Meeting Title: Brainforge AI Integration Discussion Date: 2025-12-08 Meeting participants: Clarence Stone, Uttam Kumaran


WEBVTT

1 00:01:26.050 00:01:26.980 Uttam Kumaran: Thanks, Andrew.

2 00:01:28.940 00:01:30.290 Clarence Stone: Hey, what’s up?

3 00:01:30.290 00:01:31.629 Uttam Kumaran: Hey, how are ya?

4 00:01:32.420 00:01:35.230 Clarence Stone: Good! Dude, it’s, it’s giving me a rough week.

5 00:01:35.830 00:01:36.670 Uttam Kumaran: Why?

6 00:01:36.820 00:01:41.079 Clarence Stone: I have so much, so many meetings that are stacked on top of each other.

7 00:01:41.270 00:01:44.460 Uttam Kumaran: I know.

8 00:01:44.730 00:01:49.669 Uttam Kumaran: Yeah, tell me about it. So they actually cleared up for me. A couple things got moved, so…

9 00:01:49.800 00:01:51.519 Uttam Kumaran: And actually get some work done.

10 00:01:55.630 00:02:02.310 Uttam Kumaran: Yeah, she didn’t say that… she didn’t, like, accept the invite, so I’m not sure…

11 00:02:03.290 00:02:06.559 Uttam Kumaran: I may be able to give you this time back.

12 00:02:07.560 00:02:08.289 Clarence Stone: Sweet.

13 00:02:28.080 00:02:30.010 Clarence Stone: I haven’t even looked this person up.

14 00:02:34.320 00:02:36.050 Uttam Kumaran: She’s a PM at Bain.

15 00:02:38.160 00:02:39.150 Clarence Stone: Nice.

16 00:02:42.120 00:02:42.800 Uttam Kumaran: Yeah, what’s…

17 00:03:22.100 00:03:28.929 Uttam Kumaran: Dude, you know what my urge is now? It’s like, we have all these, like, expenses for, like, these apps, like, Clockify, and I’m like…

18 00:03:29.670 00:03:32.440 Uttam Kumaran: I wanna just build it myself.

19 00:03:32.440 00:03:33.550 Clarence Stone: Yeah, yeah.

20 00:03:33.550 00:03:35.889 Uttam Kumaran: But it’s, like, not enough money

21 00:03:36.310 00:03:44.119 Uttam Kumaran: Like, the savings is not worth me building it just yet. But, like, when we get to a certain scale, it will be, for sure.

22 00:03:44.640 00:03:46.000 Clarence Stone: Yeah.

23 00:03:47.170 00:03:52.379 Uttam Kumaran: Like, Zoom is not something, like, I’m trying to replace, but there’s a few other apps that we use.

24 00:03:52.580 00:03:54.409 Uttam Kumaran: that I’m like, oh…

25 00:03:55.250 00:04:00.259 Uttam Kumaran: Like, we had a half-decent product manager, we could just build a couple things that we need.

26 00:04:00.710 00:04:02.150 Uttam Kumaran: And own it, you know?

27 00:04:02.310 00:04:06.320 Clarence Stone: Yeah, I think it fits for a lot of things that…

28 00:04:06.430 00:04:11.649 Clarence Stone: like, even operate. I can imagine a future where you guys figure out how to do it yourself.

29 00:04:12.200 00:04:14.640 Uttam Kumaran: Yeah, yeah, exactly. Yeah.

30 00:04:16.100 00:04:21.010 Clarence Stone: But there’s no way you’re gonna host a cloud conferencing platform, it even worth it.

31 00:04:21.019 00:04:21.989 Uttam Kumaran: Yeah, yeah, yeah, exactly.

32 00:04:21.990 00:04:22.670 Clarence Stone: Right, yeah.

33 00:04:22.670 00:04:23.310 Uttam Kumaran: Exactly.

34 00:04:27.090 00:04:34.380 Uttam Kumaran: I mean, any vendor we use, like, as a means to help our future selves, I just make sure that they have really good APIs.

35 00:04:37.040 00:04:42.730 Uttam Kumaran: You know, because eventually, I don’t think we’ll spend much time in operating, like, the interface will be through chat or…

36 00:04:43.590 00:04:46.649 Uttam Kumaran: It’ll… it just may be holding the relationships.

37 00:04:47.500 00:04:50.009 Uttam Kumaran: You know, but we’ll be writing to the system, yeah.

38 00:04:50.150 00:04:55.350 Clarence Stone: I sent you that link to pieces, you should check it out. It is a really cool idea.

39 00:04:55.350 00:04:57.100 Uttam Kumaran: Yeah, you do… did you do a demo?

40 00:04:57.450 00:05:00.249 Clarence Stone: So I was testing it out on my own yesterday.

41 00:05:00.880 00:05:02.639 Clarence Stone: And it works pretty well.

42 00:05:04.360 00:05:16.040 Clarence Stone: It dug through all my Teams, chats, and my vicinity emails related to Casper, and gave me a really good set of insights, and I hadn’t even started using it yet.

43 00:05:22.880 00:05:25.080 Uttam Kumaran: Where’d you send that? Oh, in the leadership channel?

44 00:05:25.080 00:05:25.770 Clarence Stone: Yeah.

45 00:05:29.220 00:05:30.720 Uttam Kumaran: Oh, they’re using it to Casper?

46 00:05:31.930 00:05:36.830 Clarence Stone: No, no, no, I used it for all my Casper work to see if it would work, and it worked.

47 00:05:36.830 00:05:43.100 Uttam Kumaran: So this is done at the individual level, so it’s for the end user to track what they’ve been doing on a daily basis.

48 00:05:46.780 00:05:48.449 Uttam Kumaran: Andrew, how’d you find these guys?

49 00:05:49.630 00:05:54.600 Clarence Stone: I was interested because their foundation is local AI.

50 00:05:55.470 00:05:56.309 Uttam Kumaran: Huh.

51 00:05:57.340 00:06:03.120 Clarence Stone: And, when I looked at who was applying, long-term memory capabilities, this.

52 00:06:03.120 00:06:03.490 Uttam Kumaran: Yeah.

53 00:06:03.490 00:06:04.410 Clarence Stone: came up.

54 00:06:13.330 00:06:14.670 Uttam Kumaran: Interesting.

55 00:06:15.220 00:06:20.780 Uttam Kumaran: I mean, I’ll try it, I don’t… if, like… I haven’t burn all…

56 00:06:20.780 00:06:21.320 Clarence Stone: cursor?

57 00:06:21.320 00:06:26.710 Uttam Kumaran: Granola has… like, yeah, I mean, for me, it’s like, if it’s better than granola, then I would just turn off granola.

58 00:06:33.560 00:06:39.750 Clarence Stone: Yeah, oh, interesting, so it runs on all of your… They’re so good.

59 00:06:40.810 00:06:42.650 Uttam Kumaran: Oh, you used it for stand-up, it was good?

60 00:06:42.980 00:06:43.830 Clarence Stone: Yeah.

61 00:06:44.080 00:06:47.160 Clarence Stone: That’s why I was like, hey, maybe, maybe this is cool.

62 00:06:49.430 00:06:54.279 Clarence Stone: My only thing is, at your enterprise level, I would rather all these conversations be…

63 00:06:54.490 00:07:00.329 Clarence Stone: sent back. So, I don’t know how this architecture works as a whole, I just tried the free option.

64 00:07:00.740 00:07:03.089 Uttam Kumaran: Well, that’s what my point is that,

65 00:07:03.680 00:07:11.870 Uttam Kumaran: like, I just want to build a granola, like, the platform is pretty close to it also, and basically, like, imagine granola except, like.

66 00:07:12.170 00:07:14.169 Uttam Kumaran: It had access to everything.

67 00:07:14.980 00:07:15.890 Uttam Kumaran: You know?

68 00:07:16.140 00:07:19.559 Clarence Stone: Yeah. Like, it had the MCP connectors, it would.

69 00:07:19.560 00:07:26.140 Uttam Kumaran: Because, dude, ultimately, like, this is my… my point is that, like, the agents have to start making decisions for you.

70 00:07:26.270 00:07:29.640 Clarence Stone: Yeah. And I’m not actually, like, so interested in, like, getting…

71 00:07:29.640 00:07:42.120 Uttam Kumaran: Like, if we design the whole thing with the human in mind, then you never get to a reality where, like, an agent is looking at an activity stream of stuff, and is like, oh, okay, a stand-up just finished.

72 00:07:42.460 00:07:44.220 Uttam Kumaran: Like, I should probably go…

73 00:07:44.480 00:07:50.320 Uttam Kumaran: send a note, or, like, update a ticket. If I… if I build everything, like, locally per laptop.

74 00:07:50.950 00:07:59.129 Uttam Kumaran: then I don’t get any benefit of the fact that, like, the system is gonna start making decisions. And so, like, I’m… our ergonomics need to be with that in mind.

75 00:07:59.570 00:08:06.670 Clarence Stone: Well, the way this works is it, like, it’s just holding the context locally, so the.

76 00:08:06.670 00:08:07.550 Uttam Kumaran: Yeah, yeah, yeah.

77 00:08:07.550 00:08:12.180 Clarence Stone: But it’s gonna use, you know, like any Frontier

78 00:08:12.390 00:08:14.399 Clarence Stone: API that you throw at it.

79 00:08:16.300 00:08:17.710 Uttam Kumaran: Hmm, okay.

80 00:08:17.980 00:08:27.609 Clarence Stone: So, the only downside is, I don’t want that… that context to be local in your case. I want all the employees to use it, and you to be able to dig into it as well.

81 00:08:27.740 00:08:31.080 Clarence Stone: Right, you should be able to talk to, like, MyPiece’s co-pilot.

82 00:08:31.080 00:08:32.000 Uttam Kumaran: Yeah, yeah, yeah.

83 00:08:32.020 00:08:33.159 Clarence Stone: been working on.

84 00:08:34.270 00:08:36.639 Uttam Kumaran: I mean, this is where I think granola will head.

85 00:08:37.409 00:08:40.340 Uttam Kumaran: But the problem with granola…

86 00:08:40.710 00:08:49.070 Uttam Kumaran: It’s just what’s gonna be the problem with all these things, dude? You can’t link… like, your furthest integration is, like, what you’re biasing for. Like, for us.

87 00:08:49.370 00:08:52.709 Uttam Kumaran: Like, at what point can I bring my Slack messages into Granola?

88 00:08:52.830 00:08:54.420 Uttam Kumaran: I don’t know, 2 years?

89 00:08:54.970 00:08:56.009 Uttam Kumaran: What are they gonna build out?

90 00:08:56.010 00:08:57.659 Clarence Stone: I haven’t seen peeved into it, man.

91 00:08:57.900 00:08:59.190 Clarence Stone: It’s crazy.

92 00:08:59.190 00:09:02.690 Uttam Kumaran: I know, but Slack, you can’t facilitate this via MCP, like.

93 00:09:02.690 00:09:03.840 Clarence Stone: Oh!

94 00:09:03.840 00:09:06.289 Uttam Kumaran: You’re not, like, I’m talking about, I wanna…

95 00:09:06.480 00:09:14.520 Uttam Kumaran: like, I want to basically run rag over all of the Slack messages when I ask a question. It’s not as simple as, like.

96 00:09:14.850 00:09:17.770 Uttam Kumaran: go find the Slack message that’s relevant. It’s like…

97 00:09:18.000 00:09:21.690 Uttam Kumaran: you have to do a loop through everything. We have to, like, basically…

98 00:09:21.890 00:09:26.219 Uttam Kumaran: Vectorize and run pre-filtering to, like, store relevant.

99 00:09:26.220 00:09:26.920 Clarence Stone: Slack?

100 00:09:26.930 00:09:30.390 Uttam Kumaran: Slack things, right? That’s, like, I don’t know when anybody’s gonna do that.

101 00:09:30.620 00:09:36.609 Uttam Kumaran: It’s just gonna take a while, because all the basic use cases, like, Prep for a meeting,

102 00:09:37.170 00:09:43.749 Uttam Kumaran: Yeah, like… pull a relevant Google Drive doc that you, like, already link, like, that’s fine, but…

103 00:09:44.080 00:09:47.410 Uttam Kumaran: That’s, like, that’s whatever. That’s, like, table stakes.

104 00:09:50.000 00:09:56.849 Uttam Kumaran: You know, because also, look, I want to… for example, like, the client hub is a really good example of, like, something we’re trying to build, which is, like.

105 00:09:57.040 00:10:08.019 Uttam Kumaran: I want a single point where you can ask a question about a client. All meetings, docs, code bases, everything is pre-loaded. And this is the thing, I don’t have to support

106 00:10:08.140 00:10:12.740 Uttam Kumaran: N number of integrations. I just have to support the integrations that I know Brainford uses.

107 00:10:13.650 00:10:21.200 Clarence Stone: Yeah, so, right now, it does what you’re describing for my Gmail and my Teams chats. It is… it is ragging.

108 00:10:22.970 00:10:23.480 Uttam Kumaran: Great.

109 00:10:23.480 00:10:29.879 Clarence Stone: I just found this image, I just threw it in your chat. I think they’re gonna… I mean, they’re already connecting to GitHub.

110 00:10:30.220 00:10:32.740 Clarence Stone: So I think they’re eventually gonna expand that.

111 00:10:42.700 00:10:51.989 Clarence Stone: the weird part is that it plugs directly into the coding agent, and it actually helps you keep track of the work that you’ve done as well on top of that. Like, so, like.

112 00:10:52.180 00:10:59.580 Clarence Stone: if you get a linear ticket, you work on it in cursor, right, Pieces already knows that you finished it. So in chat…

113 00:10:59.580 00:11:00.340 Uttam Kumaran: Oh, interesting.

114 00:11:00.340 00:11:01.300 Clarence Stone: mini chat.

115 00:11:03.580 00:11:09.670 Uttam Kumaran: Yeah, I mean, I would probably… I mean, like, dude, eventually, like… this is why I’m sure, like, Palantir does this, is, like.

116 00:11:09.960 00:11:14.070 Uttam Kumaran: you just have a Palantir machine, and Palantir, they just ship their own pieces, like…

117 00:11:14.350 00:11:14.800 Clarence Stone: Yeah.

118 00:11:14.800 00:11:20.749 Uttam Kumaran: This is where, like, I would just ship a Brainforge extension, where I’d say, install the Brainforge extension, what’s this gonna do is it’s gonna, like.

119 00:11:20.880 00:11:26.400 Uttam Kumaran: Suck up all your stuff. And this is how you actually get employees to want to use

120 00:11:26.570 00:11:34.490 Uttam Kumaran: those, like, remote certificate things, right? As you say, like, these are all the benefits you get, versus, like, we’re just spying on you.

121 00:11:34.940 00:11:39.910 Uttam Kumaran: I don’t even care about the spying, I guess you probably have to do it at some point for compliance, but like, yeah, I don’t…

122 00:11:40.230 00:11:47.900 Uttam Kumaran: I actually care. For me also, dude, I want to prove that all of our teams are using all the features, like…

123 00:11:47.900 00:11:48.560 Clarence Stone: Yep.

124 00:11:48.960 00:11:56.309 Uttam Kumaran: like, Gabe’s KPIs should be… Hey, at least once a week, people should be using the platform for something.

125 00:11:56.650 00:12:01.650 Uttam Kumaran: And then it should be once a day. Or platform or platform-related features, like, you know?

126 00:12:02.150 00:12:11.469 Clarence Stone: Yeah, so maybe a good exercise for Dave is to, like, look at a lot of these different types of implementations, and then figure out what the best mix is for Brainforge’s platform.

127 00:12:11.720 00:12:12.470 Uttam Kumaran: Yeah.

128 00:12:12.470 00:12:24.749 Clarence Stone: it’s really a little bit of all of these things, right? I think for, like, cursor capability back into the cloud, you might still have to have some sort of local memory storage, and then… but it should always push back up.

129 00:12:24.940 00:12:30.510 Uttam Kumaran: No, but dude, Cursor, everybody has Cursor installed, and I get the cursor analytics on what people are calling.

130 00:12:31.630 00:12:36.239 Clarence Stone: Yeah, but the MCP integrator connects what you did in linear.

131 00:12:40.550 00:12:46.419 Uttam Kumaran: Well, I guess, like, even longer term, dude, like, what I’m thinking about, like, if you’re talking about executing code.

132 00:12:46.730 00:12:50.530 Uttam Kumaran: More of my concern is getting the ticket right.

133 00:12:51.570 00:12:52.370 Clarence Stone: Yes.

134 00:12:52.370 00:13:07.879 Uttam Kumaran: Because, like, if the ticket isn’t good enough, then the AI can’t take it, and for every ticket, basically, we’re… one of the things on Gabe’s roadmap is a ticket groomer, where basically it goes through and makes sure that the ticket has enough context.

135 00:13:08.110 00:13:12.850 Uttam Kumaran: Unfortunately, we don’t have enough people creating tickets for me to do that yet, but…

136 00:13:13.110 00:13:15.150 Uttam Kumaran: Basically, whenever you create a ticket.

137 00:13:15.280 00:13:18.709 Uttam Kumaran: You will have to go through the grooming agent first.

138 00:13:18.920 00:13:22.390 Uttam Kumaran: And then ideally, like, there’s, like, a nice loop where, like.

139 00:13:22.960 00:13:31.050 Uttam Kumaran: grooming agent considers a ticket well-groomed if it’s actually able to be built by AI, you know, partially. So the AI takes the first stab at, like.

140 00:13:31.480 00:13:32.870 Uttam Kumaran: The plan, basically.

141 00:13:34.660 00:13:41.820 Clarence Stone: Yeah, so check this out. It actually pulls… Pieces Integration pulled this entire work activity

142 00:13:42.250 00:13:47.069 Clarence Stone: And actually created that from the ticket that they were working on.

143 00:13:50.280 00:13:54.119 Uttam Kumaran: Mmm… I see, okay, okay.

144 00:13:54.120 00:14:00.570 Clarence Stone: Yeah, so it’s aware of the ticket that you’re working on, and then even preps your cursor rules.

145 00:14:01.830 00:14:02.720 Uttam Kumaran: Hmm.

146 00:14:02.720 00:14:03.499 Clarence Stone: To fit into it.

147 00:14:03.500 00:14:04.920 Uttam Kumaran: I see, I see. Yeah, yeah, yeah.

148 00:14:04.920 00:14:12.419 Clarence Stone: really neat. I was really surprised how well that worked. So, I, like, I think this piece will be at least very helpful for

149 00:14:12.590 00:14:14.250 Clarence Stone: DreamForge’s workflow.

150 00:14:14.510 00:14:15.240 Uttam Kumaran: Yeah.

151 00:14:18.820 00:14:22.670 Uttam Kumaran: Yeah, the other part is we just, like, not all of our people think about

152 00:14:22.990 00:14:27.620 Uttam Kumaran: doing… using AI for their work every day, so most of it I had to, like, kind of force.

153 00:14:28.800 00:14:32.310 Uttam Kumaran: You know, like, I can’t trust people to use the AI.

154 00:14:33.390 00:14:34.480 Uttam Kumaran: just yet.

155 00:14:36.210 00:14:38.409 Uttam Kumaran: On their own volition.

156 00:14:39.820 00:14:41.360 Uttam Kumaran: Like, I have to push it.

157 00:14:41.360 00:14:42.100 Clarence Stone: Yeah.

158 00:14:43.520 00:14:44.650 Uttam Kumaran: That’s the tough part.

159 00:14:45.620 00:14:48.940 Uttam Kumaran: Like, it has to be always kind of running in the background somewhere.

160 00:14:49.840 00:14:53.779 Uttam Kumaran: Because ultimately, dude, my goal is to, like, if we’re able to take

161 00:14:54.220 00:15:00.730 Uttam Kumaran: the easiest 20-30% of tickets, pass it to AI, then that’s immediate cost savings.

162 00:15:02.310 00:15:07.710 Uttam Kumaran: So you have to back that problem up into, okay, how do you identify the 20 or 30% of tickets that are tough?

163 00:15:08.160 00:15:12.050 Uttam Kumaran: Okay, you need to start to… Understand what toughness means.

164 00:15:12.230 00:15:16.229 Uttam Kumaran: Right? So you have to start from, like, okay, is a ticket groomed?

165 00:15:16.420 00:15:26.640 Uttam Kumaran: okay, then, have we tried? Do we have a definition? And also, the lovely thing is, I’m not a do-everything type of place. We only… we have a finite amount of tasks that we do.

166 00:15:27.190 00:15:29.910 Uttam Kumaran: Maybe it’s a couple hundred, I don’t know, but it’s like…

167 00:15:30.580 00:15:33.599 Uttam Kumaran: it’s finite, you know, and I’m sure there’s a bell curve.

168 00:15:35.730 00:15:41.789 Uttam Kumaran: And so, as long as we just get good at building work streams that can take on the 20 or 30% of things.

169 00:15:43.280 00:15:47.889 Uttam Kumaran: I mean, like, I would spend money on that easily, because…

170 00:15:48.050 00:15:57.280 Uttam Kumaran: Then we have… this is how you truly can go from one engineer on 3 clients, and one on five, and one on 7. It’s like, 50% of the work is being done by AI.

171 00:15:57.400 00:15:57.960 Uttam Kumaran: Yeah.

172 00:15:57.960 00:16:00.999 Clarence Stone: It’s really just PR requests and code reviews at that point.

173 00:16:01.000 00:16:01.440 Uttam Kumaran: Yeah.

174 00:16:01.440 00:16:13.379 Clarence Stone: you do the heavy lifting, you know, that the AI did do. I guess you start with a classifier, right? You might as well just, like, skip, like, you know, any of the t-shirt sizing, if you’re doing any of that.

175 00:16:13.380 00:16:21.380 Uttam Kumaran: Oh, exactly, dude. We actually started the kind of, like, classifier thing. It just isn’t done yet. Like, basically, every ticket needs to get

176 00:16:22.620 00:16:33.449 Uttam Kumaran: Basically, I told them to take every other… take every past ticket and classify them so we can start to see trends. And ideally, we should agree on, like, what are the 500 types of work that we do.

177 00:16:33.650 00:16:36.709 Uttam Kumaran: And, like, someone needs to, like, come up with that taxonomy.

178 00:16:37.330 00:16:38.120 Uttam Kumaran: You know?

179 00:16:39.240 00:16:40.339 Uttam Kumaran: And then basically, like.

180 00:16:40.340 00:16:45.940 Clarence Stone: You can just sling it over to AI, because if the classifier says it’s easy, you know, Yeah.

181 00:16:45.940 00:16:49.409 Uttam Kumaran: Yeah, so you… and then also, like, based on the classification.

182 00:16:49.630 00:16:55.179 Uttam Kumaran: you need to have certain requirements in the ticket in order for it to be done by… and all these, you can just focus on, like.

183 00:16:55.720 00:16:59.170 Uttam Kumaran: The bottom of the barrel, shit, like… update.

184 00:16:59.580 00:17:01.420 Uttam Kumaran: The dashboard title.

185 00:17:01.590 00:17:02.499 Uttam Kumaran: You know? Yep.

186 00:17:02.710 00:17:06.630 Uttam Kumaran: create metric. Like, those are the things where

187 00:17:07.369 00:17:12.290 Uttam Kumaran: But see, this is the thing, I don’t want to trust someone to think that AI can do it.

188 00:17:12.490 00:17:13.839 Uttam Kumaran: Because I think most of the…

189 00:17:14.220 00:17:23.620 Uttam Kumaran: People will think, oh, AI can’t do it. In fact, I’m seeing it, because most of my company doesn’t use codecs. When they can get their whole job done in 5 minutes, they just use codecs.

190 00:17:23.869 00:17:25.090 Uttam Kumaran: Like I said, too.

191 00:17:25.640 00:17:26.890 Clarence Stone: I love it.

192 00:17:26.890 00:17:32.339 Uttam Kumaran: Literally, I will be in there, I’m like, did you try giving this to Codex? They’re like, oh, I don’t know, like, I haven’t tried Codex yet. I’m like…

193 00:17:33.330 00:17:35.629 Uttam Kumaran: I’m just gonna blow my fucking brains out.

194 00:17:37.440 00:17:44.669 Clarence Stone: Dude, the epic version of this would be if it went to… Code Rabbit after.

195 00:17:44.670 00:17:46.020 Uttam Kumaran: Yeah, yeah, yeah, yeah.

196 00:17:46.020 00:17:51.820 Clarence Stone: And, like, only if it failed Code Rabbit testing, it would kick back to user.

197 00:17:52.530 00:18:03.669 Uttam Kumaran: Well, dude, but that’s the problem with data work, is it’s not like, you can’t… you can’t measure the success through, like, a lot of unit testing. Like, there has to be… because you’re looking at, like, a graph, or…

198 00:18:04.620 00:18:10.120 Uttam Kumaran: I mean, like, yeah, it’s… which makes this, like, a more unique problem, because I’m not shipping, like.

199 00:18:10.800 00:18:15.019 Uttam Kumaran: ship a button that clicks this and calls this endpoint. Like, these are a little bit more nuanced problems.

200 00:18:15.020 00:18:22.329 Clarence Stone: You know what? Maybe then a non-technical person could do that, right? Like, hey, did the graph now show this?

201 00:18:22.970 00:18:24.650 Clarence Stone: Was the title change?

202 00:18:24.810 00:18:25.350 Clarence Stone: Like.

203 00:18:25.350 00:18:26.050 Uttam Kumaran: Yeah.

204 00:18:26.050 00:18:27.770 Clarence Stone: Required debt to review that.

205 00:18:28.070 00:18:30.820 Uttam Kumaran: No, no, no, it doesn’t. I mean, that’s the thing, like.

206 00:18:30.820 00:18:32.500 Clarence Stone: Anybody actually beat that.

207 00:18:32.500 00:18:33.160 Uttam Kumaran: Yeah.

208 00:18:33.790 00:18:43.570 Uttam Kumaran: That’s what… but that’s what I’m saying, like, I want the… it’s actually easier to review code than it is to write it, and it’s easier to review these changes, so then that’s why, like, the average engineer

209 00:18:44.070 00:18:52.900 Uttam Kumaran: Their job is more writing tickets and reviewing code and suggesting fixes versus actually doing it, and then they can take on more clients per person.

210 00:18:53.270 00:18:54.600 Clarence Stone: Yep, I like that.

211 00:18:56.020 00:19:05.990 Uttam Kumaran: Or, you know, if there’s a logic… if there’s a natural, like, hey, I can’t… I can’t hold more than 7 clients in my brain, then I can actually just get, like, more junior people

212 00:19:06.720 00:19:08.740 Uttam Kumaran: Right, you lower the rate of the people.

213 00:19:09.470 00:19:17.490 Uttam Kumaran: I don’t need, like, super, super technical people on every client. I can run it with, like, cheaper people.

214 00:19:20.810 00:19:21.520 Clarence Stone: Yeah.

215 00:19:23.210 00:19:29.219 Uttam Kumaran: Yeah, dude, if you, if you, if you buy me back, like, 20 hours a week, I’ll build this for you by end of January.

216 00:19:30.790 00:19:36.479 Uttam Kumaran: I swear, I swear. I’ve had this stuff written down for, like, a year and a half.

217 00:19:36.480 00:19:39.160 Clarence Stone: Oh, you got something going on this Friday? What is this?

218 00:19:39.160 00:19:39.970 Uttam Kumaran: Yeah.

219 00:19:40.300 00:19:44.470 Clarence Stone: Okay, let’s see. Okay, I’m gonna try my best to make it.

220 00:19:45.290 00:19:46.240 Uttam Kumaran: No problem if you can.

221 00:19:46.240 00:19:48.450 Clarence Stone: I’m still over.

222 00:19:48.840 00:19:49.489 Uttam Kumaran: It’s not a break.

223 00:19:49.490 00:19:50.210 Clarence Stone: Yeah.

224 00:19:50.210 00:19:53.290 Uttam Kumaran: I mean, I’m gonna be working right up, right up too, right now.

225 00:20:03.290 00:20:05.759 Uttam Kumaran: Okay, well, I guess this is a dud.

226 00:20:05.760 00:20:07.459 Clarence Stone: Yeah, it’s all good.

227 00:20:07.910 00:20:09.250 Uttam Kumaran: Alright, dude, well…

228 00:20:09.250 00:20:11.580 Clarence Stone: Well, I’ll catch up with you in, like, the top of the hour.

229 00:20:11.580 00:20:12.359 Uttam Kumaran: Yeah, yeah.

230 00:20:12.360 00:20:14.619 Clarence Stone: Okay, okay, perfect. Perfect. Alright. Cool.

231 00:20:14.840 00:20:15.720 Clarence Stone: Very good.