Meeting Title: Uttam <> Atharv Date: 2024-08-06 Meeting participants: Atharv Gudi, Uttam Kumaran


WEBVTT

1 00:08:48.210 00:08:49.149 Atharv Gudi: Hey! With them!

2 00:08:49.440 00:08:50.690 Uttam Kumaran: Hey! How are you?

3 00:08:50.840 00:08:51.840 Atharv Gudi: Good, good.

4 00:08:54.640 00:08:56.272 Uttam Kumaran: Good. Yeah, I’ve been.

5 00:08:56.990 00:08:58.897 Uttam Kumaran: I’ll show you what I’m working on.

6 00:09:00.340 00:09:09.784 Uttam Kumaran: we’re we’re doing a bunch of. I have a couple of like sales calls that I’m involved in today. And one of the things that I wanted to do for a long time is put together.

7 00:09:10.470 00:09:38.809 Uttam Kumaran: is put together really nice resumes for the folks on our team, so that the process when I go after like a new client is, I send them profiles of folks that were that we have on the team right? Because our business is, it’s not one is no longer about me, so like people can find my profile. But it’s actually about our team. And so one of the things I’ve been doing and I I send a message over to the team is, I’m I asked for everybody’s resumes, cause I’m like creating really nicely formatted profiles.

8 00:09:38.810 00:09:49.619 Uttam Kumaran: And so the process is. And I and I literally just sent this. That was a note I was writing before this. I’m talking to someone, and they’re like, Hey, we’re looking for a data scientist with like, XYZ

9 00:09:49.710 00:10:14.460 Uttam Kumaran: needs. And we have this ABC sort of problem. And I’m like, cool. Here’s a couple of people on our team that could fit this that could fit this mold that you need and so one like before to give you a sense of what it was like before. Before I just take the call, and I’m like sort of figuring out as I go. I know where I was. I was all I was kind of like pitching myself to. Now I’m I’m like

10 00:10:14.520 00:10:32.639 Uttam Kumaran: I’m not really for hire, like. Personally, I will work like you could see me how I’m involved in everything. But I’ll do stuff capacity. I’m not really like planning on going as a dedicated engineer, if I if I can. You know, make that a reality. So what I have is

11 00:10:32.912 00:10:50.627 Uttam Kumaran: I’m making profiles for everybody in Figma. So that and send them over being like, Hey, these are a couple of people on our team. These all have a really great information. The other thing is, here’s an example of like what one looks like for, like a like a don’t maybe a competitor, but like another kind of person in our space.

12 00:10:50.900 00:11:07.869 Uttam Kumaran: Usually they just send some people just send their resume. These guys are a little bit above, and that they have a little bit of summary page. And for me, I always think about like what differentiates us in, like every step of the way. And one I’m like. This looks ugly. So I want to make one that looks

13 00:11:07.930 00:11:16.389 Uttam Kumaran: good. And then I just wanted to have a little bit more flavor, like a summary and kind of like this is.

14 00:11:16.420 00:11:23.951 Uttam Kumaran: I was just looking at theirs. It’s like tough to read. There’s experiences here. There’s also experiences here.

15 00:11:25.230 00:11:29.350 Uttam Kumaran: like multitasking as an area of strength is like.

16 00:11:29.700 00:11:35.550 Uttam Kumaran: you know, it’s just like, resume padding. It’s like classic, just like trying to fill this up. So

17 00:11:35.886 00:11:38.840 Uttam Kumaran: yeah, that’s kind of like what I was just working on. So.

18 00:11:39.930 00:11:40.600 Atharv Gudi: Okay.

19 00:11:43.930 00:11:46.249 Atharv Gudi: you do have my resume already. Right?

20 00:11:46.610 00:11:55.570 Uttam Kumaran: I think I do. I’m gonna be putting I these. I just needed these guys done urgently. But I I’m gonna have everybody’s in a format like this.

21 00:11:55.880 00:11:56.260 Atharv Gudi: Okay.

22 00:11:58.822 00:12:12.450 Uttam Kumaran: But also people are feel people can definitely like I can design these without the brain forge logo. If you want to take this. The nice thing about figma is like, I just create it really quickly. And then I can just export this as a Pdf, so it’s it’s really nice.

23 00:12:13.200 00:12:13.950 Atharv Gudi: Okay.

24 00:12:18.240 00:12:22.879 Atharv Gudi: Where do I come in again? I think I’ve lost a bit of the I’m not.

25 00:12:23.620 00:12:25.819 Atharv Gudi: So I’m supposed to.

26 00:12:26.610 00:12:30.410 Uttam Kumaran: Oh, no, this isn’t anything up to do with you. I was. It’s just what I was working on.

27 00:12:30.780 00:12:34.419 Uttam Kumaran: Oh, yeah, this is not all of you. I was sharing what I was doing today.

28 00:12:34.420 00:12:35.640 Atharv Gudi: Okay.

29 00:12:36.330 00:12:40.310 Atharv Gudi: this makes a lot. Yeah, this makes it really easy for the

30 00:12:40.530 00:12:55.239 Atharv Gudi: oh, speaking of which I was I was wondering, I, at the end of the internship. Usually we get in India. We usually just get a certificate of some sort of state, and to attend the internship. Is it possible to get one for me as well at.

31 00:12:55.240 00:13:03.820 Uttam Kumaran: Yeah, I’ve actually, I’ve actually given a couple out. So I have. I can give you not only like an offer letter. But also, yeah, like a letter of

32 00:13:03.870 00:13:05.940 Uttam Kumaran: completion or something like that. I mean.

33 00:13:06.477 00:13:29.980 Uttam Kumaran: let me know. Like, if there’s anything in particular, I mean, kind of the way I was thinking about. And we should. We could talk about that today. One of the things I was thinking about is one it would be great for folks to kind of do a little bit of like a write up of like stuff you learn. Of course, I think today we’re gonna a little bit about how you can continue to, you know, work closely with us. But that was what I was thinking. One, I think.

34 00:13:29.990 00:13:54.830 Uttam Kumaran: for us is where we’ll publish it on our blog and publish on Linkedin. I think. Second is you could. I think it’s a good asset for you to have to be able to link back to our blog, to publish on your own Linkedin, and to like kind of hype yourself up, and then every company will like like that and stuff like that. So let’s plan on. Let’s plan on doing that. I’ll just take some notes.

35 00:14:06.910 00:14:07.820 Uttam Kumaran: hmm!

36 00:14:17.160 00:14:23.439 Atharv Gudi: Think I’m also getting in touch with Ryan tomorrow we have. But I have planned something tomorrow to meet up and finish

37 00:14:24.630 00:14:25.470 Atharv Gudi: the okay.

38 00:14:25.470 00:14:26.130 Uttam Kumaran: Basically.

39 00:14:42.760 00:14:46.190 Uttam Kumaran: So yeah, I’m gonna do the offer letter.

40 00:15:07.330 00:15:13.380 Uttam Kumaran: Okay, cool. And then, so you mentioned that? August. So August 19th

41 00:15:15.400 00:15:18.129 Uttam Kumaran: week of August 19.th You’re gonna be out.

42 00:15:18.130 00:15:19.040 Atharv Gudi: Yeah, so.

43 00:15:19.040 00:15:25.719 Uttam Kumaran: I think today I just want to plan a couple of things for the next like 2 weeks. And then also, let’s

44 00:15:25.770 00:15:28.240 Uttam Kumaran: like, I’ll kind of give you a sense of

45 00:15:28.970 00:15:44.940 Uttam Kumaran: I think, where there there’s still like opportunity to kind of like work together, I mean my goal, and then I’ll kind of talk to you about like I’m thinking of. I think some other folks, some other interns, are also interested. So I’ll kind of give you a sense like what I’m thinking.

46 00:15:46.550 00:15:53.179 Uttam Kumaran: so kind of like the project between now and then. I kind of would love some help.

47 00:15:54.609 00:16:01.800 Uttam Kumaran: exploring some of these AI like ideas. I have. A couple of them.

48 00:16:02.964 00:16:09.919 Uttam Kumaran: Tend to be around automating both our engineering and our project management tasks

49 00:16:10.000 00:16:32.690 Uttam Kumaran: to be like very broad which is on the engineering side, like to walk you through like what engineers do like. They get requirements from the project manager. They have some understanding of code. They then may need to engage with new technologies. They write some code, it gets reviewed, and then it gets put into production at each of those stages there are.

50 00:16:32.870 00:16:49.470 Uttam Kumaran: There’s a need to reference, some sort of knowledge base. There’s need to talk to the team. There’s a lot of information brokering that happens. And so one of the things that I’m really interested in using AI for is to speed all those wheels up.

51 00:16:49.982 00:16:59.929 Uttam Kumaran: And one of the things that I kind of started writing down here. And I’ll send it to you. I’m in the utham Adarv

52 00:16:59.990 00:17:01.140 Uttam Kumaran: notion, Doc.

53 00:17:01.140 00:17:03.160 Atharv Gudi: Yeah, okay, and take a look at.

54 00:17:03.160 00:17:03.750 Uttam Kumaran: Yeah.

55 00:17:25.550 00:17:27.230 Uttam Kumaran: So I kind of just started like

56 00:17:27.430 00:17:36.490 Uttam Kumaran: throwing in notes here about like an AI project manager and an AI engineer.

57 00:17:36.490 00:17:37.380 Atharv Gudi: Okay.

58 00:17:38.570 00:17:40.995 Uttam Kumaran: What you’re gonna see is

59 00:17:43.910 00:17:50.588 Uttam Kumaran: what you’re gonna see in both of these docs is you’re gonna see? Like just like links to

60 00:17:51.670 00:17:59.620 Uttam Kumaran: basically, like, basically like links to different articles that I’ve been reading about how to use chat Gpt for like code reviews

61 00:17:59.660 00:18:07.720 Uttam Kumaran: or automatically commenting on like Github issues on the project manager side. A lot of it is around

62 00:18:08.304 00:18:10.480 Uttam Kumaran: like managing github issues.

63 00:18:11.197 00:18:13.770 Uttam Kumaran: I’ve had some ideas there about like

64 00:18:14.377 00:18:32.622 Uttam Kumaran: one of the things that Nico does every week is, he writes emails to clients about what the things we’ve done. A lot of that is already in issues. So can we like somehow generate that email? So kind of like the things that I’m interested in exploring is not really the

65 00:18:32.990 00:18:49.540 Uttam Kumaran: like, I’m definitely interested in like using some of these used cases to figure some of these technologies out. But more importantly, like we’re, we’re pat. I’ve passed the point where I’ve like I get a Claude and chat Gbt and stuff like that work, and I’m sure that you know you’ve used it, and you’re kind of familiar with, just like

66 00:18:49.560 00:18:51.349 Uttam Kumaran: you’re asking a knowledge base.

67 00:18:51.733 00:19:15.840 Uttam Kumaran: That knowledge base may or may not have access to the Internet, like perplexity has access to Internet. But you’re basically like telling it something. You’re giving it some text content. You’re asking it for clear instruction. This is going to be more tailored towards our business, and probably more specific in the actual asks that we have meaning we’re not going to have like an we don’t need like an open, ended Chatbot. We’re asking like, what the weather is right? We’re asking like a summary of

68 00:19:15.840 00:19:32.189 Uttam Kumaran: an issue. We’re asking a question on top of our code base, and those are. I think we can segment probably into like 5 or 10 key workflows or key, like solutions. That we need and one of the things that I think a good place to start

69 00:19:32.190 00:19:39.319 Uttam Kumaran: is and I. This is what I thought. But I think chatting over our code base is like an interesting place to start.

70 00:19:39.661 00:20:02.540 Uttam Kumaran: One is, we do ask a lot of questions already. Over our code base about like, how do I fix this problem? The second thing is, I already have a clod project that I shared, which is basically I’ve uploaded all of our code base into clod, and you can ask questions over it. I don’t know if you had a if you had a chance to like play around with that at all.

71 00:20:02.540 00:20:12.019 Atharv Gudi: Not with. I have not really been using Claude at all. I think Nico has been using more of it. He shows me how to use it, but I use chat gpt pretty much all the time instead.

72 00:20:12.740 00:20:32.690 Uttam Kumaran: Try. I think you should have a a invite to Claude or let me know if you don’t, because you should try using it. I created like Claude projects. I can think of some like custom Gpts. But the context window on Claude is a lot bigger. And so I’ve actually just uploaded like our entire code base into there. And I’ve been using it

73 00:20:32.710 00:20:35.809 Uttam Kumaran: to like, basically cut my dev time in like half.

74 00:20:35.980 00:20:38.699 Uttam Kumaran: And basically, that’s the goal is like.

75 00:20:38.990 00:20:48.019 Uttam Kumaran: I just want everybody to be able to speed up the amount of work they’re able to do. And so one thing I would suggest is to explore

76 00:20:48.090 00:20:51.100 Uttam Kumaran: is to explore the existing cloud project.

77 00:20:52.820 00:20:54.029 Uttam Kumaran: I’m just gonna write that.

78 00:20:54.030 00:20:57.279 Atharv Gudi: Claudinate again. I don’t think I have received any.

79 00:20:57.850 00:20:58.540 Uttam Kumaran: Okay.

80 00:20:59.000 00:21:00.019 Uttam Kumaran: let me check.

81 00:21:05.940 00:21:09.771 Uttam Kumaran: but you’ll see that I’ve set up a bunch of projects.

82 00:21:10.240 00:21:13.959 Uttam Kumaran: and so I think it’d be really cool for you to check that out.

83 00:21:19.410 00:21:20.160 Atharv Gudi: Okay.

84 00:21:36.280 00:21:37.499 Atharv Gudi: yeah. I got it. Now.

85 00:21:38.140 00:21:38.730 Uttam Kumaran: Okay.

86 00:21:45.490 00:21:48.619 Uttam Kumaran: she said. She should see under projects on the left, like

87 00:21:49.510 00:21:55.630 Uttam Kumaran: we’ve been. I’ve been creating a ton of different projects like the pool parts to go. Engineer project has

88 00:21:57.100 00:22:00.910 Uttam Kumaran: like custom instructions. But basically, I’ve uploaded like a lot of our

89 00:22:02.440 00:22:07.809 Uttam Kumaran: I’ve uploaded a lot of our code base already into there, and I use it to kind of ask questions.

90 00:22:11.030 00:22:15.719 Uttam Kumaran: so one thing that I’m interested in interested in exploring is doing this

91 00:22:16.050 00:22:21.390 Uttam Kumaran: a little in a little bit of a smarter way. Basically the problem with this. Now.

92 00:22:21.848 00:22:24.829 Uttam Kumaran: and I’m just gonna I’ll write this in the notes.

93 00:23:18.470 00:23:21.569 Uttam Kumaran: so there’s a couple. So if you just Google writing

94 00:23:22.110 00:23:23.090 Uttam Kumaran: like

95 00:23:24.395 00:23:53.734 Uttam Kumaran: chat Gpt over code base or AI over code base, you’re gonna get a bunch of different things. A lot of people have wrote wrote a lot of different stuff over it. One is, I would check out basically as many of the articles that I put in the AI engineer page as possible. These are a lot of things about how to use chat Bt with Github, but I also want us to explore this tool called flow wise. Wise is like, and basically like an Llm builder that’s open source.

96 00:23:54.337 00:23:58.970 Uttam Kumaran: that I’ve talked to a lot of friends about that are seeing a lot of success on it.

97 00:23:59.393 00:24:24.729 Uttam Kumaran: The problem is, I just don’t have any time to like go fix, figure it out. So I’m hoping that you can maybe poke around there and leverage flow wise to go see whether there’s any opportunity to. You know. Query our code base and write questions over it. Another great resource for this is I. I subscribe to this guy on Youtube, who does a bunch of different, like flow wise related tutorials.

98 00:24:25.210 00:24:27.069 Uttam Kumaran: I think that would be really

99 00:24:28.130 00:24:29.450 Uttam Kumaran: helpful to

100 00:24:29.520 00:24:33.709 Uttam Kumaran: take a look at I’m just going to put in like

101 00:24:34.630 00:24:38.168 Uttam Kumaran: I’m just gonna put in one of these in here. But

102 00:24:38.700 00:24:41.269 Uttam Kumaran: I’ll let you take a look at his profile

103 00:24:41.610 00:24:43.660 Uttam Kumaran: and basically like

104 00:24:43.910 00:24:47.070 Uttam Kumaran: the overarching theme of this again, just to.

105 00:24:47.200 00:24:53.269 Uttam Kumaran: you know, kind of like beat a dead horse is like my goal is to improve everybody’s efficiency, but ideally automate

106 00:24:53.290 00:24:55.790 Uttam Kumaran: as much of the business as possible.

107 00:24:56.314 00:25:05.429 Uttam Kumaran: And so leveraging these tools is going to be our moat for quite a while. Because a lot of the bigger companies aren’t going to be able to to do this.

108 00:25:05.450 00:25:12.890 Uttam Kumaran: So there’s things around like and kind of think, think, tell you a little bit about the long tail of like where this could go.

109 00:25:13.307 00:25:17.079 Uttam Kumaran: An example of where I probably see us. Heading is

110 00:25:17.230 00:25:33.710 Uttam Kumaran: when we create a Github issue. You know, the Github issues typically has like, here’s a problem. Here’s like the acceptance criteria. Here’s like the things that you need to look at ideally. AI should be able to reference, our code base reference, the issue, and then also propose a solution.

111 00:25:33.850 00:25:37.700 Uttam Kumaran: Right? And that’s basically automating like an engineer.

112 00:25:37.810 00:26:05.199 Uttam Kumaran: To be quite honest, I think. Is it going to be right all the time? No, do I care not, really, because I think over time it will get better and better at it. But those the sorts of directions that I want to help I want to head towards, and the risks that I want to take, which is, if we get really good at just defining the requirements of a project. Really well, AI should totally be able to propose, like what the solution is, or at least give you a proactive sense of where to search a lot of the problems that I see in current engineering is

113 00:26:05.738 00:26:15.830 Uttam Kumaran: there’s so much information brokering like to give you a sense of what happens. I get a text or an email from a client that’s like, Hey, this is a problem. I have to go

114 00:26:15.860 00:26:40.829 Uttam Kumaran: to Nico and be like, Hey, this is a problem. We didn’t have to wait until Monday to plan out the problem, and then we have to explain it to the team to be like, here’s the problem. And here’s how it’s laid out. Right? So Nico has to spend time building the goals acceptance. That person then reads that it’s like cool. I can do it. 2 days later they come back. They’re like, I kind of forgot, like what we were even talking about. They didn’t re ask, Hey, what’s this about?

115 00:26:40.830 00:27:00.730 Uttam Kumaran: And then they and then they start working on it. And then they have to be like, Okay, they spend hours thinking, looking at the code base right? Like, think about how ugly of a process that is. And and I’m even telling you that. And we move pretty quickly, like I’ve I’ve been part of a lot of engineering teams like we’re we move pretty good. But think about how much

116 00:27:00.740 00:27:07.570 Uttam Kumaran: there still is room to improve there right instead, if it could go from, I get an inbound request

117 00:27:07.660 00:27:15.429 Uttam Kumaran: to then me and Nico structure the requirements. And then basically, the engineer just works with the AI to then solve the problem.

118 00:27:15.500 00:27:18.190 Uttam Kumaran: There’s no back and forth, right? And so

119 00:27:18.350 00:27:25.369 Uttam Kumaran: I I want to explore tools like flow wise. And there’s a lot of other tools that I think we can, you know, just slack me about. But

120 00:27:25.830 00:27:38.189 Uttam Kumaran: basically to try to automate that process on the project manager side. There’s a lot of stuff that goes into not only writing those requirements right like. For example, if we get an inbound like, hey? This column is like broken.

121 00:27:38.380 00:28:06.049 Uttam Kumaran: Nico has to then find a table where it’s broken. Write that in another thing, and then make sure that that has all the context needed. Right? That’s something that AI could totally do the other thing. And and not only can it do. We’re already doing it like we use cloud projects to do some of this. So it’s it works. But like it’s just not integrated into like a workflow. Really, nicely. The second thing is summarizing all the work we’ve done for clients right on Friday. Me and Nico spend like an hour. We talk through everything we’ve done in the week.

122 00:28:06.080 00:28:10.509 Uttam Kumaran: and then we write an email. And then I review that email. And then it gets sent

123 00:28:10.590 00:28:19.039 Uttam Kumaran: like ideally, we get it. We get that email drafted. And then we just basically make updates that we need. And the emails are already saying the issues that we finished

124 00:28:19.250 00:28:21.140 Uttam Kumaran: right? And so

125 00:28:21.220 00:28:36.800 Uttam Kumaran: you can see a kind of the picture I’m painting, which is just like there’s a lot of gaps to fill here. The the problem that I’ve realized is actually the problem with AI is not really that the technology doesn’t exist. It’s just the medium isn’t really great meaning

126 00:28:37.110 00:28:41.820 Uttam Kumaran: like, I can use Chat Gbt, my personal life to just easily instead of Google.

127 00:28:41.860 00:28:42.505 Uttam Kumaran: But

128 00:28:43.580 00:28:51.029 Uttam Kumaran: like, these are specific things where you have to reference a code base reference, a Github issue. Put it in slack right? There’s like these integrations.

129 00:28:51.509 00:28:57.500 Uttam Kumaran: And so it’s going to be a mix of flow, wise and a mix of zapier that we make this happen, you know, for now.

130 00:29:00.290 00:29:05.159 Atharv Gudi: I think now that I’m start, I’m starting to get a hang of what’s the issue here?

131 00:29:05.470 00:29:07.209 Atharv Gudi: I feel like I still might

132 00:29:07.230 00:29:14.559 Atharv Gudi: want to, you know. Talk to Patrick, and see how what his side of things are, what Nico side of things are before I

133 00:29:14.750 00:29:18.550 Atharv Gudi: also. Still, I just tried cloud trace. I think I’ll

134 00:29:19.120 00:29:22.080 Atharv Gudi: sit and simmer with Claude for a bit and look at yeah.

135 00:29:22.080 00:29:25.689 Uttam Kumaran: The the thing I would urge you to do is just play around like.

136 00:29:25.690 00:29:26.490 Atharv Gudi: Yeah, we.

137 00:29:26.490 00:29:29.969 Uttam Kumaran: Applaud. We could pay for a chat we could pay for whatever.

138 00:29:30.290 00:29:42.239 Uttam Kumaran: I just want to see from your perspective, like, Oh, this is working. This isn’t working like the. And then the only thing I’ll also tell you is the problem that I’m finding when I talk about AI stuff

139 00:29:42.290 00:29:43.410 Uttam Kumaran: is.

140 00:29:43.700 00:29:47.779 Uttam Kumaran: and you may find this like, but people aren’t yet convinced

141 00:29:47.850 00:29:49.370 Uttam Kumaran: that it could do this.

142 00:29:49.560 00:29:52.099 Uttam Kumaran: And that’s not something that I want to negotiate

143 00:29:52.520 00:30:04.590 Uttam Kumaran: like. I’m not here to convince people that this is possible. It’s actually like. My belief is that this is possible. And so what the the thing you’ll find when you talk to folks is that they may not

144 00:30:04.950 00:30:10.349 Uttam Kumaran: you. It’s talk to folks about what their problems are. Don’t ask them about how to solve it.

145 00:30:10.360 00:30:20.450 Uttam Kumaran: and that’s the real role, as like a great product person is like, you don’t talk to them. You you just want to hear like you want to hear the disdain. You want to hear the real issues. If you ask them to solve it, then

146 00:30:20.560 00:30:22.199 Uttam Kumaran: then they should have solved. They would have solved it.

147 00:30:22.200 00:30:22.540 Atharv Gudi: Yeah.

148 00:30:22.540 00:30:27.539 Uttam Kumaran: Right? So that’s the thing I realized in AI is, I spend less time asking people like.

149 00:30:27.600 00:30:30.010 Uttam Kumaran: Oh, how can we do this? Instead? It’s like.

150 00:30:30.240 00:30:41.119 Uttam Kumaran: Oh, there’s a clear like information brokering that takes hours and hours right? And that’s hundreds of dollars that we’re spending. And every time we do that. Think about it. We’re running multiple projects

151 00:30:41.170 00:30:50.020 Uttam Kumaran: like it’s just like so much wasted time. That instead, it’s like, I think the goal is my goal, for you would be just to.

152 00:30:50.070 00:30:53.679 Uttam Kumaran: I’m expecting you to just poke around and like poke holes and try these things

153 00:30:53.720 00:31:04.829 Uttam Kumaran: like you can connect it to slack, connect it to Zapier. Whatever you want to do and just play around. I I do think it’s worth meeting with both of them. I think Nico is definitely more more bought in

154 00:31:05.289 00:31:17.160 Uttam Kumaran: but yeah, that’s where I’m like a little bit like hard headed on, like, I know that this is working. And so it’s like we just need to take cracks at it, you know, and like, take shots on goal. So.

155 00:31:17.670 00:31:23.319 Atharv Gudi: I think it’s less of the people don’t believe it works. It just seems a bridge too far.

156 00:31:23.630 00:31:31.080 Atharv Gudi: and breaking it up into bits and saying, this could go here, this could go here and this whole. This fits. The puzzle is more of the

157 00:31:31.520 00:31:35.223 Atharv Gudi: issue in a lot of cases. I but that might just.

158 00:31:35.560 00:31:49.479 Uttam Kumaran: Integrations. Like, for example, the things I’m describing, those will be companies that build software that do that. But it’s gonna take another 2 years. Yeah, like, I can’t wait 2 years for someone else to build like

159 00:31:49.910 00:32:04.850 Uttam Kumaran: the Github integration. And this like this is what we’re gonna do. But also, the nice thing is they’re gonna build software that works for a lot of people. This just has to work for us, meaning it doesn’t really need our front end. We just have a couple of things. We’re stringing along.

160 00:32:05.251 00:32:10.398 Uttam Kumaran: And the risk is the risk is, it’s all upside, actually, like there’s no risk. So

161 00:32:10.720 00:32:11.590 Atharv Gudi: Yeah.

162 00:32:11.590 00:32:12.240 Uttam Kumaran: Yeah.

163 00:32:13.870 00:32:23.629 Atharv Gudi: I think that my perspective is also simply might just be because I’m from a Cs heavy school. So the people I work with are confident in AI

164 00:32:24.350 00:32:30.059 Atharv Gudi: as opposed to maybe everyone in the business world who I don’t interact with as much as.

165 00:32:30.060 00:32:47.159 Uttam Kumaran: But you know it’s it’s just like anything right. And I think as we get older, you get more stubborn. But people who have been working for a long time. They they get used to the way things are going. And again, like, I still think our company is filled with people that are very flexible. We use a lot of really, really new technologies.

166 00:32:47.350 00:32:48.190 Uttam Kumaran: But

167 00:32:48.500 00:32:56.860 Uttam Kumaran: at some point like these things like, I’ve I just spend so much my time reading and playing around with the tools that I’m I’m convinced.

168 00:32:56.940 00:32:59.140 Uttam Kumaran: I’m sure if they went through that same path

169 00:32:59.230 00:33:15.809 Uttam Kumaran: they would see that. But again it’s up to some of us who can see that it’s working to really make that a reality. And for some of the folks in the team chatting with Claude, and being like pasting in your code, and then being like this is not enough.

170 00:33:16.240 00:33:42.080 Uttam Kumaran: Right? It’s like, how do we get over the hump where it’s like this is like, Oh, this is like magic, like, for example, like Nico, could put an issue up, and then it comments in like, probably, Look, look in here. Let me know if you have any other questions right? Like we’re basically inventing like another engineer. Or it’d be like another version of me, right? Because a lot of questions I get because I just have know the code base really well is like, where’s this thing? Where’s this thing? Where’s this thing

171 00:33:42.130 00:33:46.850 Uttam Kumaran: like, think about right. What’s the value of that? And so that’s kind of like

172 00:33:47.750 00:33:48.810 Uttam Kumaran: instead of

173 00:33:49.580 00:34:08.760 Uttam Kumaran: I don’t want to convince people and be like, tell people. They’re wrong. Instead, it’s just like show. We need to go one layer deeper like, if Claude and projects, isn’t it? It’s 1 layer deeper. And again. The nice thing is like, we’re gonna this is the things that’s really gonna help scale our company. So this is like where I want to spend a little bit more time.

174 00:34:09.134 00:34:13.820 Uttam Kumaran: Especially because I think when you go back to school, you’re gonna see a ton of ton of stuff about

175 00:34:13.980 00:34:15.670 Uttam Kumaran: this this year for sure.

176 00:34:16.110 00:34:20.190 Atharv Gudi: Yeah, I think this is one thing that I will probably just

177 00:34:20.270 00:34:22.190 Atharv Gudi: grab my head around

178 00:34:22.230 00:34:23.500 Atharv Gudi: after

179 00:34:25.300 00:34:27.209 Atharv Gudi: after I finish the sequel, Fluff.

180 00:34:27.719 00:34:28.329 Uttam Kumaran: Yeah.

181 00:34:28.719 00:34:30.749 Atharv Gudi: Which is also not going to be, I think.

182 00:34:30.789 00:34:33.989 Atharv Gudi: what I’m hoping for with this week’s work

183 00:34:34.059 00:34:38.119 Atharv Gudi: before I move on to you. Know you exploring cloud is to

184 00:34:38.459 00:34:40.249 Atharv Gudi: set up a basic.

185 00:34:41.489 00:34:48.789 Atharv Gudi: the code. The config files are never gonna be complete. They’re gonna change with project. They’re gonna change with everything. But I

186 00:34:49.659 00:34:52.689 Atharv Gudi: like, so far the config files have been serving me.

187 00:34:52.949 00:34:56.459 Atharv Gudi: Yeah, yeah. But tomorrow, after tomorrow, I want them to serve

188 00:34:56.599 00:34:58.059 Atharv Gudi: Brain Forge.

189 00:34:58.060 00:35:00.190 Uttam Kumaran: Yeah, no, that’s perfect.

190 00:35:00.270 00:35:03.851 Uttam Kumaran: I I would suggest to use Claude in that process.

191 00:35:04.210 00:35:04.750 Atharv Gudi: Right.

192 00:35:04.750 00:35:05.480 Uttam Kumaran: You know so.

193 00:35:05.775 00:35:06.070 Atharv Gudi: Sorry.

194 00:35:06.070 00:35:13.569 Uttam Kumaran: Yeah, talk. Talk to Ryan. He’s gonna be opinionated about it. And the nice thing is you can send a note out like, once you guys get to a good place.

195 00:35:13.620 00:35:34.610 Uttam Kumaran: send a note out saying, Hey, can I get some more eyes to review this? Here’s what we’re thinking, right? The the problem is like, everything is about the medium. So like, if you’re to send a Pr request, some people review. But if if you send like a note in slack, which is like, Hey guys, wrote this notion, Doc, about how we’re doing sequel. Fluff, if anyone has opinions about like

196 00:35:34.760 00:35:40.110 Uttam Kumaran: like the way columns look the way sequel is formatted the way Yaml is formatted. Now’s the time.

197 00:35:40.290 00:35:42.420 Uttam Kumaran: If you send that, you’re going to get a good opinions.

198 00:35:43.280 00:35:57.989 Uttam Kumaran: So yeah, let’s do that. And then I’m going to be I’m going to be a little bit out Thursday, Friday. But maybe let’s just I’m just gonna put some time, so maybe we can slack like Thursday morning, and then

199 00:35:58.400 00:36:01.340 Uttam Kumaran: let’s catch up again about AI stuff next week.

200 00:36:01.340 00:36:02.840 Atharv Gudi: Yeah, for sure. Yeah.

201 00:36:03.360 00:36:04.100 Uttam Kumaran: Okay.

202 00:36:04.460 00:36:05.290 Atharv Gudi: Alright!

203 00:36:06.330 00:36:09.639 Uttam Kumaran: Okay, cool. Alright. Yeah. Let me know how it goes talking to Ryan.

204 00:36:09.640 00:36:11.730 Atharv Gudi: Yes, definitely. Alright.

205 00:36:11.810 00:36:13.079 Atharv Gudi: Okay. Thank you.