Meeting Title: Brainforge <> Lilo: 30 min sync Date: 2026-02-20 Meeting participants: zacfromson, Uttam Kumaran, Pranav Narahari, Bobby Palmieri, Clarence Stone
WEBVTT
1 00:01:01.570 ⇒ 00:01:02.330 Uttam Kumaran: Hey, Zach.
2 00:01:04.000 ⇒ 00:01:05.259 zacfromson: Hey, how’s it going?
3 00:01:05.260 ⇒ 00:01:06.130 Uttam Kumaran: Hey, good.
4 00:01:22.110 ⇒ 00:01:26.350 zacfromson: I’m just gonna go on mute for now. We’re gonna have a coffee shop, unfortunately, so…
5 00:01:26.550 ⇒ 00:01:27.749 Uttam Kumaran: Oh, okay, no worries.
6 00:01:51.410 ⇒ 00:01:56.269 Uttam Kumaran: Fair enough. Just waiting for Bobby, and then waiting for Clarence on our side, too.
7 00:02:06.760 ⇒ 00:02:07.460 Bobby Palmieri: Hey, guys.
8 00:02:07.780 ⇒ 00:02:08.470 Uttam Kumaran: A.
9 00:02:09.050 ⇒ 00:02:09.570 Bobby Palmieri: anymore.
10 00:02:11.250 ⇒ 00:02:18.119 Uttam Kumaran: Cool, we just have probably… well, at least one more person from our side joining, but we can go ahead and get started. Yeah, I mean, I just wanted to sort of…
11 00:02:18.250 ⇒ 00:02:35.529 Uttam Kumaran: maybe take this call to sort of reset a little bit. I mean, I tried to go through in detail a bit in my message. I mean, I’d be happy to go through sort of line by line on that if we’d like to as well. I just felt like there’s a lot said in that message that I felt like the majority of it just wasn’t…
12 00:02:35.800 ⇒ 00:02:42.290 Uttam Kumaran: wasn’t correct from what, you know, I’m aware of on this project. So… Yeah.
13 00:02:42.480 ⇒ 00:02:43.880 Bobby Palmieri: Yeah, look, I mean, I think…
14 00:02:44.440 ⇒ 00:02:50.189 Bobby Palmieri: Like, I mean, just looking at where we’re at, right? Like, we’re… we’re 50K in.
15 00:02:50.310 ⇒ 00:02:53.820 Bobby Palmieri: And… Like, what shift is…
16 00:02:54.040 ⇒ 00:03:13.320 Bobby Palmieri: MetaMCP, Klaviyo MCP, one Slack report, and, like, the nano Banana tool that we built and passed over the code. Like, I know there’s a lot going on, and, like, you know, there’s been some shift in things there, but, like, we do need to be moving faster, like, we’re, you know.
17 00:03:13.570 ⇒ 00:03:22.609 Bobby Palmieri: half of $100,000 into this, and, like, there’s not a ton that’s been shipped and or functional, and, like, now I thought we…
18 00:03:22.930 ⇒ 00:03:34.970 Bobby Palmieri: aligned, you know, on Friday, and now it feels like, you know, come to this week, and, like, not much has been… like, we’ve burned a week, you know, for being honest, and I think it’s just a little frustrating that it took the time…
19 00:03:34.970 ⇒ 00:03:53.160 Uttam Kumaran: I don’t think we’re being honest in that, like, I’d like to push back. Like, we’ve… we not only came in in December and inherited, like, a system that… none of which was working, there was actually code around, like, crypto and a bunch of other random stuff. We turned that all around. You’re… you’re failing to mention all the deployment, all of the admin, like.
20 00:03:53.260 ⇒ 00:04:03.339 Uttam Kumaran: we’ve shipped, like, legitimate software here. Additionally, like, we are moving very fast. I think you’d be hard-pressed, like, I’m not sure how much software…
21 00:04:03.340 ⇒ 00:04:13.470 Uttam Kumaran: you know, building, you know, you’ve been involved in, but in my career, this is the fastest I’ve seen something like this ever come out. And so, especially at, you know, our current contract size, like, we are moving
22 00:04:13.520 ⇒ 00:04:32.940 Uttam Kumaran: you know, at a rate that I’m actually very comfortable with. Like, I… I… I’m really close with Pranav on this, but Pranav is leading this project. But now, you know, as I’ve sort of seen the last two weeks, it’s clear that, like, at our… at our contract size, like, what you’re asking from us is just not possible. You know, we’re… we’ve actually been able to develop
23 00:04:32.970 ⇒ 00:04:52.860 Uttam Kumaran: tons and tons of features, and you may say that 3-month time span is short, but, you know, I would say that I’d push back on that, like, this is an incredibly fast timeline for what we’ve built. We’re also building on, like, frontier-level technology for you. There’s not many other development firms that are doing this type of work.
24 00:04:52.950 ⇒ 00:05:09.250 Uttam Kumaran: at the price point that we are offering. I know the price may seem high, but it’s all relative. Like, this is, like, extreme frontier AI technology that we’re developing on, and I have to pay people, talented people, to come do this work.
25 00:05:09.290 ⇒ 00:05:26.459 Uttam Kumaran: As part of Brainforge, and, like, I really just want to push back on that. Like, we’ve developed a lot here. It’s not just the things that you’ve mentioned, and I’m comfortable saying that if you feel like you could AI what we’ve done in a weekend, then I’m happy for you to go do that. Like, that’s not what we’ve done here at all.
26 00:05:26.920 ⇒ 00:05:28.190 Bobby Palmieri: Hello, good.
27 00:05:28.420 ⇒ 00:05:37.090 Bobby Palmieri: I’m not trying to make this combative, like, you know, I think if we… and honestly, calling a spade a spade, like, if we look at the initial scope, like, it was…
28 00:05:37.310 ⇒ 00:05:54.440 Bobby Palmieri: supposed to be done in 30K was all of the MCP’s Slack reports, forecasting, and a storyboard, which we parlayed into Nanobanana, right? So, like, we do have the expectation that all of this was, you know, kind of the original language in the contract was, like.
29 00:05:54.830 ⇒ 00:05:57.040 Bobby Palmieri: In 3 months, or 30K, like…
30 00:05:57.040 ⇒ 00:06:13.469 Uttam Kumaran: And I’d love to review that original contract, because the scope we put in there was… when you do a scope, we walk into a system that we have no awareness of, so it is a guiding document. Like, we didn’t know the state of all of the GitHub, we didn’t know the state of all the features you’ve asked for.
31 00:06:13.470 ⇒ 00:06:22.169 Uttam Kumaran: You also know that you’ve asked for several features that are outside of that scope. So, I’m wondering what… should we have just pushed back on that when that was asked?
32 00:06:23.370 ⇒ 00:06:27.859 Bobby Palmieri: Yeah, like, and I’d love to… I mean, maybe we can review that, of just, like…
33 00:06:27.860 ⇒ 00:06:29.020 Uttam Kumaran: We need to review that, yeah.
34 00:06:29.020 ⇒ 00:06:45.389 Bobby Palmieri: what was in that scope, and look, I think we’re trying to be collaborative here, and, you know, but I do think that there’s an expectation on our end that, like, more of this would have been completed for 30K, and I think, like, we do have weeks where things do get shipped, and then, like.
35 00:06:45.390 ⇒ 00:06:57.159 Bobby Palmieri: this week, you know, feels like a wasted week, and we’re just, you know, the invoices just keep coming in, and, like, we’re not getting any closer to having, like, the forecasting tool done, and things like that as well. And, you know, all of these…
36 00:06:57.160 ⇒ 00:07:01.410 Uttam Kumaran: Did you see that we shipped the… we shipped 5 things that are coming out this week?
37 00:07:01.520 ⇒ 00:07:13.600 Uttam Kumaran: And we’d mentioned when we talked earlier this week that we have to spend time getting organized on the priority list. That time doesn’t come from additional budget. Like, that is all part of the price you’re paying.
38 00:07:14.270 ⇒ 00:07:21.920 Bobby Palmieri: Totally, I’m aligned with that, but, you know, I think in terms of, like, you know, what was shipped this week, I’m just not sure if, like.
39 00:07:22.180 ⇒ 00:07:29.389 Bobby Palmieri: Those are… those weren’t in the priority list in that sense, in terms of, like, the Nano Banana API error resolved, like.
40 00:07:29.590 ⇒ 00:07:32.249 Bobby Palmieri: There wasn’t an error to be resolved, it was…
41 00:07:33.040 ⇒ 00:07:36.329 Bobby Palmieri: Nano Banana being down, right? Like… But isn’t that.
42 00:07:36.330 ⇒ 00:07:39.310 Uttam Kumaran: That’s an error to resolve, like, we’re maintaining the system.
43 00:07:41.290 ⇒ 00:07:43.090 Bobby Palmieri: I thought it was just down…
44 00:07:43.490 ⇒ 00:07:45.850 Bobby Palmieri: Like, it was just an error from that perspective.
45 00:07:45.850 ⇒ 00:07:54.959 Uttam Kumaran: But do you know that we have to go investigate? Like, when you send a message saying it’s down, someone on our side has to investigate, and we now implemented a backup API so that it doesn’t go down?
46 00:07:55.290 ⇒ 00:07:57.049 Uttam Kumaran: Like, there’s alternatives.
47 00:07:57.050 ⇒ 00:08:00.790 Bobby Palmieri: But that was on… Thursday, right? Like, look.
48 00:08:01.570 ⇒ 00:08:10.749 Uttam Kumaran: But I also want to address, like, some of your other points. You mentioned that we are potentially getting financial kickbacks from vendors, like, that’s completely false.
49 00:08:10.750 ⇒ 00:08:11.150 Bobby Palmieri: We mentioned.
50 00:08:11.150 ⇒ 00:08:24.549 Uttam Kumaran: the air by feet feature. We’ve completely… we’re super forward with the fact that we were making that decision to adhere to timelines. Like, I just would like to go line by line, because I just feel like a lot of the things that were said there were incorrect.
51 00:08:24.910 ⇒ 00:08:37.240 zacfromson: And I think, you know, even on the AirByte piece, like, I thought we were completely aligned on there, and then, like, we’re like, hey, like, we’re not using AirByte. Like, that was, like, a discussion we all had, so I think we were just shocked to find out we’re not using AirByte, because we did.
52 00:08:37.240 ⇒ 00:08:38.100 Uttam Kumaran: That was a month ago.
53 00:08:39.270 ⇒ 00:08:40.240 Bobby Palmieri: Yeah, I get it.
54 00:08:40.240 ⇒ 00:08:40.650 zacfromson: Yeah.
55 00:08:40.659 ⇒ 00:08:52.539 Bobby Palmieri: Look, we explicitly… I know you guys have the recordings, like, we explicitly talked about, like, hey, everyone that we’ve talked to that’s building is on AirByte, we just feel comfortable on AirByte, it has all of that.
56 00:08:52.540 ⇒ 00:09:04.900 Uttam Kumaran: But it wasn’t working for us, which is why we… like, the Shopify connector was not working, and then we said, hey, in order to continue to build this, we can switch to another vendor. Would you like us to wait or build?
57 00:09:05.000 ⇒ 00:09:07.409 Uttam Kumaran: And we went forward with building it.
58 00:09:07.630 ⇒ 00:09:23.169 Bobby Palmieri: the first message that I have on Polytomic is, hey, we’ve got Shopify connected here. I… to be honest, I don’t really give a shit, but it was something that we, like, specifically talked about, and then all of a sudden we’re on this connector that we’ve never heard of.
59 00:09:24.280 ⇒ 00:09:25.710 Uttam Kumaran: You gotta understand, like, your…
60 00:09:25.710 ⇒ 00:09:26.240 zacfromson: That’s not…
61 00:09:26.240 ⇒ 00:09:40.859 Uttam Kumaran: How have you never heard of it? It’s like, we build these systems, like, we can run every tool we use by you, but I’m telling you that from our experience, and you’re hiring us for our expertise, like, we recommend tools for the job that they do.
62 00:09:40.950 ⇒ 00:09:46.460 Uttam Kumaran: we tried AirByte, it wasn’t working for 2 weeks. Like, I will go look at the.
63 00:09:46.460 ⇒ 00:09:46.969 zacfromson: He tells them.
64 00:09:46.970 ⇒ 00:09:47.769 Uttam Kumaran: When we made that proposal.
65 00:09:47.770 ⇒ 00:09:58.820 zacfromson: We can look at it. I think something’s shocking, like, you know, I just, like, just from my perspective, like, I’ve spent time with Sawyer, the reason why we’re all here together, and, like, general context is being used… is building on Airbite, and they have everything, like, up or running.
66 00:09:58.820 ⇒ 00:10:01.780 Uttam Kumaran: Sawyer isn’t a data engineer either.
67 00:10:01.780 ⇒ 00:10:04.839 zacfromson: Sure. He’s not, but his team is.
68 00:10:05.300 ⇒ 00:10:10.720 Uttam Kumaran: It’s a startup that they started, like, 3 months ago. I know their whole team, too.
69 00:10:10.970 ⇒ 00:10:12.199 zacfromson: Yeah, I’m… yeah, I guess…
70 00:10:12.200 ⇒ 00:10:16.069 Bobby Palmieri: Okay. I think we’re losing the point here, like, look…
71 00:10:16.160 ⇒ 00:10:32.610 Bobby Palmieri: wanted to come to this call, like, obviously, you want to be, you know, combative on each point. Let’s go line by line. I’m happy to discuss our qualms with it, and how we can get to, you know, a solution here. Like, wanted to come open, but if you want to go line by line, let’s do it. Where do you want to start?
72 00:10:32.610 ⇒ 00:10:49.699 Clarence Stone: Bobby, I think that’s a great point. I don’t think it’s, you know, gonna be productive for us to kind of go in circles here. I’m actually interested in figuring out what you would like to see as a change, right, and how we can actually support that. And by the way, nice to meet y’all. I’m Clarence.
73 00:10:50.430 ⇒ 00:11:14.490 Clarence Stone: I am head of labs at, Brainforge. I just started a few months ago. What we do, actually, is build internal tools for ourselves, our company OS, and all the actual, you know, cool things that you’re seeing, our frontier-level stuff, that I’m actually kind of shy about sharing, because they’re a little too powerful, and, you know, honestly, I’m here because I’m wondering if I can get this team back to continue to help me build what we’re building internally, so…
74 00:11:14.760 ⇒ 00:11:39.479 Clarence Stone: Like, if there’s not value that we’re bringing to you today, then, you know, that’s the alternative. And I don’t want it to go there, right? Absolutely, because, like, this team has put in a lot of work on building things for you, and it seems like Pranav and team have, you know, always had tons of smiles and said that things are going really well over at Lilo, so this was, like, sort of a, you know, context switch for me, and I’m just here to figure out, you know.
75 00:11:39.480 ⇒ 00:12:04.309 Clarence Stone: where that disconnect went to be, because, you know, I finally got a chance to take a look at your new request, and by the way, really forward-thinking, amazing, right? You’re making the right moves on infrastructure, you’re making the right moves on how connectors should work between client to client, but I’m going to be completely honest with you. I’ve spent years at EY as the GTM lead for AI, and the pricing that you’re getting is just not going to be possible
76 00:12:04.310 ⇒ 00:12:14.010 Clarence Stone: We cannot, you know, have this current team deliver at that rate if, you know, this is the kind of frontier attack that you want.
77 00:12:14.150 ⇒ 00:12:22.349 Clarence Stone: So this is why they’re spinning in circles, right? They’re looking at things that they just haven’t seen. This is stuff that’s… you’re probably seeing just on X.
78 00:12:22.680 ⇒ 00:12:28.940 Clarence Stone: Right? So, like, with that, all of that in mind, what I do want to ask is, what is the change you want to see?
79 00:12:28.940 ⇒ 00:12:45.210 Clarence Stone: Right? And is it fine if we continue with the current SOW, or do we want to start pricing in the value of those new builds, and reconfigure our strategy? Because what this team really needs is direction, so that they can go heads down and start doing productive things for you.
80 00:12:45.860 ⇒ 00:12:54.560 Bobby Palmieri: Yeah, look, and I think that’s the conversation that we had on Friday, and we put together the priority list of, like, where we want to…
81 00:12:54.700 ⇒ 00:13:03.669 Bobby Palmieri: spend our time, right? And I feel as though that, like, we put that together, and now we’re here a full week later, and 5K later, and…
82 00:13:03.970 ⇒ 00:13:06.119 Bobby Palmieri: I don’t feel as though that we’re pushing.
83 00:13:06.120 ⇒ 00:13:23.050 zacfromson: Yeah, I think… I think there’s just… we really just want to get a line clearance, simply put. I feel like we’re… we put together the document, the goal is kind of, like, come back to us, get us some… some info back on that, understand, like, what we can scope for the next 30 days, like, what you guys are actually working on. I think right now we don’t have a lot of visibility and understanding, and, like.
84 00:13:23.050 ⇒ 00:13:31.850 zacfromson: we felt like we did our part from last week’s conversation, and then, like, over the past week, I kind of felt like there maybe was a breakdown, maybe we’re misunderstanding, like, I think, like, we’re, you know…
85 00:13:32.020 ⇒ 00:13:32.659 Clarence Stone: Yeah, so…
86 00:13:32.660 ⇒ 00:13:37.269 zacfromson: We have been happy, to your point. There is all smiles, like, we are making progress. I think right now, we just have, like, a…
87 00:13:37.270 ⇒ 00:13:53.109 zacfromson: there’s just a misalignment in our visibility and, like, progress plan of what’s happening. Like, we just need more visibility on that. I think there needs to be a little bit of project management stuff, too, and I’ve noted it prior, that, like, yeah, there has been some scope creep, and, like, we’ve been mindful and honest of that, like, sometimes Bobby’s throwing something new at you guys, like.
88 00:13:53.560 ⇒ 00:13:55.630 zacfromson: And by the way, and I think that’s right.
89 00:13:55.630 ⇒ 00:13:58.719 Clarence Stone: You’re the CEO, you should always be pushing the boundaries, right?
90 00:13:58.720 ⇒ 00:14:04.879 zacfromson: So that’s why we just want to get aligned in that. That’s why we want to get aligned in a backlog, or a Gantt chart.
91 00:14:04.880 ⇒ 00:14:25.600 zacfromson: this is how this affects the project. So, sure, you want to do this, Bobby, but that’s gonna F up X, Y, and Z as a result, right? And, like, that’s just getting blurred lines there, and I feel like we just need better visibility in the project plan, and again, something could break, something could take longer. We just need that communicated so we know how we’re pacing and what’s going on, and I think, like, other side frustration is, like, yep, some of this stuff is pushed, and we’re like, oh crap, like.
92 00:14:25.600 ⇒ 00:14:31.500 zacfromson: you know, we can’t get any of our, you know, not any of it, but, like, you know, we still don’t have Shopify set up, can’t get Google connected.
93 00:14:31.500 ⇒ 00:14:47.609 zacfromson: Having issues with connecting Meta, and Klaviyo MCP’s hallucinating like crazy, so we still don’t even have… we can’t even use that product at all, Clarence. Like, we’re trying to roll it out with our team, and, like, we can’t even use it, which I think is, like, causing a little bit of frustration between me and Bobby as well, that, like.
94 00:14:47.610 ⇒ 00:14:54.710 zacfromson: The product doesn’t really work currently, and we can’t get everything connected, so it’s, like, that’s also, like, a frustration of what we’re seeing.
95 00:14:54.710 ⇒ 00:14:55.279 Bobby Palmieri: That was so funny.
96 00:14:55.280 ⇒ 00:14:55.890 zacfromson: here.
97 00:14:56.110 ⇒ 00:15:04.239 Bobby Palmieri: Yeah, I also think, like, just reviewing, like, the initial scope, and like, you know, we’re not holding you fully accountable to that. Like, I know things have shifted, but, like.
98 00:15:04.500 ⇒ 00:15:09.490 Bobby Palmieri: in writing, 30K, all MCP Slack reports.
99 00:15:10.120 ⇒ 00:15:24.869 Bobby Palmieri: forecasting, which we’ve deviated… we’ve dropped down considerably, because we’re not even using a model anymore for that. And then, you know, I felt as though that the message, you know, yesterday, like, just kind of, like.
100 00:15:25.150 ⇒ 00:15:39.110 Bobby Palmieri: fully out of left field, just in terms of, like, you know, hey, this is gonna be another, you know, 6 weeks in this sense, or whatever, from that perspective. When it’s just not really what we…
101 00:15:39.440 ⇒ 00:15:51.420 Bobby Palmieri: aligned on, or, like, any of the feedback that, like, we had given. Like, to be honest, like, I was pretty shocked on the comments on, like, the forecasting, because I went through the doc last night to make sure that, like.
102 00:15:52.580 ⇒ 00:15:59.090 Bobby Palmieri: My, my feedback was, like, relatively minimal. Like, I actually thought that we were, like.
103 00:15:59.350 ⇒ 00:16:18.910 Bobby Palmieri: pretty close on forecasting. Like, I’d love to know where it’s, like, a complete shift there, you know, in that regard. Like, I actually think that, like, if we look at the forecasting review doc that I put together, like, it’s like, hey, this scrolls too far, we need to be able to set this metric, like.
104 00:16:19.710 ⇒ 00:16:27.179 Bobby Palmieri: I didn’t think that it was anything that we hadn’t chatted about when we look… when, like, we make note of, like, data flow dependencies that are new.
105 00:16:28.700 ⇒ 00:16:36.930 Clarence Stone: Okay, so a couple things. It feels like we’re jumping around a couple different topics, and I definitely want to address every single one of them, right, Zach and Bobby? So here, here’s one.
106 00:16:37.040 ⇒ 00:16:38.099 Clarence Stone: as these…
107 00:16:38.100 ⇒ 00:17:02.189 Clarence Stone: like, AI tools and capabilities get stacked on top of each other, we now have a maintenance job on top of building, you know, what’s happening next, right? And this team is a little bit too small to be handling both of those things, and then throwing in some scoping on those changes. It’s not saying they don’t have the capabilities to do it, but from, like, this bracket, like, we… you know, if it was me coming from EY, I would
108 00:17:02.190 ⇒ 00:17:17.919 Clarence Stone: I would have told you, you have to wait two weeks for us to scope it. And this is the kind of deal you’re going to get from other vendors. So what I’ve told this team is that they’ve actually been over-delivering and not looking at stability, right? So what I’m hearing from you is, one, there’s a bunch of things I originally wanted.
109 00:17:17.920 ⇒ 00:17:28.989 Clarence Stone: Right? And they’re not working perfectly. So maybe we need to actually take a step back and say, hey, how do we do a stabilization phase and a planning phase, right, so that we know exactly what you want going forward.
110 00:17:28.990 ⇒ 00:17:41.730 Clarence Stone: Right? And we’re actually effectively staffing this team. Because the biggest problem from our side is that this team is trying to do all three of these things for you. It’s sort of like, you know, building the train as the train is moving.
111 00:17:41.730 ⇒ 00:17:54.109 Clarence Stone: Right? And, you know, things are kind of disappearing, and, you know, the train only gets faster. AI’s only accelerating for us. So does that concept make sense? Like, hey, let’s sit down and talk about, like, the stabilization features we need.
112 00:17:54.110 ⇒ 00:18:13.359 Clarence Stone: Right? And then let’s prioritize the things that you need next, and if we have to plus up our team to hit your timelines, I’m more than happy to add that additional value for you. But, like, that is taking people from my team, so, like, that’s how I eventually just got involved in all of this and taking a look at what we’re building here and the value analysis behind it.
113 00:18:14.340 ⇒ 00:18:17.339 Bobby Palmieri: Yeah, no, I’m aligned there. I think, like, if…
114 00:18:17.760 ⇒ 00:18:31.079 Bobby Palmieri: just putting it out there, like, we’re 50K in, and, like, I just don’t think we have a lot to show for it. There’s a lot that’s in the works and getting close, but, like, we need to… and, like, the whole conversation on Friday was, like.
115 00:18:31.610 ⇒ 00:18:43.160 Bobby Palmieri: we need prioritization to, like, start putting these things in the done column, and then Thursday comes around, and, like, we check in, and it’s, like, nothing’s been done. And I’m happy to go through what you…
116 00:18:43.160 ⇒ 00:18:44.080 Clarence Stone: Okay, yeah.
117 00:18:44.490 ⇒ 00:19:06.779 Clarence Stone: I think our biggest disagreement here is that, like, we hear comments like, nothing’s been done, and it frustrates the team because they’ve actually been building for you, right? So, I just want to be sensitive about being kind to everybody who’s working. I’ve not been part of that, so, like, I’m just going to make that point. But if you feel like nothing’s been done, let’s talk about how we can resolve that feeling for you, right?
118 00:19:07.550 ⇒ 00:19:08.330 Bobby Palmieri: Yeah, that’s…
119 00:19:08.330 ⇒ 00:19:09.890 zacfromson: Yeah, yeah, I think, I think just…
120 00:19:09.890 ⇒ 00:19:25.190 Clarence Stone: These are next steps, and before I pass you off to Pranav to actually do this analysis, what I’m going to tell Pranav is, like, right here, hey, let’s take a look at what we owe Lilo from the original contract, what’s not working, and actually focus on fixing that right away.
121 00:19:25.190 ⇒ 00:19:45.769 Clarence Stone: Right? And then, Bobby and Zach, you’ve given a list of priorities to us, I’m gonna have the team actually fully scope it, and if we find out that we actually need to plus up the team in order to support, like, the uptime, the availability, and the accuracy of models on top of the delivering new features, then you’re gonna have to actually, like, we’re gonna have to really revisit that SOW.
122 00:19:45.770 ⇒ 00:19:52.789 Clarence Stone: So, like, that’s really the situation right now. I mean, we’ve got a little menu of options, right? Deliver what we promised you.
123 00:19:52.790 ⇒ 00:20:07.860 Clarence Stone: Phase 2’s extensions and the things that you wanted there, and then your scope changes. So, like, out of all those three things from the menu, I’m interested in hearing from you what exactly you want to land at, because, I need this team back, actually. We’re pulling really cool stuff.
124 00:20:10.090 ⇒ 00:20:10.930 Bobby Palmieri: That works.
125 00:20:11.650 ⇒ 00:20:14.259 Clarence Stone: Cool. Alright, Pranav, you know what to do, man.
126 00:20:20.160 ⇒ 00:20:22.139 Pranav Narahari: Yeah, so go over the SOW right now.
127 00:20:24.400 ⇒ 00:20:25.770 Pranav Narahari: Oh, you’re, you’re muted.
128 00:20:28.410 ⇒ 00:20:29.160 Bobby Palmieri: You’re muted.
129 00:20:29.160 ⇒ 00:20:30.299 Pranav Narahari: You’re muted, you’re muted.
130 00:20:34.460 ⇒ 00:20:35.520 Pranav Narahari: Hello?
131 00:20:36.240 ⇒ 00:20:38.190 Bobby Palmieri: I think he might be talking to someone else.
132 00:20:38.310 ⇒ 00:20:38.839 Pranav Narahari: It’s hard.
133 00:20:38.840 ⇒ 00:21:02.739 Clarence Stone: Oh, I’m talking to you guys, but I went on mute. So I was saying, hey Pranav, like, go ahead and, you know, listen to Zach and Bobby, take notes of all the issues they’re experiencing with these MCP connectors. That’s priority one. It’s a bad look to have any of these connectors broken, right? There’s business processes that are behind it, and we should be dedicated to making sure there’s uptime and availability for all the capabilities that we originally promised.
134 00:21:02.740 ⇒ 00:21:15.970 Clarence Stone: Right, from there, let’s have another conversation of how you want to prioritize these next steps. Let’s wipe away all the existing SOWs and phase changes. We’re going to just create a brand new SOW to target exactly what you want. And if I have to plus up that team, I will.
135 00:21:16.970 ⇒ 00:21:18.580 zacfromson: Yeah, I think… That sounds great.
136 00:21:21.580 ⇒ 00:21:23.760 Bobby Palmieri: Do you want me to drive that, or…
137 00:21:23.760 ⇒ 00:21:38.180 Clarence Stone: So, Pranav, like, I would start by saying, hey, Zach, you rattled off a lot of things. I’m interested in hearing exactly, like, where the broken pieces are, right? Klaviyo Hallucinating, would love to know more. Do you have any screenshots? Like, I want.
138 00:21:38.180 ⇒ 00:21:48.119 zacfromson: Sure, yeah, I’ve actually sent a few over the past couple days that I’ve been noticing and seeing in the Brainforge Slack, and I’ve been flagging those instances as they occur.
139 00:21:48.370 ⇒ 00:21:51.560 zacfromson: And, like, even one of the most recent ones, too, like,
140 00:21:51.600 ⇒ 00:22:02.909 zacfromson: Sorry if there’s some background noise, unfortunately, I’m a coffee shop, but, you know, I even pulled, like, the IDs, like, for example, I was like, hey, like, can you, like, pull… can you, like, analyze this flow? I was like, hey, there’s 10 emails in here. There’s actually 12 active.
141 00:22:02.940 ⇒ 00:22:12.180 zacfromson: So, like, I went and pulled, like, all of the IDs that are in there versus the IDs that it’s showing, and I’m just, like, not sure why it’s missing, like, two of the emails, for example. So, like.
142 00:22:12.390 ⇒ 00:22:20.959 zacfromson: for… it’s, like, under-reporting in that instance, for example. So that’s, like, one… one instance. We could look at that one specifically, if that would be interesting, or.
143 00:22:20.960 ⇒ 00:22:31.639 Pranav Narahari: Yeah, and on that one, too, I can talk about what we shipped this week for that. So you mentioned hallucination, we had a whole discussion in that thread about how we can mitigate that.
144 00:22:31.640 ⇒ 00:22:45.659 Pranav Narahari: Bobby had some ideas. He thought, like, maybe just updating to Sonnet 4.6 would change it. We did a quick spike on, like, what 4.6 was doing. I think the… your guys’ understanding was, like, we were using Haiku, but we’re actually already on Sonnet.
145 00:22:45.670 ⇒ 00:22:53.329 Pranav Narahari: And so the context window wasn’t an issue there. The real issue after what we looked into, after we did a spike on that was,
146 00:22:53.520 ⇒ 00:23:13.379 Pranav Narahari: the temperature being one issue, potentially, and then also just, the system prompt. And so, we created a patch on that, Casey actually already merged that today, and so, yeah, I think we did a discussion on that, we kind of pushed that change as well. And so, yeah, that’s one that we’ve definitely been tracking, and I think we’re…
147 00:23:13.400 ⇒ 00:23:18.959 Pranav Narahari: pretty aligned on that one, too. Like, I’ve been trying my best, at least, to message you guys on that channel.
148 00:23:19.240 ⇒ 00:23:19.940 Bobby Palmieri: I think…
149 00:23:20.020 ⇒ 00:23:21.709 zacfromson: And I think overall, like.
150 00:23:23.390 ⇒ 00:23:27.480 Bobby Palmieri: I think the solution that we’re looking for is this doc. Zach and I spent
151 00:23:27.650 ⇒ 00:23:32.569 Bobby Palmieri: Couple hours over the weekend, last weekend, like, Putting this together, like…
152 00:23:32.750 ⇒ 00:23:39.780 Bobby Palmieri: This is the dock that has… everything… Ranked in order of, like.
153 00:23:40.130 ⇒ 00:23:47.730 Bobby Palmieri: priority of, like, where we want you guys to spend your time, right? Like, I think that, yeah, there’s a lot of things, like, I don’t know, there’s, like.
154 00:23:47.940 ⇒ 00:23:50.249 Bobby Palmieri: I don’t know, yeah, there’s a lot.
155 00:23:50.250 ⇒ 00:24:00.179 Pranav Narahari: I have a quick question on that, Bobby, just so I’m clear. So you sent over this doc, as well as, a Slack message that listed 6 bullet points, 4 priorities.
156 00:24:00.550 ⇒ 00:24:06.029 Pranav Narahari: I was under the assumption that those are the priorities, and this is just, like, a list with no ordering.
157 00:24:06.160 ⇒ 00:24:06.820 Pranav Narahari: So…
158 00:24:06.820 ⇒ 00:24:08.040 Bobby Palmieri: They’re the same, no?
159 00:24:08.400 ⇒ 00:24:09.549 Pranav Narahari: No, they’re different.
160 00:24:10.680 ⇒ 00:24:13.400 Bobby Palmieri: The ones I highlighted were the ones that you mentioned.
161 00:24:13.400 ⇒ 00:24:15.770 Pranav Narahari: And then you didn’t mention,
162 00:24:15.990 ⇒ 00:24:22.420 Pranav Narahari: If you scroll all the way to the bottom, I think I put in… Yeah, the n…
163 00:24:22.880 ⇒ 00:24:27.470 Pranav Narahari: All the way down… yeah, I added one thing, yeah, for the knowledge base.
164 00:24:28.440 ⇒ 00:24:37.000 Pranav Narahari: I didn’t see that anywhere else. Or maybe that was in there, but see, they’re just, like, in different locations, it’s not all at the top. So that’s just one thing I wanted to mention.
165 00:24:37.000 ⇒ 00:24:39.940 Bobby Palmieri: And these are the… these are the bullets that I…
166 00:24:40.000 ⇒ 00:24:45.980 Bobby Palmieri: had sent through, right? So, data warehouse stand-up, you know, forecasting finalized.
167 00:24:46.040 ⇒ 00:25:01.200 Bobby Palmieri: Meta dashboard, cron jobs for, you know, prompt management, and knowledge base. Yeah. You know, in that regard, I think, like, these two things are, to my understanding, like, done.
168 00:25:01.420 ⇒ 00:25:18.279 Bobby Palmieri: That we just want to get kind of across the finish line. Okay. Yeah, hopefully, like, I guess these bullets do line up, like, close to this. This is the order of… of priorities. You know, I would say. I think that the goal of kind of having this
169 00:25:18.290 ⇒ 00:25:28.529 Bobby Palmieri: like, longer things, like, if there’s blockers, or, like, Casey has additional time, or whatever to, like, you know, I think we want to get all of these done, so, like, you know, if…
170 00:25:28.530 ⇒ 00:25:46.340 zacfromson: Yeah, and I think some of it, too, was, like, our conversation last, like, hey, let’s just dump in all of our ideas so we’re not just talking about these things hypothetically, and then it’s at least all documented. You guys can review them and be like, hey, like, this is what’s feasible, so we can align on those expectations. I think right now, it’s kind of like, hey, yeah, we’ll work on this, and we… we have no idea, and again, like, I know, I think, like.
171 00:25:46.340 ⇒ 00:25:49.130 zacfromson: you guys are noting, like, hey, like, sometimes we might, like, up…
172 00:25:49.290 ⇒ 00:26:01.780 zacfromson: a projection by 20%, and, like, maybe we get it done faster, so we could produce further, or maybe we need to go back and fix a bug when we can reduce that time. I think that was part of the conversation, like, hey, we’re gonna up-budget everything by 20%, so we don’t
173 00:26:01.970 ⇒ 00:26:12.919 zacfromson: over-promise and under-deliver, was, I think, some of the discussions. I think, you know, just trying to use this as their guiding light to move forward, I think would be helpful. And I think just trying to clean up a couple of the things that maybe aren’t in here, where it’s like.
174 00:26:13.030 ⇒ 00:26:21.770 zacfromson: I felt like Klaviyo was working, and now it’s, like, the MCP is not, right? And a couple of these other things, like, we thought we were good to connect Google, and then we’re realizing it’s not, and then, like, you know, we’re…
175 00:26:21.770 ⇒ 00:26:42.529 zacfromson: should be good to get meta, and then… and then it’s not. And again, like, hey, like, if we don’t need to be on AirByte, and Polyatomic is fine, that’s great, but it’s not working either, so it almost… it almost feels like neither work, to some extent, so I, you know, we’re not saying, hey, we have to work on AirByte. I think the frustrating thing is, like, hey, AirByte’s not working, Polyatomic’s not working, we’re just, like.
176 00:26:42.530 ⇒ 00:26:47.510 zacfromson: it just… we can’t even use our MCP, that we have, like, Done, but, like…
177 00:26:47.840 ⇒ 00:26:58.919 zacfromson: don’t have anything connected, and, like, anything that’s, like, confidently being used to work, so I think that’s, like, the other thing, too, there, where it’s like, hey, we’re frustrated because we’ve gotten a couple other people who said, use AirByte.
178 00:26:58.920 ⇒ 00:27:11.059 zacfromson: you guys said PolyTomic, you know, I don’t want to get in a huge fight about that. I think the thing is just, like, it’s not working either. So it’s like, you guys have an issue with AirByte, and you’re like, hey, let’s do PolyTomic, and then, like, this isn’t working, you can’t connect anything.
179 00:27:11.060 ⇒ 00:27:13.980 zacfromson: there, so I think, you know, even from the top, just, like.
180 00:27:14.190 ⇒ 00:27:28.000 zacfromson: for the top piece of the data warehouse, like, you don’t have to go to AirByte if you guys really think that that’s not the best option, and Polycomic is, like, first-in-class best option, but we gotta find a way to make it actually work, because right now, I think, like, from a connector perspective, like, it’s…
181 00:27:29.220 ⇒ 00:27:31.629 zacfromson: it’s not performing.
182 00:27:31.630 ⇒ 00:27:35.940 Clarence Stone: Exactly, are you guys cool if we continue with Polytomic, then?
183 00:27:36.290 ⇒ 00:27:37.670 Clarence Stone: Look.
184 00:27:37.670 ⇒ 00:27:38.779 zacfromson: If we can make it work.
185 00:27:38.780 ⇒ 00:27:45.530 Clarence Stone: should we roll back? Like, you know, if you guys are… if you’re confident, if you guys are very confident we can get it to work.
186 00:27:45.530 ⇒ 00:28:00.779 zacfromson: let’s do it. I think our frustration right now, where that’s coming from, and that comment’s coming from, is that, like, hey, like, this is gonna work, and now, like, nothing’s working with it that well. Like, we’re having issues with it, so I think we’re questioning, like, is this the right thing to be doing, because it’s not working? So, if…
187 00:28:00.890 ⇒ 00:28:11.080 zacfromson: everyone is dead set, like, yes, this is better than AirByte, right? Like, we tell people what software to use, and we know this is the best in class. If you’re like, hey, Zach, this is the best goddamn software you can use, then, like.
188 00:28:11.240 ⇒ 00:28:19.259 zacfromson: tell us that, but we also need to make sure you guys can make it work, because the problem is it’s not working, so we’re not, like, dead set there.
189 00:28:19.260 ⇒ 00:28:37.630 Clarence Stone: confidence, like, this is what’s gonna happen next week. The team is going to look through every single error, and you will slowly see all your MCPs validated, tested, right, Pranav? And actually replicated for those issues that you’re highlighting specifically. So we will replicate it, we’ll show you that it’s resolved, with screenshots and proof.
190 00:28:37.800 ⇒ 00:28:47.039 Clarence Stone: Right? Priority one is to get you guys back online and operating. I know what it’s like when you rely on tools and they don’t work, and that’s not acceptable on our part.
191 00:28:47.700 ⇒ 00:28:49.570 Bobby Palmieri: It’s just, yeah, like…
192 00:28:49.570 ⇒ 00:29:06.669 Clarence Stone: Number two, though, Bobby, like, we also have to scope out the new things you’re asking us, right? And, like, and I want you to sign off on it and make sure that you’re happy with it, because every time there’s any changes in priorities or focus, this team doesn’t have any sort of PM built in.
193 00:29:06.670 ⇒ 00:29:31.619 Clarence Stone: And they are going to have to use your allocated time to actually, you know, scope those things out. And it’s not an easy thing to do. So, what I’m asking is, like, we will create a final approach for you to say, hey, these are the things you’re asking for, these are the things we can do, right, in our bucket, and these are the things we’re going to need additional help and support for. And you guys can pick from those menu of options on the velocity capabilities and things that
194 00:29:31.620 ⇒ 00:29:33.000 Clarence Stone: You want out of us.
195 00:29:33.370 ⇒ 00:29:37.540 Clarence Stone: Cool? Is that okay to see that by, like, Wednesday or Thursday of next week?
196 00:29:37.540 ⇒ 00:29:46.400 Bobby Palmieri: Yeah, I think… totally. I think… Within regards to that, Most of this is not… new stuff.
197 00:29:46.600 ⇒ 00:29:50.740 Bobby Palmieri: Right, so forecasting was in the original scope of work from November.
198 00:29:51.660 ⇒ 00:29:56.760 Bobby Palmieri: Google app, Google Slack reports, all in the original scope. And…
199 00:29:58.130 ⇒ 00:30:04.249 Bobby Palmieri: Also, I think part of the frustration on our end, got a lot of shit at 70% of the way done.
200 00:30:05.070 ⇒ 00:30:11.209 Bobby Palmieri: Right? And I think the conversation last week was… Let’s get this priority list.
201 00:30:12.160 ⇒ 00:30:17.510 Bobby Palmieri: So that we can move… A thing to 100% done.
202 00:30:17.650 ⇒ 00:30:22.819 Bobby Palmieri: and move on, right? So I think that’s what we want, is just, like, both sides…
203 00:30:22.820 ⇒ 00:30:47.509 Clarence Stone: What we did, like, de-conflict my confusion, I’m actually looking at these original SOWs. The first SOW actually didn’t include anything about Claude skills, MCPs, or customized prompts, and we’ve slowly, like, gone to that second contract that does include it, and now we’re actually going towards, like, really bespoke architecture, where you’re splitting out skills to be specific to a certain client, and then aggregating all that data back into one dashboard, which, by the way, Bobby, that’s
204 00:30:47.510 ⇒ 00:30:51.239 Clarence Stone: That’s brilliant, that’s how things should be done, but that is not easy to do.
205 00:30:51.520 ⇒ 00:31:03.020 Clarence Stone: So, like, each of these leaps are additional difficulties and challenges for the team, and the reason Pranav was spinning on this is because it takes a long time to figure out how you’re gonna stitch all those things together, and pun intended.
206 00:31:03.020 ⇒ 00:31:09.910 Bobby Palmieri: Totally, I guess the big thing is that, like, we’re… we’re now 50K and not 30K, and from the original scope, and I…
207 00:31:09.910 ⇒ 00:31:16.550 zacfromson: I think we’re 30K, and we’re about to start into our next… the next 20K, so we’re 30K in, just so that’s clear, there’s no misalignment there.
208 00:31:16.550 ⇒ 00:31:20.970 Clarence Stone: We’ll fun out those three weeks, right? And then we’ll talk about what happens next.
209 00:31:21.430 ⇒ 00:31:21.960 Bobby Palmieri: Right.
210 00:31:21.960 ⇒ 00:31:22.500 zacfromson: Yep.
211 00:31:22.690 ⇒ 00:31:25.960 Bobby Palmieri: I think… yeah, we’re saying the same thing. I think…
212 00:31:26.160 ⇒ 00:31:28.110 Bobby Palmieri: What we’re saying is, is that we’re…
213 00:31:28.250 ⇒ 00:31:33.529 Bobby Palmieri: we just want to get stuff done in the done column. Last week, when we chatted, we weren’t…
214 00:31:33.830 ⇒ 00:31:47.990 Bobby Palmieri: like, you know, penny-pinching for saying, like, hey, we should be done with this on what we paid. I think our… our call was, is like, hey, we’re okay with what we’ve paid so far and where we’re at, but we… we as a collective.
215 00:31:48.430 ⇒ 00:32:06.999 Bobby Palmieri: need to be better aligned on what the priorities are to… to get things put in the done column. So I don’t… I’m not saying that, like, hey, we need to be done with everything, because I understand that Claude Skills is, you know, a different approach than maybe what we thought, and, like, that’s why we want to work with you guys, is to be able to… to do that, but I think
216 00:32:07.100 ⇒ 00:32:20.489 Bobby Palmieri: Last week, we took the… the onus on us that I have pushed us in different directions. We need to resort to the mean of, like, staying laser-focused. Zach and I spent a bunch of time
217 00:32:20.690 ⇒ 00:32:28.429 Bobby Palmieri: getting you that focus, and I think… I feel as though we burned this week. Like, we’re frustrated on that.
218 00:32:28.430 ⇒ 00:32:42.829 Clarence Stone: So let’s fix that, right? Let’s take all those 70 percents to 100, and we will reprioritize all the wishlist items, Bobby, because, like, the reason why I’m actually on this call is I finally looked at your requirements, I’m like, somebody gets it.
219 00:32:42.830 ⇒ 00:32:50.779 Clarence Stone: They actually know what they’re doing here, and asking us for things that are really actually smart, well-organized, and how agents should be built.
220 00:32:51.060 ⇒ 00:33:06.340 Clarence Stone: But guess what? I mean, that kind of format just really, you know, hit the ground running in November, maybe December of last year. We barely have the people that know that in depth enough to be able to deliver that for you, and you guys are occupying most of that team. So I just want to make it clear that, like.
221 00:33:06.340 ⇒ 00:33:20.169 Clarence Stone: You know, our team wants to always help you, and they just got a little bit distracted saying, hey, there’s a new thing, we gotta do that too, and totally dropped the ball on that 70%, so you know what? We’re gonna solve that 70%, and then we’re gonna take a look at scoping it fully.
222 00:33:21.300 ⇒ 00:33:29.299 zacfromson: Yeah, and I think, I think the last thing, too, on that, on top of that, I guess, like, Pranav, since you’re owning this, I think if we start derailing stuff.
223 00:33:29.960 ⇒ 00:33:37.710 zacfromson: tell us. Yeah. I think that’d be the only last ask from you going forward, is being like, hey guys, this sounds cool, but this is gonna F up the plan.
224 00:33:38.380 ⇒ 00:33:52.209 zacfromson: let’s backlog this. I think that’s kind of, like, the… will help keep us on track going forward, just, like, don’t be afraid to be like, hey guys, I think, you know, I guess, in the nicest way, you’ve kind of… you guys have been very nice, and it’s been kind of, yes, yes, yes, but I think, like.
225 00:33:52.210 ⇒ 00:34:04.460 zacfromson: it’s okay to say no, and it’s not, no, we can’t do this, but no, this is gonna act off our timeline that we want to meet for you guys to deliver on, like, before we get excited about what’s next, like, let’s parlay this from the past. I think that’s…
226 00:34:04.550 ⇒ 00:34:05.930 zacfromson: Been a little bit of the…
227 00:34:06.120 ⇒ 00:34:25.420 zacfromson: you guys have been too nice, in terms of saying yes to everything, and, like, just chomping off too much, and then it’s derailing the project, and then we’re stuck at… everything’s at 70% of the way done instead of finishing it. So, I think that’s what was the goal of our backlog, that’s the goal of our plan forward, and, like, we’ll ask you stuff, that’s fine, we can talk about it, but if you’re like, hey, this is not for this week, like, I need to finish X, Y, and Z, then just
228 00:34:25.520 ⇒ 00:34:44.009 zacfromson: say that. Like, don’t be scared to, like, push back. I think we’re all trying to make this work, clearly. So I think, like, just being bluntly honest with us, like, just call a spade a spade. That’s not the issue. I think the issue is, like, we’re frustrated because we’re 70% done, and then you’re frustrated because we’re sending you on these goose chases to do all this new shit, and then we’re all butting heads for…
229 00:34:44.270 ⇒ 00:34:59.570 zacfromson: stupid reasons when, like, we all should just be on the same page. I think getting you guys to understand where we’re at, understand all the bugs, get all these things in the done column, and then working forward is fine. New things that come up, if we want to reprioritize them, that’s fine, but it’s going to come at a cost of X, Y, and Z, and just being aligned on that will
230 00:34:59.570 ⇒ 00:35:08.810 zacfromson: I think synthesize and fix all the issues that we have. Like, we’ve made good progress, but nothing is done, and we just said, like, let’s just take the next few weeks to get what we have started finished.
231 00:35:08.820 ⇒ 00:35:10.799 zacfromson: And then we can work on net new from there.
232 00:35:10.820 ⇒ 00:35:15.649 Clarence Stone: Not next few weeks, the next week, they’re gonna do it.
233 00:35:15.880 ⇒ 00:35:16.900 Clarence Stone: Right. Okay.
234 00:35:17.240 ⇒ 00:35:20.160 Clarence Stone: We have to get things sorted. And number two.
235 00:35:20.160 ⇒ 00:35:44.280 Clarence Stone: I think it’d be cool if you guys also shared the vision that you want to work towards, right? I know there’s a tab of priorities, but Bobby, like, what you’re putting together is really cool, and I think the team really needs to know where you want to take this, right? So as you’re seeing cool things, new solutions, and items that you’re kind of getting inspired by, drop it for the team so that they actually know where you want to progress this platform, because if you notice Pranav’s note.
236 00:35:44.280 ⇒ 00:35:51.920 Clarence Stone: Like, the platforms and the solutions and the technology that we pick, that’s always going to be driven by where you want to take this.
237 00:35:51.920 ⇒ 00:36:07.220 Clarence Stone: Right? So it’s okay to throw new things at us and say, this is what’s top of mind, but we’re still gonna go with, you know, your prioritization list otherwise, right? And, like, I don’t want it to be like you can’t share your vision or what you want to do with us, right? That context is important.
238 00:36:07.760 ⇒ 00:36:09.580 Bobby Palmieri: Totally. Yeah, I’m all in there.
239 00:36:09.920 ⇒ 00:36:11.389 Clarence Stone: Cool. Awesome.
240 00:36:12.530 ⇒ 00:36:15.650 Clarence Stone: Gents, I hate to do this, but I have another call.
241 00:36:15.830 ⇒ 00:36:17.330 zacfromson: I think we’re… I think we’re good, we’re to go.
242 00:36:17.330 ⇒ 00:36:18.069 Clarence Stone: That’s okay.
243 00:36:18.260 ⇒ 00:36:19.460 zacfromson: Yeah, I think we’re at a good point.
244 00:36:19.620 ⇒ 00:36:23.209 Bobby Palmieri: Yeah, I guess just to round this out, and you can jump, Clarence, is just, like.
245 00:36:23.820 ⇒ 00:36:25.520 Bobby Palmieri: I just want to make sure that we’re…
246 00:36:25.660 ⇒ 00:36:29.280 Bobby Palmieri: we’re working on the priorities that we’ve outlined. I think.
247 00:36:29.280 ⇒ 00:36:29.760 zacfromson: Yeah.
248 00:36:29.760 ⇒ 00:36:34.789 Bobby Palmieri: Chatted last week that, that… Up until that point, it was my fault.
249 00:36:35.210 ⇒ 00:36:44.590 Bobby Palmieri: And if we’re being… and, like, I think this week we were hoping that we would be better aligned on what those priorities are, and I think we were still drifting away. So what’s needed?
250 00:36:44.590 ⇒ 00:36:58.509 zacfromson: getting a call recap and a plan for next week in writing, so we’re all synthesized and aligned when we meet next week, we’re not like, we said this, he said this on this call. It’s like, it’s written, it’s done, we can agree, say a checkbox in Slack, we’re on the same page, and then…
251 00:36:58.510 ⇒ 00:37:04.959 zacfromson: we know what’s gonna happen, I think just, like, have it in writing, have it done, we have a good plan, we’re not… we know what we’re expecting, you guys know what you’re doing.
252 00:37:04.960 ⇒ 00:37:08.510 zacfromson: Our teams are fully one-to-one, eye-to-eye aligned, and…
253 00:37:08.510 ⇒ 00:37:26.760 zacfromson: I think we’ll be in a good spot, and I think we could run that as a sprint each week, like, here’s the game plan for the week, here’s what to expect on Friday, then we can do that, and then if you guys get blockers, problems, issues, we can get those discovered and explained to us on Friday, and then reset for the following week. I think just understanding what our weekly sprint is going to be will keep us in a really good spot going forward.
254 00:37:27.040 ⇒ 00:37:30.109 Clarence Stone: Absolutely. We’re absolutely on the same page on that one, guys.
255 00:37:30.110 ⇒ 00:37:36.570 zacfromson: Cool. Alright, we can let you hop. I know you’re quite a bit over. I think we’re all fairly aligned, and I think we’re all feeling… hopefully feeling better,
256 00:37:36.740 ⇒ 00:37:37.589 zacfromson: But maybe not.
257 00:37:37.590 ⇒ 00:37:49.300 Clarence Stone: It got my attention, so if there’s any time you guys are stuck, I mean, I’ll have them throw me into that chat, too. Feel free to ping me if you run into some AI errors, or just really thinking about, like, strategy overall. I’m happy to answer questions related to that, too.
258 00:37:49.550 ⇒ 00:37:50.170 Bobby Palmieri: Awesome.
259 00:37:50.410 ⇒ 00:37:52.799 zacfromson: Alright, appreciate it. My dad was recurse.
260 00:37:57.670 ⇒ 00:38:08.690 Pranav Narahari: Cool, guys. Is there anything else that we wanted to talk about? There was one thing I was gonna mention, but don’t know if we want to talk about it now or Monday, up to y’all. I’m actually out Monday, so it’d be Tuesday.
261 00:38:09.070 ⇒ 00:38:11.630 Pranav Narahari: Yeah.
262 00:38:11.630 ⇒ 00:38:16.590 Bobby Palmieri: I guess the conversation didn’t start out great. Like, Pranav, I guess, like.
263 00:38:17.170 ⇒ 00:38:19.789 Bobby Palmieri: We’re very confident in, like, what you’re…
264 00:38:19.830 ⇒ 00:38:36.959 Bobby Palmieri: building for us. I think the frustration is… is the… the alignment on that, and like, I just want to kind of clear the air there. Like, I think we’re… you know, the call on Friday, I thought we were, you know, fully aligned there. So, you know, I think we’re… up until that point, you know, we’ve been…
265 00:38:36.960 ⇒ 00:38:45.749 Bobby Palmieri: super happy with what you’ve built for us. I think it’s just, like, we just need you to… to guide us better on, like, getting these in the done column, because
266 00:38:46.360 ⇒ 00:38:50.299 Bobby Palmieri: We… we see the… the work, but, like.
267 00:38:50.330 ⇒ 00:38:53.469 Pranav Narahari: our team isn’t able to use it. Totally. Right?
268 00:38:53.470 ⇒ 00:39:01.189 zacfromson: It’s funny, like, we’re in your shoes at other times. Yeah, we came in a little hot, I think we’re frustrated, but, like, we are also agency-side and client side, and we have client.
269 00:39:01.190 ⇒ 00:39:01.740 Pranav Narahari: 100%.
270 00:39:01.740 ⇒ 00:39:09.259 zacfromson: So, we’re in your shoes often, and, like, I… you know, it’s our… now we’re being, you know, the client that’s frustrated, and, like, you know, we deal with that on a daily basis.
271 00:39:09.260 ⇒ 00:39:09.710 Pranav Narahari: And that’s how.
272 00:39:09.710 ⇒ 00:39:10.710 zacfromson: Yeah, I should be.
273 00:39:10.710 ⇒ 00:39:21.219 Pranav Narahari: Yes, and so one thing I wanted to just kind of, like, why I want to stay on the call right now, if we can, like, for another 20-ish minutes, is, like, earlier this week, I did kind of come back to you guys with, like, that
274 00:39:21.470 ⇒ 00:39:28.899 Pranav Narahari: comments on priority lists, like, so Bobby, on Monday, or maybe the weekend, like, you sent over the message about, like, hey, we took
275 00:39:28.900 ⇒ 00:39:44.260 Pranav Narahari: like, the feedback from that Friday meeting, that 15-minute call with Utam, got that list out, and here you guys go. And so, I looked at that, talked to Sam, talked to Casey, tried estimating things based on, like, the feedback, or based on the descriptions you guys gave.
276 00:39:44.690 ⇒ 00:39:47.640 Pranav Narahari: I just felt like there were certain decisions made there that I was like.
277 00:39:47.820 ⇒ 00:39:51.000 Pranav Narahari: I don’t understand exactly why you guys need this, and it’s almost like…
278 00:39:51.260 ⇒ 00:40:01.950 Pranav Narahari: I want to, like Clarence was saying, like, where are we trying to go with this product? Let me help guide you guys there a little bit more. Zach, like, you brought up a good point, too, like.
279 00:40:02.050 ⇒ 00:40:16.490 Pranav Narahari: I think this is something I’ve learned in the last couple weeks, like, you guys have so many ideas, I need to guys kind of let you guys know, like, okay, let’s focus on this. Instead of just dropping it at 70% and starting back at 0% on a new feature, let’s bring the 70% to 100, so then we can just, like, move on.
280 00:40:16.490 ⇒ 00:40:26.050 zacfromson: That was my biggest complaint, I guess, internally here with Bobby, too, to be like, hey, like, you know, you gotta chill, because I think you’re derailing it, like, you know, just something we’ve even discussed internally. So, I don’t disagree with you, and it’s like.
281 00:40:26.050 ⇒ 00:40:26.440 Pranav Narahari: consideration.
282 00:40:26.440 ⇒ 00:40:34.870 zacfromson: We’re pretty blunt, so you can just call a spade a fucking spade with me, and you’re never gonna hurt my feelings, so fuck off, no, like…
283 00:40:35.470 ⇒ 00:40:45.630 zacfromson: Like, this is what we need to be doing, like, push us, like, yes, we’re hungry, we work hard, we push hard, and we’re gonna push you a little bit, and, you know, obviously trying to do that respectfully, but.
284 00:40:45.630 ⇒ 00:40:46.719 Pranav Narahari: Oh, and I love that, too.
285 00:40:46.720 ⇒ 00:40:50.799 zacfromson: It’s fine to push… it’s fine to push back on us, and like, yeah, like, don’t be scared. I think the worst…
286 00:40:51.030 ⇒ 00:41:02.600 zacfromson: the worst thing to do is be scared to man adjusts back, and then, like, be like, I don’t understand why the hell you guys want to do this, or this doesn’t make any sense, or this is gonna fuck up my whole week, like, let’s not do this. And then we’re like, okay, cool, sounds good for Nob, like…
287 00:41:02.660 ⇒ 00:41:11.360 zacfromson: let’s fix this and get on the same page, and I think that right now, let’s just clean up what we have. Like, I think if we can get all the MCPs working, get forecasting going with a bit of a model.
288 00:41:11.490 ⇒ 00:41:12.090 Pranav Narahari: And I want to…
289 00:41:12.090 ⇒ 00:41:30.559 zacfromson: behind it as well, like, then we’re all gonna be in a good spot. I think, like, that will solve so many business use cases for us, then we can improve it from there. Like, you know, getting Claude’s skills for prompts will be a plus one against that to make it a little bit better, but it, like, you know, I think, like, those are the things that we can be using effectively now to layer on top of, so…
290 00:41:30.580 ⇒ 00:41:47.029 zacfromson: I think we’re… we have a lot of ideas of how we want to automate the business, and I think some people probably come to you guys and be like, hey, we don’t really know what we want. We want to use AI. We have a shit ton of ideas, and like, a lot of it isn’t, you know, I know, you guys are probably seeing this on X. These aren’t really things we’re even seeing on X. We’re investigating our own business.
291 00:41:47.030 ⇒ 00:41:53.299 zacfromson: and we’re like, hey, like, these are problems that we have, or things that we’re missing. Like, how can, like, the MCP, and then with skills.
292 00:41:53.340 ⇒ 00:42:06.450 zacfromson: basically clone… I think the problem is, like, you know, we’re almost 80 people here at our agency, and, like, as we’re doing that, like, me and Bobby’s brain is shrinking, so it’s like, these clawed skills are me and Bobby’s brain that we can then give to every employee, right? And that’s why they’re critical.
293 00:42:06.450 ⇒ 00:42:06.960 Pranav Narahari: do.
294 00:42:06.960 ⇒ 00:42:08.119 zacfromson: We can give you that contact.
295 00:42:08.120 ⇒ 00:42:13.469 Pranav Narahari: you’re saying, like, is completely true, like, about how you guys are doing some of, like.
296 00:42:13.610 ⇒ 00:42:23.439 Pranav Narahari: the most innovative stuff all in one project, and so it is, like, really, like, frontier stuff that we’re working on here, like, for you guys, which is exciting. It’s obviously, it’s, like.
297 00:42:23.720 ⇒ 00:42:30.879 Pranav Narahari: It’s difficult, and it’s going to be, like, a lot of stuff, but it’s exciting for us too. So, I’m very passionate about it as well.
298 00:42:31.330 ⇒ 00:42:31.789 zacfromson: I think that’s.
299 00:42:31.790 ⇒ 00:42:32.620 Pranav Narahari: Get more.
300 00:42:32.620 ⇒ 00:42:38.059 zacfromson: A disruption for us. I, you know, I guess for us, we’re like, we just want this stuff, and we want to get it, like, I guess you, like.
301 00:42:38.060 ⇒ 00:42:38.430 Pranav Narahari: Totally.
302 00:42:38.430 ⇒ 00:42:51.730 zacfromson: term that hasn’t really been used is, like, frontier, and honestly, I didn’t even know what you guys meant until maybe clicking now. Like, no one’s doing this, I guess is what you guys are saying, like, this is new investigative stuff, where it’s like, if we wanted something built differently, that there’s, like, a playbook for this, but there’s, like, not…
303 00:42:51.850 ⇒ 00:43:01.699 zacfromson: as many… there’s people that can do this, and there’s people that are building this stuff, but I think, to your point, like, we have so many errhenius asks that are very inconsistent in terms of their,
304 00:43:01.860 ⇒ 00:43:21.459 zacfromson: a fluency together that’s kind of leaving you guys, and like, we have to do all this R&D investigation to stand this up, because there isn’t documentation, playbooks, and 20 different products out there that are already doing it, and that’s what’s causing you guys the most problems in speed to development, because you don’t have any idea what you’re doing, like, respectfully, to some extent, because it hasn’t been done before. You’re like, we gotta figure out how to do this stuff,
305 00:43:21.460 ⇒ 00:43:41.439 Pranav Narahari: A structural thing that I think would be great for us, just for on a weekly basis, weekly cadence. And we discussed this on that Friday call last week, was just like, let’s have a sync on the Monday, Tuesday beginning of the week to just align on what we want to do for the upcoming week. That day, bring all the ideas in, and then I will kind of help with direction on, like.
306 00:43:42.200 ⇒ 00:43:53.390 Pranav Narahari: because there’s a few things there, like… and Bob, you started this, but we didn’t get to finish it, like, let’s look at that, sheet again. Like, where you’re showing the priorities. Let’s go through a couple of them if we can right now.
307 00:43:53.390 ⇒ 00:44:01.110 zacfromson: And I think we’re all for, like, a Monday… a Monday. I know this week you’re at Monday, but, like, in future weeks, like, a Monday morning stand-up is cool for us. We do that.
308 00:44:01.110 ⇒ 00:44:01.870 Pranav Narahari: It needs to be really.
309 00:44:01.870 ⇒ 00:44:02.190 zacfromson: Interesting.
310 00:44:02.190 ⇒ 00:44:06.119 Pranav Narahari: three of us. But let’s go to, like, forecasting feedback.
311 00:44:10.220 ⇒ 00:44:18.470 Pranav Narahari: And so, this area right here, Meta, Google, Klaviyo, yeah, you might notice the theme is looking a little bit different there.
312 00:44:18.950 ⇒ 00:44:26.959 Pranav Narahari: So, yeah, let’s scroll down a little bit, the Meta Google Klaviyo. Now, this is the area where I wanted to talk to you a little bit about how
313 00:44:27.120 ⇒ 00:44:32.709 Pranav Narahari: Setting up these connections is going to be paramount for actually standing up this forecasting feedback page.
314 00:44:32.770 ⇒ 00:44:52.270 Pranav Narahari: So, even Zach, like, when you were discussing, like, polyatomic, all these connections not working with, like, the data and stuff, the MCPs and Polytomic, or Airbyte, whatever we end up using as the ETL tool, they’re existing separately. So the MCPs aren’t going through, like, Polytomic, AirByte, etc. The MCPs are separate of that.
315 00:44:52.430 ⇒ 00:44:57.730 Pranav Narahari: So… with Klaviyo, like, having issues, that has nothing to do with, polyatomic.
316 00:44:57.730 ⇒ 00:44:59.159 Bobby Palmieri: Yeah, yeah, yeah.
317 00:44:59.160 ⇒ 00:45:09.970 Pranav Narahari: So, I just noticed there was, like, some confusion there, too. You guys have, like, a lot of these ideas, which are great. I just want to make sure, like, we’re… we understand, like, okay, developing this will result in Y, you know?
318 00:45:09.970 ⇒ 00:45:13.899 zacfromson: Yeah, and that might be my… I think Bobby has a better understanding of the data warehouse stuff than I do, so maybe.
319 00:45:13.900 ⇒ 00:45:21.539 Pranav Narahari: Yeah, and that’s totally fine, I just want to make sure to, you know, just another example, us just all being in sync on this stuff, like, what is going to affect what?
320 00:45:21.540 ⇒ 00:45:38.769 Pranav Narahari: Now this is a great example. MetaGoogle Klaviyo. Like, these are for any type of forecasting feature, for it to be done correctly, and like you guys said, not just for you two, right? You guys can kind of see, like, where this is building up to. The people at y’all’s agency probably don’t care as much if it’s not giving them the output that they need.
321 00:45:38.850 ⇒ 00:45:42.899 Pranav Narahari: And so, for a forecasting feedback, like, page.
322 00:45:43.140 ⇒ 00:45:55.750 Pranav Narahari: they need to… it needs to be connected to the data warehouse. It can’t be pulling in the information from an MCP, because it’s not going to be as reliable, and it’s going to break very often. The whole reason for a data warehouse is that it is not querying the data from the source.
323 00:45:56.250 ⇒ 00:46:00.929 Pranav Narahari: at will, when somebody goes to that, that page. Which is why, like…
324 00:46:01.840 ⇒ 00:46:05.620 Bobby Palmieri: Yeah, I’m fully aligned on that. I guess my…
325 00:46:05.620 ⇒ 00:46:25.170 Pranav Narahari: That’s the reason why we can’t work on this yet. That’s why this is, like, blocked. We haven’t been able to set up the connection with Meta, we haven’t been able to set up the connection with Google, we only have Shopify set up, and to discuss that a little bit further, like, yeah, AirByte, we were running into that block. We first initially tried with AirByte because you guys were, like.
326 00:46:25.240 ⇒ 00:46:39.620 Pranav Narahari: saying, like, okay, our competitors are using this, or maybe other people in your space were using this for, like, similar type of, like, data needs. And so we’re like, okay, you guys are pretty sure on this. We haven’t used Air… or we’ve tried AirByte in the past,
327 00:46:39.710 ⇒ 00:46:57.570 Pranav Narahari: we’ve… our preference has always been polyatomic for these type of things, but we first tried AirByte. When we tried to pull in the data, we were having a ton of issues. We went straight to Mother Duck, who we’re using as our data warehouse provider. They were saying, yeah, this is, like, a… an issue that’s been open since November of last year.
328 00:46:57.860 ⇒ 00:47:14.180 Pranav Narahari: for getting transaction data being queryable for, an application. And so, we realized with them, like, they’re not gonna… they’re not gonna push in the matter of days, right? Like, they have no business doing that for us. And so, it’s like, we brought this up, we were like, okay.
329 00:47:14.360 ⇒ 00:47:23.150 Pranav Narahari: let’s do it… let’s do it on Polytomic. Let’s just, for POC reasons, let’s just do it on Polyatomic, and then we present it to you guys, okay, it actually worked. Worked great.
330 00:47:23.500 ⇒ 00:47:37.660 Pranav Narahari: now, you know, if you guys really want us to use AirByte, we can go that route. We can also go the route, as well, of, like, supporting both. Like, if you guys only want to have all connections in the future run through AirByte.
331 00:47:37.850 ⇒ 00:47:41.309 Pranav Narahari: We could do that, but… I would just…
332 00:47:41.910 ⇒ 00:47:45.300 Pranav Narahari: I would just wonder why, you know, like, if… wouldn’t it just be a better…
333 00:47:45.300 ⇒ 00:47:48.480 zacfromson: I think we’re here. It is one tool.
334 00:47:48.480 ⇒ 00:47:50.520 Bobby Palmieri: I can see it on the Polycom rack.
335 00:47:50.520 ⇒ 00:47:51.330 Pranav Narahari: Okay, great.
336 00:47:51.330 ⇒ 00:47:54.390 zacfromson: You don’t have to hear about air… we don’t have to hear… we don’t have to hear about air bite again.
337 00:47:54.400 ⇒ 00:47:55.280 Bobby Palmieri: Perfect.
338 00:47:55.280 ⇒ 00:47:55.690 Pranav Narahari: Perfect.
339 00:47:55.690 ⇒ 00:47:58.339 Bobby Palmieri: I think my… my general, like.
340 00:47:58.710 ⇒ 00:48:04.130 Bobby Palmieri: frustration last night, and I apologize, I could have gone for a walk and reformatted that message.
341 00:48:04.390 ⇒ 00:48:05.680 Bobby Palmieri: the…
342 00:48:06.850 ⇒ 00:48:12.590 Bobby Palmieri: I… I totally get that we need this to come from the data warehouse, which is why that is…
343 00:48:12.880 ⇒ 00:48:18.650 Bobby Palmieri: the first… feature on this. Like, I get that that’s the blocker.
344 00:48:20.110 ⇒ 00:48:25.470 Bobby Palmieri: I read your message, and maybe I misunderstood it, as, like, this…
345 00:48:26.990 ⇒ 00:48:35.549 Bobby Palmieri: Feedback was, like, a huge deviation from the forecast, because There’s nothing in here…
346 00:48:36.100 ⇒ 00:48:41.829 Bobby Palmieri: that is different than what we already reviewed, and transparently, I thought that this feedback was actually
347 00:48:42.070 ⇒ 00:48:48.500 Bobby Palmieri: really light. I think when I went through it, me and you got on a call two Fridays ago, we looked at it.
348 00:48:48.730 ⇒ 00:48:54.300 Bobby Palmieri: Like, it’s there. Like, there’s no… we had some questions on, like, the logic here.
349 00:48:54.300 ⇒ 00:49:02.350 Pranav Narahari: I think I have the message you’re referring to, right? Like, I said, there’s small and large lifts that are, presented in this document, and the large lifts are…
350 00:49:04.250 ⇒ 00:49:05.180 Bobby Palmieri: Yeah, go ahead.
351 00:49:05.180 ⇒ 00:49:07.359 Pranav Narahari: The large lifts are essentially, like.
352 00:49:07.370 ⇒ 00:49:19.620 Pranav Narahari: there all… there has to be this whole foundation of data that’s coming through the data warehouse that we’ve been kind of blocked on, and I’ve been trying to, like, get y’all to push a little bit more on that with, like, setting up the connections, getting us some time to, like.
353 00:49:19.620 ⇒ 00:49:33.989 Pranav Narahari: figure out what is the issue you guys are running with. I really thought before this whole Airbuy situation that you guys were good at Polytomic. Like, you know, like, you’re happy to, like, have things working, you know? I didn’t think… I think you guys were basically, like, tool agnostic, like, didn’t really…
354 00:49:33.990 ⇒ 00:49:37.709 Bobby Palmieri: Until… until we had issues with Polytomic getting connected to…
355 00:49:38.060 ⇒ 00:49:46.150 Bobby Palmieri: meta, which I think… I think I was able to get resolved last night. I don’t know if it was me or them, but I think I have one account pulled up there.
356 00:49:46.150 ⇒ 00:49:57.880 zacfromson: things that aren’t working, I think, just got us frustrated. And I think that’s where that all came out of left field from. So yes, like, we aren’t… we don’t necessarily care about the tooling, you know, we’re here for you guys to advise us. I think we’re, like.
357 00:49:58.650 ⇒ 00:50:00.019 zacfromson: or this isn’t working.
358 00:50:00.020 ⇒ 00:50:06.290 Pranav Narahari: Let me know, because I can help a lot with those type of things, whether it’s, like, with, some…
359 00:50:06.290 ⇒ 00:50:15.809 zacfromson: We can go about that differently. We can go about the conversation differently. Yes, I think, you know, Bobby came in hot. I read the message, I was like, send it. I also was having a pretty bad day, so, you guys can.
360 00:50:15.810 ⇒ 00:50:16.740 Pranav Narahari: How good it happens.
361 00:50:16.740 ⇒ 00:50:18.190 zacfromson: It’s like,
362 00:50:18.190 ⇒ 00:50:37.720 zacfromson: I know Udom is overly pissed at us, but it is what it is. I, you know, we’re in his shoes often. I know he was not happy to speak to us today, obviously. So, yeah, we came in hot, and that’s, you know, I think, you know, all of our bad. We just need to do a better job of communicating, quite frankly, and I think we’ll be in a better spot, and some of it is on us for just being a poor communicator yesterday, which I think obviously just threw
363 00:50:37.720 ⇒ 00:50:41.240 zacfromson: more fuel in the fire here, and I think hearing some of that back from you as well, like.
364 00:50:41.240 ⇒ 00:50:53.519 zacfromson: maybe I’m not as privy to some of the conversations you and Bobby have without me, but, like, even just hearing, hey, like, this is some of the things we’re running into, like, that’s helpful context that, at least personally I didn’t have, that’s giving me more clarity on what’s actually happening.
365 00:50:53.520 ⇒ 00:50:53.930 Pranav Narahari: Yeah.
366 00:50:53.930 ⇒ 00:51:02.959 zacfromson: Which is like, okay, I get it, like, before, I’m just like, I don’t get it. I just didn’t get what was going on. I just hear, like, it’s not working, and I’m like, that’s not a really great answer for me. Like, I don’t get this.
367 00:51:02.960 ⇒ 00:51:03.330 Bobby Palmieri: Hmm.
368 00:51:03.330 ⇒ 00:51:21.930 Pranav Narahari: Totally. And, like, part of my message, like, on Tuesday, which was, like, let’s sync. Let’s, like, talk about these priorities so I can kind of discuss this not now, but back then. We… I think what we need to do in the future, which will just, like, save us a ton of headache, save us just, like, a ton of time, too, and just have us moving more efficiently, let’s have that sink, you know?
369 00:51:21.930 ⇒ 00:51:38.480 zacfromson: Yeah, we start getting into heavy client calls midweek, so if we can do those on Monday, I know this week’s tough, and we can slip it in on Tuesday, but if we can do them on Monday, you’re gonna get a lot more of our time, our attention. Like, you’re not gonna hear from Bobby ever on a Wednesday. He’s booked from 8am until 6pm.
370 00:51:38.480 ⇒ 00:51:45.869 zacfromson: through and through, like, he hopes a call ends early so you can even just take a pass. But I think, like, with that in mind, like, you’re not gonna really get a lot from him.
371 00:51:45.890 ⇒ 00:51:57.150 zacfromson: on a Wednesday, for example. So if we can get a line for the start of the week, we’re not burning cycles. So if we’re not meeting until Tuesday at 1, we just burn a day and a half on our cycle, right, for you guys to be productive throughout the week. And I’m sure, like.
372 00:51:57.160 ⇒ 00:52:12.659 zacfromson: this isn’t your… the only thing you guys are working on, right? So it’s like you’re trying to balance out your hours throughout your days, right? So I think if we can be aligned for Monday, it gives you the most possible time to be productive throughout the week, and probably is the most free day to get Bobby’s attention outside of Friday, which is, I think, a good time for us to close out the week.
373 00:52:14.090 ⇒ 00:52:15.319 Bobby Palmieri: The one question.
374 00:52:15.320 ⇒ 00:52:16.170 zacfromson: But we’ll get…
375 00:52:16.510 ⇒ 00:52:17.249 Bobby Palmieri: Yeah, go ahead.
376 00:52:18.080 ⇒ 00:52:22.860 zacfromson: No, I think, like, after… after next week, we’ll… we can just get a stand-up on Mondays, which will get us at a good…
377 00:52:23.160 ⇒ 00:52:24.190 zacfromson: Sync cycle.
378 00:52:24.550 ⇒ 00:52:30.579 Bobby Palmieri: What is… I guess the one… the one that I do want to just clarify to make sure that we are on the same…
379 00:52:31.060 ⇒ 00:52:32.619 Bobby Palmieri: page, which I think…
380 00:52:33.250 ⇒ 00:52:44.199 Bobby Palmieri: still has me a little bit worried is, like you said, important context. These capabilities weren’t in our original 8-week scope of work. Spent the past 2 days deep scoping to understand.
381 00:52:44.560 ⇒ 00:52:49.490 Bobby Palmieri: the… Data flow dependencies from warehouse to forecasting.
382 00:52:50.530 ⇒ 00:52:53.300 Bobby Palmieri: I think we’re on the same page that those were…
383 00:52:54.050 ⇒ 00:52:56.680 Bobby Palmieri: Like, that was the understanding to start, no?
384 00:52:57.910 ⇒ 00:53:04.709 Pranav Narahari: Yeah, so… let’s bring in, like, Utam for talking about, like, things like SOW and stuff like that.
385 00:53:04.710 ⇒ 00:53:09.240 Bobby Palmieri: That’s fine if we wanna… like, I’m not… I guess just from your end, like.
386 00:53:09.890 ⇒ 00:53:15.650 Bobby Palmieri: This is not… this is what you’ve been building, though, right? Like, we didn’t… this feedback doesn’t…
387 00:53:16.670 ⇒ 00:53:20.619 Bobby Palmieri: It doesn’t move in a different direction than what you’ve been building.
388 00:53:21.120 ⇒ 00:53:29.300 Pranav Narahari: No, I think… but from our just conversation from, like, last Friday, we just discussed, like, how there’s a bunch of ad hoc things that we are saying yes to.
389 00:53:29.440 ⇒ 00:53:32.739 Pranav Narahari: And so, going forward, like, I mean, I know what you guys want. That’s not…
390 00:53:32.740 ⇒ 00:53:37.909 Bobby Palmieri: No, I guess… yeah, but what I’m… I guess what I’m trying to say is, like, yes, I… throw all the ad hoc shit to…
391 00:53:37.910 ⇒ 00:53:40.490 zacfromson: The scope of work is all fucked up because we did a bunch of other.
392 00:53:40.490 ⇒ 00:53:41.230 Bobby Palmieri: Yeah, and I’m not worried.
393 00:53:41.230 ⇒ 00:53:53.499 zacfromson: So I think from an hour’s perspective, but I think Bobby’s big question is, like, there’s clarity on this project, regardless of scope of work moving forward. Like, there’s not, like, oh crap, I didn’t understand that you didn’t expect this. Like, this is still the same expectation.
394 00:53:53.500 ⇒ 00:53:54.019 Pranav Narahari: Oh, you legit.
395 00:53:54.020 ⇒ 00:53:55.200 zacfromson: our outlook.
396 00:53:55.320 ⇒ 00:53:56.040 zacfromson: Okay.
397 00:53:56.240 ⇒ 00:53:59.160 Pranav Narahari: Yeah, so, let’s go back to that Stitch backlog.
398 00:53:59.560 ⇒ 00:53:59.980 Bobby Palmieri: Yeah.
399 00:53:59.980 ⇒ 00:54:05.460 Pranav Narahari: And… let’s go to knowledge bases. We kind of discussed knowledge bases, but I want to know…
400 00:54:05.760 ⇒ 00:54:08.679 Pranav Narahari: Knowledge bases, how I think of it, is it’s a solution.
401 00:54:09.100 ⇒ 00:54:13.019 Pranav Narahari: Let’s talk about, like, what is the problem that you’re seeing? What is the problem that you feel like.
402 00:54:13.250 ⇒ 00:54:13.800 Bobby Palmieri: Your people are.
403 00:54:13.800 ⇒ 00:54:26.619 Pranav Narahari: agency are having, and why do you think knowledge bases are gonna solve it? Because I don’t know if it’s the right solution for you guys. That’s my thing. Like, do you just want a data store of, like, files that you can use as additional context, or do you actually want a true knowledge base that
404 00:54:26.770 ⇒ 00:54:35.739 Pranav Narahari: that’s a pretty big lift. We’re creating embeddings for a various type of media, and then you’re having it used as, like, a source of truth.
405 00:54:35.740 ⇒ 00:54:38.910 Bobby Palmieri: Who… I guess the…
406 00:54:40.020 ⇒ 00:54:47.539 Bobby Palmieri: We can… we can chat about knowledge base, but it’s lower on our priority list, so I do want to just make sure that forecasting
407 00:54:47.820 ⇒ 00:54:50.299 Bobby Palmieri: like, we’re aligned on that. Was there…
408 00:54:50.830 ⇒ 00:54:57.230 Bobby Palmieri: questions and things within this. I know Meta is a blocker, Google’s a blocker. I can’t.
409 00:54:57.230 ⇒ 00:54:57.870 Pranav Narahari: Yeah.
410 00:54:57.870 ⇒ 00:55:05.390 Bobby Palmieri: I actually can’t see it anymore, because I think Mother Duck’s free trial is over. Yeah. But… Like, I…
411 00:55:05.780 ⇒ 00:55:07.409 Bobby Palmieri: I sent this.
412 00:55:08.420 ⇒ 00:55:10.399 Bobby Palmieri: Being like, oh shit, we’re like…
413 00:55:10.710 ⇒ 00:55:13.750 Bobby Palmieri: Pretty close to this being done.
414 00:55:13.910 ⇒ 00:55:23.540 Bobby Palmieri: Like, I thought this feedback was like, hey, Pranav, you got a 95 on the test, like, here’s a few things we need to clean up, and obviously need to get the data warehouse.
415 00:55:23.540 ⇒ 00:55:24.140 Pranav Narahari: Right.
416 00:55:24.440 ⇒ 00:55:27.690 Bobby Palmieri: But, like, I think this is, like, when we align, like, you’ve built.
417 00:55:27.690 ⇒ 00:55:44.869 Pranav Narahari: The missing component is just that middle spot, like, where we see, like, the four… the three different other connections of data, and that is just kind of being blocked, not really by me, like, I’m… I need to get that connection from either Polytomic or AirByte, and that’s kind of something we’ve been discussing for the last couple weeks.
418 00:55:44.870 ⇒ 00:55:49.239 Bobby Palmieri: But you knew that was… you knew that was part of the forecasting tool, right?
419 00:55:49.870 ⇒ 00:55:50.779 Pranav Narahari: Yes, yeah, yeah, yeah.
420 00:55:50.780 ⇒ 00:55:52.080 Bobby Palmieri: Okay, okay, cool.
421 00:55:52.080 ⇒ 00:55:59.669 Pranav Narahari: From our discussions, not from the SOW, just, like, I know that’s something you guys want, based on just, like, what you showed in the Replet.
422 00:55:59.670 ⇒ 00:56:00.050 Bobby Palmieri: Okay.
423 00:56:02.120 ⇒ 00:56:13.869 Pranav Narahari: Yeah, going forward, like, I know… I think for that POC, what we discussed was just, like, okay, let’s just bring the Shopify data, and then we’ll have columns for Meta, and… or ads in general, Meta and Google.
424 00:56:13.920 ⇒ 00:56:25.320 Pranav Narahari: And then now, from that feedback, I was like, okay, this makes sense. I know what you’re referring to when you say Meta, Google, Klaviyo, because, like, I’ve seen, like, in that Replit, like, there’s additional tabs.
425 00:56:25.550 ⇒ 00:56:27.549 Pranav Narahari: But it’s not something we can execute on right now.
426 00:56:29.020 ⇒ 00:56:32.809 Bobby Palmieri: Because we’re blocked on it. But… That’s a problem.
427 00:56:32.810 ⇒ 00:56:34.130 zacfromson: polyatomic issue.
428 00:56:34.130 ⇒ 00:56:34.640 Bobby Palmieri: Yeah.
429 00:56:34.640 ⇒ 00:56:37.410 zacfromson: So we need them… we need their support to fix this.
430 00:56:37.710 ⇒ 00:56:43.669 Pranav Narahari: Yeah, and let us help with that, you know? Like, if you’re having issues with that, it’s like, we’ve made…
431 00:56:43.790 ⇒ 00:56:49.599 Pranav Narahari: dozens and dozens of connections with Polytomic. And when I say connections, I mean, like.
432 00:56:49.640 ⇒ 00:57:03.830 Pranav Narahari: source data connections, into Polytomic. So, like, we understand that. If I don’t know it, I have tons of contacts within Brainforge that have used Polytomic for an array of other projects to help with that. And so…
433 00:57:04.170 ⇒ 00:57:06.370 Pranav Narahari: Yeah, like, definitely, like, just, like.
434 00:57:06.670 ⇒ 00:57:09.090 Pranav Narahari: pull me in more for those type of things, just be like, hey.
435 00:57:09.090 ⇒ 00:57:12.869 Bobby Palmieri: Yeah, we did flag it, but I can lean on you more to get it resolved.
436 00:57:13.150 ⇒ 00:57:24.400 Pranav Narahari: Yeah, and pull me into huddles, you know? I don’t have as many meetings as you. If you have, like, a second, like, you know, you already went to the bathroom, and you don’t need that break for going to the bathroom again, hit me up, you know? Like, throw, like, just call me
437 00:57:24.730 ⇒ 00:57:27.400 Pranav Narahari: I’ll be able to pick it up more often than not.
438 00:57:27.400 ⇒ 00:57:33.180 Bobby Palmieri: Cool. The knowledge base is… you know, the MCP
439 00:57:33.370 ⇒ 00:57:39.450 Bobby Palmieri: right now lacks the brand knowledge, right? So… Tone of voice…
440 00:57:39.580 ⇒ 00:57:49.980 Bobby Palmieri: goals, top performing ads, right? So, like, if I go to Claude… We have a project for…
441 00:57:52.480 ⇒ 00:57:55.509 Pranav Narahari: The main thing that I would need to know is,
442 00:57:55.710 ⇒ 00:57:58.539 Pranav Narahari: How much data would be in this knowledge base?
443 00:57:58.970 ⇒ 00:57:59.580 Bobby Palmieri: Yeah, so what?
444 00:57:59.580 ⇒ 00:58:05.130 Pranav Narahari: Could it be, like, a few different files, like, 10? Or could it be hundreds? .
445 00:58:05.670 ⇒ 00:58:07.309 zacfromson: It’s not tons, like…
446 00:58:07.310 ⇒ 00:58:09.460 Pranav Narahari: Then, yeah, knowledge base is not what you need.
447 00:58:09.580 ⇒ 00:58:14.500 Pranav Narahari: knowledge base is not what you need. You kind of just need, like, additional files that you can just upload
448 00:58:14.990 ⇒ 00:58:15.880 Pranav Narahari: context.
449 00:58:15.880 ⇒ 00:58:20.270 zacfromson: They’re like, we’d love to upload, like, here’s, like, the voice brand document, like, here are, like.
450 00:58:20.640 ⇒ 00:58:29.579 zacfromson: 20 past emails and 10 past emails that we’ve written, so you can, like, see the tone of voice and the emotion of the brand, so I was like, we eventually get to the copywriting aspect of this, that that is…
451 00:58:29.580 ⇒ 00:58:41.929 zacfromson: that that is there. So a lot of this is, like, guidelines and materials that the MCP can learn the brand, because right now, I guess, like, it just scans the website and maybe online, and then makes its own interpretation, but a lot of brands have, like, very…
452 00:58:41.940 ⇒ 00:58:50.020 zacfromson: Finite ways that we need to speak on them, like detailed brand guidelines and words that they want to use, and we want to make sure it has that information that it likely wouldn’t know otherwise.
453 00:58:50.020 ⇒ 00:58:52.420 Pranav Narahari: Right. And we can do that without…
454 00:58:53.400 ⇒ 00:58:58.119 Pranav Narahari: I mean, I understand what you guys mean by knowledge base, we can call it a knowledge base within the app itself.
455 00:58:58.230 ⇒ 00:59:05.420 Pranav Narahari: But that’s where I needed more clarity. I was like, okay, do you guys have thousands of files that we need to create embeddings for and do a similarity search?
456 00:59:05.650 ⇒ 00:59:06.200 Pranav Narahari: all the.
457 00:59:06.200 ⇒ 00:59:06.820 zacfromson: No.
458 00:59:06.820 ⇒ 00:59:07.490 Pranav Narahari: Maybe jargon…
459 00:59:07.490 ⇒ 00:59:19.559 zacfromson: Basically, this is, yeah, it’s called the knowledge base here in Cloud, and why we’re using that term, but basically you can just upload a bunch of Word documents, PDFs, images, and it will, you know, Excel sheets, and it will, like, review and scan all of that data.
460 00:59:19.790 ⇒ 00:59:20.270 Bobby Palmieri: Mmm.
461 00:59:20.270 ⇒ 00:59:25.920 Pranav Narahari: They probably support, like, you being able to, like, have, like, thousands of documents in there.
462 00:59:25.920 ⇒ 00:59:27.589 zacfromson: We can, but we, you know, I think we’re using
463 00:59:27.910 ⇒ 00:59:30.359 zacfromson: of our capacity in most cases, so it’s like, yeah.
464 00:59:30.360 ⇒ 00:59:30.720 Bobby Palmieri: We’re good.
465 00:59:30.720 ⇒ 00:59:32.340 zacfromson: But we don’t need all of that.
466 00:59:33.180 ⇒ 00:59:41.460 Bobby Palmieri: is, like, winning ads, ad comments, a review list, you know, so, like, I could say, you know.
467 00:59:41.460 ⇒ 00:59:57.740 Bobby Palmieri: write me an email, for our fast motion launch, look at the 5-star reviews CSV, and pull in relevant, you know, reviews to feature in the email. And then it will… I don’t think it got everything that I said, but,
468 00:59:57.840 ⇒ 01:00:02.160 Bobby Palmieri: You know, it will pull in, like, you know, we’ll review that, we’ll use the
469 01:00:02.290 ⇒ 01:00:19.129 Bobby Palmieri: copywriter skill, it will use that CSV, and then it has that that it can pull from. So I think it’s, like, brand overview of, like, what the brand does, probably uploading, like, our CAC and, like, ROAS targets, reviews, etc.
470 01:00:20.850 ⇒ 01:00:21.480 Pranav Narahari: Perfect.
471 01:00:22.480 ⇒ 01:00:25.420 Pranav Narahari: I have to jump. I have a 230, but…
472 01:00:25.420 ⇒ 01:00:26.729 Bobby Palmieri: They didn’t have calls.
473 01:00:27.470 ⇒ 01:00:31.429 Pranav Narahari: No, no, I have calls today. I have calls today, unfortunately, yeah.
474 01:00:31.430 ⇒ 01:00:33.319 Bobby Palmieri: No, that works.
475 01:00:33.320 ⇒ 01:00:34.410 Pranav Narahari: But, yeah…
476 01:00:34.410 ⇒ 01:00:40.270 Bobby Palmieri: Let’s find the sink on… on Tuesday when you’re… when you’re back. I think the biggest thing is, is, like, we just want, like.
477 01:00:40.690 ⇒ 01:00:42.120 Bobby Palmieri: ruthless…
478 01:00:42.310 ⇒ 01:00:54.999 Bobby Palmieri: dedication to, like, whatever’s 1 and 2 until those get put in the done column. I know that, like, there’ll be blockers so we can go down the list, but, like, I think, you know, our big understanding from the call
479 01:00:55.020 ⇒ 01:01:06.049 Bobby Palmieri: last Friday was like, you guys just need to know what our top priorities are, and, like, now you… now you have that, hold us accountable to, like, not straying away from that.
480 01:01:06.210 ⇒ 01:01:14.220 Pranav Narahari: And yeah, there… I think we need to have a little bit more just conversation on that, just because we talked about MCPs too, we didn’t get to discuss that right now.
481 01:01:14.550 ⇒ 01:01:29.990 Pranav Narahari: we’ll have a little bit of conversation just about, like, okay, what exactly are the priorities? What is, like, at 70% still that we need to ship to 100 before we get started on, like, the rest of the data warehouse stuff? If that’s the case. If you want us to just, like, table all that stuff until we get the data warehouse and the forecasting stuff going, like.
482 01:01:29.990 ⇒ 01:01:35.750 Pranav Narahari: let’s just finish up that conversation on Tuesday, but overall, I agree with what you’re saying, Bobby.
483 01:01:35.750 ⇒ 01:01:41.250 Bobby Palmieri: The only other component that I would add to that, that I think Zach and I lack the context of is, like.
484 01:01:41.420 ⇒ 01:01:49.850 Bobby Palmieri: who’s working on this project? And, like, is it… is it multiple people? Is it just you? Does Casey, like, chime in from time to time? Like…
485 01:01:49.850 ⇒ 01:02:06.409 Bobby Palmieri: are there things that we can work in parallel, or is it truly one item at a time? But let’s… let’s chat through that on Tuesday, in that regard, because I do think there’s small things that, like, we do want to get done, like, you know, we need to be able to select the Google Ads account, like, it’s not the biggest priority, but it’s…
486 01:02:06.640 ⇒ 01:02:10.970 Bobby Palmieri: a short fix that will put that in the done column, right? I think we really need to.
487 01:02:10.970 ⇒ 01:02:16.189 Pranav Narahari: ideas on that, too. And so, just a short answer… like, the short answer on that is, like.
488 01:02:16.330 ⇒ 01:02:30.469 Pranav Narahari: Casey has some time to dedicate to this project, I’m going to be the main developer on this, and Sam is, like, the one that has the most technical acumen to, like, kind of help with, like, the big decisions where I feel like I need somebody to bounce him off of.
489 01:02:30.570 ⇒ 01:02:31.430 Pranav Narahari: Now it…
490 01:02:31.800 ⇒ 01:02:34.080 Bobby Palmieri: And… Yeah.
491 01:02:34.080 ⇒ 01:02:35.589 Pranav Narahari: more about that on Tuesday, then.
492 01:02:36.160 ⇒ 01:02:38.120 Pranav Narahari: Sounds good. Alright guys, see ya.
493 01:02:38.430 ⇒ 01:02:39.390 zacfromson: Tuesday. Bye.