Meeting Title: Eden AI: Project Plan Technical Implementation Working Session Date: 2026-03-24 Meeting participants: Awaish Kumar, Samuel Roberts, Pranav Narahari
WEBVTT
1 00:00:59.640 ⇒ 00:01:00.650 Samuel Roberts: Anyways…
2 00:01:06.790 ⇒ 00:01:07.400 Awaish Kumar: Hello.
3 00:01:08.280 ⇒ 00:01:09.130 Samuel Roberts: How are you?
4 00:01:09.560 ⇒ 00:01:11.079 Awaish Kumar: Wait, how about you?
5 00:01:11.860 ⇒ 00:01:12.970 Samuel Roberts: Then, alright.
6 00:01:15.380 ⇒ 00:01:18.190 Samuel Roberts: Just trying to find that… Mom.
7 00:01:19.380 ⇒ 00:01:23.569 Samuel Roberts: technical information draft that I have for this project the other day.
8 00:01:31.530 ⇒ 00:01:32.570 Samuel Roberts: So…
9 00:01:34.380 ⇒ 00:01:38.850 Samuel Roberts: Yeah, I wasn’t sure if we wanted to do this here, or async, or how you wanted to best handle this.
10 00:01:38.970 ⇒ 00:01:41.439 Samuel Roberts: How familiar are you with this project so far?
11 00:01:41.560 ⇒ 00:01:43.290 Awaish Kumar: I don’t know.
12 00:01:43.290 ⇒ 00:01:48.060 Samuel Roberts: Okay, okay, that’s fine. No, I got a little bit of context last week, I think, before I was,
13 00:01:48.270 ⇒ 00:01:52.490 Samuel Roberts: out Thursday, Friday, so I’m just trying to make sure I have all the… everything I need here.
14 00:01:56.380 ⇒ 00:01:57.550 Awaish Kumar: Okay, yeah.
15 00:01:58.390 ⇒ 00:02:00.180 Samuel Roberts: where I put that…
16 00:02:03.360 ⇒ 00:02:04.409 Samuel Roberts: Where is it?
17 00:02:07.060 ⇒ 00:02:07.959 Samuel Roberts: That’s not.
18 00:02:21.600 ⇒ 00:02:25.450 Samuel Roberts: So, there… the… One sec.
19 00:02:25.550 ⇒ 00:02:29.730 Samuel Roberts: It’s not how I want to view this.
20 00:02:32.700 ⇒ 00:02:33.209 Samuel Roberts: Mmm…
21 00:02:37.310 ⇒ 00:02:38.030 Samuel Roberts: file.
22 00:02:41.510 ⇒ 00:02:46.440 Samuel Roberts: So let me share you this link, because I don’t know if you have the repo down, but let me…
23 00:02:46.870 ⇒ 00:02:53.389 Samuel Roberts: So this was what, based on the SOW, I had kind of started to throw together.
24 00:02:53.680 ⇒ 00:02:57.110 Samuel Roberts: So, effectively.
25 00:02:57.800 ⇒ 00:03:03.940 Samuel Roberts: what this project is, is… well, there’s two parts to it, right? You saw in the Notion?
26 00:03:05.740 ⇒ 00:03:11.710 Samuel Roberts: I don’t know if you’ve heard the notion yet or not, but… Maybe we start there.
27 00:03:21.560 ⇒ 00:03:26.129 Awaish Kumar: Yeah, we can start. I don’t have a crucial lingo.
28 00:03:26.430 ⇒ 00:03:34.080 Samuel Roberts: Let me show you the note. This is the notion that, Pronounced sense.
29 00:03:34.690 ⇒ 00:03:40.470 Samuel Roberts: Earlier… And so there’s… this is a… so, I mean, I haven’t done a lot of…
30 00:03:40.780 ⇒ 00:03:47.100 Samuel Roberts: work for Eden yet, so this is, like, a new… new client for me, but I guess the COO…
31 00:03:47.230 ⇒ 00:03:50.969 Samuel Roberts: Wants this kind of, like.
32 00:03:51.790 ⇒ 00:03:54.550 Samuel Roberts: Command Center is what we’re calling it right now?
33 00:03:54.750 ⇒ 00:04:00.890 Samuel Roberts: Where he can pull in… Data…
34 00:04:01.440 ⇒ 00:04:18.059 Samuel Roberts: from Google Workspace and Slack to kind of discover themes across the business, is what it says here. So, I think the one you and I need to think about is this first part of the project, the MCP and data access. I think what… let me just double check what, how he phrased it here.
35 00:04:18.730 ⇒ 00:04:26.929 Samuel Roberts: basically, yeah, I’m here to give you context, and then we need your help on the milestones that pertain to data. Particularly, I’d like to know if we can leverage
36 00:04:27.480 ⇒ 00:04:28.930 Samuel Roberts: GWS…
37 00:04:30.240 ⇒ 00:04:36.229 Samuel Roberts: CLI, as opposed to data pipelines, just an idea, not sure if it’d be the right choice. Okay, so…
38 00:04:38.200 ⇒ 00:04:45.479 Samuel Roberts: There are 1, 2, 3… the first 3 milestones are, I think, what we need to talk about here.
39 00:04:50.120 ⇒ 00:04:52.360 Awaish Kumar: Okay, for the GWS CLI, you…
40 00:04:52.510 ⇒ 00:04:55.390 Awaish Kumar: You can basically read, like, data from.
41 00:04:55.910 ⇒ 00:04:56.810 Samuel Roberts: Right.
42 00:04:56.810 ⇒ 00:05:01.720 Awaish Kumar: Gmail, and the Sheets, right? We can’t… you can’t read the BigQuery data from there.
43 00:05:03.720 ⇒ 00:05:05.480 Samuel Roberts: Okay, so I think…
44 00:05:05.830 ⇒ 00:05:10.870 Samuel Roberts: Let’s… maybe we should pull Pranav into this, because he has a little more context on these different milestones now.
45 00:05:13.100 ⇒ 00:05:13.730 Awaish Kumar: Boom.
46 00:05:14.720 ⇒ 00:05:18.870 Samuel Roberts: Because this is a little different than what I had, I think, but…
47 00:05:18.980 ⇒ 00:05:26.080 Awaish Kumar: Yeah, Command Center said you need to leverage data from Google Workspace, And Slack. It doesn’t say…
48 00:05:26.270 ⇒ 00:05:31.310 Awaish Kumar: the unique BigQuery data. So far, yes, you can use…
49 00:05:31.920 ⇒ 00:05:35.540 Awaish Kumar: GS, GWCLI, because I think,
50 00:05:35.800 ⇒ 00:05:39.089 Awaish Kumar: We already are doing for Brain Forge.
51 00:05:40.730 ⇒ 00:05:43.800 Awaish Kumar: like, Utam and Brill already worked on…
52 00:05:43.930 ⇒ 00:05:49.729 Awaish Kumar: Something like that, where they’re reading data from Google Docs directly in Scars.
53 00:05:49.730 ⇒ 00:05:55.040 Samuel Roberts: Okay, so I think he wants to be able to do this in his Gemini, I think, right? Yeah, chat with Gemini.
54 00:05:55.360 ⇒ 00:05:58.700 Samuel Roberts: So that’s, I think, the question down on milestone 1.
55 00:06:00.990 ⇒ 00:06:05.969 Awaish Kumar: But what I’m trying to say is that if you just need Google Workspace data.
56 00:06:06.140 ⇒ 00:06:13.610 Awaish Kumar: Obviously, you can use, Upsili…
57 00:06:14.380 ⇒ 00:06:19.649 Awaish Kumar: to… like, we, like, we need to move this data to S3 or something?
58 00:06:19.870 ⇒ 00:06:21.809 Awaish Kumar: For Gemini to read, or…
59 00:06:23.300 ⇒ 00:06:28.529 Samuel Roberts: That’s, I think, what we have to kind of think about a little bit here. I don’t know the best way to go about that, yeah.
60 00:06:28.870 ⇒ 00:06:32.229 Awaish Kumar: So, it doesn’t mention what I need to look at, like, is it…
61 00:06:32.410 ⇒ 00:06:38.720 Awaish Kumar: It’s just the same C people’s data access, and you just… what data access? Like, you just need Google Workspace and Slack, or anything else?
62 00:06:39.470 ⇒ 00:06:42.350 Samuel Roberts: I think that’s all we’re looking for is, yeah, Google Workspace and Slack.
63 00:06:42.910 ⇒ 00:06:46.990 Awaish Kumar: Yeah, I can help with Google Workspace, data, like.
64 00:06:47.120 ⇒ 00:06:53.279 Awaish Kumar: Like, we can employ the same techniques you are using right now in the AI tool, like, I can help you with… Right.
65 00:06:53.450 ⇒ 00:06:56.869 Awaish Kumar: the data, but the… how we are bringing in Slack.
66 00:06:57.060 ⇒ 00:07:02.159 Awaish Kumar: So, I remember that we brought… we were bringing this data from Slack using…
67 00:07:08.010 ⇒ 00:07:08.660 Samuel Roberts: Okay.
68 00:07:08.930 ⇒ 00:07:15.280 Awaish Kumar: The previous one we… In the beginning, we started with NATN, which are directly read Flag.
69 00:07:15.870 ⇒ 00:07:17.120 Samuel Roberts: Right, right, right.
70 00:07:18.060 ⇒ 00:07:23.660 Awaish Kumar: And then store it in a… as, like, in storage, right? Google Cloud storage or something.
71 00:07:25.260 ⇒ 00:07:29.589 Samuel Roberts: So the question, I think, then becomes… so you’re saying Google Workspace, we can use the CLI,
72 00:07:29.790 ⇒ 00:07:32.410 Samuel Roberts: And figure out how to connect that to Gemini, basically.
73 00:07:32.830 ⇒ 00:07:33.500 Awaish Kumar: Precise.
74 00:07:33.500 ⇒ 00:07:34.930 Samuel Roberts: Black, though, we’re not sure.
75 00:07:35.570 ⇒ 00:07:37.310 Awaish Kumar: For the AI to work.
76 00:07:37.730 ⇒ 00:07:45.559 Awaish Kumar: Actually, I’m asking, like, I’m trying to understand, like, if you are building something like… Gemini,
77 00:07:45.730 ⇒ 00:07:49.190 Awaish Kumar: Where do you need this data? Like, how will it…
78 00:07:49.340 ⇒ 00:07:58.509 Awaish Kumar: where Gemini can take the data from. Can Gemini read from S3, or will you be connecting it to some database or something?
79 00:07:58.510 ⇒ 00:08:02.630 Samuel Roberts: I don’t know if I have all that context here yet.
80 00:08:04.790 ⇒ 00:08:11.639 Awaish Kumar: So, for AI to read all those docs, it needs to be somewhere where AI can read it, maybe in Google.
81 00:08:11.990 ⇒ 00:08:17.450 Awaish Kumar: somewhere, right? And then I can figure out, okay, how can I move, select, data.
82 00:08:17.640 ⇒ 00:08:22.680 Awaish Kumar: from Slack to… Google Cloud story, like, and the Google Workspace data…
83 00:08:23.230 ⇒ 00:08:29.040 Awaish Kumar: to Google Cloud Storage, or some… it’s already on storage. It’s already on drive, so it may be…
84 00:08:30.420 ⇒ 00:08:37.429 Awaish Kumar: You can directly… I need to know, like, what… how… how the f- and the…
85 00:08:38.440 ⇒ 00:08:43.529 Awaish Kumar: How’s the end user? Like, the… The EI, or Gemini?
86 00:08:44.090 ⇒ 00:08:49.080 Awaish Kumar: How it will read that data, like, where it wants its data need to be.
87 00:08:51.980 ⇒ 00:08:52.570 Samuel Roberts: Okay.
88 00:08:53.010 ⇒ 00:08:54.050 Samuel Roberts: I…
89 00:08:54.730 ⇒ 00:09:02.719 Samuel Roberts: I’m not 100% sure what the, like… because it says Gemini, so I don’t know if that means we’re building a… if it’s just right in his Gemini, in his Google…
90 00:09:03.210 ⇒ 00:09:06.660 Samuel Roberts: Or if we need to build something on top of a Gemini agent.
91 00:09:07.760 ⇒ 00:09:09.979 Samuel Roberts: Let me… let’s get Pranav in here.
92 00:09:11.290 ⇒ 00:09:14.459 Samuel Roberts: If he can’t look at that. He actually had the conversations.
93 00:09:14.860 ⇒ 00:09:19.129 Samuel Roberts: For this SOW, so he might have a little more context.
94 00:09:25.960 ⇒ 00:09:26.700 Samuel Roberts: Stolen.
95 00:09:49.000 ⇒ 00:09:51.110 Awaish Kumar: Okay, I think we also had…
96 00:09:51.310 ⇒ 00:09:57.099 Awaish Kumar: and pipeline in Dextile, from where I was actually pulling
97 00:09:57.300 ⇒ 00:10:05.850 Awaish Kumar: data for Slack, and then throwing it, to the… Super exciting.
98 00:10:06.730 ⇒ 00:10:11.579 Samuel Roberts: Yes, that’s right. Yeah, for… for the… the platform, Brainforge, you mean? Our Slack?
99 00:10:12.550 ⇒ 00:10:22.120 Awaish Kumar: Yeah, I… yeah, for the… I don’t know after you arrived, I just, like, kind of moved away from AI team, but previously…
100 00:10:22.120 ⇒ 00:10:22.880 Samuel Roberts: Sam.
101 00:10:22.880 ⇒ 00:10:28.869 Awaish Kumar: It’s working, with Casey, I built the pipeline for them.
102 00:10:29.100 ⇒ 00:10:33.640 Awaish Kumar: at that moment, when I… what I did was I actually wrote a Python script which reads data.
103 00:10:33.810 ⇒ 00:10:38.079 Awaish Kumar: I think it’s already… I think it’s still running in Dexter, but yeah, like, then it moves.
104 00:10:38.080 ⇒ 00:10:42.519 Samuel Roberts: Yeah, I think you’re right. I’ve seen that, I just haven’t dealt with it a ton. Aprilf.
105 00:10:43.710 ⇒ 00:10:50.590 Awaish Kumar: So that just moves the data from those channels into… Superbase, and from there.
106 00:10:50.590 ⇒ 00:11:02.740 Samuel Roberts: Right, so then it’s just a RAG system from there, I think, where we embed that, and then retrieve that. Okay. So, Pranav, we’re just talking… so for the first three milestones, that’s what we’re here to discuss, right?
107 00:11:03.450 ⇒ 00:11:09.359 Pranav Narahari: I believe… yes, the first three, those are the data pertaining ones. Yes.
108 00:11:09.360 ⇒ 00:11:14.530 Samuel Roberts: Yeah, MCP and data access. So, this chat with Gemini…
109 00:11:14.820 ⇒ 00:11:19.640 Samuel Roberts: So, Oakes seems to think we can probably use the… the GWS CLI for that.
110 00:11:19.880 ⇒ 00:11:27.380 Samuel Roberts: My question, I think, is I’m a little not sure now. What… this chat with Gemini, is that gonna be, like, right in his Google?
111 00:11:27.800 ⇒ 00:11:28.880 Samuel Roberts: Like, is it…
112 00:11:29.920 ⇒ 00:11:34.830 Samuel Roberts: kind of the way we’re doing it with ABC, where it’s, like, ends their Google Chat, but this one’s just with Gemini?
113 00:11:36.040 ⇒ 00:11:39.610 Samuel Roberts: Is that the idea here? Or are we building another, like, interface?
114 00:11:40.570 ⇒ 00:11:59.559 Pranav Narahari: No, we’re not building another interface, at least not for these milestones. Okay. So, the idea with this whole MCP server is just so that we can have these milestones for them. The end product is, like, maybe they’ll use the MCP server as well. I mean, yeah, they’ll use the…
115 00:12:00.200 ⇒ 00:12:05.769 Pranav Narahari: yeah, the MCP server, I guess, in their… integrated into their Gemini chat, but…
116 00:12:05.940 ⇒ 00:12:17.649 Pranav Narahari: if you look at, like, Project 2, like, the final handoff is, like, they’re just gonna get, like, automated reports within their email. So same backend, pulling from the same data.
117 00:12:17.720 ⇒ 00:12:23.589 Pranav Narahari: But this is just a way for them to, like, query that data, and just show that, okay, we’re…
118 00:12:23.590 ⇒ 00:12:41.289 Pranav Narahari: we’ve… we’re able to pull in all the data across Eden, and you’re able to access it using this MCP server integrated into their Gemini chat. So, I think… are you, like, familiar with just kind of, like, you know, Gemini? You can just, like, use them the same way as ChatGPT, like, I’m talking about, like, that specific portal. Like, they have a…
119 00:12:41.290 ⇒ 00:12:51.900 Samuel Roberts: Yeah, so that’s what I’m… I’m making sure that that is what we’re using. Yeah, I think so, okay. So then the question becomes… so, you want an MCP server specifically? Because I wasn’t sure what MCP plus data access meant here.
120 00:12:51.900 ⇒ 00:12:55.580 Pranav Narahari: No, I mean, so I was just thinking that,
121 00:12:56.900 ⇒ 00:13:00.709 Pranav Narahari: Yeah, maybe MCP is not the right way to do this.
122 00:13:01.160 ⇒ 00:13:05.370 Pranav Narahari: Essentially, we just need to be able to… Have an additional, like.
123 00:13:05.630 ⇒ 00:13:13.339 Pranav Narahari: integration, probably a custom integration within Gemini to access all of this data that, we pull in.
124 00:13:14.260 ⇒ 00:13:14.900 Samuel Roberts: Okay.
125 00:13:14.900 ⇒ 00:13:17.879 Pranav Narahari: Yeah, so I’m happy to change the verbiage there, like, if…
126 00:13:17.880 ⇒ 00:13:24.360 Samuel Roberts: I just want to make sure… Yeah, no, I’m just trying to understand what we, like, were thinking overall. So, yeah, I don’t know enough about, like.
127 00:13:24.700 ⇒ 00:13:37.350 Samuel Roberts: yet about the way the Gemini, like, we were just in GCP for, you know, ABC stuff, and there’s different ways to build, like, Google has all kinds of stuff, so there might be a way to build, like, a Gemini add-on or something that might be MCP, might be…
128 00:13:37.450 ⇒ 00:13:43.279 Samuel Roberts: I don’t know how they handle skills, like Markdown-type stuff that, like, you know, Cloud Code can handle and stuff like that, but…
129 00:13:43.590 ⇒ 00:13:44.450 Samuel Roberts: Yup.
130 00:13:44.600 ⇒ 00:13:47.799 Samuel Roberts: I just don’t know if it can even get a CLI access for…
131 00:13:48.350 ⇒ 00:13:51.010 Samuel Roberts: Google Workspace, or if we… you know what I mean? Like…
132 00:13:52.090 ⇒ 00:13:55.079 Samuel Roberts: I think there’s a little more to understand there.
133 00:13:55.420 ⇒ 00:13:57.950 Pranav Narahari: You don’t know if they’re able to give us CLI access?
134 00:13:57.950 ⇒ 00:14:03.350 Awaish Kumar: What do you need? Like, I… I… I… we already have access to Google,
135 00:14:04.080 ⇒ 00:14:07.279 Awaish Kumar: workspace. Like, we already have access to…
136 00:14:08.120 ⇒ 00:14:14.360 Awaish Kumar: Like, where we use Robert’s email, she provided trident.com, which already has a lot of…
137 00:14:15.120 ⇒ 00:14:16.139 Awaish Kumar: Access to a lot of.
138 00:14:16.140 ⇒ 00:14:24.559 Samuel Roberts: Right, right, no, I’m just… what I’m saying is, like, so, if that Gemini chat, like a, you know, you go to, whatever it is, gemini.google.com or whatever for the workspace, right?
139 00:14:24.800 ⇒ 00:14:25.410 Pranav Narahari: Yup.
140 00:14:26.900 ⇒ 00:14:29.530 Samuel Roberts: Does that already just have Google Workspace access?
141 00:14:30.650 ⇒ 00:14:34.659 Pranav Narahari: It does have it, yeah, integrated already. But…
142 00:14:34.660 ⇒ 00:14:39.490 Samuel Roberts: Is that piece… Separate then, or do we need to work… okay, go ahead, continue.
143 00:14:39.880 ⇒ 00:14:42.999 Pranav Narahari: I think we need to assess
144 00:14:43.860 ⇒ 00:15:02.780 Pranav Narahari: what is the… what that integration looks like, because for this project, we don’t need it to… it can’t be scoped to the individual. We need to create our own scoping, maybe after the fact. And for this initial project that we’re working on, it needs to be just org-wide, because it’s only going to be used by Danny.
145 00:15:02.930 ⇒ 00:15:09.129 Pranav Narahari: So Danny wants to see what is happening in every, like, nook and cranny of the organization, not just in.
146 00:15:09.130 ⇒ 00:15:09.450 Samuel Roberts: Right.
147 00:15:09.450 ⇒ 00:15:13.890 Pranav Narahari: Certain poems, certain things that he has, maybe…
148 00:15:14.410 ⇒ 00:15:30.119 Pranav Narahari: formal access to. So, like, just because he’s added to a document, that shouldn’t be the only context. He should be getting activity from, like, other people’s calendar events as well, other people’s activity in Google Drive, other people’s, Slack messages.
149 00:15:31.090 ⇒ 00:15:33.869 Samuel Roberts: Right, okay, okay.
150 00:15:34.590 ⇒ 00:15:35.960 Samuel Roberts: So, Slack…
151 00:15:36.490 ⇒ 00:15:45.949 Pranav Narahari: So Slack is completely different, but for Google Workspace, like, there is already built-in integrations that he’s using, however, the scope of the stuff. Yeah. Yes. Yeah.
152 00:15:46.430 ⇒ 00:15:47.120 Samuel Roberts: Okay.
153 00:15:47.710 ⇒ 00:15:48.700 Samuel Roberts: So…
154 00:15:50.140 ⇒ 00:15:55.910 Samuel Roberts: we still want to ingest the data for everything, and store that somewhere, probably, then the CLI probably is not enough.
155 00:15:56.770 ⇒ 00:16:01.989 Pranav Narahari: Yeah, I guess that’s kind of, like, my question to you guys, like…
156 00:16:02.260 ⇒ 00:16:06.599 Samuel Roberts: I mean, I’m thinking there’s probably all kinds of access issues with Google, because I don’t know…
157 00:16:07.100 ⇒ 00:16:13.940 Samuel Roberts: I’m not sure, from an admin perspective, how you, like, what… Can be granted across.
158 00:16:14.760 ⇒ 00:16:15.859 Awaish Kumar: We do?
159 00:16:16.070 ⇒ 00:16:18.629 Awaish Kumar: We just… I think we just need…
160 00:16:19.270 ⇒ 00:16:22.120 Awaish Kumar: Ex… access to, like, the…
161 00:16:22.230 ⇒ 00:16:30.420 Awaish Kumar: the Google Workspace, right? We can ask them to give us a service account key, which has access to
162 00:16:30.710 ⇒ 00:16:35.960 Awaish Kumar: Everything in Google Workspace, and just use that service count key, I think we can…
163 00:16:36.220 ⇒ 00:16:46.800 Awaish Kumar: We can access everything, so… Okay. How you are going to integrate it, I’m not sure if you want… if you need to move that… like, I don’t think we need to move it, right? You…
164 00:16:47.520 ⇒ 00:16:50.770 Samuel Roberts: You just need access, you think, and then we can use the CLI with the service account.
165 00:16:52.120 ⇒ 00:17:03.160 Awaish Kumar: you can use, obviously, CLI with that service account key to connect to Google Workspace, and then it will have access to everything the key, like, that account gets access to.
166 00:17:03.320 ⇒ 00:17:10.339 Awaish Kumar: And that could be, like, more than what Danny has, or… but they just will have to create a key which has a more broader…
167 00:17:11.990 ⇒ 00:17:12.640 Samuel Roberts: Okay.
168 00:17:13.079 ⇒ 00:17:14.169 Awaish Kumar: Yeah, no.
169 00:17:14.170 ⇒ 00:17:16.250 Samuel Roberts: You’re not thinking we need to move that anywhere for this?
170 00:17:16.929 ⇒ 00:17:20.579 Awaish Kumar: I don’t think we should move, because, like, there will be a lot of documents.
171 00:17:22.109 ⇒ 00:17:23.139 Awaish Kumar: Those things, it’s…
172 00:17:23.999 ⇒ 00:17:30.329 Awaish Kumar: if everything… if it can be integrated with Gemini, I think that should be the best way to do that, otherwise… Okay.
173 00:17:30.429 ⇒ 00:17:37.699 Awaish Kumar: we have to think about the… all the injection process, so moving all the documents from Drive to maybe
174 00:17:38.679 ⇒ 00:17:42.079 Awaish Kumar: Like, at wherever you need, cloud storage, or superbase, whatever.
175 00:17:46.020 ⇒ 00:17:46.680 Samuel Roberts: Okay.
176 00:17:48.590 ⇒ 00:17:49.200 Pranav Narahari: Yeah, so…
177 00:17:49.200 ⇒ 00:17:49.580 Samuel Roberts: more.
178 00:17:49.580 ⇒ 00:18:02.740 Pranav Narahari: What I would think about here is, how can we… the milestone that I really want to show them, and I believe we can do this easiest with their own Gemini, by creating, like, some custom integration.
179 00:18:02.870 ⇒ 00:18:12.100 Pranav Narahari: is just, they’re able to now query data from across the organization that exists in Google Workspace and Slack.
180 00:18:12.290 ⇒ 00:18:18.809 Pranav Narahari: So, if it becomes too difficult to do that within Gemini itself.
181 00:18:18.940 ⇒ 00:18:23.689 Pranav Narahari: Then we can maybe talk about how we can put it into something else.
182 00:18:24.410 ⇒ 00:18:27.520 Pranav Narahari: And then… so that’s one thing.
183 00:18:27.520 ⇒ 00:18:30.430 Samuel Roberts: If we need to put it in… oh, yeah, sorry, go ahead and finish the next slide.
184 00:18:30.730 ⇒ 00:18:32.259 Pranav Narahari: Well, you can…
185 00:18:32.720 ⇒ 00:18:38.570 Samuel Roberts: Yeah, I was just gonna say, so I think… I think that if it’s… so… the access… I guess…
186 00:18:38.800 ⇒ 00:18:46.619 Samuel Roberts: I’m wondering… with that service account, potentially, Is it worth…
187 00:18:46.940 ⇒ 00:18:50.699 Samuel Roberts: Getting, like, the metadata of actions or whatever out.
188 00:18:51.120 ⇒ 00:18:56.690 Samuel Roberts: Or… Like, obviously, like, not storing all the documents and everything, but, like.
189 00:18:57.220 ⇒ 00:19:01.279 Samuel Roberts: Or is it just worth trying to see if the CLI has the right
190 00:19:01.730 ⇒ 00:19:08.079 Samuel Roberts: way to access it. You know, it’s the difference between, like, the MCPs for Lilo versus the ingestion, right?
191 00:19:09.530 ⇒ 00:19:10.980 Samuel Roberts: I’m not sure…
192 00:19:11.420 ⇒ 00:19:16.010 Samuel Roberts: Like, at the end of the day, I feel like for these reports, we’re gonna want to ingest something.
193 00:19:16.150 ⇒ 00:19:16.980 Samuel Roberts: Right?
194 00:19:18.130 ⇒ 00:19:19.239 Samuel Roberts: and store it.
195 00:19:19.240 ⇒ 00:19:24.850 Pranav Narahari: Yeah, I guess that’s… That’s where I’m…
196 00:19:24.850 ⇒ 00:19:25.600 Samuel Roberts: You know what I mean? Like…
197 00:19:25.870 ⇒ 00:19:37.010 Pranav Narahari: Like, I’m kind of deferring to you guys to, like, make that decision. Does it make more sense to use the CLI, or does it make more sense to, like, yeah, have, like, a data warehouse, like.
198 00:19:37.140 ⇒ 00:19:38.090 Pranav Narahari: And just…
199 00:19:38.090 ⇒ 00:19:38.450 Samuel Roberts: Okay.
200 00:19:38.450 ⇒ 00:19:48.339 Pranav Narahari: like, probably, like, set a BQ and do this. Okay. Yeah, at the end of the day, I think we understand the two, like, the three of us, just, like.
201 00:19:48.720 ⇒ 00:19:51.350 Pranav Narahari: what the outcome needs to be.
202 00:19:51.350 ⇒ 00:19:55.100 Samuel Roberts: That’s what I’m trying to just make sure, like, I have the full picture here, because… so…
203 00:19:56.190 ⇒ 00:20:01.179 Samuel Roberts: these milestones… and again, I think this one, like, we just kind of made these milestones. These are not even necessarily the, like.
204 00:20:01.550 ⇒ 00:20:03.249 Samuel Roberts: promised things, right?
205 00:20:05.850 ⇒ 00:20:19.180 Samuel Roberts: if we’re gonna, like… I’m just trying to, like, I’m trying to build towards the end goal, and, like, what’s along the way. I don’t just want to, like, create waypoints that it’s like, oh, we ingested it so that it would satisfy milestone 2, but really it doesn’t matter in the end, you know what I mean?
206 00:20:19.540 ⇒ 00:20:25.850 Pranav Narahari: Okay, yeah, yeah, I’m happy to change the milestones up, and so maybe let me share my screen, and, like, let’s just go through the…
207 00:20:26.330 ⇒ 00:20:28.590 Pranav Narahari: Maybe that might help a little bit, too.
208 00:20:29.800 ⇒ 00:20:32.329 Samuel Roberts: Okay, yeah. Cause I’m just… I just, I think…
209 00:20:32.870 ⇒ 00:20:34.589 Samuel Roberts: The question about, like, is it…
210 00:20:35.270 ⇒ 00:20:50.259 Samuel Roberts: important that he is in Gemini, or is it important that we just ingest this data? Because, like, I don’t know if Gemini has… Gemini obviously can have access to the Google Workspace, but I don’t know how we give his Gemini, potentially, the service account access that Oasis is talking about.
211 00:20:53.620 ⇒ 00:20:54.560 Pranav Narahari: Gotcha.
212 00:20:54.820 ⇒ 00:20:58.199 Samuel Roberts: So I think, I think… I’m just trying to make sure, like, if that… yeah.
213 00:20:58.820 ⇒ 00:21:03.299 Pranav Narahari: I guess there just needs to be, like, a little bit of research here, then, to, like, figure out, like…
214 00:21:03.300 ⇒ 00:21:03.800 Samuel Roberts: Definitely.
215 00:21:03.800 ⇒ 00:21:07.240 Pranav Narahari: Yeah, cause I, like, I don’t have that answer either.
216 00:21:07.240 ⇒ 00:21:13.870 Samuel Roberts: Yeah, well, because, like, I think the difference here is, like, ChatGPT versus, like, a cursor agent running on your computer, right?
217 00:21:14.080 ⇒ 00:21:14.550 Pranav Narahari: Right.
218 00:21:14.550 ⇒ 00:21:18.490 Samuel Roberts: Like, using chat in the browser, It can’t execute.
219 00:21:18.630 ⇒ 00:21:19.970 Samuel Roberts: CLI commands.
220 00:21:20.190 ⇒ 00:21:24.899 Samuel Roberts: But we could easily set up a… Virtual system that…
221 00:21:25.010 ⇒ 00:21:32.540 Samuel Roberts: we could give an agent access to that CLI, but, like, I don’t think Gemini necessarily would just have that on, like, the Gemini.google or whatever.
222 00:21:33.450 ⇒ 00:21:37.610 Samuel Roberts: So, like, if it has to be… in Gemini.
223 00:21:38.250 ⇒ 00:21:40.050 Pranav Narahari: No, I don’t think it needs to be Gemini.
224 00:21:40.050 ⇒ 00:21:48.469 Samuel Roberts: That’s what I’m saying, I’m just trying to make sure I understand, because I think, like, Milestone 1 is, like, integrated into Gemini, I just wasn’t sure how critical that is, or if it’s, like, we just have a separate
225 00:21:48.850 ⇒ 00:21:51.269 Samuel Roberts: Way to access it and show that, you know?
226 00:21:51.630 ⇒ 00:21:56.339 Pranav Narahari: Let’s think about it like this, they just need the functionality.
227 00:21:56.340 ⇒ 00:21:57.280 Samuel Roberts: Okay.
228 00:21:58.610 ⇒ 00:22:10.349 Pranav Narahari: And so I kind of came up with these milestones, and this is kind of like a back and forth that I’ve had with, like, Utam and B, where it’s like, I am coming up with milestones before I fully understand the technical implementation, and so there’s going to need.
229 00:22:10.350 ⇒ 00:22:11.229 Samuel Roberts: No, totally, totally.
230 00:22:11.230 ⇒ 00:22:14.220 Pranav Narahari: just need to change, but I think it’s,
231 00:22:14.670 ⇒ 00:22:30.970 Pranav Narahari: it’s just like, okay, I have, like, a broad understanding with my technical background of, like, what the milestones can be. However, I think with you two, and especially, like, away here, too, with, like, the understanding of, like, the data, and then Sam, with you, like, the understanding of, like, AI,
232 00:22:31.460 ⇒ 00:22:39.740 Pranav Narahari: we can really think about the optimal technical approach, and then, yeah, I’m happy to change up the milestones to reflect that.
233 00:22:40.470 ⇒ 00:22:43.750 Samuel Roberts: That’s fine. I just… I wanted… I was trying to really understand the, like.
234 00:22:44.720 ⇒ 00:22:50.019 Samuel Roberts: the final deliverable. If, like, milestones are fungible, like, the end goal is really just the…
235 00:22:50.510 ⇒ 00:22:53.569 Samuel Roberts: The data to generate the reports, right?
236 00:22:53.570 ⇒ 00:22:54.640 Pranav Narahari: Yep. Yep.
237 00:22:54.640 ⇒ 00:23:02.690 Samuel Roberts: Okay, that’s where I was just a little bit, like, I didn’t know what was critical along the way, and it seemed like not much is critical along the way, as long as…
238 00:23:02.850 ⇒ 00:23:07.250 Samuel Roberts: we can… Get the data, generate the reports.
239 00:23:07.530 ⇒ 00:23:08.360 Samuel Roberts: In the end.
240 00:23:08.720 ⇒ 00:23:10.980 Awaish Kumar: I haven’t eaten.
241 00:23:11.200 ⇒ 00:23:16.220 Awaish Kumar: Whatever you will be, like, using,
242 00:23:18.320 ⇒ 00:23:21.700 Awaish Kumar: Like, can Gemini connect to Google Cloud storage?
243 00:23:25.330 ⇒ 00:23:29.079 Samuel Roberts: Just as that, like, Gemini Google, I don’t… I don’t know offhand.
244 00:23:31.080 ⇒ 00:23:33.000 Awaish Kumar: Like, how, like.
245 00:23:33.000 ⇒ 00:23:37.550 Samuel Roberts: I’m sure… I’m sure it can, but I don’t know how it’s scoped, which is part of the problem here, like…
246 00:23:38.340 ⇒ 00:23:41.979 Awaish Kumar: Like, how will COVID be feeding data?
247 00:23:42.110 ⇒ 00:23:50.469 Awaish Kumar: for example, from… if we talk about Slack, if I have a script which runs and gets all the data from all the channels.
248 00:23:50.640 ⇒ 00:23:52.660 Awaish Kumar: That exposed it to…
249 00:23:53.210 ⇒ 00:24:00.830 Awaish Kumar: some database or a file system, whatever, then how you are going to connect that with Gemini, do you know that?
250 00:24:02.900 ⇒ 00:24:08.100 Samuel Roberts: No, and that’s what I’m kind of getting at. I’m not sure if it’s important that Gemini even has that or not.
251 00:24:08.100 ⇒ 00:24:14.920 Awaish Kumar: I think we need to go into research more then, like, do… Yeah. …understanding on… how, basically.
252 00:24:15.610 ⇒ 00:24:20.519 Awaish Kumar: the final Gemini will be reading it, right? If it will be reading it
253 00:24:20.640 ⇒ 00:24:29.820 Awaish Kumar: From storage, if we’ll be reading it from maybe… directly from Google, I think the… Right, right.
254 00:24:30.300 ⇒ 00:24:36.989 Samuel Roberts: Right, it might… it might have access built in for itself, but probably not for Slack, and so, okay, I see what you’re saying there.
255 00:24:37.170 ⇒ 00:24:41.399 Awaish Kumar: Because, like… It’s not like, I just… I just made those.
256 00:24:41.400 ⇒ 00:24:43.619 Samuel Roberts: I haven’t done anything with this print off yet, I just created.
257 00:24:43.620 ⇒ 00:24:46.039 Awaish Kumar: Oh, go ahead.
258 00:24:46.270 ⇒ 00:24:49.150 Awaish Kumar: When you do a direct integration, Hmm.
259 00:24:50.010 ⇒ 00:24:59.849 Awaish Kumar: it will use your user, it can’t be a super user, right? So, I just looked it up. It is possible to create a service count and…
260 00:25:00.230 ⇒ 00:25:10.100 Awaish Kumar: basically impersonate people, so who has the most access? So our server is count can access all the files, for example, CEO can access.
261 00:25:10.290 ⇒ 00:25:17.009 Awaish Kumar: And then… We use that service count key to basically feed data to Gemini.
262 00:25:18.950 ⇒ 00:25:19.520 Samuel Roberts: Okay.
263 00:25:19.520 ⇒ 00:25:20.940 Awaish Kumar: That, that is for Google.
264 00:25:21.060 ⇒ 00:25:22.360 Awaish Kumar: Workspace for it.
265 00:25:23.510 ⇒ 00:25:33.080 Awaish Kumar: For other one, we can do similar, like, we can… we have to run a pipeline which brings data from Slack, stores it into storage, and then use same service count key to read for…
266 00:25:34.720 ⇒ 00:25:35.739 Awaish Kumar: Let me know.
267 00:25:36.650 ⇒ 00:25:39.789 Samuel Roberts: Right, so either way, Slack’s gonna have to get stored, I think, no matter what.
268 00:25:40.260 ⇒ 00:25:43.760 Awaish Kumar: Slack is going to be stored, and… yeah.
269 00:25:43.940 ⇒ 00:25:47.950 Awaish Kumar: And maybe if we stored it on Google Cloud Storage, that will be…
270 00:25:48.670 ⇒ 00:25:55.590 Samuel Roberts: Right, that might simplify… yeah, totally. Okay. That makes some sense. All right, let me… let me maybe do a little bit of digging into how Gemini…
271 00:25:56.080 ⇒ 00:25:59.859 Samuel Roberts: what, like, what tools it can have access to, like, is MCP the right thing?
272 00:26:01.410 ⇒ 00:26:10.020 Samuel Roberts: But either way, we probably need to set up a pipeline for the Slack data, which probably means a Slack app has to get installed, right?
273 00:26:11.390 ⇒ 00:26:11.920 Samuel Roberts: Okay.
274 00:26:12.640 ⇒ 00:26:13.200 Awaish Kumar: That is…
275 00:26:13.200 ⇒ 00:26:13.880 Samuel Roberts: Okay.
276 00:26:14.210 ⇒ 00:26:15.690 Awaish Kumar: We already have to do that.
277 00:26:16.060 ⇒ 00:26:18.310 Samuel Roberts: Yeah, yeah, that makes sense. Okay.
278 00:26:20.690 ⇒ 00:26:30.599 Samuel Roberts: So then, we may need to take a little, like, a break and reconvene, or just async this. I will… now that I understand a little bit more… so, like.
279 00:26:32.350 ⇒ 00:26:35.849 Awaish Kumar: I think, Sam, we can start, like, a Notion doc.
280 00:26:36.610 ⇒ 00:26:37.490 Awaish Kumar: Okay.
281 00:26:37.600 ⇒ 00:26:44.139 Awaish Kumar: And, and, like, we all fill in whatever our findings are, with our name.
282 00:26:44.550 ⇒ 00:26:45.780 Awaish Kumar: Okay, yeah, let’s do that.
283 00:26:45.780 ⇒ 00:26:46.320 Samuel Roberts: that.
284 00:26:46.320 ⇒ 00:26:50.159 Awaish Kumar: And then we can… based on that, we can make a decision, okay?
285 00:26:50.160 ⇒ 00:26:51.690 Samuel Roberts: That sounds good. Yeah, Pranav, can you…
286 00:26:51.690 ⇒ 00:26:52.160 Pranav Narahari: Oh, brother.
287 00:26:52.160 ⇒ 00:26:52.889 Samuel Roberts: Down a little bit.
288 00:26:53.360 ⇒ 00:26:56.000 Pranav Narahari: Yeah, I just want to add one thing, too, so…
289 00:26:56.000 ⇒ 00:26:56.510 Samuel Roberts: Yeah.
290 00:26:56.840 ⇒ 00:27:03.869 Pranav Narahari: I know I said, like, no MCP server, and that’s true. I think, though, with the reports that may…
291 00:27:04.630 ⇒ 00:27:10.410 Pranav Narahari: We might take a certain approach with reports, the…
292 00:27:10.680 ⇒ 00:27:19.850 Pranav Narahari: I also want the technical implementation, if possible, and we can discuss, like, what is possible in terms of the technical approach, but…
293 00:27:20.650 ⇒ 00:27:23.469 Pranav Narahari: If you’re able to query the data
294 00:27:23.820 ⇒ 00:27:31.399 Pranav Narahari: and get some type of… like, within a chat interface as well, which I know is a different type of…
295 00:27:31.770 ⇒ 00:27:45.759 Pranav Narahari: technical approach, because that’s the whole thing that happened with Lilo, right? It was like, we decided we wanted to create a data warehouse for forecasting, but with that same data, it’s not gonna be the most… it’s not gonna be live data, per se, it could be a day old.
296 00:27:46.890 ⇒ 00:27:47.510 Samuel Roberts: Sure.
297 00:27:48.800 ⇒ 00:27:56.199 Pranav Narahari: That’s something that I want to consider as well, and then we can talk about, if we want to do that, what are…
298 00:27:57.200 ⇒ 00:28:03.700 Pranav Narahari: what are the limitations? Like, I think it’s an easy… it’s easy for me to disagree to, like…
299 00:28:03.850 ⇒ 00:28:16.970 Pranav Narahari: Yeah, we could do that, but the data that it’s going to be querying from is going to be less than… it could be up to 24 hours still. If that’s the case, like, I think that’s okay, considering what they’re trying to do with this.
300 00:28:16.970 ⇒ 00:28:17.650 Samuel Roberts: Yeah, yeah.
301 00:28:17.650 ⇒ 00:28:27.340 Pranav Narahari: I’m sure things might change where they need to be querying live data, but then that’ll be a different scope of work. So I’m not too worried about, like, us being able to hit that target, but…
302 00:28:27.420 ⇒ 00:28:41.449 Pranav Narahari: Yeah, so those are the two things. I think we should be able to assess, like, overall themes with all of this data, and then the second thing is, if we want to query the data, like, in a chat interface, we should be able to do that as well.
303 00:28:43.310 ⇒ 00:28:43.960 Awaish Kumar: Okay.
304 00:28:45.070 ⇒ 00:28:47.149 Samuel Roberts: Okay, cool. Yeah, I think…
305 00:28:47.580 ⇒ 00:28:51.989 Samuel Roberts: Yeah, so we should probably do a little… little spike each on some of this, then, so…
306 00:28:51.990 ⇒ 00:28:52.460 Pranav Narahari: Yeah.
307 00:28:52.780 ⇒ 00:28:55.419 Pranav Narahari: And what we can do, too, is,
308 00:28:56.400 ⇒ 00:29:09.869 Pranav Narahari: think about it as if, like, if this, like, say, you know, with the service account, we’re able to get, like, the access to the entire Eden organization’s data, then we’re able to do X, Y, and Z. But it’s…
309 00:29:10.230 ⇒ 00:29:29.650 Pranav Narahari: It’s, dependent on the service account being able to do this, and then let me, or I can loop one of you guys in to talk with our sponsor at Eden to figure out whether that’s the case or not. And, yeah, I have a meeting actually scheduled today with, Adam at Eden, so…
310 00:29:29.810 ⇒ 00:29:34.159 Pranav Narahari: He should be able to figure all that, or at least get us the things that we need to do that.
311 00:29:34.160 ⇒ 00:29:34.560 Samuel Roberts: Alright.
312 00:29:34.560 ⇒ 00:29:36.160 Pranav Narahari: research. Cool.
313 00:29:36.400 ⇒ 00:29:39.249 Samuel Roberts: I was gonna say, I don’t know who Adam is in the…
314 00:29:39.250 ⇒ 00:29:44.720 Pranav Narahari: Yeah, yeah, I don’t think… maybe, Oasia, you’ve met him, but he’s, he’s just, like, our of contact.
315 00:29:45.880 ⇒ 00:29:46.570 Samuel Roberts: Okay.
316 00:29:46.780 ⇒ 00:29:47.360 Pranav Narahari: Yeah.
317 00:29:47.900 ⇒ 00:29:49.990 Samuel Roberts: Okay, what time is that?
318 00:29:50.260 ⇒ 00:29:54.550 Samuel Roberts: meeting, just in case I need, like, to get anything together for you to ask him, kind of stuff.
319 00:29:55.400 ⇒ 00:29:57.190 Pranav Narahari: Yes, let me check…
320 00:30:00.980 ⇒ 00:30:04.449 Pranav Narahari: It’s at, 2.30 EST.
321 00:30:05.450 ⇒ 00:30:10.410 Samuel Roberts: Okay, I’m just gonna put that on my calendar so we have a… 2…
322 00:30:13.830 ⇒ 00:30:14.730 Samuel Roberts: Okay.
323 00:30:15.210 ⇒ 00:30:16.430 Samuel Roberts: Okay.
324 00:30:19.880 ⇒ 00:30:28.640 Samuel Roberts: Alright, so I would say let’s make a, like, another doc off of this notion, I guess, down probably in the technical approach. I’ll just do like this. How do I…
325 00:30:29.300 ⇒ 00:30:31.850 Samuel Roberts: Have. New page, okay.
326 00:30:32.460 ⇒ 00:30:46.180 Samuel Roberts: Okay, cool. So let’s just work in that file, Aish. Just collect some stuff. I’m… I’m thinking I will dig into the Gemini, service account.
327 00:30:47.000 ⇒ 00:30:52.179 Samuel Roberts: the Google Workspace stuff. What else do we need to know? Let me make a quick listen. So, Gemini…
328 00:30:52.540 ⇒ 00:30:54.350 Awaish Kumar: Yeah, I think that’s an important section.
329 00:30:54.540 ⇒ 00:30:57.569 Samuel Roberts: Gemini Access Workspace Access…
330 00:30:58.700 ⇒ 00:31:06.870 Samuel Roberts: service account, and then, but Slack, we’re pretty sure it’s just, we’re gonna have to ingest that into a pipe… it’s gonna be the same thing as our platform, effectively, right, Awish?
331 00:31:07.060 ⇒ 00:31:12.829 Awaish Kumar: Like, we already have… we are already using Chromic for Eden, we can just use that for Slack.
332 00:31:13.490 ⇒ 00:31:14.090 Samuel Roberts: Okay.
333 00:31:14.630 ⇒ 00:31:15.410 Pranav Narahari: Nice.
334 00:31:16.270 ⇒ 00:31:24.229 Samuel Roberts: polyatomic, to, google Storage somewhere. So, like, BigQuery or something, whatever we need for that, probably.
335 00:31:24.620 ⇒ 00:31:27.340 Awaish Kumar: Yeah, if there’s no direct integration, then yes.
336 00:31:27.340 ⇒ 00:31:28.530 Samuel Roberts: Yeah. Okay.
337 00:31:28.630 ⇒ 00:31:34.819 Samuel Roberts: Alright, I will do a little bit of that. I’ll probably ping one or both of you at different times, as I start to…
338 00:31:35.560 ⇒ 00:31:36.790 Samuel Roberts: I don’t need, sales.
339 00:31:36.790 ⇒ 00:31:37.360 Awaish Kumar: Yeah.
340 00:31:37.650 ⇒ 00:31:44.069 Pranav Narahari: So, yeah, feel free to pull me into anything, or if you need additional context, or just, like.
341 00:31:44.270 ⇒ 00:31:56.400 Pranav Narahari: bounce ideas off of, like, I’m around. I just have one interview at 2 ESTs, so right before my call with, the Eden, like, point of contact, so… until then, I’m chillin’.
342 00:31:56.540 ⇒ 00:31:57.200 Pranav Narahari: Yeah.
343 00:31:57.200 ⇒ 00:31:58.890 Samuel Roberts: Alright, cool, yeah, I think,
344 00:32:00.580 ⇒ 00:32:05.229 Samuel Roberts: Yeah, I mean, unless you want to dig into any of this stuff too, it might help speed things up a little bit.
345 00:32:06.320 ⇒ 00:32:12.750 Pranav Narahari: Yeah, let me know, I’m still kind of, doing some of the ABC stuff, so we have all those deliverables to, like…
346 00:32:12.750 ⇒ 00:32:18.550 Samuel Roberts: Yeah, oh, we have a meeting after this. Yeah, no, you’re right, and I… I was… I… yeah, there’s other stuff I had to figure out there, too. Okay.
347 00:32:18.890 ⇒ 00:32:22.520 Samuel Roberts: Still waiting to see if Tim has enabled billing yet.
348 00:32:22.750 ⇒ 00:32:28.720 Samuel Roberts: Okay. All right, let me dig a little bit of this. I’ll be talking to you in a little bit anyway, on that meeting, so…
349 00:32:29.490 ⇒ 00:32:30.160 Pranav Narahari: Perfect.
350 00:32:30.390 ⇒ 00:32:33.700 Samuel Roberts: I’ll see what I got till then. Cool, alright. Thank you guys.
351 00:32:34.750 ⇒ 00:32:35.500 Pranav Narahari: See you guys.
352 00:32:35.720 ⇒ 00:32:37.220 Samuel Roberts: Alright. Bye.