 Uttam Kumaran: Okay, cool. The only other thing Miguel, maybe
287 00:26:12.796 ⇒ 00:26:27.209 Uttam Kumaran: we kind of talked about this today was, we have Abigail and Barb, who are interns that I’m like, I I’m just like running out of bandwidth. I can’t really manage their day to day, so I’m trying to decide like what to do with them.
288 00:26:27.240 ⇒ 00:26:30.220 Uttam Kumaran: and I had came up with like 2 ideas, one
289 00:26:30.250 ⇒ 00:26:34.549 Uttam Kumaran: like we can kind of train them on the data side, although I do think it’s going to be
290 00:26:35.520 ⇒ 00:26:38.639 Uttam Kumaran: like. It’s gonna be a little bit hard, especially
291 00:26:38.850 ⇒ 00:26:45.329 Uttam Kumaran: like we already tried with a darb, and it was a little bit difficult. The second option, I’m thinking Miguel, is like.
292 00:26:45.640 ⇒ 00:26:49.970 Uttam Kumaran: if you have a curriculum in mind that we can start training people on AI stuff
293 00:26:52.620 ⇒ 00:26:57.349 Uttam Kumaran: like Zapier relevance, and we can give them some sample projects to work on
294 00:26:57.360 ⇒ 00:27:01.150 Uttam Kumaran: that could be actually really great cause. Then you can get some more help.
295 00:27:01.920 ⇒ 00:27:02.340 joshuadeveyra: Yeah.
296 00:27:02.340 ⇒ 00:27:03.769 Uttam Kumaran: What do you think about that.
297 00:27:05.180 ⇒ 00:27:07.940 joshuadeveyra: Yeah, yeah, sure, you mean.
298 00:27:08.190 ⇒ 00:27:09.600 joshuadeveyra: okay, yeah, yeah.
299 00:27:10.750 ⇒ 00:27:14.600 Uttam Kumaran: So what that what that means in actuality is like, I want to put together
300 00:27:14.840 ⇒ 00:27:17.980 Uttam Kumaran: like a document which is basically like how to
301 00:27:18.230 ⇒ 00:27:21.159 Uttam Kumaran: like a month long curriculum of like.
302 00:27:21.260 ⇒ 00:27:25.270 Uttam Kumaran: Here’s how you would go. Here’s like sample projects to go learn
303 00:27:25.300 ⇒ 00:27:26.670 Uttam Kumaran: Zapier
304 00:27:26.720 ⇒ 00:27:28.830 Uttam Kumaran: relevance. Open AI.
305 00:27:29.170 ⇒ 00:27:31.930 Uttam Kumaran: But I kind of need some help thinking about like
306 00:27:32.180 ⇒ 00:27:34.799 Uttam Kumaran: things that are actually accomplishable.
307 00:27:35.070 ⇒ 00:27:36.940 Uttam Kumaran: you know, in a week or 2 weeks.
308 00:27:36.940 ⇒ 00:27:37.590 joshuadeveyra: Actually.
309 00:27:38.220 ⇒ 00:27:38.880 Uttam Kumaran: Yeah.
310 00:27:40.014 ⇒ 00:27:43.649 joshuadeveyra: Do we like? Is that like urgent, do we?
311 00:27:43.730 ⇒ 00:27:47.189 joshuadeveyra: Because we’re like we’re already working on the Hpi project right.
312 00:27:48.060 ⇒ 00:27:51.010 Uttam Kumaran: That’s not super urgent, but if it takes.
313 00:27:51.010 ⇒ 00:27:52.039 joshuadeveyra: And I’m also looking more.
314 00:27:52.040 ⇒ 00:27:53.049 Uttam Kumaran: If we can.
315 00:27:53.130 ⇒ 00:27:55.650 Uttam Kumaran: No, no, if we can send them something.
316 00:27:55.900 ⇒ 00:27:56.330 joshuadeveyra: Yeah.
317 00:27:58.760 ⇒ 00:28:06.900 Uttam Kumaran: like just to like. I’ll maybe I’ll message you. But like I’ll message you today. But if we can send them something just to like work on for the next week or 2.
318 00:28:06.930 ⇒ 00:28:08.369 Uttam Kumaran: That would be great.
319 00:28:09.590 ⇒ 00:28:10.570 joshuadeveyra: Okay. Yeah.
320 00:28:10.570 ⇒ 00:28:11.589 Nicolas Sucari: So if you found like.
321 00:28:11.590 ⇒ 00:28:13.289 joshuadeveyra: Marketing girl, right.
322 00:28:14.180 ⇒ 00:28:18.059 Uttam Kumaran: Yeah, Abigail is helping on sales stuff. But she’s actually a math major.
323 00:28:18.150 ⇒ 00:28:22.179 Uttam Kumaran: So she’s technical. And Adharv is also like a computer science major. So they’re both like
324 00:28:22.250 ⇒ 00:28:26.090 Uttam Kumaran: they both should have enough context to like, get this sort of stuff done.
325 00:28:27.110 ⇒ 00:28:28.220 joshuadeveyra: Okay, yeah. Sure.
326 00:28:32.950 ⇒ 00:28:42.849 joshuadeveyra: Yeah. When? Yeah. Just a side note, when I think, was it Patrick who was showing like the Dbt stuff earlier? I was like, yeah, I can’t relate anywhere.
327 00:28:43.170 ⇒ 00:28:43.910 Uttam Kumaran: Yeah.
328 00:28:44.650 ⇒ 00:28:49.660 joshuadeveyra: I was like I I mean, I didn’t tell you with them right like SQL. Is like my mortal enemy.
329 00:28:50.600 ⇒ 00:28:53.289 Uttam Kumaran: Yeah, it’s different. It’s tough. But
330 00:28:53.680 ⇒ 00:28:56.959 Uttam Kumaran: you know, it’s I would say the AI stuff is also
331 00:28:57.160 ⇒ 00:28:59.899 Uttam Kumaran: so it’s like difficult in different ways. So.
332 00:29:01.510 ⇒ 00:29:03.110 joshuadeveyra: But yeah, okay, sure.
333 00:29:03.890 ⇒ 00:29:05.449 Uttam Kumaran: Okay, so let me slack you about it.
334 00:29:08.120 ⇒ 00:29:09.420 Uttam Kumaran: Okay, great.
335 00:29:11.340 ⇒ 00:29:13.930 Uttam Kumaran: Okay. Thanks, guys. We’ll talk in slack.
336 00:29:14.520 ⇒ 00:29:15.839 joshuadeveyra: Okay, thanks. Everyone. Bye, bye.
337 00:29:15.840 ⇒ 00:29:17.180 Nicolas Sucari: Thanks, thanks, bye, bye.