Meeting Title: AI Service Standup Date: 2026-02-02 Meeting participants: Samuel Roberts, Uttam Kumaran, Mustafa Raja, Casie Aviles, Clarence Stone, Pranav Narahari, Gabriel Lam, Amber Lin
WEBVTT
1 00:00:18.750 ⇒ 00:00:19.800 Uttam Kumaran: Hello.
2 00:00:20.850 ⇒ 00:00:21.810 Samuel Roberts: Ayy.
3 00:00:23.270 ⇒ 00:00:23.910 Mustafa Raja: Hey.
4 00:00:24.890 ⇒ 00:00:26.110 Samuel Roberts: How’s your weekends?
5 00:00:27.150 ⇒ 00:00:28.629 Mustafa Raja: Good. How are you? Good, dude.
6 00:00:28.630 ⇒ 00:00:31.079 Uttam Kumaran: I’m like, I crush stuff on…
7 00:00:31.350 ⇒ 00:00:33.190 Samuel Roberts: I was looking at that, yeah, man.
8 00:00:33.190 ⇒ 00:00:38.699 Uttam Kumaran: Yeah, dude, I’m like… I’m the best, like, I’m… I’m going crazy. It’s so fun.
9 00:00:41.720 ⇒ 00:00:45.839 Uttam Kumaran: I, I’m learning a lot in the process,
10 00:00:46.090 ⇒ 00:00:50.149 Uttam Kumaran: And we just gotta… we’ll think about some… we’ll… we can talk during my update.
11 00:00:50.150 ⇒ 00:00:52.089 Samuel Roberts: Yeah, I was just looking at your update there.
12 00:00:52.830 ⇒ 00:00:53.530 Uttam Kumaran: Yeah.
13 00:00:54.150 ⇒ 00:00:56.790 Samuel Roberts: Yeah, that’s sick.
14 00:00:57.150 ⇒ 00:00:59.779 Samuel Roberts: Yeah, I was… I played around with,
15 00:01:00.580 ⇒ 00:01:02.600 Samuel Roberts: open claw this weekend a little bit.
16 00:01:03.610 ⇒ 00:01:06.909 Samuel Roberts: Nice. You saw Claudebot got renamed, like, twice.
17 00:01:08.140 ⇒ 00:01:13.920 Uttam Kumaran: Yeah, I think I’m gonna… I think I’m gonna skip it. Yeah.
18 00:01:14.030 ⇒ 00:01:15.049 Uttam Kumaran: Yeah. I would recommend.
19 00:01:15.050 ⇒ 00:01:27.709 Samuel Roberts: that at this point, because it took me a while to get it set up and working in a way that I felt comfortable up and working, in a way that I felt comfortable that it wasn’t just on my machine. Server here running something I can virtualize machines with, and so I tried to set it up in there, and it was not…
20 00:01:28.290 ⇒ 00:01:29.520 Samuel Roberts: Working at first.
21 00:01:30.440 ⇒ 00:01:45.190 Uttam Kumaran: Yeah, for me, I’m not doing a lot of, like… I’m not doing a lot of personal AI, like, all of it is so focused on the company, so there’s not, like, a company benefit. Like, in my personal life, I’m just… I’m super, like, normie, like, I’m just ChatGPT.
22 00:01:45.570 ⇒ 00:01:50.180 Samuel Roberts: I mean, you could… you could… you could set it up for business stuff if you wanted, but, you know…
23 00:01:50.540 ⇒ 00:01:53.890 Uttam Kumaran: Oh, I could, but, like, dude, we have such a backlog that, like.
24 00:01:53.890 ⇒ 00:01:58.039 Samuel Roberts: That’s what I’m saying, like, yeah, I don’t know if this is the right way to go about it, but it was interesting, at least, to be…
25 00:01:58.040 ⇒ 00:02:03.659 Uttam Kumaran: Yeah, like, I was playing with Conductor, actually, it’s probably the first net new thing that I was using.
26 00:02:03.720 ⇒ 00:02:07.840 Samuel Roberts: Because I’m having the challenge that I want to work on multiple…
27 00:02:08.270 ⇒ 00:02:09.819 Uttam Kumaran: Things at a time.
28 00:02:10.630 ⇒ 00:02:16.819 Uttam Kumaran: And you can’t do that in cursor, because they don’t support multiple
29 00:02:17.070 ⇒ 00:02:20.559 Uttam Kumaran: branches. They don’t support multiple repos to do work trees.
30 00:02:21.380 ⇒ 00:02:22.810 Samuel Roberts: Yeah, it’s so annoying.
31 00:02:23.060 ⇒ 00:02:32.860 Uttam Kumaran: Yeah, there’s this thing called Conductor, where you can do that, and so I played around with it a bit, but I’m gonna do more… spend more time this week on it.
32 00:02:33.070 ⇒ 00:02:33.790 Samuel Roberts: Sweet.
33 00:02:36.160 ⇒ 00:02:38.500 Samuel Roberts: Hello, everyone. Welcome, welcome.
34 00:02:39.530 ⇒ 00:02:40.740 Pranav Narahari: Hey, good morning.
35 00:02:42.760 ⇒ 00:02:44.119 Samuel Roberts: Everyone have a good weekend?
36 00:02:46.050 ⇒ 00:02:46.770 Uttam Kumaran: Yes.
37 00:02:48.530 ⇒ 00:02:51.009 Samuel Roberts: We saw… I saw your weekend, at least, so…
38 00:02:51.010 ⇒ 00:02:59.259 Uttam Kumaran: It’s just Saturday, I didn’t do anything yesterday. I went… I went roller skating after my, like, 7-hour vibe coding marathon.
39 00:02:59.260 ⇒ 00:03:00.050 Samuel Roberts: Amazing.
40 00:03:03.120 ⇒ 00:03:03.690 Samuel Roberts: weeds.
41 00:03:04.170 ⇒ 00:03:06.579 Samuel Roberts: Anybody else do anything exciting?
42 00:03:06.800 ⇒ 00:03:08.400 Samuel Roberts: Or fun, or anything?
43 00:03:12.360 ⇒ 00:03:13.190 Samuel Roberts: Cool.
44 00:03:13.450 ⇒ 00:03:13.990 Samuel Roberts: True.
45 00:03:14.770 ⇒ 00:03:15.500 Samuel Roberts: Exciting, exciting.
46 00:03:15.500 ⇒ 00:03:16.790 Uttam Kumaran: morning today. Yes.
47 00:03:16.790 ⇒ 00:03:17.270 Samuel Roberts: Fixed up.
48 00:03:17.270 ⇒ 00:03:17.850 Uttam Kumaran: everyone.
49 00:03:18.230 ⇒ 00:03:24.749 Samuel Roberts: Alright, I guess let’s, let’s jump in, start with Lilo, in case Amber joins later for ABC.
50 00:03:24.990 ⇒ 00:03:31.000 Samuel Roberts: So I was catching up this morning on the error they were seeing.
51 00:03:32.640 ⇒ 00:03:36.899 Samuel Roberts: I think… Navy and I were just discussing it, so I’ll talk to you.
52 00:03:38.370 ⇒ 00:03:45.690 Samuel Roberts: everyone else, I guess, who might be interested, but it’s basically, they have too many skills that are getting passed through. It’s not anything about the API keys.
53 00:03:47.860 ⇒ 00:03:55.190 Samuel Roberts: So, I guess we’re pulling all the skills from Anthropic, but it can only handle sending 8 of them at a time, and they have 9?
54 00:03:57.050 ⇒ 00:03:58.510 Samuel Roberts: So,
55 00:03:58.960 ⇒ 00:04:04.499 Samuel Roberts: Two things about that. One, we have to limit that, and I gotta… we gotta figure out with them how best to handle that, because I don’t know if they’re gonna want…
56 00:04:04.720 ⇒ 00:04:07.350 Samuel Roberts: Just, like, one skill for chat or something.
57 00:04:07.530 ⇒ 00:04:22.890 Samuel Roberts: Or multiple, or how that’s gonna get organized. And then two, I gotta make sure… we gotta make sure that the errors that are coming back in the API get passed all the way down, because I think sometimes certain ones are, and other ones just get caught, and if it doesn’t get a good response, it just prints that very unhelpful
58 00:04:23.070 ⇒ 00:04:24.639 Samuel Roberts: Message that they saw.
59 00:04:24.840 ⇒ 00:04:27.539 Samuel Roberts: That kind of misleads where we’re going, so…
60 00:04:28.800 ⇒ 00:04:33.019 Samuel Roberts: try to work on those today. But besides that, how are we…
61 00:04:33.490 ⇒ 00:04:35.970 Samuel Roberts: Are we doing CSO-wise and EP-wise?
62 00:04:39.340 ⇒ 00:04:46.389 Pranav Narahari: Yeah, I think, CSOIs were doing good, as… I think, just the main concern for them is just that
63 00:04:46.650 ⇒ 00:04:50.280 Pranav Narahari: They can, release production to everybody on… in, like.
64 00:04:50.390 ⇒ 00:04:57.680 Pranav Narahari: In Lilo. So, I think we’re… we’re tracking well on that. We’ll just get this patched, like, this morning.
65 00:04:58.280 ⇒ 00:05:03.490 Pranav Narahari: Did… so… did you want to, like, tackle that, like, Claude Skills thing?
66 00:05:03.740 ⇒ 00:05:07.060 Pranav Narahari: update, and then I can get this, UI bug.
67 00:05:08.790 ⇒ 00:05:20.290 Samuel Roberts: Yeah, yeah, I’ll send a message, because I just want to clarify with them how they’re going to use the skills, especially if they’re gonna add more, before I, like, fully just patch, like, just filtering out down to 8, in case there’s some missing.
68 00:05:20.870 ⇒ 00:05:21.430 Pranav Narahari: Yep.
69 00:05:21.890 ⇒ 00:05:27.419 Samuel Roberts: But I’ll send them a message and I’ll get that resolved, and then I’ll… I think I also have Cursor working on some error stuff right now, so…
70 00:05:27.780 ⇒ 00:05:29.429 Samuel Roberts: That should hopefully help out.
71 00:05:29.750 ⇒ 00:05:35.920 Samuel Roberts: But other than the UI stuff, yeah, I think just changing it so they don’t see anything for that. I didn’t see your last message, I saw it came through as I was getting.
72 00:05:35.920 ⇒ 00:05:42.150 Pranav Narahari: Yeah, so I was thinking maybe it just makes sense to, like, patch it so it looks how it did before.
73 00:05:42.150 ⇒ 00:05:46.280 Uttam Kumaran: However, it just, like, shows the unencrypted, like, masked key.
74 00:05:46.280 ⇒ 00:05:55.499 Pranav Narahari: I think that way, in the future, they don’t have to come through us to, like, you know, cross-check, are they using the right API key and the right account? .
75 00:05:55.500 ⇒ 00:05:56.329 Samuel Roberts: That’s fine, yeah.
76 00:05:56.650 ⇒ 00:05:57.400 Pranav Narahari: Okay.
77 00:05:59.710 ⇒ 00:06:04.320 Samuel Roberts: Yeah, especially because it’s, like, an internal tool, so I’m less worried about anything leaking too, too much, but I guess…
78 00:06:04.740 ⇒ 00:06:07.540 Samuel Roberts: I wasn’t even sure if it was worth doing that at first, but… Okay, cool.
79 00:06:10.000 ⇒ 00:06:14.250 Samuel Roberts: I know we have a call later as well, right? Just like a team sync?
80 00:06:15.600 ⇒ 00:06:17.239 Pranav Narahari: Oh, yeah, I just…
81 00:06:17.660 ⇒ 00:06:25.110 Casie Aviles: I set that up, could be just, you know, to get us aligned every week, every start of the week.
82 00:06:25.370 ⇒ 00:06:29.809 Samuel Roberts: No, that’s perfect. I was gonna say, we can… we can push things like the Gantt and linear till then.
83 00:06:31.560 ⇒ 00:06:32.609 Samuel Roberts: And then go to…
84 00:06:32.610 ⇒ 00:06:34.150 Uttam Kumaran: contract. When is that?
85 00:06:35.080 ⇒ 00:06:38.990 Samuel Roberts: That is on, 12 Eastern, it looks like.
86 00:06:39.940 ⇒ 00:06:43.510 Uttam Kumaran: Okay… Okay, yeah, I wanted to meet…
87 00:06:43.710 ⇒ 00:06:47.360 Uttam Kumaran: Y’all to talk about just, like, how to…
88 00:06:47.480 ⇒ 00:06:54.180 Uttam Kumaran: use AI for development more, but maybe I can… we can do that midweek, so I don’t distract you guys right now.
89 00:06:54.390 ⇒ 00:07:05.010 Uttam Kumaran: Okay. But I’m… I’m trying… I’m… I’m starting to do a lot, on the platform, and I feel like I learned a lot of ways to probably give you guys another, like.
90 00:07:05.220 ⇒ 00:07:08.319 Uttam Kumaran: 20-40% efficiency, so I wanted to share that.
91 00:07:08.700 ⇒ 00:07:17.950 Samuel Roberts: Sweet, yeah, that would definitely… yeah, that would sound… that sounds good, because I think, like, I’m definitely, like, leaning into Cursor so much, but I know that I forget about codecs sometimes and things, like…
92 00:07:18.110 ⇒ 00:07:19.740 Samuel Roberts: I’m sure there’s other ways to do things.
93 00:07:21.720 ⇒ 00:07:22.500 Samuel Roberts: Cool.
94 00:07:23.540 ⇒ 00:07:26.060 Samuel Roberts: Alright, anything else, Lilo-wise?
95 00:07:26.510 ⇒ 00:07:30.509 Casie Aviles: Yeah, I can just… In terms of, like, timeline, so…
96 00:07:30.770 ⇒ 00:07:39.810 Casie Aviles: I know that for this week, we should… our focus should… would be the data warehouse, so we want to fix… we want to finalize that, and…
97 00:07:40.980 ⇒ 00:07:43.820 Casie Aviles: Have that wire, do the front-end wiring.
98 00:07:44.260 ⇒ 00:07:46.360 Casie Aviles: So I know that we are,
99 00:07:46.750 ⇒ 00:07:56.500 Casie Aviles: we’ve made some, like, we showed them, like, a quick demo last week, and they pointed out there were some discrepancies, so I think that’s something we would polish for this week.
100 00:07:56.960 ⇒ 00:08:00.170 Casie Aviles: To get the numbers as close,
101 00:08:00.800 ⇒ 00:08:05.969 Casie Aviles: And then… I believe there are two, like, ad hoc stuff as well.
102 00:08:06.130 ⇒ 00:08:08.049 Casie Aviles: Which, I believe those are the…
103 00:08:08.810 ⇒ 00:08:11.259 Casie Aviles: The things that they push, so…
104 00:08:11.620 ⇒ 00:08:11.950 Samuel Roberts: Yeah.
105 00:08:11.950 ⇒ 00:08:12.480 Casie Aviles: Right, this week.
106 00:08:12.480 ⇒ 00:08:19.030 Samuel Roberts: another thing to discuss at some point, maybe, which I’m looping you into, like, as they’re adding things and want us to, like.
107 00:08:19.740 ⇒ 00:08:25.510 Samuel Roberts: add the backend features and stuff, like, how does that all fit into scope for us? Like…
108 00:08:25.900 ⇒ 00:08:36.510 Samuel Roberts: Yeah, perfect. I throw a couple things together, and I, like, we can definitely take a look at them, but, like, technically that’s not, like, listed, you know, like, doing that is, but, like, depending on how big the features become, I just don’t want to, like…
109 00:08:37.460 ⇒ 00:08:40.890 Samuel Roberts: You know, one, get off track, and then two, if there’s, you know.
110 00:08:41.809 ⇒ 00:08:51.869 Uttam Kumaran: No, this is a great question. Maybe, if you guys, during your meeting today, want to just list what all those additional features are, and then Pranav, during, like, CSO leads, let’s chat about that.
111 00:08:52.080 ⇒ 00:08:52.490 Pranav Narahari: Okay.
112 00:08:52.490 ⇒ 00:08:56.349 Uttam Kumaran: and then I’ll… I’ll show you, like, how do you weave that in.
113 00:08:56.700 ⇒ 00:08:58.029 Pranav Narahari: Cool, yeah, that’d be great.
114 00:08:58.470 ⇒ 00:09:02.719 Samuel Roberts: Yeah, I think there’s just two right now, so it shouldn’t be too crazy, but I know they’re gonna be adding more, so…
115 00:09:02.950 ⇒ 00:09:14.500 Pranav Narahari: Yeah, I did, like, when I was rethinking what, like, I said in the meeting, which was like, you know, I felt like we were kind of ahead of schedule, so I was like, okay, these two ad hoc things we could probably.
116 00:09:14.500 ⇒ 00:09:15.220 Samuel Roberts: Yeah.
117 00:09:15.220 ⇒ 00:09:28.639 Pranav Narahari: just do, but I think we should have a formalized, like, process for ad hoc things, you know, because in the future, you know, this is definitely gonna happen again, like, we need to set, like, kind of like a precedent, like, when…
118 00:09:28.850 ⇒ 00:09:41.420 Pranav Narahari: we get ad hoc requests that are out of scope, essentially, we need to be like, okay, well, when that happens, we do X, Y, and Z. Just to make it clear with, like, Bobby and Zach, like, okay, this is our process.
119 00:09:41.710 ⇒ 00:09:42.230 Samuel Roberts: Yeah.
120 00:09:42.610 ⇒ 00:09:44.970 Samuel Roberts: Okay, yeah, we can discuss that later, and then we can…
121 00:09:45.260 ⇒ 00:09:49.709 Pranav Narahari: Yeah, in the CSO meeting, I think that would be… that’d be a great thing to discuss.
122 00:09:50.140 ⇒ 00:09:51.440 Samuel Roberts: Perfect. Okay, cool.
123 00:09:52.160 ⇒ 00:09:57.449 Samuel Roberts: Anything else? We’ll sync later, and then we’ll be…
124 00:09:58.590 ⇒ 00:10:01.839 Samuel Roberts: Good for that. Alright, cool. Let’s jump ABC…
125 00:10:04.910 ⇒ 00:10:09.170 Samuel Roberts: I’m actually gonna throw it to you, Amber, just because I think there were things happening Friday I didn’t catch up on.
126 00:10:10.030 ⇒ 00:10:15.790 Amber Lin: Yeah, so, main thing is one on zip codes and one on the central dock.
127 00:10:15.890 ⇒ 00:10:19.360 Amber Lin: We have a meeting later, and…
128 00:10:19.730 ⇒ 00:10:23.219 Amber Lin: I think after the stand-ups, and we’ll discuss how
129 00:10:23.650 ⇒ 00:10:34.559 Amber Lin: we should weave that into this sprint, and maybe the next sprint, to make sure we get that done. And then I need us to have another meeting of…
130 00:10:34.690 ⇒ 00:10:42.630 Amber Lin: How we can help them improve their central docs, because that’s going to be, extra hours.
131 00:10:42.630 ⇒ 00:10:55.389 Amber Lin: So, I want to make sure if, when that meeting should happen. I currently have it for today, it’s called ABC Sync, but I don’t know, Utam, if that’s a good time for you, or if you want to move it to a different.
132 00:10:55.390 ⇒ 00:11:02.660 Uttam Kumaran: No, let’s do… let’s do that time. Yeah, I just… if you can throw stuff in a Notion Doc, I can prep, and then, yeah, I think that’ll… that should be fine.
133 00:11:02.880 ⇒ 00:11:04.730 Amber Lin: Cool, okay, I’ll keep it then.
134 00:11:06.190 ⇒ 00:11:09.970 Amber Lin: Yeah, that’s all for me. I think that will be the planning.
135 00:11:10.190 ⇒ 00:11:13.790 Amber Lin: That will do, so I won’t talk more in this stand-up.
136 00:11:14.240 ⇒ 00:11:16.949 Samuel Roberts: Yeah, okay, that sounds good. Yeah, Monday, we’re all kidding.
137 00:11:17.200 ⇒ 00:11:19.980 Samuel Roberts: Getting ready for the week, so all these meetings sound good.
138 00:11:20.090 ⇒ 00:11:21.410 Samuel Roberts: Alright,
139 00:11:22.540 ⇒ 00:11:28.149 Samuel Roberts: Yeah, I don’t have any other early updates there. I think that kind of covered most of what we’re gonna handle EBC-wise.
140 00:11:29.030 ⇒ 00:11:30.520 Samuel Roberts: So I guess we’ll jump to platform.
141 00:11:32.670 ⇒ 00:11:34.140 Mustafa Raja: Yeah, I agree.
142 00:11:34.380 ⇒ 00:11:50.940 Mustafa Raja: I had two things. So, I worked on two spikes for ABC last week, one for Cloud SQL, so I believe we need, some permissions, and, I’m not aware of if they’re aware of the pricing, so, let me know if…
143 00:11:50.960 ⇒ 00:11:56.579 Mustafa Raja: If we should ping them, that we need these rules, and this is the pricing for Cloud SQL.
144 00:11:56.660 ⇒ 00:12:05.039 Mustafa Raja: Since there’s… there are multiple plans, and we might want to ask them, you know, which one should… should we choose, you know?
145 00:12:05.440 ⇒ 00:12:14.209 Mustafa Raja: So maybe we could sync on that. I could come up with a draft and, you know, ping in the channel. If it looks good, I can send that.
146 00:12:16.070 ⇒ 00:12:28.220 Mustafa Raja: And then there’s one other, ticket that I worked on, that’s for how do we, grab the workflow response to calculate latency, and then…
147 00:12:28.560 ⇒ 00:12:32.089 Mustafa Raja: send it to, Slack or whatever, right?
148 00:12:32.090 ⇒ 00:12:51.940 Mustafa Raja: for that, I believe in the master instance, we have not implemented, the entry point yet, right? That we had in the, in our current, cloud run, right? So, that we can do with the custom API route the master offers, and then
149 00:12:51.940 ⇒ 00:12:55.160 Mustafa Raja: transfer whatever we have in Cloud Run right now.
150 00:12:55.340 ⇒ 00:13:01.120 Mustafa Raja: And that’s how it’s going to work. I have the tickets linked in your DM? Yeah.
151 00:13:01.120 ⇒ 00:13:03.739 Samuel Roberts: I saw them, I just hadn’t gotten a chance to go through them yet. Thank you for…
152 00:13:03.740 ⇒ 00:13:05.890 Mustafa Raja: Okay, yeah, that’s all from me.
153 00:13:06.270 ⇒ 00:13:12.120 Samuel Roberts: Okay, yeah, I’ll take a look at those, I’ll get back to you. We may even want to sync on that during the ABC sync then, just in case.
154 00:13:12.310 ⇒ 00:13:15.900 Samuel Roberts: So if there’s some pricing and stuff then, then we have to make…
155 00:13:16.330 ⇒ 00:13:19.189 Samuel Roberts: Make sure it gets surfaced to them. Cool.
156 00:13:19.830 ⇒ 00:13:20.950 Samuel Roberts: Thank you, Mustafa.
157 00:13:21.150 ⇒ 00:13:22.160 Samuel Roberts: Anything else?
158 00:13:24.280 ⇒ 00:13:25.459 Samuel Roberts: Let’s do platform.
159 00:13:25.690 ⇒ 00:13:26.530 Samuel Roberts: Eternal.
160 00:13:26.530 ⇒ 00:13:28.500 Uttam Kumaran: Okay, yeah, so I…
161 00:13:28.500 ⇒ 00:13:29.820 Gabriel Lam: Yeah.
162 00:13:30.800 ⇒ 00:13:39.110 Uttam Kumaran: Yeah, I guess I’m, Gabe, I’m gonna start to just lead platform updates, because I’m… I’m gonna be, like, shipping basically, like, a ton of stuff.
163 00:13:39.330 ⇒ 00:13:44.210 Uttam Kumaran: So I’ll just lead, and then I can kick past you for anything on…
164 00:13:44.500 ⇒ 00:14:00.269 Uttam Kumaran: If there’s any, like, sales or marketing, but I think overall, me and you will get a bit tighter. So kind of overall, I think this was, like, the first last two weeks, I was able to actually just, like, start to kind of fulfill what our original vision was, which is, like.
165 00:14:00.410 ⇒ 00:14:17.939 Uttam Kumaran: someone outside the AI team was able to sort of start to improve the platform in several different ways. And I think it’s working out really, really well. So I was able to get out, like, a ton of stuff. So several new MCPs.
166 00:14:17.940 ⇒ 00:14:33.030 Uttam Kumaran: I also was adding a ton of stuff to the vault, several, like, nice-to-have features for the platform, Slack Assistant. So I feel really confident that, you know, even with a few hours of my time, I’m able to get out a lot of progress.
167 00:14:33.040 ⇒ 00:14:34.450 Uttam Kumaran: I think,
168 00:14:34.660 ⇒ 00:14:41.730 Uttam Kumaran: I do have some open questions, but I think overall, like, from my standpoint, like, I feel really great. I think…
169 00:14:41.830 ⇒ 00:14:44.219 Uttam Kumaran: I will probably continue to be…
170 00:14:44.600 ⇒ 00:14:51.180 Uttam Kumaran: probably some of the more, like, high… highest leverage person outside the AI team, shipping, like, end-to-end stuff.
171 00:14:51.370 ⇒ 00:15:06.220 Uttam Kumaran: But I think slowly, I’ll be working with Gabe on, like, how do we grow that, right? I think some people are starting to push stuff into Vault and into playbooks, but I’m gonna urge them to get into actually developing, like, features and apps. Yeah. So…
172 00:15:06.560 ⇒ 00:15:13.269 Uttam Kumaran: That being said, I think, like, my first question is just, like, I’m gonna be… potentially shipping…
173 00:15:13.390 ⇒ 00:15:18.030 Uttam Kumaran: like, 3 to 10 PRs Like, every day.
174 00:15:18.250 ⇒ 00:15:23.650 Uttam Kumaran: So… how do… Like, how does reviews keep up?
175 00:15:23.760 ⇒ 00:15:26.779 Uttam Kumaran: And how does… how does this team stay in the loop?
176 00:15:26.880 ⇒ 00:15:29.020 Uttam Kumaran: I think, like, my… my ask…
177 00:15:29.180 ⇒ 00:15:41.100 Uttam Kumaran: for you, Sam, is to think about, one, like, what updates you require from platform team in this meeting, or outside, and I’m happy to adhere to whatever you need. And then second is, like.
178 00:15:41.630 ⇒ 00:15:43.180 Uttam Kumaran: I… how do we…
179 00:15:43.790 ⇒ 00:15:47.790 Uttam Kumaran: How do we, like… yeah, how do we get that much code out, you know, every day?
180 00:15:47.930 ⇒ 00:15:54.390 Uttam Kumaran: So maybe I’ll… I’ll pass it to you, and then I have… I have some thoughts, but that… I would like to have a discussion on that briefly.
181 00:15:55.510 ⇒ 00:15:57.249 Samuel Roberts: Yeah, I mean, I think…
182 00:15:58.620 ⇒ 00:16:08.250 Samuel Roberts: you know, doing these reviews definitely takes a little bit of time. I’m wondering, like, the cursor bot is pretty good at catching some things, but I wonder if there’s… I mean, I know there are other tools for doing, like, bigger…
183 00:16:08.520 ⇒ 00:16:10.400 Samuel Roberts: reviews? I don’t know.
184 00:16:12.120 ⇒ 00:16:23.610 Samuel Roberts: what’s the current, like, cutting edge of that, because I haven’t been following that as much, but there might be worth something that’s just, like… because sometimes I’ll just take it and pass it into cursor and do the first passive review that way, but, like, that still takes time for me.
185 00:16:23.610 ⇒ 00:16:26.649 Uttam Kumaran: Well, this is where, like, you tell me what you would like, I’ll build it for you.
186 00:16:26.650 ⇒ 00:16:32.069 Samuel Roberts: That’s what I’m saying, I gotta do a little more research, because I know, like, there’s, like, Code Rabbit, and we talked about graphite and things, I think that…
187 00:16:32.070 ⇒ 00:16:51.589 Uttam Kumaran: So, yeah, in terms of an out-of-the-box tool, nothing’s gonna be better than, like, BugBot or Codex right now. They’re all… they’re all just like this. But, if instead you’re like, hey, I need the review output to be in a certain way, like, I can write a job that takes the PR and goes and runs a review.
188 00:16:51.800 ⇒ 00:16:57.439 Uttam Kumaran: Otherwise, every PR going to the platform is going through both BugBot and Kodaks.
189 00:16:57.600 ⇒ 00:16:58.190 Samuel Roberts: Yeah.
190 00:16:58.490 ⇒ 00:16:59.950 Uttam Kumaran: You know.
191 00:17:01.280 ⇒ 00:17:10.669 Samuel Roberts: Okay, yeah, I’ll think about, like, the form that has to take then. I think the one I just looked at that you had looked pretty good. Like, the format, at least. I mean, the code itself looked good, too.
192 00:17:11.180 ⇒ 00:17:13.239 Samuel Roberts: There are a couple of small things,
193 00:17:13.420 ⇒ 00:17:18.699 Samuel Roberts: But I think you… you were… you’re getting on the right track there. I think… I mean, we’ve… I was mentioning it earlier, you mentioned it earlier, the Loom…
194 00:17:19.619 ⇒ 00:17:20.239 Uttam Kumaran: Yes. Documentation.
195 00:17:20.240 ⇒ 00:17:38.459 Samuel Roberts: stuff. I think you’re on the right track there, too. I probably just need to really sit down and think, like, what that looks like, and then adding some things to, like, even, like, AgentsMD to make sure that, like, documents get updated, and documentation, rather, gets updated by the coding agents, and all that gets put into the PR description. Like, I think we can…
196 00:17:38.990 ⇒ 00:17:41.000 Samuel Roberts: Solidify that a little bit more.
197 00:17:41.850 ⇒ 00:17:49.160 Samuel Roberts: But, I mean, the one you did just now looked pretty good. I was able to follow, like, everything, because sometimes jumping into those, you know, code…
198 00:17:49.680 ⇒ 00:17:50.980 Samuel Roberts: or AI-generated.
199 00:17:51.210 ⇒ 00:17:57.409 Uttam Kumaran: Yeah, a lot of the smaller ones are gonna be fine, some of… but I will try to do Loom as well, so it’s just easy for you.
200 00:17:57.790 ⇒ 00:18:06.679 Samuel Roberts: Yeah, I think that’s definitely helpful in terms of, like, testing, too. Like, showing that it’s been tested is always good, too, because, like, if there’s gonna be 10 PRs, I’m not gonna be able to, like.
201 00:18:06.840 ⇒ 00:18:11.429 Samuel Roberts: Do that. And then we can also spread it out a little bit, too, depending on how big the PRs are.
202 00:18:13.810 ⇒ 00:18:16.760 Samuel Roberts: You know, I don’t think I necessarily need to be the only gatekeeper here.
203 00:18:17.410 ⇒ 00:18:21.600 Uttam Kumaran: Yeah, so that’s… so that’s, again, that’s… I would… I’ll just defer to you, like.
204 00:18:21.600 ⇒ 00:18:21.980 Samuel Roberts: Yeah.
205 00:18:21.980 ⇒ 00:18:25.889 Uttam Kumaran: Whatever you want to do, I think I’m fine with. But for me.
206 00:18:26.060 ⇒ 00:18:30.520 Uttam Kumaran: I just want to try to maintain some type of high velocity, like, I really…
207 00:18:30.760 ⇒ 00:18:31.330 Samuel Roberts: Certainly.
208 00:18:31.330 ⇒ 00:18:34.180 Uttam Kumaran: don’t want to have PRs sitting for, like.
209 00:18:34.440 ⇒ 00:18:38.960 Uttam Kumaran: a few days, but I also know that, like, it can be a lot, so…
210 00:18:39.160 ⇒ 00:18:44.179 Uttam Kumaran: Tell me how to operate, and again, like, I’m the first Like, business user here.
211 00:18:44.380 ⇒ 00:18:50.470 Uttam Kumaran: as people get smarter and we train people up, more people are gonna be pushing PRs, and they’re gonna be worse than mine.
212 00:18:50.870 ⇒ 00:18:58.599 Uttam Kumaran: Yeah. Because I… like, you guys know me, so I’m… my stuff you can think about is, like, I’m the most enabled person here.
213 00:18:58.600 ⇒ 00:18:59.070 Samuel Roberts: Right.
214 00:18:59.070 ⇒ 00:19:14.109 Uttam Kumaran: Everything, if we don’t set the principles, every… it’s gonna… think about, like, the nth marketing person, like, shipping something. It’s gonna be really garbage. So I really want us to just, like, try to have some principles so that any net new person shipping stuff can adhere to those.
215 00:19:14.480 ⇒ 00:19:26.149 Uttam Kumaran: Because imagine, like, someone on the business side just being, build this, and then it builds, and they ship the PR, like, it could get really gnarly. Like, I’m not doing that. Much more, I’m a little bit more exercise, so…
216 00:19:26.490 ⇒ 00:19:29.129 Samuel Roberts: Yeah, but it’s only gonna go down from here.
217 00:19:29.130 ⇒ 00:19:31.179 Uttam Kumaran: You’re only gonna go down from here, yeah.
218 00:19:31.180 ⇒ 00:19:33.960 Samuel Roberts: Yeah, no, I hear you, I hear you, that’s a really good, good point. Okay.
219 00:19:33.960 ⇒ 00:19:37.950 Uttam Kumaran: Because I’m making sure that there are PR descriptions, I’m looking through all the code.
220 00:19:38.170 ⇒ 00:19:40.620 Uttam Kumaran: I’m testing everything really well.
221 00:19:43.050 ⇒ 00:19:47.560 Uttam Kumaran: a lot of folk folks are just not gonna do that. So either we need to just think about, like.
222 00:19:47.820 ⇒ 00:19:50.509 Uttam Kumaran: what is… what is the MVP PR?
223 00:19:51.330 ⇒ 00:20:05.699 Uttam Kumaran: for the most part, I think people are going to ship up something that works, but then the code is probably not going to be written well, that maybe it needs to get… this team’s job is going to be probably more of, like, polish it and get it out, right? Right. You know, that’s sort of what we talked about.
224 00:20:06.240 ⇒ 00:20:15.190 Samuel Roberts: Yeah, and I’m just… I’m worried about, like, edge cases and thinking through a few other things that, like, might not be the first. Even if you’re doing a loom and you’re testing it, you know, you’re testing the perfect path kind of thing.
225 00:20:16.370 ⇒ 00:20:16.750 Uttam Kumaran: Yeah.
226 00:20:18.480 ⇒ 00:20:21.139 Samuel Roberts: Okay, no, this is… this is good stuff. I mean, I think…
227 00:20:22.650 ⇒ 00:20:24.760 Samuel Roberts: You can give it a little more thought,
228 00:20:25.730 ⇒ 00:20:28.800 Samuel Roberts: And put something together. Give it some more thought, don’t perfect PR.
229 00:20:28.800 ⇒ 00:20:32.910 Uttam Kumaran: Yeah, don’t, don’t… don’t give it too much thought, but like…
230 00:20:33.290 ⇒ 00:20:33.659 Samuel Roberts: Yeah, yeah.
231 00:20:33.660 ⇒ 00:20:35.760 Uttam Kumaran: Use me as a guinea pig, because I’m gonna be.
232 00:20:35.760 ⇒ 00:20:36.140 Samuel Roberts: Totally.
233 00:20:38.620 ⇒ 00:20:43.589 Uttam Kumaran: Okay, cool. My other question was, like, should we just combine all the repos?
234 00:20:46.600 ⇒ 00:20:56.380 Uttam Kumaran: I don’t know, like, I’m a user, and I was, like, getting frustrated because I’m having to bounce back between a bunch of things. I wonder, like, what the risk is in just combining everything into platform.
235 00:20:56.510 ⇒ 00:21:01.130 Uttam Kumaran: Because ultimately, you guys, you guys can put the app In a subfolder.
236 00:21:01.760 ⇒ 00:21:07.080 Samuel Roberts: That’s… yeah, we could… I mean, I was thinking about that a little bit in terms of, like, we’re already adding other apps to the platform repo.
237 00:21:07.080 ⇒ 00:21:10.979 Uttam Kumaran: Yeah, and like, I’m also… we’re talking to people, like, in the business about, like.
238 00:21:11.560 ⇒ 00:21:16.080 Uttam Kumaran: Repos and things like that, and people are getting confused at, like, the most basic things.
239 00:21:16.330 ⇒ 00:21:16.710 Samuel Roberts: Right.
240 00:21:16.710 ⇒ 00:21:20.890 Uttam Kumaran: So… I’m almost like maybe we rely on code owners.
241 00:21:21.270 ⇒ 00:21:23.789 Uttam Kumaran: And just shove everything into one repo.
242 00:21:25.110 ⇒ 00:21:30.209 Uttam Kumaran: And then, I also think, like, monorepo structure is gonna be more and more meaningful.
243 00:21:31.330 ⇒ 00:21:35.660 Uttam Kumaran: But, yeah, I don’t know, Gabe or anybody, if anyone has any, like, thoughts?
244 00:21:37.840 ⇒ 00:21:50.819 Clarence Stone: Yeah, so for context handling, I think we could really benefit from a monorepo. I, like, I don’t have a problem switching around, it’s the fact that, like, prompting and holding onto the context is difficult.
245 00:21:52.980 ⇒ 00:21:53.460 Uttam Kumaran: Okay.
246 00:21:53.460 ⇒ 00:22:03.010 Samuel Roberts: Yeah, I mean… So we’re… I mean, for the Vault and the playbook, you’re saying, basically, like, get those as, like, other folders at, like, a high level? Yeah. So it’d be, like, Vault, Playbook, apps…
247 00:22:03.440 ⇒ 00:22:04.240 Uttam Kumaran: Yes.
248 00:22:04.420 ⇒ 00:22:05.010 Uttam Kumaran: Yeah.
249 00:22:05.010 ⇒ 00:22:20.400 Samuel Roberts: Yeah, I… my first thought is no, but just because that’s kind of my, like, oh, no, we should keep it separate, but I can see the benefit, and I’m just thinking, how big is this gonna get? It’s mostly just text files, markdown files, right? We’re not really storing much of else.
250 00:22:20.610 ⇒ 00:22:22.360 Samuel Roberts: In these repos, right?
251 00:22:22.360 ⇒ 00:22:24.869 Uttam Kumaran: And it’s probably not the worst thing.
252 00:22:25.910 ⇒ 00:22:31.490 Samuel Roberts: you know, just syncing the repos down would get theoretically slower, but if it’s just text, I’m not that worried. It’s not like we’re.
253 00:22:31.490 ⇒ 00:22:33.550 Uttam Kumaran: Can I give you some other,
254 00:22:34.790 ⇒ 00:22:38.579 Uttam Kumaran: potentially ways to nudge you, like, as we start to use…
255 00:22:38.860 ⇒ 00:22:43.380 Uttam Kumaran: codecs and trigger jobs directly from Slack.
256 00:22:44.510 ⇒ 00:22:45.860 Samuel Roberts: having the…
257 00:22:46.100 ⇒ 00:22:49.919 Uttam Kumaran: having the AI have to decide which repo to go make changes in?
258 00:22:49.920 ⇒ 00:22:51.300 Samuel Roberts: That’s a good point, yeah.
259 00:22:51.300 ⇒ 00:22:52.740 Uttam Kumaran: It’s gonna be something.
260 00:22:54.030 ⇒ 00:22:59.099 Uttam Kumaran: And yeah, I was… I thought about this for, like, 30 minutes, I couldn’t come up with a good reason not to.
261 00:22:59.600 ⇒ 00:23:03.099 Uttam Kumaran: Or, like, a good enough reason, apart from, yeah, I mean, it’s…
262 00:23:03.200 ⇒ 00:23:05.719 Uttam Kumaran: shoving everything in one place, but I feel like…
263 00:23:06.860 ⇒ 00:23:08.970 Uttam Kumaran: It’s actually nicer to see everything there.
264 00:23:08.970 ⇒ 00:23:15.220 Samuel Roberts: Yeah, especially, like, my only other thought would be just, like, protecting the code, but if we’re getting people in there anyway, and that’s, like.
265 00:23:15.220 ⇒ 00:23:17.359 Uttam Kumaran: Well, you just code… just code owner…
266 00:23:17.360 ⇒ 00:23:27.650 Samuel Roberts: I don’t think it’s that big… yeah, I think that’s fine. I’m not as worried about that. It’s more just, like, you know, my gut is, like, none of the non-technical people shouldn’t be in the code, but, like, we’re already trying to get to that point anyway, so I’m already.
267 00:23:27.650 ⇒ 00:23:28.160 Uttam Kumaran: Fighting that.
268 00:23:28.160 ⇒ 00:23:31.049 Samuel Roberts: like, instinct. I think it’s, yeah, I…
269 00:23:33.100 ⇒ 00:23:35.210 Uttam Kumaran: And, like, let me give you another example.
270 00:23:35.210 ⇒ 00:23:35.760 Samuel Roberts: Yeah.
271 00:23:35.760 ⇒ 00:23:37.889 Uttam Kumaran: Sorry, I’m just, like, I really.
272 00:23:37.890 ⇒ 00:23:39.530 Samuel Roberts: No, no, you’re good, you’re good.
273 00:23:39.530 ⇒ 00:23:56.119 Uttam Kumaran: Another example is, like, I’m gonna… I’m gonna move some Webflow to Astral or whatever, this week, I think. And imagine the ability to go from, like, a meeting transcript to a case study.
274 00:23:56.600 ⇒ 00:23:57.820 Uttam Kumaran: a landing page.
275 00:23:59.710 ⇒ 00:24:01.449 Uttam Kumaran: All with, like.
276 00:24:01.940 ⇒ 00:24:03.819 Samuel Roberts: Just prompting directly.
277 00:24:04.200 ⇒ 00:24:08.939 Uttam Kumaran: So, right now, if we were to do that, I would have to hop across probably 3 repos.
278 00:24:08.940 ⇒ 00:24:09.490 Samuel Roberts: Right.
279 00:24:10.030 ⇒ 00:24:15.710 Uttam Kumaran: So instead, it’s like, for example, if a marketing person wants to ship a new landing page, I want that to be, like.
280 00:24:16.540 ⇒ 00:24:20.619 Uttam Kumaran: like… fifth grade, you know? And I think it’s…
281 00:24:20.870 ⇒ 00:24:28.870 Uttam Kumaran: that sort of, like, cross-collaboration and, like, execution is only possible right now. So that’s, like, my goal by the end of the week. I’m gonna move some stuff.
282 00:24:29.290 ⇒ 00:24:31.580 Uttam Kumaran: from Webflow into platform.
283 00:24:31.800 ⇒ 00:24:34.449 Samuel Roberts: And then try to basically enable a, like.
284 00:24:34.450 ⇒ 00:24:36.690 Uttam Kumaran: Chat to landing page flow.
285 00:24:38.000 ⇒ 00:24:44.529 Uttam Kumaran: And then we’re, like… there’s, like, sort of, like, okay, cool, I can actually, like, cut a lot of costs on the Webflow development side, but I do.
286 00:24:44.530 ⇒ 00:24:45.860 Samuel Roberts: Yeah, yeah, definitely.
287 00:24:45.860 ⇒ 00:24:47.419 Uttam Kumaran: So that’s just a couple reasons.
288 00:24:47.420 ⇒ 00:24:55.020 Samuel Roberts: No, I think that does make sense. I think the only, like, big thing to worry about is, like, tooling around the monorepo and making sure that, like.
289 00:24:55.550 ⇒ 00:24:56.320 Samuel Roberts: You know.
290 00:24:56.960 ⇒ 00:25:02.329 Samuel Roberts: There’s, you know… There’s different tools for that, and just maybe diving into that.
291 00:25:03.060 ⇒ 00:25:05.370 Uttam Kumaran: Yeah, so that’s where I would… I would nudge you, because…
292 00:25:05.370 ⇒ 00:25:06.289 Samuel Roberts: Yeah, that’s fine.
293 00:25:06.290 ⇒ 00:25:11.989 Uttam Kumaran: Yeah, like, I… I would… I’m thinking about it, and I’m sort of being a good citizen, but, like.
294 00:25:12.230 ⇒ 00:25:17.980 Uttam Kumaran: It’s, I sort of… I want to make sure you guys own this repo.
295 00:25:17.980 ⇒ 00:25:23.810 Samuel Roberts: Yeah, the only other thing that jumps to mind is there’s just gonna be a lot of extra commits in terms of, like, how we’re adding.
296 00:25:23.910 ⇒ 00:25:25.300 Uttam Kumaran: Transcripts, right?
297 00:25:26.040 ⇒ 00:25:28.960 Uttam Kumaran: Well, I just… I actually just did all of it, so there’s, like.
298 00:25:29.110 ⇒ 00:25:30.989 Uttam Kumaran: All of it are in there.
299 00:25:30.990 ⇒ 00:25:33.549 Samuel Roberts: No, I know, but, like, every meet… is every meeting a new commit?
300 00:25:34.450 ⇒ 00:25:39.009 Uttam Kumaran: Well, yes, directly, the main.
301 00:25:39.470 ⇒ 00:25:39.970 Samuel Roberts: Yeah.
302 00:25:39.970 ⇒ 00:25:42.139 Uttam Kumaran: Again, it’s in one folder that, like.
303 00:25:42.140 ⇒ 00:25:44.359 Samuel Roberts: Yeah, that’s what I’m saying, as long as… yeah, and again, like.
304 00:25:44.360 ⇒ 00:25:45.050 Uttam Kumaran: need to be.
305 00:25:45.050 ⇒ 00:25:49.270 Samuel Roberts: Railway has the watch folder, so it’s not like it’s gonna be, like, trying to redeploy things or anything. I think we’re…
306 00:25:49.560 ⇒ 00:25:51.420 Samuel Roberts: I think I’m convincing myself that it’s okay.
307 00:25:51.710 ⇒ 00:25:52.410 Uttam Kumaran: Okay.
308 00:25:52.650 ⇒ 00:25:55.810 Uttam Kumaran: I mean, I want to hear the objections, because I’m.
309 00:25:55.810 ⇒ 00:25:56.650 Samuel Roberts: Yeah, no, I mean…
310 00:25:56.650 ⇒ 00:25:58.239 Uttam Kumaran: Use case from the user side.
311 00:25:58.370 ⇒ 00:25:59.960 Uttam Kumaran: And…
312 00:26:00.140 ⇒ 00:26:12.239 Uttam Kumaran: I’m also trying to pave the way for everybody behind me who is gonna be really fucking confused, not only just about, like, what it is this whole thing is, but also, like, how to use GitHub and stuff, and so…
313 00:26:12.310 ⇒ 00:26:13.990 Samuel Roberts: Yeah, I was having trouble with GitHub.
314 00:26:13.990 ⇒ 00:26:20.850 Uttam Kumaran: of, like, okay, the least enabled business person at Brainforge needs to be in here, and, like.
315 00:26:21.010 ⇒ 00:26:23.750 Uttam Kumaran: where are they gonna get tripped up on? And…
316 00:26:23.750 ⇒ 00:26:24.440 Samuel Roberts: Yeah.
317 00:26:24.680 ⇒ 00:26:28.400 Uttam Kumaran: I’ve been explaining it to tons of people, and I’m just seeing that, you know?
318 00:26:29.030 ⇒ 00:26:33.319 Samuel Roberts: No, you’re absolutely right. Like, the simpler we can nuke it for those kind of people, the better.
319 00:26:33.320 ⇒ 00:26:50.919 Pranav Narahari: I would say two, like, things to consider, they’re not objections, is just that when we’re deploying these apps, we should just be careful for what are we creating as the… I guess, like, the deployed directory, just so that we’re not rebuilding every app, because that can be…
320 00:26:50.920 ⇒ 00:26:51.889 Samuel Roberts: See, that’s exactly what I was thinking, too.
321 00:26:51.890 ⇒ 00:26:52.579 Uttam Kumaran: What does that mean?
322 00:26:53.000 ⇒ 00:26:54.300 Pranav Narahari: So basically, when…
323 00:26:54.300 ⇒ 00:26:55.020 Samuel Roberts: These are happening.
324 00:26:55.020 ⇒ 00:26:58.299 Pranav Narahari: in, like, certain folders.
325 00:26:58.540 ⇒ 00:27:15.110 Pranav Narahari: you’re gonna have, like, tons of apps in just one repo, right? If you don’t supply, like, a directory in, like, this, like, I think it’s called, like, the, like, the build directory or something, build destination. Yeah. Then essentially it’ll just rebuild every single app that
326 00:27:15.450 ⇒ 00:27:32.429 Pranav Narahari: that is within that direct… that’s within that repo, because what it normally does is just looks for any code changes. If there’s a code change, then it rebuilds. So you have to basically say, like, only if there’s changes in this, like, root folder, then you rebuild the app.
327 00:27:32.430 ⇒ 00:27:36.379 Pranav Narahari: What I was saying about, like, watch folders and railway, like, it’s… Yeah, watch folder is the right term.
328 00:27:36.380 ⇒ 00:27:47.480 Samuel Roberts: It’s… it’s… it’s doable, we just have to be a little more diligent, probably, and we’re already being pretty diligent, like, I’ve noticed this on Leela, because we have a bunch of apps deploying, and I want to redeploy MCP, so we kind of already have a sense of that. I think…
329 00:27:47.990 ⇒ 00:27:50.590 Samuel Roberts: One thing to do with the platform, especially if we’re putting it
330 00:27:50.860 ⇒ 00:27:57.700 Samuel Roberts: on railway is just defining those, like, railway configs really well, or even Docker files, whatever we want that’s just gonna be, like.
331 00:27:57.960 ⇒ 00:27:59.630 Samuel Roberts: Pointing everything to the right spot.
332 00:28:00.000 ⇒ 00:28:06.470 Samuel Roberts: I’m not too worried about that, because it can handle it, but we just have to be… diligent about.
333 00:28:06.470 ⇒ 00:28:07.040 Pranav Narahari: Yeah.
334 00:28:07.040 ⇒ 00:28:09.539 Samuel Roberts: it the right way. That’s less of a concern in terms of, like.
335 00:28:09.710 ⇒ 00:28:18.210 Samuel Roberts: anything else but just making sure that it’s set up the right way, and honestly, like, once they’re deployed and set up the watch folders, it’s not that bad. It’s not, like, anything we have to maintain.
336 00:28:18.460 ⇒ 00:28:22.860 Samuel Roberts: Yeah, exactly. It’s actually pretty quick. I think the… one other thing that I thought of, too, is just… Yeah.
337 00:28:22.860 ⇒ 00:28:28.129 Pranav Narahari: we should have certain, and I don’t know if this is already there,
338 00:28:28.360 ⇒ 00:28:31.870 Pranav Narahari: Like, certain, like, push, like,
339 00:28:32.600 ⇒ 00:28:35.319 Pranav Narahari: Like, you shouldn’t be able to push onto certain branches.
340 00:28:35.480 ⇒ 00:28:36.340 Samuel Roberts: Yeah, I was gonna say…
341 00:28:36.340 ⇒ 00:28:37.619 Pranav Narahari: All those requirements, yeah.
342 00:28:37.620 ⇒ 00:28:41.069 Samuel Roberts: I was thinking, like, yeah, locking down some of the branches.
343 00:28:41.420 ⇒ 00:28:55.279 Samuel Roberts: better than we have been, because we were doing that on Lilo, and I think this makes, you know, now that… if we’re going to be combining things, that makes a lot of sense. We might even want to do some more, I mean, it might not be the biggest deal, but pre… pre-commit hooks or something,
344 00:28:55.280 ⇒ 00:28:56.060 Pranav Narahari: Yeah.
345 00:28:56.060 ⇒ 00:29:02.869 Samuel Roberts: Just to, like, make sure that, like, if someone’s changed code, it knows… you know what I mean? Like, it might just, like, throw an error or something, so it doesn’t, like…
346 00:29:03.450 ⇒ 00:29:06.490 Samuel Roberts: Commit a file that shouldn’t have been changed or something.
347 00:29:06.730 ⇒ 00:29:10.090 Pranav Narahari: Yeah, that’s my worry, I think, like, the…
348 00:29:10.210 ⇒ 00:29:13.420 Pranav Narahari: What’s gonna happen is that the…
349 00:29:13.650 ⇒ 00:29:17.609 Pranav Narahari: the PR and, like, the QA process might just be…
350 00:29:17.880 ⇒ 00:29:30.510 Pranav Narahari: is gonna get longer now, because there’s people that aren’t as technical shipping code, and so we want to do things like that, like having certain, like, pre-commit hooks, certain, like, linting checks, just so, like.
351 00:29:30.510 ⇒ 00:29:36.399 Uttam Kumaran: Yeah, I think pre-commit hook, linting checks, as well as, like, auto-labeling on PRs, like.
352 00:29:36.840 ⇒ 00:29:39.959 Uttam Kumaran: Some PRs are gonna be, like, cake, and…
353 00:29:39.960 ⇒ 00:29:40.670 Samuel Roberts: Exactly, exactly.
354 00:29:40.670 ⇒ 00:29:47.629 Uttam Kumaran: guys should rely more and more on the… on, like, the AI reviews. Some shit, like I have, I’m like.
355 00:29:47.930 ⇒ 00:29:56.389 Uttam Kumaran: like, for example, some of these, I, like, okay, I’m like, it’s 90% there. Sam will look and be like, it’s pretty good. There’s some stuff where I’m like, I got it 60%,
356 00:29:56.970 ⇒ 00:30:03.219 Uttam Kumaran: And, like, for example, like, the Zoom… I, like, did a lot of work on, like, moving Zoom from N8N Windmill.
357 00:30:03.340 ⇒ 00:30:07.109 Uttam Kumaran: to Mastra, but I’m like, yo, this is, like, so complicated, like…
358 00:30:07.410 ⇒ 00:30:11.430 Uttam Kumaran: I got a 50%, and I’m like, okay, I’m gonna hand it off, right? So…
359 00:30:11.430 ⇒ 00:30:15.039 Samuel Roberts: Right. I would… so there’s gonna be differences like that, you know?
360 00:30:15.620 ⇒ 00:30:32.650 Uttam Kumaran: Because what we’re… again, like, what we’re democratizing here is not shipping end-to-end. So I want to be really clear, like, I don’t need people to, like, learn software engineering, they are not going to learn. What I need them to… what I want to democratize is the ability to go from idea to, like, prototype.
361 00:30:32.780 ⇒ 00:30:33.450 Samuel Roberts: Yep.
362 00:30:33.450 ⇒ 00:30:43.880 Uttam Kumaran: That’s what the goal is. So, like, they should… people should be able to run stuff locally, get something that’s working, because idea is the bottleneck right now. Or, like, idea to prototype is the bottleneck.
363 00:30:44.040 ⇒ 00:30:53.839 Uttam Kumaran: But that’s what we’re democratizing. I want to be… I’m gonna reiterate this many times, because the rest of the world is not thinking this way right now. It’s like, I don’t need people to ship perfect code.
364 00:30:53.840 ⇒ 00:30:54.190 Samuel Roberts: Right.
365 00:30:54.190 ⇒ 00:31:01.070 Uttam Kumaran: I don’t need them to know software engineering, and I don’t need them to be able to push domain and, like, move… move, like.
366 00:31:01.410 ⇒ 00:31:03.120 Uttam Kumaran: Like, that fast.
367 00:31:03.300 ⇒ 00:31:06.959 Uttam Kumaran: like, 24-hour, 48-hour SLA on stuff is fine.
368 00:31:07.080 ⇒ 00:31:11.000 Uttam Kumaran: I needed to just, like, open up the amount of people that can push ideas.
369 00:31:11.160 ⇒ 00:31:14.419 Uttam Kumaran: To somewhere, and then we’ll take it from there, you know?
370 00:31:14.790 ⇒ 00:31:16.349 Samuel Roberts: That’s… that’s a good way to put it.
371 00:31:17.340 ⇒ 00:31:27.050 Samuel Roberts: Yeah, the only thing that comes to mind right now, beyond what, like, Pranak brought up, is just environment variables and managing that across
372 00:31:27.340 ⇒ 00:31:28.640 Samuel Roberts: More and more people.
373 00:31:29.110 ⇒ 00:31:31.140 Samuel Roberts: It doesn’t really matter.
374 00:31:31.140 ⇒ 00:31:32.689 Uttam Kumaran: Yeah, so I’ve been using…
375 00:31:32.690 ⇒ 00:31:33.310 Samuel Roberts: Yeah.
376 00:31:33.600 ⇒ 00:31:35.379 Uttam Kumaran: I’ve been using 1Pass CLI.
377 00:31:36.300 ⇒ 00:31:42.140 Samuel Roberts: Okay, cool. I was gonna ask if, like, we… yeah, okay, cool. I was reading a little bit about that this weekend with my CloudBot stuff, so…
378 00:31:42.140 ⇒ 00:31:48.300 Uttam Kumaran: I would probably suggest, yeah, either we use Railway, To pull down, or…
379 00:31:48.300 ⇒ 00:31:50.960 Samuel Roberts: That’s, that’s why I set up for it. Yeah, okay.
380 00:31:52.020 ⇒ 00:31:55.389 Samuel Roberts: Yeah, no, I think it’s definitely solvable, I just wanted to mention it.
381 00:31:55.790 ⇒ 00:31:58.340 Samuel Roberts: Because, like, most people, like, if they’re just putting things in the…
382 00:31:58.520 ⇒ 00:32:02.829 Samuel Roberts: Playbook or vaults at any given time don’t need it, but if they… when they want to do something.
383 00:32:03.240 ⇒ 00:32:06.239 Samuel Roberts: in the apps, they’re gonna need that for even running locally.
384 00:32:06.470 ⇒ 00:32:07.410 Samuel Roberts: Yeah.
385 00:32:07.890 ⇒ 00:32:12.920 Samuel Roberts: Okay. No, this is… This is good. I think, I think…
386 00:32:13.560 ⇒ 00:32:15.780 Samuel Roberts: I’m not gonna say yes right now.
387 00:32:16.130 ⇒ 00:32:21.659 Samuel Roberts: But I’m getting there, and this probably will by the end of the day. That sounds good.
388 00:32:21.840 ⇒ 00:32:24.999 Samuel Roberts: And make a plan for it, too, just make sure we have everything in place.
389 00:32:25.140 ⇒ 00:32:26.670 Samuel Roberts: In terms of moving stuff around.
390 00:32:27.140 ⇒ 00:32:37.850 Uttam Kumaran: Okay, cool. I know we’re out of time here, so I’ll send a couple of updates in the channel, just for some remaining things, and then maybe later this week, I can… I’ll start to record looms for each of my PRs, so there’s those demos there.
391 00:32:38.300 ⇒ 00:32:38.980 Samuel Roberts: Perfect.
392 00:32:39.130 ⇒ 00:32:39.930 Samuel Roberts: Cool.
393 00:32:40.540 ⇒ 00:32:41.850 Samuel Roberts: Alright, sounds good.
394 00:32:42.500 ⇒ 00:32:43.000 Uttam Kumaran: Okay.
395 00:32:44.180 ⇒ 00:32:47.700 Samuel Roberts: Yeah, anything else? We are at time, but just in case there’s anything we missed.
396 00:32:49.260 ⇒ 00:32:50.360 Samuel Roberts: Alright, cool.
397 00:32:50.570 ⇒ 00:32:54.590 Samuel Roberts: I’ll see most of you later on various calls. If you need anything from me, let me know.
398 00:32:54.940 ⇒ 00:32:56.460 Samuel Roberts: Otherwise, have a good day.
399 00:32:57.220 ⇒ 00:32:57.980 Uttam Kumaran: Thank you.