Meeting Title: AI Team Weekly Planning Date: 2025-03-03 Meeting participants: Janna Wong, Uttam Kumaran, Miguel De Veyra, Casie Aviles
WEBVTT
1 00:04:57.100 ⇒ 00:04:57.800 Miguel de Veyra: Good morning!
2 00:05:04.510 ⇒ 00:05:05.100 Casie Aviles: Ping.
3 00:09:19.210 ⇒ 00:09:20.323 Miguel de Veyra: Hey, teams?
4 00:09:24.590 ⇒ 00:09:27.539 Miguel de Veyra: He said. This is gonna be running a few minutes later.
5 00:09:30.010 ⇒ 00:09:33.326 Miguel de Veyra: Nothing’s offline now he’s back.
6 00:09:37.970 ⇒ 00:09:40.619 Miguel de Veyra: Oh, he’s offline again. Oh, there you go!
7 00:09:41.470 ⇒ 00:09:42.280 Miguel de Veyra: Hey, Tim.
8 00:09:42.530 ⇒ 00:09:46.569 Uttam Kumaran: Hey, guys, hey, Tim? How’s the weekend?
9 00:09:49.340 ⇒ 00:09:50.510 Miguel de Veyra: It was nice.
10 00:09:51.060 ⇒ 00:09:52.373 Miguel de Veyra: More figurines.
11 00:09:55.500 ⇒ 00:09:56.940 Uttam Kumaran: How many more?
12 00:09:57.840 ⇒ 00:10:01.870 Miguel de Veyra: I got like 1, 2, 3, 4, 5 more, 6.
13 00:10:02.260 ⇒ 00:10:02.760 Uttam Kumaran: Well.
14 00:10:02.760 ⇒ 00:10:04.470 Miguel de Veyra: 5, 5 more. Yeah.
15 00:10:08.420 ⇒ 00:10:10.439 Miguel de Veyra: And yeah, pretty fun.
16 00:10:11.130 ⇒ 00:10:13.540 Uttam Kumaran: Jana Casey, how about you guys? What do you guys do.
17 00:10:15.790 ⇒ 00:10:16.360 Casie Aviles: And.
18 00:10:16.360 ⇒ 00:10:17.500 Janna Wong: Much fun, man.
19 00:10:18.910 ⇒ 00:10:23.710 Casie Aviles: Yeah, not much. Just caught up watching some movies.
20 00:10:25.130 ⇒ 00:10:26.460 Uttam Kumaran: Nice what movies.
21 00:10:28.866 ⇒ 00:10:38.670 Casie Aviles: It’s just, you know, it’s just the one, the Warhammer one that well, it’s not really a movie. I guess, that Miguel recommended. So I think it was secret level.
22 00:10:38.670 ⇒ 00:10:39.690 Casie Aviles: Oh, yeah.
23 00:10:40.530 ⇒ 00:10:42.539 Uttam Kumaran: I thought there was a Warhammer movie.
24 00:10:43.400 ⇒ 00:10:44.000 Miguel de Veyra: No, no.
25 00:10:44.000 ⇒ 00:10:46.120 Uttam Kumaran: There’s a there’s a game for sure.
26 00:10:47.510 ⇒ 00:10:52.609 Miguel de Veyra: I mean, there’s a lot that the last one was like 2010.
27 00:10:52.840 ⇒ 00:10:57.889 Miguel de Veyra: They’re working on something where 10 of the cavill isn’t like main actor.
28 00:10:59.100 ⇒ 00:11:02.419 Uttam Kumaran: Oh, no way. Okay. So it’s so it’s a big movie.
29 00:11:02.580 ⇒ 00:11:04.070 Miguel de Veyra: Yeah, yeah, it’s with Amazon.
30 00:11:04.890 ⇒ 00:11:05.790 Uttam Kumaran: Oh! Damn!
31 00:11:08.570 ⇒ 00:11:14.129 Miguel de Veyra: But yeah, were you able to work during the weekend with them, or the girlfriend didn’t allow you.
32 00:11:14.820 ⇒ 00:11:17.059 Uttam Kumaran: A little bit. Why.
33 00:11:17.543 ⇒ 00:11:22.689 Miguel de Veyra: I saw you online a bit. But then I was like my girl, I was gonna message. But then my girlfriend was like.
34 00:11:22.800 ⇒ 00:11:23.609 Miguel de Veyra: you have no problem.
35 00:11:25.630 ⇒ 00:11:26.070 Miguel de Veyra: Ain’t gonna.
36 00:11:26.070 ⇒ 00:11:30.969 Uttam Kumaran: Yeah, you could always. You could always message me. The company is my 1st love. So
37 00:11:32.539 ⇒ 00:11:37.570 Uttam Kumaran: I met the company before I met my girlfriend, so no.
38 00:11:37.570 ⇒ 00:11:38.820 Miguel de Veyra: You guys are fairly new.
39 00:11:38.820 ⇒ 00:11:39.680 Uttam Kumaran: Huh!
40 00:11:39.680 ⇒ 00:11:40.909 Miguel de Veyra: You guys are fairly new.
41 00:11:40.910 ⇒ 00:11:43.639 Uttam Kumaran: Like 2 years or a year and a half.
42 00:11:44.284 ⇒ 00:11:45.700 Miguel de Veyra: Okay. Okay. I see.
43 00:11:46.664 ⇒ 00:11:52.049 Uttam Kumaran: Yeah, if I come to the Philippines I’ll I’ll ask if she wants to join up, and we can come. See? You guys.
44 00:11:53.872 ⇒ 00:11:56.529 Casie Aviles: Oh, are you coming this year or not?
45 00:11:56.530 ⇒ 00:11:59.339 Uttam Kumaran: I wanna try to come later this year if I can.
46 00:11:59.650 ⇒ 00:12:00.700 Miguel de Veyra: Oh, okay. Okay.
47 00:12:01.927 ⇒ 00:12:03.480 Uttam Kumaran: But let’s see.
48 00:12:04.510 ⇒ 00:12:10.850 Uttam Kumaran: Yeah. I was working Saturday. Got a bunch done and a little bit yesterday we have our new
49 00:12:11.411 ⇒ 00:12:18.020 Uttam Kumaran: pm, I’m sort of starting his trial today, so I’ll introduce him during the
50 00:12:18.470 ⇒ 00:12:22.760 Uttam Kumaran: all company meeting later. And then.
51 00:12:22.760 ⇒ 00:12:24.770 Miguel de Veyra: I had started with the Am.
52 00:12:25.480 ⇒ 00:12:29.839 Uttam Kumaran: Yeah, it’ll be short. I just wanna introduce him. And then
53 00:12:29.970 ⇒ 00:12:32.969 Uttam Kumaran: just talk a little bit about changes this week. But
54 00:12:33.470 ⇒ 00:12:36.068 Uttam Kumaran: yeah, so that’s going well. And then,
55 00:12:36.880 ⇒ 00:12:44.805 Uttam Kumaran: yeah, we just had a bunch of work last week on data. We’re on data clients. And there’s just, of course, a lot of other stuff. So but yeah, I
56 00:12:45.400 ⇒ 00:12:54.350 Uttam Kumaran: you know, we’re coming up on what was supposed to be our last week for ABC. However, last week of like the 1st phase.
57 00:12:54.470 ⇒ 00:13:09.940 Uttam Kumaran: so kind of like the 2 things. One, the positive thing is like, I feel really proud of the progress we’ve made. I think that there’s we’re not 100% finished with all the things that we set out to do. But there was also a lot of new things that we discovered in the process.
58 00:13:10.488 ⇒ 00:13:18.119 Uttam Kumaran: Which, of course, you know, we have to just sort of roll with the punches. I think I think we’ll do a more formal.
59 00:13:18.420 ⇒ 00:13:23.775 Uttam Kumaran: I know we haven’t been doing like a retro every week. It’s really just been because of.
60 00:13:25.100 ⇒ 00:13:34.379 Uttam Kumaran: it’s really just been as a function of my time. But when we do a retro, I think the biggest thing that I’ll think about is like, Yeah, I feel like we
61 00:13:34.690 ⇒ 00:13:37.310 Uttam Kumaran: it. We probably needed like a month and a half, I think
62 00:13:37.875 ⇒ 00:13:45.559 Uttam Kumaran: to get everything done. I think we did a really good job with the 1st agent, but as I sent some messages today, there’s still a couple more things.
63 00:13:46.086 ⇒ 00:13:56.229 Uttam Kumaran: But yeah, I guess we can, you know. Is there any? I guess Miguel or Jana, do you guys have any other sort of feedback? And then then we can probably just talk about stuff for this week.
64 00:13:58.830 ⇒ 00:14:01.139 Miguel de Veyra: Yeah, I think
65 00:14:02.610 ⇒ 00:14:10.100 Miguel de Veyra: there should. Definitely. I mean, yeah, majority of the blockers honestly, was their documents. Even now it’s not done right?
66 00:14:10.930 ⇒ 00:14:12.317 Miguel de Veyra: So I think,
67 00:14:13.360 ⇒ 00:14:27.099 Miguel de Veyra: the next clients we have, especially if it’s like similar or the same, almost similar we should have. We did this before Utam with the Polish Company, where Casey and I met them. We allocated. I think it was a week. No.
68 00:14:28.110 ⇒ 00:14:31.419 Miguel de Veyra: Casey, just for Information guide collection.
69 00:14:32.450 ⇒ 00:14:33.300 Casie Aviles: Yeah, yeah.
70 00:14:33.530 ⇒ 00:14:35.670 Miguel de Veyra: Yeah, they just give us, yeah.
71 00:14:35.670 ⇒ 00:14:42.010 Casie Aviles: They they changed their onboarding process so they could gather the in the needed info.
72 00:14:43.390 ⇒ 00:14:47.589 Uttam Kumaran: Okay, cool. Yeah. I I agree, I think, that really held us back. I mean.
73 00:14:48.130 ⇒ 00:14:54.970 Uttam Kumaran: what’s our first, st their 1st client, big client? There, I think we’ll get better for sure. Okay, I hear you guys.
74 00:14:55.900 ⇒ 00:14:57.860 Miguel de Veyra: Yeah. And then
75 00:14:58.290 ⇒ 00:15:05.799 Miguel de Veyra: for feedback. Yeah, I think that’s pretty much it. Context was godsend. To be honest, it’s, you know, a lifesaver
76 00:15:06.160 ⇒ 00:15:08.200 Miguel de Veyra: made everything a lot easier.
77 00:15:08.420 ⇒ 00:15:09.150 Uttam Kumaran: Yeah.
78 00:15:09.720 ⇒ 00:15:16.030 Miguel de Veyra: But yeah, this week me and Casey were gonna be testing out basically rag versus context.
79 00:15:16.290 ⇒ 00:15:16.610 Uttam Kumaran: Okay.
80 00:15:16.610 ⇒ 00:15:19.069 Miguel de Veyra: Advanced rag, not just the basic rag.
81 00:15:20.350 ⇒ 00:15:27.850 Miguel de Veyra: And then one of the few things. Oh, shit! Is this still feedback, or can I continue with discovery?
82 00:15:27.850 ⇒ 00:15:29.560 Uttam Kumaran: Yeah, yeah, anything that’s fine.
83 00:15:29.560 ⇒ 00:15:34.270 Miguel de Veyra: Okay, yeah. And then, yeah, I’ve been like twiddling around with the
84 00:15:35.090 ⇒ 00:15:39.519 Miguel de Veyra: prompt adjustments. And then I’m not sure if they sent already the oh, by the way, with them.
85 00:15:40.200 ⇒ 00:15:46.090 Miguel de Veyra: yeah, they did the yeah, it’s 1 of the stuff I’ll be working on.
86 00:15:46.970 ⇒ 00:15:51.492 Uttam Kumaran: So I think one thing that’s helpful is also to look at the
87 00:15:52.100 ⇒ 00:15:56.220 Uttam Kumaran: The list that we send that I sent this morning. I think the couple things one
88 00:15:56.480 ⇒ 00:16:04.289 Uttam Kumaran: we still haven’t done evals, and we still haven’t sort of focused on the response. Times. I assume the rag work is related to the response. Times, right.
89 00:16:04.290 ⇒ 00:16:05.260 Miguel de Veyra: Yes, yes.
90 00:16:05.260 ⇒ 00:16:15.690 Uttam Kumaran: Okay? And then, yeah, I I you know, I I think we just probably need some time spent on the Eval work. How do you think about that, and like
91 00:16:16.466 ⇒ 00:16:18.949 Uttam Kumaran: you know, trying to put some effort towards that.
92 00:16:19.700 ⇒ 00:16:23.460 Miguel de Veyra: Oh, yeah, the that one
93 00:16:24.380 ⇒ 00:16:31.590 Miguel de Veyra: I am not is the evil work, the one where it’s connected directly to the Google stuff. Right? The golden data set.
94 00:16:31.590 ⇒ 00:16:37.699 Uttam Kumaran: Well, basically, we just have to put that into vellum and vellum has an evaluation function like feature.
95 00:16:38.301 ⇒ 00:16:42.230 Miguel de Veyra: Yeah, yeah, I think Casey showed it to me before we just have to basically
96 00:16:43.130 ⇒ 00:16:45.010 Miguel de Veyra: technically migrate the data.
97 00:16:46.131 ⇒ 00:16:56.079 Miguel de Veyra: Yeah, I guess I will work on that today, so we can just get stuff done for the Eva. I feel like it’s been, you know, sideline for the for the longest time.
98 00:16:57.110 ⇒ 00:17:00.381 Uttam Kumaran: Yeah. So that’s the biggest things. Is that
99 00:17:02.380 ⇒ 00:17:05.817 Uttam Kumaran: that’s the biggest thing that I want to make sure that
100 00:17:07.069 ⇒ 00:17:18.387 Uttam Kumaran: we can. Yeah, we can get the evals up. And then I’m gonna be, I’m gonna be. We’re gonna arrive at the next set of proposals later this afternoon in terms of the
101 00:17:19.507 ⇒ 00:17:28.180 Uttam Kumaran: getting paid like on an ongoing basis. But I really do think that if we can work through
102 00:17:28.854 ⇒ 00:17:32.009 Uttam Kumaran: the Evals, that would be very, very helpful.
103 00:17:32.410 ⇒ 00:17:40.489 Uttam Kumaran: So I don’t know, like I guess, Miguel, it’s up to you. I think we should just leverage the whole team for ABC this week and try to push everything through.
104 00:17:43.030 ⇒ 00:17:44.980 Miguel de Veyra: You mean, like all hands on deck.
105 00:17:44.980 ⇒ 00:17:45.820 Uttam Kumaran: Yes.
106 00:17:46.190 ⇒ 00:17:49.120 Miguel de Veyra: Oh, yeah, definitely. I think that would be a good idea.
107 00:17:50.060 ⇒ 00:17:50.770 Uttam Kumaran: Okay.
108 00:17:50.930 ⇒ 00:17:53.979 Miguel de Veyra: So, yeah, we can close the next part of the proposal.
109 00:17:54.380 ⇒ 00:17:55.325 Uttam Kumaran: Okay,
110 00:17:56.990 ⇒ 00:18:03.059 Uttam Kumaran: And then I so I guess I also gave you the the embed link for real.
111 00:18:04.480 ⇒ 00:18:05.220 Miguel de Veyra: Yep.
112 00:18:05.400 ⇒ 00:18:11.580 Miguel de Veyra: I just don’t know where we put it yet. Is it the client Slash Demo cause? He told me we’re not going to use that.
113 00:18:13.310 ⇒ 00:18:27.090 Uttam Kumaran: Yeah, I guess I was thinking about doing in like a password protected area. We could. Also, it’s like, not the highest priority. Honestly. So that’s okay. I think the highest priority is getting them something they can look at in terms of the dashboard.
114 00:18:32.157 ⇒ 00:18:33.320 Uttam Kumaran: so how do.
115 00:18:33.320 ⇒ 00:18:36.120 Miguel de Veyra: I can just iframe it to you.
116 00:18:36.120 ⇒ 00:18:38.489 Uttam Kumaran: But the agent to update information.
117 00:18:40.840 ⇒ 00:18:54.160 Miguel de Veyra: The agent to update. I want to run some tests because I think we’re if we’re all hands on deck. I think we’re gonna finish around Thursday or early as possible is Wednesday. Then I wanna run some tests on how we can update some stuff.
118 00:18:54.760 ⇒ 00:18:55.400 Uttam Kumaran: Okay.
119 00:18:55.932 ⇒ 00:19:01.729 Miguel de Veyra: But yeah, they seem to really like the Bible idea. So I’m not sure if it’s a good idea to split things up
120 00:19:03.690 ⇒ 00:19:05.420 Miguel de Veyra: because we sold them, you know.
121 00:19:05.970 ⇒ 00:19:11.120 Miguel de Veyra: Hey? We should put all information in one stuff so it doesn’t get lost. And then
122 00:19:11.610 ⇒ 00:19:12.519 Miguel de Veyra: what do you think.
123 00:19:15.620 ⇒ 00:19:17.380 Uttam Kumaran: Yeah, that makes sense.
124 00:19:23.140 ⇒ 00:19:28.530 Miguel de Veyra: They really like the section stuff. I mean, I didn’t really clean it as much as I wanted to, because I’m not really the document, Guy.
125 00:19:29.100 ⇒ 00:19:31.909 Miguel de Veyra: but you know I think they loved it.
126 00:19:32.980 ⇒ 00:19:41.209 Uttam Kumaran: The organization. Yeah, I don’t know. I would. I would just ask Dude. I would just ask, chat, gpt, like, what’s the best way to organize something like this.
127 00:19:41.949 ⇒ 00:19:42.999 Miguel de Veyra: Yeah. Yeah. Okay.
128 00:19:44.330 ⇒ 00:19:49.270 Uttam Kumaran: So we want to work on. So let me just note down what those
129 00:19:49.570 ⇒ 00:19:52.177 Uttam Kumaran: what the pieces we’re working on are.
130 00:19:54.330 ⇒ 00:19:59.369 Uttam Kumaran: the. So where are we gonna work on the agent to update information? So that’s part of this
131 00:19:59.680 ⇒ 00:20:01.290 Uttam Kumaran: organization piece.
132 00:20:06.580 ⇒ 00:20:08.720 Miguel de Veyra: Agent update information.
133 00:20:11.860 ⇒ 00:20:17.249 Miguel de Veyra: We want this agent to update information to be the one updating right? Not just basically help.
134 00:20:18.320 ⇒ 00:20:20.400 Uttam Kumaran: No, we want a separate agent for.
135 00:20:20.400 ⇒ 00:20:21.160 Miguel de Veyra: Yeah, yeah.
136 00:20:21.160 ⇒ 00:20:22.600 Uttam Kumaran: Janice basically.
137 00:20:23.070 ⇒ 00:20:32.039 Miguel de Veyra: I mean, do we want it to directly update the document, or just Hey, give us the information, and then it will write something about you know
138 00:20:32.450 ⇒ 00:20:35.539 Miguel de Veyra: about it. And then Janice will just copy paste. Basically.
139 00:20:36.480 ⇒ 00:20:39.539 Uttam Kumaran: I would prefer it to go update the document.
140 00:20:41.190 ⇒ 00:20:45.960 Miguel de Veyra: Now, what what if we do it face by face? We 1st phase by phase, but.
141 00:20:46.310 ⇒ 00:20:49.170 Uttam Kumaran: The problem is still, the bigger problem is still there.
142 00:20:49.610 ⇒ 00:20:50.320 Uttam Kumaran: Yeah, yeah.
143 00:20:50.835 ⇒ 00:20:51.350 Miguel de Veyra: Cause.
144 00:20:51.350 ⇒ 00:20:54.139 Miguel de Veyra: I think we just do the the.
145 00:20:54.140 ⇒ 00:21:00.169 Uttam Kumaran: I don’t want phase, but like I don’t want phase by phase to be like next week, like, I don’t think this is that hard
146 00:21:00.900 ⇒ 00:21:06.160 Uttam Kumaran: like. If you figure out the Google sheets. If you figure out the Google docs like right back.
147 00:21:06.160 ⇒ 00:21:06.750 Miguel de Veyra: See it.
148 00:21:07.970 ⇒ 00:21:08.640 Uttam Kumaran: I don’t know. I don’t.
149 00:21:10.130 ⇒ 00:21:12.040 Miguel de Veyra: Yeah. The right? Okay.
150 00:21:12.040 ⇒ 00:21:28.719 Uttam Kumaran: Basically, there’s a couple of things you could do. One is like, if you can go to a specific spot and update, that’s great. The other thing you could do is basically take a section, rewrite it and put it back. I don’t know. It’s this is a tough problem.
151 00:21:28.780 ⇒ 00:21:45.179 Uttam Kumaran: The other thing, we, the other. The other solution is to move everything to like super base. And then you can make the update in super base and then write everything back to Google docs with a python script. So there’s no, there’s a intermediate layer, because.
152 00:21:45.400 ⇒ 00:21:46.110 Miguel de Veyra: Yeah.
153 00:21:46.540 ⇒ 00:21:47.989 Uttam Kumaran: That may be the best.
154 00:21:51.070 ⇒ 00:21:51.534 Miguel de Veyra: The
155 00:21:53.760 ⇒ 00:21:59.669 Miguel de Veyra: So it’s basically per section. It’s going to be like a recorded super base, or wherever.
156 00:22:01.000 ⇒ 00:22:03.059 Uttam Kumaran: I think so. Yeah, I mean.
157 00:22:03.060 ⇒ 00:22:04.949 Miguel de Veyra: Yeah, that could be the way to go.
158 00:22:06.930 ⇒ 00:22:12.130 Uttam Kumaran: I mean, also your retreat, like you’re gonna you guys are, gonna find out you’re gonna have to do that for the rag. Anyways.
159 00:22:15.300 ⇒ 00:22:17.339 Miguel de Veyra: Yeah, okay, yeah, we’ll look into it.
160 00:22:21.420 ⇒ 00:22:23.109 Uttam Kumaran: Okay, let me just write that down.
161 00:22:30.940 ⇒ 00:22:34.650 Miguel de Veyra: Sorry, Jen, a quick question. Do you have access to the golden sheet already?
162 00:22:37.840 ⇒ 00:22:38.480 Janna Wong: Sorry.
163 00:22:40.710 ⇒ 00:22:45.490 Miguel de Veyra: The golden data sheet, 3%.
164 00:22:45.490 ⇒ 00:22:46.490 Janna Wong: Than docs.
165 00:22:46.650 ⇒ 00:22:48.030 Miguel de Veyra: Yeah. Yeah. Yeah. Let me check.
166 00:22:48.030 ⇒ 00:22:49.050 Janna Wong: Yeah, I think so.
167 00:22:49.050 ⇒ 00:22:50.140 Miguel de Veyra: Okay. Okay.
168 00:22:50.140 ⇒ 00:22:52.360 Uttam Kumaran: Yeah, it looks like this still, like
169 00:22:52.600 ⇒ 00:22:56.999 Uttam Kumaran: I mean, but this is my. But this is my question, like, why aren’t these filled out.
170 00:22:58.230 ⇒ 00:22:59.310 Miguel de Veyra: Which ones.
171 00:22:59.960 ⇒ 00:23:04.270 Uttam Kumaran: Like the expected answers.
172 00:23:06.086 ⇒ 00:23:07.869 Miguel de Veyra: We’re using proper answer, though.
173 00:23:08.990 ⇒ 00:23:10.610 Uttam Kumaran: But what is to follow?
174 00:23:11.557 ⇒ 00:23:14.779 Miguel de Veyra: They don’t know the answer to. They’re gonna give it to us eventually.
175 00:23:14.780 ⇒ 00:23:15.610 Uttam Kumaran: Oh!
176 00:23:17.730 ⇒ 00:23:21.189 Miguel de Veyra: Because they have to update the documents. I was on a call with them.
177 00:23:21.470 ⇒ 00:23:24.429 Uttam Kumaran: Okay. But like, let’s take a look at line 48.
178 00:23:24.920 ⇒ 00:23:25.720 Miguel de Veyra: Okay, it would.
179 00:23:29.610 ⇒ 00:23:30.390 Miguel de Veyra: Yeah.
180 00:23:33.250 ⇒ 00:23:39.559 Uttam Kumaran: If the if, If the Csr asks, How do I schedule an annual rodent inspection?
181 00:23:40.000 ⇒ 00:23:42.050 Uttam Kumaran: This is the answer.
182 00:23:45.310 ⇒ 00:23:50.160 Miguel de Veyra: Yes, this one, the S.
183 00:23:50.570 ⇒ 00:23:51.969 Uttam Kumaran: Schedule, the follow up!
184 00:23:52.950 ⇒ 00:23:56.989 Miguel de Veyra: Yeah, basically, the answer is like, all 6, 7 is wrong.
185 00:23:57.440 ⇒ 00:24:00.347 Uttam Kumaran: I know, dude, but like that’s not good enough.
186 00:24:02.410 ⇒ 00:24:09.009 Uttam Kumaran: They’re they’re. This is the problem they’re looking at. They’re looking at the agent answers and just saying it’s correct or not
187 00:24:09.130 ⇒ 00:24:15.640 Uttam Kumaran: like they should be providing they should be writing down what they expect. Because how can we?
188 00:24:15.810 ⇒ 00:24:20.679 Uttam Kumaran: How can we judge the agent on its own work. It’s always going to be correct.
189 00:24:23.200 ⇒ 00:24:27.420 Uttam Kumaran: like, I don’t want them to critique the answer. I want them to write like
190 00:24:28.080 ⇒ 00:24:34.340 Uttam Kumaran: the see the the answers from 57 down are much better.
191 00:24:35.560 ⇒ 00:24:37.459 Miguel de Veyra: Yeah, yeah, this was from the, I think.
192 00:24:37.460 ⇒ 00:24:38.680 Uttam Kumaran: For everything.
193 00:24:50.930 ⇒ 00:24:52.100 Uttam Kumaran: See what I mean.
194 00:24:52.100 ⇒ 00:24:53.079 Miguel de Veyra: Yeah, yeah, yeah.
195 00:24:54.450 ⇒ 00:24:57.320 Uttam Kumaran: Dude. I don’t want to keep talking about this Doc. Bro.
196 00:24:58.440 ⇒ 00:25:03.780 Uttam Kumaran: Like what is, what’s I don’t understand like? Is that not clear like? Why can’t they?
197 00:25:04.450 ⇒ 00:25:10.379 Uttam Kumaran: I thought you already met with them like, why can’t they write down the answer they expect for each of these.
198 00:25:11.519 ⇒ 00:25:16.510 Miguel de Veyra: Yeah, I mean, they told me that they will put it here because they have access to this one also.
199 00:25:21.380 ⇒ 00:25:23.820 Uttam Kumaran: And everything to follow is like what.
200 00:25:25.086 ⇒ 00:25:27.090 Miguel de Veyra: They don’t know the answer, so they’ll
201 00:25:27.310 ⇒ 00:25:32.749 Miguel de Veyra: you know. I think they’ll update the docs. It’s either they update the docs or they have to look to find the do. The document.
202 00:25:33.620 ⇒ 00:25:41.520 Uttam Kumaran: Okay, yeah, I. Okay, I want everything to be more like the stuff towards the end.
203 00:25:46.350 ⇒ 00:25:48.259 Uttam Kumaran: like 57 onwards.
204 00:25:49.100 ⇒ 00:25:52.350 Miguel de Veyra: Okay? I can chat like a follow up email.
205 00:25:53.540 ⇒ 00:25:58.649 Miguel de Veyra: just a reminder to answer it. Maybe I can delete the or hide this agent. Answer, no.
206 00:26:00.110 ⇒ 00:26:19.379 Uttam Kumaran: No leave it, but, like they, I want. The prompt for them is that they need to just fill out what they would. What if if they were asked the question, what they would respond with, that’s it. Because I’m telling you, Eval, do not work like this like you can’t tell it. You can’t say
207 00:26:19.900 ⇒ 00:26:25.839 Uttam Kumaran: you can’t be like. Yes, this is the correct answer. Like we can’t fit. We can’t judge it. We can’t judge it based on that.
208 00:26:26.480 ⇒ 00:26:32.870 Uttam Kumaran: or if it is, it’s fine, but like it can’t be. Every single one can’t be. This is correct. This is correct. Like
209 00:26:33.200 ⇒ 00:26:38.739 Uttam Kumaran: is the length appropriate, like is the tone appropriate like those are things we can’t figure out.
210 00:26:39.380 ⇒ 00:26:40.340 Miguel de Veyra: Okay. Okay.
211 00:26:40.340 ⇒ 00:26:40.985 Uttam Kumaran: So.
212 00:26:45.450 ⇒ 00:26:46.410 Miguel de Veyra: Okay.
213 00:26:47.420 ⇒ 00:26:48.499 Uttam Kumaran: Yeah, I think.
214 00:26:48.500 ⇒ 00:26:50.440 Miguel de Veyra: Schedule. Another meeting with them this week.
215 00:26:50.840 ⇒ 00:26:56.459 Uttam Kumaran: Yeah, if you can just talk to them. And then just I need like.
216 00:26:56.600 ⇒ 00:26:59.749 Uttam Kumaran: for example, correct answer is not good enough, like.
217 00:27:00.700 ⇒ 00:27:07.200 Uttam Kumaran: I really want them to say whether, like everything from the length to the tone, so that like, this is what I don’t.
218 00:27:08.000 ⇒ 00:27:09.185 Uttam Kumaran: I don’t know.
219 00:27:15.180 ⇒ 00:27:22.000 Uttam Kumaran: yeah. This is why I just wanna make sure like, I don’t want to keep talking about this this document dude like, I want this to be finished up
220 00:27:24.140 ⇒ 00:27:31.870 Uttam Kumaran: and then. So then tell me what’s up like? When can I expect to see the Eval work like, can we talk about that tomorrow like? Do you think we’ll be able to
221 00:27:32.360 ⇒ 00:27:34.349 Uttam Kumaran: put that all there today.
222 00:27:35.340 ⇒ 00:27:41.109 Miguel de Veyra: Yeah, we should be able to put it all in there today. Then come tomorrow. We should be able to discuss it. Yeah.
223 00:27:41.340 ⇒ 00:27:42.619 Uttam Kumaran: Okay, let’s do that.
224 00:27:42.840 ⇒ 00:27:46.269 Uttam Kumaran: Yeah, and ask them, just say anything that’s to follow, like we need it.
225 00:27:47.140 ⇒ 00:27:48.830 Miguel de Veyra: Okay. And then.
226 00:27:48.830 ⇒ 00:27:51.319 Uttam Kumaran: Keep pushing them, and then anything that’s like
227 00:27:51.840 ⇒ 00:27:56.390 Uttam Kumaran: correct answer is fine. If they really are like, this is a fine answer.
228 00:27:56.530 ⇒ 00:27:56.850 Miguel de Veyra: Yeah.
229 00:27:56.850 ⇒ 00:28:01.180 Uttam Kumaran: But I’m I actually really am much more interested in them.
230 00:28:02.430 ⇒ 00:28:03.440 Uttam Kumaran: Like.
231 00:28:03.650 ⇒ 00:28:12.140 Uttam Kumaran: I mean, I’m okay to if you if you think they’re not gonna be able to do it. Like, for example, see? After 57, they literally wrote all the answers, because none of these have
232 00:28:12.290 ⇒ 00:28:14.019 Uttam Kumaran: the agent answers.
233 00:28:14.720 ⇒ 00:28:16.020 Uttam Kumaran: So, oh, yeah, yeah.
234 00:28:16.020 ⇒ 00:28:19.819 Uttam Kumaran: I don’t think they’re like, I think you just need to explain it again to them.
235 00:28:20.250 ⇒ 00:28:20.870 Miguel de Veyra: Okay.
236 00:28:20.870 ⇒ 00:28:21.630 Uttam Kumaran: Yeah.
237 00:28:21.980 ⇒ 00:28:31.989 Uttam Kumaran: like what? Exactly we’re looking for, because I’m not. I’m not looking for the fact that our answer is right. I’m looking for what the right answer is. Otherwise we’ll never be able to have a source of truth.
238 00:28:32.620 ⇒ 00:28:38.940 Miguel de Veyra: Yup, okay, the ones from 57 below. That’s that. All came from their FAQ. Docs.
239 00:28:40.980 ⇒ 00:28:42.910 Uttam Kumaran: Oh, they didn’t even write these.
240 00:28:42.910 ⇒ 00:28:44.179 Miguel de Veyra: No, no! I wrote it.
241 00:28:44.720 ⇒ 00:28:49.710 Uttam Kumaran: Dude. Okay. What the fuck, then what are they even doing on the stock? They didn’t do anything.
242 00:28:49.860 ⇒ 00:28:51.479 Miguel de Veyra: No, no, they haven’t done anything.
243 00:28:51.680 ⇒ 00:28:58.190 Uttam Kumaran: Okay. Then I’m gonna send an email this morning. And I’m gonna say we can’t. We’re we’re super delayed on getting answers for these.
244 00:28:58.450 ⇒ 00:29:00.429 Miguel de Veyra: Yeah, yeah, so can I do.
245 00:29:00.430 ⇒ 00:29:10.980 Miguel de Veyra: I? I tried explaining to them already, like again and again. But I don’t really want to push event, because every time I hop on a call with her like Oh, of course we understand. We’ll do it
246 00:29:11.760 ⇒ 00:29:14.549 Miguel de Veyra: like beyond that I don’t know what to do.
247 00:29:14.550 ⇒ 00:29:23.279 Uttam Kumaran: Okay, okay, see? But that’s what I want to hear from you. Right? Like, if when I go onto this, Doc and I see that it’s not filled out. I’m like.
248 00:29:24.510 ⇒ 00:29:31.839 Uttam Kumaran: I’m like, what is the problem because we’ve been doing? I’ve been talking about this document for 2 weeks. I’m like, I don’t want to talk about this document anymore.
249 00:29:34.790 ⇒ 00:29:41.999 Uttam Kumaran: But I want to know from you like, Hey, the client is not giving us what we need.
250 00:29:42.530 ⇒ 00:29:42.910 Miguel de Veyra: And I can.
251 00:29:42.910 ⇒ 00:29:51.759 Uttam Kumaran: Take a different answer to the to this moment. I’m like yo. Miguel isn’t figuring this document out, so I need to know that, like yo, they are not doing this.
252 00:29:52.220 ⇒ 00:29:54.610 Uttam Kumaran: There’s a different set of steps for me to take.
253 00:29:55.260 ⇒ 00:29:58.210 Miguel de Veyra: Yeah. Okay. Okay, yeah. I’ll inform next time.
254 00:29:58.580 ⇒ 00:30:00.120 Uttam Kumaran: Okay. Okay. Good. See? That’s.
255 00:30:00.120 ⇒ 00:30:01.519 Miguel de Veyra: Try covering them too.
256 00:30:01.870 ⇒ 00:30:25.139 Uttam Kumaran: No, yeah, don’t cover them like the client will. The client will screw us on stuff like this where they don’t give us the right answer, we’re not able to do our jobs. And then we fail. And then they think we failed. Right? I wanna I’m gonna send an email today. That’s gonna make it very clear that we are still delayed on getting these. And that’s the reason why we’re we’re not. We can’t finish this this week.
257 00:30:25.270 ⇒ 00:30:30.799 Uttam Kumaran: right. And that’s that’s not being mean. That’s just. I’m being honest like I can’t do my job.
258 00:30:31.060 ⇒ 00:30:32.840 Miguel de Veyra: Like, if we don’t have the data.
259 00:30:32.840 ⇒ 00:30:37.509 Uttam Kumaran: Yeah, you know. And so I’ll say that. But that yeah, that’s what I want to hear.
260 00:30:37.630 ⇒ 00:30:44.969 Uttam Kumaran: I mean, this similarly happened. This happened with Eddie, too. Right where this is what happens. Clients could slow our job is to move fast.
261 00:30:45.160 ⇒ 00:30:51.918 Uttam Kumaran: Alright. That’s why they hire us. So I’ll send an email. I’ll send an email right after this and and ask for this.
262 00:30:52.550 ⇒ 00:30:53.230 Miguel de Veyra: Okay.
263 00:30:56.750 ⇒ 00:30:57.770 Uttam Kumaran: Okay. Awesome.
264 00:30:59.010 ⇒ 00:31:02.390 Miguel de Veyra: And then wait. Sorry. What’s the thing? Again?
265 00:31:03.170 ⇒ 00:31:09.099 Miguel de Veyra: Agent of bit information. Less than 10. Second response time. Okay, yeah. I think the rest are.
266 00:31:11.190 ⇒ 00:31:15.419 Uttam Kumaran: Yeah, the response time definitely really needs to happen.
267 00:31:15.750 ⇒ 00:31:16.710 Miguel de Veyra: Yeah, yeah.
268 00:31:24.750 ⇒ 00:31:27.030 Uttam Kumaran: Okay, cool.
269 00:31:28.240 ⇒ 00:31:38.169 Uttam Kumaran: Yeah, I think I think it’s fine. We we will probably just focus on everything ABC related. And yeah, if the 3 of you guys can focus on that. That’d be amazing. And
270 00:31:38.870 ⇒ 00:31:41.379 Uttam Kumaran: yeah, let’s talk. We’ll talk every day about it. And
271 00:31:41.920 ⇒ 00:31:42.260 Miguel de Veyra: Yeah.
272 00:31:42.260 ⇒ 00:31:42.780 Uttam Kumaran: And.
273 00:31:42.780 ⇒ 00:31:44.290 Miguel de Veyra: How to get it started that this week.
274 00:31:44.470 ⇒ 00:31:45.430 Uttam Kumaran: Okay. Awesome.
275 00:31:46.290 ⇒ 00:31:49.700 Miguel de Veyra: And it’s been named it already.
276 00:31:49.700 ⇒ 00:31:50.240 Uttam Kumaran: That’s nice.
277 00:31:50.680 ⇒ 00:31:52.190 Uttam Kumaran: Okay, thank you.
278 00:31:52.520 ⇒ 00:31:53.840 Miguel de Veyra: Thank you. Bye-bye.
279 00:31:53.840 ⇒ 00:31:54.179 Uttam Kumaran: Bye.