Meeting Title: Brainforge x Micah Project Sync Date: 2025-09-19 Meeting participants: Matthew’s Circleback, Rico Rejoso, Uttam Kumaran, Mustafa Raja, Samuel Roberts, Justin Breshears, Matthew Good
WEBVTT
1 00:00:38.730 ⇒ 00:00:39.909 Uttam Kumaran: There we go.
2 00:00:41.470 ⇒ 00:00:42.340 Rico Rejoso: Hi, Tom.
3 00:00:42.660 ⇒ 00:00:43.340 Rico Rejoso: How was the first?
4 00:00:44.380 ⇒ 00:00:49.350 Uttam Kumaran: Good, sorry, just got delayed by, like, 30-40 minutes, and just, like.
5 00:00:49.560 ⇒ 00:00:53.790 Uttam Kumaran: thought I’d be able to make some of these meetings, but yeah, I’m just on the way home now, so…
6 00:00:54.810 ⇒ 00:00:55.759 Rico Rejoso: No worries.
7 00:00:57.080 ⇒ 00:00:58.690 Uttam Kumaran: Thanks for covering on stuff.
8 00:01:00.100 ⇒ 00:01:01.240 Rico Rejoso: Yep, got it.
9 00:01:02.060 ⇒ 00:01:04.220 Rico Rejoso: I mean, Sam and Mustafa was there with me.
10 00:01:04.260 ⇒ 00:01:05.639 Uttam Kumaran: For the clients.
11 00:01:06.840 ⇒ 00:01:07.730 Samuel Roberts: Hello?
12 00:01:08.610 ⇒ 00:01:09.430 Uttam Kumaran: Hey, guys.
13 00:01:12.580 ⇒ 00:01:13.150 Mustafa Raja: Ayy.
14 00:01:14.420 ⇒ 00:01:15.220 Uttam Kumaran: airing.
15 00:01:18.270 ⇒ 00:01:19.530 Samuel Roberts: Tom, where are you right now?
16 00:01:20.060 ⇒ 00:01:22.019 Uttam Kumaran: I just got to Austin.
17 00:01:22.220 ⇒ 00:01:22.880 Samuel Roberts: Okay.
18 00:01:23.310 ⇒ 00:01:24.070 Uttam Kumaran: Yeah.
19 00:01:25.900 ⇒ 00:01:27.090 Samuel Roberts: Like, still at the airport?
20 00:01:28.160 ⇒ 00:01:31.540 Uttam Kumaran: Yes, I just got to the… oh, I just got…
21 00:01:32.120 ⇒ 00:01:34.440 Uttam Kumaran: to the place where the Uber’s picking me up, so…
22 00:01:36.010 ⇒ 00:01:37.279 Samuel Roberts: How far are you?
23 00:01:37.930 ⇒ 00:01:41.830 Uttam Kumaran: I am… it’s like 10 minutes in my house.
24 00:01:41.830 ⇒ 00:01:43.390 Samuel Roberts: Oh, that’s so nice.
25 00:01:43.390 ⇒ 00:01:46.510 Uttam Kumaran: Yeah, that’s why I was also, like,
26 00:01:47.570 ⇒ 00:01:55.850 Uttam Kumaran: I thought it was just a flight got, like, 30 minutes delayed, and then just all my meetings, I was like, okay, go, I’ll get home really quick, and then take off.
27 00:01:55.850 ⇒ 00:01:56.390 Samuel Roberts: M.
28 00:01:56.570 ⇒ 00:01:58.319 Uttam Kumaran: Getting jammed, but that’s okay.
29 00:01:58.320 ⇒ 00:01:58.979 Samuel Roberts: I’m like, yeah.
30 00:02:02.600 ⇒ 00:02:03.940 Uttam Kumaran: How’s the day going?
31 00:02:04.500 ⇒ 00:02:05.080 Samuel Roberts: Well…
32 00:02:06.440 ⇒ 00:02:07.289 Matthew Good: What’s up, guys?
33 00:02:07.650 ⇒ 00:02:08.509 Samuel Roberts: Hello.
34 00:02:08.820 ⇒ 00:02:09.810 Uttam Kumaran: Yay!
35 00:02:10.259 ⇒ 00:02:11.159 Matthew Good: How’s it going?
36 00:02:11.440 ⇒ 00:02:12.780 Uttam Kumaran: Good, how are you?
37 00:02:12.780 ⇒ 00:02:15.710 Matthew Good: Good. How’s, are you back from Shop Talk? How’s the conference?
38 00:02:15.710 ⇒ 00:02:28.680 Uttam Kumaran: I am… I am on my way back right now. Flight got delayed, so I’m, like, about to get picked up by the Uber and getting out, but it was great, man. Like, a lot… huge range of…
39 00:02:28.790 ⇒ 00:02:38.140 Uttam Kumaran: e-com brands, like, both doing, like, retail and digital. We met with a lot of, a lot of folks.
40 00:02:38.310 ⇒ 00:02:43.269 Uttam Kumaran: And it’s, like, not like a tech conference, like, everybody’s very, very talkative, because they’re all, like, consumer brands.
41 00:02:43.270 ⇒ 00:02:45.530 Matthew Good: So… It was.
42 00:02:45.530 ⇒ 00:02:46.920 Samuel Roberts: Fire tech people.
43 00:02:47.220 ⇒ 00:02:51.950 Uttam Kumaran: Yeah, or it’s, like, the whole thing, there’s actually, like, nice food, and there’s, like.
44 00:02:51.950 ⇒ 00:02:52.860 Matthew Good: Yeah, yeah.
45 00:02:52.860 ⇒ 00:03:05.519 Uttam Kumaran: And, like, I got free stuff, and TechConference is, like, I don’t need another, like, front zip hoodie. I don’t need any front zip hoodies, actually. I don’t want any front zip hoodies.
46 00:03:06.230 ⇒ 00:03:10.060 Uttam Kumaran: No stickers, or fronts of hoodies, or, like, AWS hats.
47 00:03:10.060 ⇒ 00:03:11.500 Samuel Roberts: It’s always the same stuff, yeah.
48 00:03:11.500 ⇒ 00:03:13.100 Uttam Kumaran: Yeah, so…
49 00:03:13.100 ⇒ 00:03:13.700 Matthew Good: Dude, yeah.
50 00:03:13.700 ⇒ 00:03:19.249 Samuel Roberts: When I was working on the hair care company, I went to a beauty, like, convention, all kinds of stuff, and I got a…
51 00:03:19.250 ⇒ 00:03:19.870 Matthew Good: that.
52 00:03:20.130 ⇒ 00:03:32.010 Samuel Roberts: It was crazy. Tons of different things, a lot of swag, like, good swag, because it was, like, everyone’s giving out samples of stuff. I felt very out of place, because I was the, like, behind-the-scenes guy at the company.
53 00:03:32.010 ⇒ 00:03:32.649 Matthew Good: Yeah, yeah.
54 00:03:32.650 ⇒ 00:03:33.670 Samuel Roberts: there, but yeah.
55 00:03:33.860 ⇒ 00:03:45.180 Matthew Good: Yeah, it’s hilarious. Yeah, I was at a… my last company, I was at Crypto Conference in Texas, and that was just, like, a surreal experience walking around. It was consensus, I don’t know if you guys had ever been… I was like, what the hell is going on, man?
56 00:03:45.180 ⇒ 00:03:46.899 Uttam Kumaran: Oh yeah, I know consensus, yeah.
57 00:03:47.690 ⇒ 00:03:49.090 Matthew Good: Yeah, it’s insane.
58 00:03:49.090 ⇒ 00:03:51.200 Uttam Kumaran: It’s also gotta be a weird cast of characters.
59 00:03:51.300 ⇒ 00:03:51.900 Uttam Kumaran: Dude…
60 00:03:51.900 ⇒ 00:03:55.080 Matthew Good: A motley cast of characters, maybe.
61 00:03:55.080 ⇒ 00:04:01.209 Samuel Roberts: you’re just, like, people are talking to you, like, I don’t even know really what you’re saying, but, like, yeah, yeah, yeah, Layer 2 abstraction, blockchain, blah blah blah.
62 00:04:02.920 ⇒ 00:04:04.180 Matthew Good: Sure.
63 00:04:04.180 ⇒ 00:04:12.040 Uttam Kumaran: Still, I have friends that are doing a lot of crypto stuff, and… I don’t know, we do fairly technical work, and so I’m not usually…
64 00:04:12.180 ⇒ 00:04:25.220 Uttam Kumaran: like, usually, like, if someone comes to me with something new, I can dig my way and figure out, like, what they’re talking about, but… Yeah. The crypto people always find, like, some words that aren’t, like, don’t exist, or…
65 00:04:25.300 ⇒ 00:04:29.930 Matthew Good: They kind of, like, run around, and I’m just trying to understand, like.
66 00:04:29.930 ⇒ 00:04:37.079 Uttam Kumaran: I get blockchain and the ledger and stuff, but then… then they skip a couple steps in their explanation.
67 00:04:37.080 ⇒ 00:04:37.540 Samuel Roberts: Yeah.
68 00:04:37.540 ⇒ 00:04:51.240 Matthew Good: Yeah, it’s like, whoa, whoa, dude, like, it’s like, it’s new money, and you’re like, okay. But it’s back. Honestly, we were just working with, stablecoin stuff is back, it’s like a new wave.
69 00:04:51.380 ⇒ 00:04:54.400 Matthew Good: So, I mean, shit, man, I don’t know. That’s…
70 00:04:54.400 ⇒ 00:04:54.860 Samuel Roberts: else.
71 00:04:54.860 ⇒ 00:04:56.920 Matthew Good: It just bugs me out.
72 00:04:57.150 ⇒ 00:05:02.160 Matthew Good: Awesome. I… am… what am I supposed to be pulling up here? .
73 00:05:03.150 ⇒ 00:05:07.549 Uttam Kumaran: Yeah, so I wanted… I think we can start. I just wanted to kind of frame…
74 00:05:07.690 ⇒ 00:05:14.849 Uttam Kumaran: this meeting. So a couple things. So one, You have a couple people…
75 00:05:15.200 ⇒ 00:05:21.019 Uttam Kumaran: I think you… I don’t know if you met… I think you met Sam before.
76 00:05:21.020 ⇒ 00:05:21.350 Matthew Good: Yeah.
77 00:05:21.350 ⇒ 00:05:26.689 Uttam Kumaran: I’m not sure, but in the beginning of this call, I think I want to just have Sam demo a…
78 00:05:26.840 ⇒ 00:05:39.890 Uttam Kumaran: demo a little bit of, like, what we’ve been working on, with Micah, and then also, you have Rika, who’s been kind of helping us progressively transitioning the project management for your project to Justin, who’s also on the call, so maybe…
79 00:05:39.890 ⇒ 00:05:40.290 Matthew Good: Okay.
80 00:05:40.290 ⇒ 00:05:53.709 Uttam Kumaran: Justin, I can have you give maybe, like, a little quick 2-minute, intro if you want, and then Mustafa and Sam, if you guys just want to just run through, kind of, what we’ve been working on for Micah, and then, yeah, I just have a couple questions after that we can chat through.
81 00:05:53.920 ⇒ 00:05:54.640 Matthew Good: Perfect.
82 00:05:54.640 ⇒ 00:06:04.999 Justin Breshears: Yeah, I’ll kick it off with a little hello. I’m Justin, brand new to Brainforge, but I’ve been leading, you know, projects and building up PMOs for.
83 00:06:05.000 ⇒ 00:06:20.920 Justin Breshears: number of years in a number of different, IT and tech spaces, so excited to be a part of the Brainforge team, and I’m gonna jump in here and help out, you know, with what we’ve got going on for you. So, excited to meet you. I’m located pretty close to Utah. I’m in Richmond, Texas, so…
84 00:06:20.920 ⇒ 00:06:21.420 Matthew Good: Okay, yeah.
85 00:06:21.420 ⇒ 00:06:22.100 Justin Breshears: verbatim.
86 00:06:22.400 ⇒ 00:06:27.260 Matthew Good: Hell yeah. Yeah, it’s great to meet you, man. I have an uncle out in, Texas, like, Houston area, outside Rosenberg, but…
87 00:06:27.260 ⇒ 00:06:30.060 Justin Breshears: Oh, yeah, so I’m, like, 45 minutes from Houston.
88 00:06:30.060 ⇒ 00:06:33.049 Matthew Good: Okay, yeah, yeah, cool. I went to Rice for a year, so I was there for a year.
89 00:06:33.050 ⇒ 00:06:35.840 Uttam Kumaran: Oh, nice. That did great.
90 00:06:35.840 ⇒ 00:06:36.610 Matthew Good: Thanks, man.
91 00:06:36.610 ⇒ 00:06:40.360 Samuel Roberts: You know… you know the… the summer humidity that we’re dealing with right now.
92 00:06:40.360 ⇒ 00:06:47.700 Matthew Good: moving in, I was like, I don’t know how people do this year-round. This is actually insane. Like, in August, I was like… It’s wild.
93 00:06:47.700 ⇒ 00:06:48.670 Samuel Roberts: That’s horrible.
94 00:06:48.670 ⇒ 00:06:49.740 Matthew Good: But, great to meet you, man.
95 00:06:49.740 ⇒ 00:06:50.399 Justin Breshears: Yeah, right back.
96 00:06:50.400 ⇒ 00:06:51.159 Matthew Good: work together.
97 00:06:51.330 ⇒ 00:06:52.890 Justin Breshears: Alright, Sam, it’s all yours.
98 00:06:53.050 ⇒ 00:06:58.440 Samuel Roberts: Yeah, so, Mustafa, if you want to share, the…
99 00:06:58.440 ⇒ 00:06:58.790 Mustafa Raja: Yeah.
100 00:06:58.790 ⇒ 00:07:04.210 Samuel Roberts: like, an example in the Slack, rather than run through one, because it can take a little time to load sometimes, I feel like.
101 00:07:04.340 ⇒ 00:07:10.580 Samuel Roberts: Because it’s doing all the generation, but I would say show one off one of the examples. But basically, we had a meeting last week.
102 00:07:10.780 ⇒ 00:07:21.489 Samuel Roberts: With Mika, and we went through… he had used it a little bit, we had… he had some good feedback. Some of the styling in Notion, I think, got improved.
103 00:07:22.380 ⇒ 00:07:32.600 Samuel Roberts: And then it now can also save… it saves as it’s doing it, so now you can… rather than having to regenerate the whole deck every time, you can just… it’ll just return.
104 00:07:32.600 ⇒ 00:07:35.029 Matthew Good: things you’ve updated, so…
105 00:07:35.030 ⇒ 00:07:42.789 Samuel Roberts: There are a few changes there. That also enables us to potentially do some other stuff, now that we’re storing it, actually, in a persistent database.
106 00:07:42.790 ⇒ 00:07:43.680 Matthew Good: Yeah.
107 00:07:43.880 ⇒ 00:07:50.450 Samuel Roberts: The other thing, that we talked about was potentially being able to Tweak the prompts over time.
108 00:07:50.450 ⇒ 00:07:50.850 Matthew Good: Yeah.
109 00:07:50.850 ⇒ 00:07:54.669 Samuel Roberts: And so, right now, or right before, everything was living in N8N.
110 00:07:55.510 ⇒ 00:08:14.639 Samuel Roberts: it would have been a little, probably more than you guys want to do to get in there and tweak the prompts and rerun it. So what we were able to do was we set up a kind of prompt, library on LangFuse, which integrates right into N8N, so now we can pull prompts, and so you can tweak prompts, you can version them, you can do a whole bunch of stuff, rerun.
111 00:08:14.640 ⇒ 00:08:15.100 Matthew Good: Thanks.
112 00:08:15.100 ⇒ 00:08:20.280 Samuel Roberts: See how it looks. We definitely want to kind of get a better kind of,
113 00:08:20.650 ⇒ 00:08:29.469 Samuel Roberts: like, evaluation loop going, but I think at this point, you can use it in Slack, it can save to Notion. We added a, rubric step.
114 00:08:30.920 ⇒ 00:08:50.529 Samuel Roberts: that he gave us, so we… now it’s just a, like, how is it looking? There’s buttons now in Slack, so you don’t have to just ask it to, like, approve and stuff. So there’s a lot of, like, little UI things that we added there. There’s probably a lot more we could do in Slack, but we focused on the Notion styling, first, so that seemed, like, more…
115 00:08:50.530 ⇒ 00:08:51.500 Matthew Good: Yeah, that’s perfect.
116 00:08:51.500 ⇒ 00:08:58.050 Samuel Roberts: I leverage, yeah. So now, yeah, this is… this is kind of what, he had showed us, the different, headings and the…
117 00:08:58.220 ⇒ 00:09:00.220 Matthew Good: Yeah, okay, cool. That’s perfect. Yeah.
118 00:09:00.220 ⇒ 00:09:15.179 Samuel Roberts: So, and then, yeah, so now the prompt stuff, like, they can be changed. Right now, you still have to just go back to Slack and probably rerun, test something. But we’re gonna get him set up, and I’ll invite you as well to those prompts, so you can see.
119 00:09:15.180 ⇒ 00:09:22.819 Samuel Roberts: But we’re definitely thinking about ways to kind of close that loop where it’s… you make a test, how would that have affected previous runs? How would that…
120 00:09:22.820 ⇒ 00:09:23.590 Matthew Good: Hmm.
121 00:09:23.590 ⇒ 00:09:31.430 Samuel Roberts: Can I just rerun just the narrative step or something like that? Because there’s a whole bunch of agents, like, sub-agents in this thing, and they all have different prompts.
122 00:09:31.480 ⇒ 00:09:45.089 Samuel Roberts: But all those prompts were based on, kind of, the examples that you provided, I think, from Claude. Yeah. And so, you know, it’s sort of… we’ve kind of pulled things away, so there’s sub-agents and different things, but now that you guys can get in there, you can really probably dial it in.
123 00:09:45.090 ⇒ 00:09:45.720 Matthew Good: Hmm…
124 00:09:45.720 ⇒ 00:09:51.480 Samuel Roberts: or your style, or, whatever it’s not providing, but yeah.
125 00:09:51.480 ⇒ 00:09:51.880 Matthew Good: Amazing.
126 00:09:51.880 ⇒ 00:09:56.119 Samuel Roberts: Kind of the big on the last week. We did not have a meeting with him this morning, I guess he was out in.
127 00:09:56.120 ⇒ 00:09:56.520 Matthew Good: Yeah.
128 00:09:56.520 ⇒ 00:10:02.629 Samuel Roberts: Didn’t get a chance to play around too much, so definitely want him to test a little bit of this, and you to test a bit of it.
129 00:10:02.740 ⇒ 00:10:15.419 Samuel Roberts: And then figure out where to… where to go next, like, what are the improvements to make? You know, little quality of life things are always good, too, but, you know, any kind of bigger things are also nice to start working towards, but…
130 00:10:15.420 ⇒ 00:10:16.010 Matthew Good: Perfect.
131 00:10:16.020 ⇒ 00:10:17.050 Samuel Roberts: Yeah.
132 00:10:17.080 ⇒ 00:10:33.569 Matthew Good: Awesome, I mean, this looks… yeah, the Notion stuff is great, that was big, just because that took a lot of time to, like, kind of manually go through. I have, so I know that, like, the initial stuff that I gave you was, like, decks that were in progress, and what I’ve had the team do… I can actually share my screen.
133 00:10:33.780 ⇒ 00:10:34.370 Samuel Roberts: Oh, cool, yeah.
134 00:10:34.370 ⇒ 00:10:35.260 Matthew Good: show.
135 00:10:35.690 ⇒ 00:10:37.409 Matthew Good: As requested.
136 00:10:38.370 ⇒ 00:10:39.030 Mustafa Raja: bone.
137 00:10:39.410 ⇒ 00:10:41.649 Mustafa Raja: I stopped sharing, so you can share again.
138 00:10:42.040 ⇒ 00:10:45.060 Matthew Good: Thank you. Where’s my… Here it is.
139 00:10:45.980 ⇒ 00:10:49.580 Matthew Good: So I put together all these deck exports from, like, prior final decks that we.
140 00:10:49.580 ⇒ 00:10:51.279 Samuel Roberts: Oh, very nice, yes.
141 00:10:51.280 ⇒ 00:11:00.220 Matthew Good: So… and I’ll delineate… this is all just in one folder right now, and some of these are, like… like, this red glass one is a VC deck, which is gonna be very different than, like.
142 00:11:00.380 ⇒ 00:11:18.489 Matthew Good: Conductor AI, which is a Series A deck, but what I’ll do is I’ll have… me or Micah will create subfolders for these, and then I’ll be able to share this folder with you to just be like, hey, these are all… these are final exported decks that we delivered to clients, and we have, like, 4 others coming up, within, like, the next probably week and a half that I’ll add to this folder.
143 00:11:19.490 ⇒ 00:11:22.349 Matthew Good: Then the idea being that, like, it’ll have… we’ll have better…
144 00:11:22.610 ⇒ 00:11:24.580 Matthew Good: I guess, better, kind of, like,
145 00:11:24.980 ⇒ 00:11:42.339 Matthew Good: data to, like, train, for lack of a better word, like, data to train that on. Yeah. Because right now, what I gave you is, like, fine, but it’s, like, they were, like, in progress, kind of, like, V1, V2s, which are, like, good enough, they were 70% of the way there, but, like, they get edited before they’re actually shipped out. Of course, yeah. But these are final decks that have been shipped out and sent to people.
146 00:11:42.540 ⇒ 00:11:47.939 Samuel Roberts: Great. And, like, in these folders, is it… is it, like, Notion, like, or is it, like, a full deck?
147 00:11:47.940 ⇒ 00:11:49.250 Matthew Good: It’s like a PDF.
148 00:11:49.510 ⇒ 00:11:50.190 Samuel Roberts: Okay, okay.
149 00:11:50.190 ⇒ 00:11:54.459 Matthew Good: We can… I can have the team, like, if there’s a certain,
150 00:11:54.710 ⇒ 00:11:59.370 Matthew Good: way you want to intake this data, I can have them, like, convert or export or do whatever.
151 00:11:59.370 ⇒ 00:12:02.609 Samuel Roberts: Yeah, we can… we can have a little… Something that you can…
152 00:12:02.610 ⇒ 00:12:03.159 Matthew Good: Yeah, let me…
153 00:12:03.160 ⇒ 00:12:05.529 Samuel Roberts: a little bit and see what we need, like, what would be best, but…
154 00:12:05.700 ⇒ 00:12:13.830 Matthew Good: Yeah, let me know the best form factor to give this to you guys, but this is, because I know that’s where, I think in the initial time, we… we just didn’t have these put together for you guys.
155 00:12:13.830 ⇒ 00:12:14.310 Samuel Roberts: Yeah.
156 00:12:14.310 ⇒ 00:12:20.170 Matthew Good: And, like, inputs will… the quality of input will give, you know, directly affect the quality of output, so,
157 00:12:20.500 ⇒ 00:12:23.020 Matthew Good: So that’s kind of what we’ve been doing on our end to get these together.
158 00:12:23.280 ⇒ 00:12:30.420 Samuel Roberts: Okay, great, that’s, that’s gonna be very helpful. I would say… Yeah, there might be…
159 00:12:30.540 ⇒ 00:12:37.739 Samuel Roberts: Are there still outlines of those, like, final… or I guess I’m wondering, like, what your process is, like, once there’s something in Notion.
160 00:12:37.740 ⇒ 00:12:38.350 Matthew Good: Like…
161 00:12:38.350 ⇒ 00:12:42.490 Samuel Roberts: Is there a final notion that then becomes that? Or I’m sure it’s iterative after that, too, but…
162 00:12:42.490 ⇒ 00:12:46.459 Matthew Good: Yeah, it’s basically… it’s actually this.
163 00:12:47.380 ⇒ 00:12:52.319 Matthew Good: So, something will happen in Notion, and then it’ll come… we’ll go into Figma, and, like, this is, like, how we take the…
164 00:12:52.320 ⇒ 00:12:52.790 Samuel Roberts: This is another…
165 00:12:52.790 ⇒ 00:12:56.830 Matthew Good: Okay. And this is the human in the loop step of, like.
166 00:12:56.950 ⇒ 00:13:15.000 Matthew Good: we got the narrative here, there was a notion that I sent over, the client also, like, these things are super iterative, so, like, they had a call with their investor that we weren’t a part of, so they were like, we have more feedback, like, here’s the loom, and here’s the deck, so now I’m like, my job is to, like, Frankenstein together, like, what we had narratively, the feedback that they gave us on the left, and then…
167 00:13:15.380 ⇒ 00:13:22.300 Matthew Good: like, put that together and what we have on the right, and then direct the design team of, like, okay, this should be here, so this is where, like, the real…
168 00:13:22.450 ⇒ 00:13:27.339 Matthew Good: like, over the line is, and where I come in. Yeah, that makes sense. Okay. But yeah.
169 00:13:27.690 ⇒ 00:13:32.280 Uttam Kumaran: And then on the left, that’s actually where the Notion piece gets into this design process.
170 00:13:32.870 ⇒ 00:13:34.830 Matthew Good: On the left is where the Notion piece.
171 00:13:34.830 ⇒ 00:13:36.840 Uttam Kumaran: Like, the screenshot of the.
172 00:13:36.840 ⇒ 00:13:37.419 Samuel Roberts: That was the…
173 00:13:37.420 ⇒ 00:13:54.339 Matthew Good: Yes, yes. Yeah, so what will happen is our designers will say, you know, I’ll go through, maybe, like, one or two, three kind of max revs of narrative, and then I will send that to them, and then they will have this, and they will literally just, like, screenshot it in. And all, like, the note… what’s in Notion is, like.
174 00:13:54.340 ⇒ 00:14:11.489 Matthew Good: I know when I send this to them, like, this is not how the copy’s exactly gonna look, because once it goes into Figma, I’m like, oh, okay, well, actually, like, we should delete this, we don’t need that, or like, you know… But I want, like, the passes at Notion are intentionally broad, because I want to have, like, everything there, and then I want to be, like, Figma is an exercise of elimination.
175 00:14:13.460 ⇒ 00:14:13.980 Samuel Roberts: Okay.
176 00:14:13.980 ⇒ 00:14:15.090 Matthew Good: That’s just, like, a…
177 00:14:15.330 ⇒ 00:14:20.510 Matthew Good: An overview of our workflow, and then this will get exported, eventually, once it’s approved.
178 00:14:21.020 ⇒ 00:14:35.179 Uttam Kumaran: So I think right now, I think the biggest thing is, like, for the upcoming decks, I think one is we would love if we can use the… the agent to generate, like, that first draft. I think I would… I would want to get a sense from you on…
179 00:14:35.230 ⇒ 00:14:41.090 Uttam Kumaran: like, where to point… where to poke at next. Like, is it taking that notion?
180 00:14:41.190 ⇒ 00:14:46.400 Uttam Kumaran: Plus, like, a person to review, and then getting to, like, deck-ready copy.
181 00:14:46.610 ⇒ 00:14:48.979 Uttam Kumaran: Or is it actually, should we…
182 00:14:49.140 ⇒ 00:15:04.550 Uttam Kumaran: start poking at stuff at the visual layer, like, for example, there are some tools that are out there now, like Gamma, for example, to actually generate slides. Like, I guess, tell me what you think, and this is kind of, like, where I want to talk a little bit about, and again, we’re just…
183 00:15:05.130 ⇒ 00:15:12.530 Uttam Kumaran: right now, just focused on, like, the deck-building process. We can talk about other opportunities in the company, but where do you… what do you think the next…
184 00:15:12.740 ⇒ 00:15:15.749 Uttam Kumaran: Thing to… to… to poke at is.
185 00:15:16.740 ⇒ 00:15:21.139 Matthew Good: Good question. I haven’t played around with Gamma. I don’t know to what extent there might be
186 00:15:22.570 ⇒ 00:15:25.530 Matthew Good: A world in which we take the notion.
187 00:15:25.530 ⇒ 00:15:28.859 Uttam Kumaran: And then we use Gamma to, like, mock up, like… It’s pretty good, dude.
188 00:15:28.860 ⇒ 00:15:30.030 Matthew Good: High-level layouts of a
189 00:15:30.410 ⇒ 00:15:40.530 Matthew Good: sides, and then… because, like, the initial design frame is, like, or design work is, like, we kind of sketch out, okay, how this is gonna flow, like, how it’s gonna look, basically, and then, like, you can see Sylvia in here is, like.
190 00:15:40.810 ⇒ 00:15:41.599 Matthew Good: you know.
191 00:15:41.600 ⇒ 00:15:47.499 Uttam Kumaran: And are these… do you have templates for slides? Like, okay, this is a graph slide, this is… or, like, it… how do they…
192 00:15:47.610 ⇒ 00:15:48.200 Uttam Kumaran: Are they?
193 00:15:48.200 ⇒ 00:15:50.639 Matthew Good: Sort of, but, like, not really. She’s been, like, you can see her right now.
194 00:15:50.640 ⇒ 00:15:51.360 Uttam Kumaran: Oh, from… yeah.
195 00:15:51.360 ⇒ 00:15:53.099 Matthew Good: Yeah, she’s building it live.
196 00:15:53.550 ⇒ 00:16:05.939 Uttam Kumaran: But are they, like, are they ref… are… I guess my question is, for her, one, is her background in, like, building these decks, or is she referencing, like, the past ones that you guys probably just have a library of, and sort of, like, going from there?
197 00:16:05.940 ⇒ 00:16:11.650 Matthew Good: She just, like, knows… she’s, like, done a bunch of decks, so she’s like, oh, this is how that would make sense, or whatnot. But ideally, right, like.
198 00:16:12.290 ⇒ 00:16:15.669 Matthew Good: in terms of, like, we only have one of her, and then we have Rafei.
199 00:16:15.870 ⇒ 00:16:16.200 Uttam Kumaran: Yeah.
200 00:16:16.200 ⇒ 00:16:22.910 Matthew Good: all the time. So, like, ideally, you want to be able to train someone, because in design school, I guess I didn’t know this, but, like, no one teaches deck design, you just kind of learn it.
201 00:16:22.910 ⇒ 00:16:23.320 Uttam Kumaran: No.
202 00:16:23.320 ⇒ 00:16:23.900 Matthew Good: outgoing.
203 00:16:23.900 ⇒ 00:16:25.260 Uttam Kumaran: Yeah.
204 00:16:25.260 ⇒ 00:16:36.209 Matthew Good: So being able to have some sort of way to, like, easily onboard someone, or, like, give them kind of, like, that knowledge without having them, like, train over and over and over again. So that might be something.
205 00:16:36.210 ⇒ 00:16:43.670 Uttam Kumaran: Give me a sense of the time, like… so, like, how much… like, say, for, like, in this example, right, we got… we have the notion.
206 00:16:43.780 ⇒ 00:16:46.340 Uttam Kumaran: How much time is it between…
207 00:16:46.850 ⇒ 00:16:51.680 Uttam Kumaran: When you hand her the notion, and then you have your first like, review.
208 00:16:51.680 ⇒ 00:16:55.680 Matthew Good: So… so what we do is, like… well, this is initial mood boarding, but…
209 00:16:55.830 ⇒ 00:16:58.830 Matthew Good: We’ll do… so this is the previous round.
210 00:16:58.980 ⇒ 00:17:13.179 Matthew Good: So, kind of, like, the first V2 narrative was here, right? And then she started… Rafae started to, like, mock out these slides of, like, okay, headlines, whatever. And then we’ll send over, like, the first 5 to 6 slides built out, like, very rough cuts.
211 00:17:13.609 ⇒ 00:17:32.369 Matthew Good: for the client as, like, visual alignment slides, to be like, hey, like, this is the mood, like, they liked, I think it was mood board C, or whatever it was. They were like, we want to go this visual direction, okay, great. Now we’re going to transpose that into, like, the first 5 or 6 slides. Does anything here feel, like, crazy off? Right? And they gave some feedback on this, like, oh, this imagery doesn’t feel right, or, like, blah blah blah, whatever, this is chaotic.
212 00:17:32.369 ⇒ 00:17:36.739 Matthew Good: So, as opposed to just, like, building out the full deck, we’ll give them, like, a first…
213 00:17:36.819 ⇒ 00:17:38.809 Matthew Good: Cut, and then we’ll build out the rest.
214 00:17:40.500 ⇒ 00:17:53.819 Uttam Kumaran: is it about, what is in that piece? Because I feel like by that point, you’ve agreed on, like, the narrative. You may not know the ordering 100%, but, like, what’s in the mood board piece?
215 00:17:54.540 ⇒ 00:17:56.099 Matthew Good: Oh, what’s in the mood board piece?
216 00:17:57.840 ⇒ 00:17:58.930 Matthew Good: You’re breaking up.
217 00:18:02.580 ⇒ 00:18:05.809 Uttam Kumaran: Sorry, so what’s in the mood board, basically?
218 00:18:06.160 ⇒ 00:18:17.740 Matthew Good: Oh, so this is just, like, a couple different visual directions in terms of, like, how data is presented, how we’re using whitespace, how we’re using gradients, tones, stuff like that. So we’ll send this over to them and then review it live.
219 00:18:17.830 ⇒ 00:18:33.990 Matthew Good: And they’ll basically say, hey, like, I think B is the direction we want to go, or, like, hey, we liked C, but, like, also we loved, like, this kind of B3 element and, like, B2, and they can, like, kind of, like, pick and choose from different ones, and then we’ll combine that into a consolidated mood board. The idea is, like.
220 00:18:34.310 ⇒ 00:18:41.339 Matthew Good: Without this, it’s, like, really hard to establish a creative direction, but we’re also not scoping for brand, so we have to figure out a way to, like.
221 00:18:41.880 ⇒ 00:18:49.009 Matthew Good: get the, like, white fence of, like, Creative Lane to swim in without doing brand work, which would take too long, because they need this deck in, like, 3 weeks.
222 00:18:49.180 ⇒ 00:18:57.239 Matthew Good: So that’s kind of the purpose of the mood board. So then when we evaluate this, they’re like, why are you changing orange? And we’re like, well, okay, well, we aligned on X, Y, and Z thing.
223 00:18:57.880 ⇒ 00:18:59.199 Matthew Good: In a prior step.
224 00:19:00.920 ⇒ 00:19:09.049 Uttam Kumaran: So I actually… so my gut instinct is that that is actually very, very important, and I don’t think, like, at this point, it’s probably worth attacking the mood board, but…
225 00:19:09.200 ⇒ 00:19:25.450 Uttam Kumaran: post-Moodboard, I feel like there’s definitely some opportunity. So if you talk about, like, what options are there on the visual side, like, and I’m reflecting a lot on how we work with our designers, and how we’ve worked with design in the past and building product, is moving from
226 00:19:25.450 ⇒ 00:19:35.220 Uttam Kumaran: wireframe to lo-fi to mid-fi, giving inspiration. The stuff we’re seeing out of Gamma is really, really good. I’ve been using it for a few weeks, and…
227 00:19:35.280 ⇒ 00:19:45.550 Uttam Kumaran: It’s extremely good. They just released, like, an API. Could be worth considering using something like Gamma for, like, a lo-fi version or, like, a rough cut of something.
228 00:19:45.550 ⇒ 00:19:46.010 Matthew Good: Hmm.
229 00:19:46.010 ⇒ 00:20:05.860 Uttam Kumaran: I don’t think it’s gonna… I don’t think it’s there to get anything, like, final, but if there is, like, and this is where, like, my question would be, if there is a long time between the notion, the mood board, and, like, a first cut, I think there’s opportunity to improve there.
230 00:20:06.240 ⇒ 00:20:12.480 Uttam Kumaran: Second is, if copy becomes a bottleneck, that is certainly an area where
231 00:20:12.550 ⇒ 00:20:28.859 Uttam Kumaran: we can play around. For example, if you need variations of things, if you want to quickly say, I don’t like how this feels, but you may spend 30 minutes or an hour sort of, like, ideating, that is something where I’m sure you’re copying and pasting back into AI to say, like, give me a different sense of this.
232 00:20:29.000 ⇒ 00:20:31.020 Uttam Kumaran: That could be an opportunity as well.
233 00:20:31.330 ⇒ 00:20:38.670 Matthew Good: Yeah, yeah, there has been times where, like, especially with a very complex product, like Sonoda, the one I just showed, they’re, like.
234 00:20:39.220 ⇒ 00:20:48.939 Matthew Good: I will input all the data into, like, Claude Console and, like, you know, kind of, like, build a knowledge base, and then I’ll just be like, okay, stop what you’re doing, and just, like, explain this to me as if I were, like, an educated 12-year-old.
235 00:20:49.000 ⇒ 00:21:06.260 Matthew Good: what exactly is going on here, so that I can just, like, reorientate myself, because, like, I’m… when you get so deep in the weeds of, like, building a slide, you’re like, okay, wait a second, am I, like, missing the actual point here? And then I’ll just, like, zoom all the way out, and Claude will do that, and be like, okay, yeah. Like, that’s just helpful for me with a super technical product.
236 00:21:06.820 ⇒ 00:21:08.870 Matthew Good: So sometimes I’ll do that. I think…
237 00:21:09.980 ⇒ 00:21:17.189 Matthew Good: Yeah, I don’t know if that was an answer to, like, what you were saying. I think the gamma lo-fi, like, we can… we can play around with it. One of the…
238 00:21:18.440 ⇒ 00:21:31.579 Matthew Good: clients will be like, you know, I could just generate… we had a client… we had a client that was like, well, you know, we sent them the rough cuts, and they were like, well, basically, I could just do this in pitch, and, like, we need to do a better job of level studying, like, yeah, but, like, that’s because these are rough cuts.
239 00:21:31.580 ⇒ 00:21:32.670 Uttam Kumaran: Yeah, yeah, yeah.
240 00:21:32.670 ⇒ 00:21:33.330 Matthew Good: You know, so, like…
241 00:21:33.330 ⇒ 00:21:35.590 Uttam Kumaran: There’s no way they can do it in pitch, so what are they gonna… yeah.
242 00:21:35.930 ⇒ 00:21:41.930 Uttam Kumaran: Not what… I don’t think that’s there. Like, but I hear you. Yeah.
243 00:21:41.930 ⇒ 00:21:44.280 Matthew Good: Yeah. So…
244 00:21:45.610 ⇒ 00:21:50.680 Matthew Good: Yeah, because a lot of these slides, especially for a super complex product, like, they wind up looking like…
245 00:21:50.870 ⇒ 00:21:58.299 Matthew Good: Something like this, which is, like, this took a lot of time to build, like, one by one.
246 00:22:00.940 ⇒ 00:22:08.170 Matthew Good: It’s like a laser company, but… so, yeah, there’s no way you could do this in, like, I don’t think in gamma or pitch or something like that, at least right now.
247 00:22:08.440 ⇒ 00:22:13.620 Uttam Kumaran: I guess my other question is, like, where do you want your designers to be spending more time, where maybe they’re…
248 00:22:13.620 ⇒ 00:22:14.510 Samuel Roberts: unbelievable.
249 00:22:14.510 ⇒ 00:22:17.310 Uttam Kumaran: They’re spending more time on alignment, or…
250 00:22:17.830 ⇒ 00:22:23.630 Uttam Kumaran: you know, not as important things. Like, maybe that’s another good way to frame, sort of, the opportunity.
251 00:22:23.990 ⇒ 00:22:24.790 Matthew Good: Yeah.
252 00:22:25.630 ⇒ 00:22:27.830 Matthew Good: It’s a good question. I mean, we just…
253 00:22:29.540 ⇒ 00:22:34.840 Uttam Kumaran: But also, 3 weeks is not a lot of time, right? Is that usually kind of, like, the turnaround between.
254 00:22:34.840 ⇒ 00:22:46.050 Matthew Good: That was a rush. Normally, we try to level set at, like, 4 to 5, but, like, this company was, like… inevitably, people are like, oh, we have a pitch, and, like, now we gotta get our shit together, and, like, you know, we need this in, like, two and a half weeks, can you do that? And it’s like…
255 00:22:46.700 ⇒ 00:22:52.130 Matthew Good: Yeah, realistically, it becomes 3, you know? So…
256 00:22:52.130 ⇒ 00:23:06.980 Uttam Kumaran: And also, I guess a follow-up on that is, like, would the unlock be, yes, we can do it in 3, and the quality stays really high, or is it more like, no, we can’t, but, like, if for a fee, we could, and you can actually deliver it? .
257 00:23:06.980 ⇒ 00:23:07.470 Matthew Good: Yeah.
258 00:23:07.470 ⇒ 00:23:07.839 Uttam Kumaran: The other option.
259 00:23:08.020 ⇒ 00:23:13.540 Matthew Good: We upcharge if the timeline’s shorter, we’ll be like, yeah, well, like, a rush fee is gonna cost you this.
260 00:23:13.810 ⇒ 00:23:18.700 Matthew Good: I think… Yeah.
261 00:23:18.890 ⇒ 00:23:33.700 Matthew Good: like, the copy is still, I think, a blocker. Like, I would love… I would love to be able to get inserted into the process, like, at this stage here, where it’s like, okay, now let me, like, go into each slide and be like, does this make sense? You know, whatever.
262 00:23:35.930 ⇒ 00:23:41.089 Matthew Good: And… I think just the quality of output
263 00:23:41.240 ⇒ 00:23:53.429 Matthew Good: like, that this… the deck agent will be able to give will be significantly higher when it has, like, actually completed decks to train off of. I think just, like, one. And then, as I’m, like, also training Micah to be, like.
264 00:23:53.610 ⇒ 00:24:12.029 Matthew Good: last week, we were doing a lot of work on, like, this is just, like, your inst… like, to improve instinct for a deck, to just be like, okay, yeah, well, this actually doesn’t make sense, like, no one would actually think this, or, like, this should go actually here, because relative to this product and, like, this particular market, or, like, they’re raising a seed extension, so, like, that means a certain thing, right?
265 00:24:12.030 ⇒ 00:24:18.050 Matthew Good: So, like, there’s a… there’s a bit of, like, coaching on just, like, a, like, keyboard feel that I’m trying to, like, give to him.
266 00:24:19.210 ⇒ 00:24:21.760 Matthew Good: But… Yeah.
267 00:24:21.960 ⇒ 00:24:23.589 Matthew Good: I don’t know if that was an answer.
268 00:24:25.810 ⇒ 00:24:26.710 Matthew Good: Or helpful.
269 00:24:30.330 ⇒ 00:24:31.530 Uttam Kumaran: No, that’s helpful.
270 00:24:34.440 ⇒ 00:24:46.919 Uttam Kumaran: Okay, so I think, kind of maybe same way, I think we can take that and kind of brainstorm. I think one is, I kind of want to continue to expand what the notion includes. Like, I think it could include almost, like.
271 00:24:47.430 ⇒ 00:24:51.689 Uttam Kumaran: Final versions of… Things, or even variations.
272 00:24:52.110 ⇒ 00:24:52.890 Matthew Good: Yeah.
273 00:24:52.890 ⇒ 00:25:05.619 Uttam Kumaran: that is, I think, something that we could solve for. Additionally, I think it could also be, like, how does the agent become an assist throughout the whole process? The next part of this process is, like, great, we have this locked in, now I need help.
274 00:25:05.800 ⇒ 00:25:24.310 Uttam Kumaran: adjusting this, you want the agent to have all… have had… when you… when it answers a question, you want it to have had all those artifacts still. The first half, the… the narratives, the transcripts, and so just having people with the additional context, and then additionally, we can just develop really great prompts for
275 00:25:24.580 ⇒ 00:25:28.350 Uttam Kumaran: doing variations, basically, and giving you those. Yeah.
276 00:25:28.600 ⇒ 00:25:38.310 Uttam Kumaran: And that way, also, you can do that live in Slack with somebody, right? And so that’s… that’s really helpful. I think we should…
277 00:25:38.450 ⇒ 00:25:39.210 Uttam Kumaran: like…
278 00:25:39.480 ⇒ 00:25:46.389 Uttam Kumaran: Gamma just released, like, a new API and a little bit of a better product. I think, like, 6 months ago, it was…
279 00:25:46.760 ⇒ 00:25:54.010 Uttam Kumaran: not great. Now it’s pretty good, and you can actually submit in colors, fonts, brand.
280 00:25:54.320 ⇒ 00:25:57.220 Uttam Kumaran: I would like us to maybe do, like.
281 00:25:57.330 ⇒ 00:26:01.570 Uttam Kumaran: Saying we could probably spend an hour or two and just, like, do a little spike on, like, what is the state of that?
282 00:26:01.570 ⇒ 00:26:02.490 Matthew Good: None.
283 00:26:02.490 ⇒ 00:26:03.180 Uttam Kumaran: corrupt.
284 00:26:03.440 ⇒ 00:26:18.820 Uttam Kumaran: I also agree. I think the biggest thing, Matt, is, like, you want to get to the point where the client is no longer, like, debating color and things, and you’re working on, sort of, the narrative. And then, second, I think the thing is, like, how can you…
285 00:26:19.990 ⇒ 00:26:26.989 Uttam Kumaran: I mean, I think this is for you to decide, is like, is the goal to shrink the time to delivery?
286 00:26:27.370 ⇒ 00:26:32.860 Uttam Kumaran: Is the goal to, like, get to more, like, high fidelity, so you can spend a majority of, like.
287 00:26:33.100 ⇒ 00:26:38.290 Uttam Kumaran: The 3 to 6 weeks discussing At that level, versus, like.
288 00:26:38.650 ⇒ 00:26:42.860 Uttam Kumaran: Okay, hey, thanks, we got it, and as a week goes while you guys are moving the chain.
289 00:26:42.860 ⇒ 00:26:45.100 Matthew Good: Or is the goal, like.
290 00:26:45.470 ⇒ 00:26:49.349 Uttam Kumaran: is the KPI, like, okay, one person now can…
291 00:26:49.990 ⇒ 00:26:57.459 Uttam Kumaran: you know, one of the ways that we think about it is, like, can someone who’s maybe not as familiar, plus AI, basically be…
292 00:26:57.610 ⇒ 00:27:06.699 Uttam Kumaran: like, able to handle that. So, can you codify what you’ve learned, and, like, what makes a good deck, what makes a bad deck, and have that person interact with AI to judge?
293 00:27:06.860 ⇒ 00:27:07.420 Uttam Kumaran: Right, like, how.
294 00:27:07.420 ⇒ 00:27:08.210 Matthew Good: Yeah.
295 00:27:08.210 ⇒ 00:27:13.240 Uttam Kumaran: Use an LLM to help them, like, they, like, plug it in and they can get feedback without going to you.
296 00:27:13.240 ⇒ 00:27:13.960 Matthew Good: Yeah.
297 00:27:14.250 ⇒ 00:27:33.619 Uttam Kumaran: In our business, like, our sales and marketing team use that, because we’ve written prompts that basically take in a lot of the feedback that I would give about things and put it into a prompt, so once they build something, they can use that to judge. And when they come to me for feedback, I usually ask, hey, did you ask, like.
298 00:27:33.830 ⇒ 00:27:36.859 Uttam Kumaran: the marketing asset reviewer what it thought.
299 00:27:36.860 ⇒ 00:27:37.670 Matthew Good: Offering.
300 00:27:38.330 ⇒ 00:27:52.520 Uttam Kumaran: And then that way, they… if they come to me with what it thought, and what they think, it’s actually also very easy for me at that point to give them feedback or, you know, to consider. So, if… if the… if your bandwidth is also the…
301 00:27:52.640 ⇒ 00:27:57.189 Uttam Kumaran: the limiting factor here, then it’s more about, like, how do we improve the judgment?
302 00:27:57.920 ⇒ 00:28:01.989 Matthew Good: Yeah, I think that would probably be it off the top of my head, where it’s like, it…
303 00:28:02.160 ⇒ 00:28:06.099 Matthew Good: to give it… make it easier for me to give feedback, and that I’m not giving the feedback
304 00:28:06.170 ⇒ 00:28:23.979 Matthew Good: over and over again, of, like, there’s too much text, there’s no… this is not clear, or, you know, one message per slide, like, because I could just riff on, like, you know, 20 different things and bullet points that would, like, make a good deck versus a bad deck, and, like, it would vary a little bit depending on, like, stage and, like, who you’re pitching, but, like, generally speaking.
305 00:28:24.370 ⇒ 00:28:35.009 Matthew Good: there’s, like, certain principles that are not gonna change, so it might be helpful to have, like, that baked into those prompts that would, like, kind of catch any of those things before I’m, like, looking at it.
306 00:28:35.260 ⇒ 00:28:39.099 Matthew Good: Or reviewing live. And before it goes to Figma, too.
307 00:28:43.060 ⇒ 00:28:45.969 Samuel Roberts: Yeah, I think also having those final decks
308 00:28:46.020 ⇒ 00:29:04.499 Samuel Roberts: gives us a little bit to play with, where there’s copy there, but there’s also, like, layout, just ideas. There’s probably, like, consistent things that we can draw out of there, maybe, and say, like, oh, okay, this is how they like to do it, this is how it should be done. Yeah. And that’s probably a little bit even just ingesting those and exploring what
309 00:29:04.730 ⇒ 00:29:23.350 Samuel Roberts: it notices and patterns there could be something that, you know, you obviously have feedback you give constantly, but, like, there might be things that, like, aren’t even noticed or thought about, and the AI doesn’t know that yet, because you guys are doing it not subconsciously, but just instinctively. Yeah. We might be able to pull out a little bit with that, now that we have that final
310 00:29:23.640 ⇒ 00:29:25.640 Samuel Roberts: final deck.
311 00:29:25.640 ⇒ 00:29:32.050 Matthew Good: Yeah. Okay, so I’ll send that folder over, too, and I’ll categorize them just in, like, mini subfolders.
312 00:29:32.540 ⇒ 00:29:37.430 Matthew Good: So I’ll send that over, I have that on my list of things to do today before I log off.
313 00:29:37.920 ⇒ 00:29:40.230 Matthew Good: And then, yeah, anything else you guys need from me?
314 00:29:41.060 ⇒ 00:29:46.500 Uttam Kumaran: Yeah, I think if we can get that box thing, and then now, Sam, I think maybe me and you can talk about…
315 00:29:46.710 ⇒ 00:29:49.210 Uttam Kumaran: How that notion can just become more…
316 00:29:49.640 ⇒ 00:29:53.579 Uttam Kumaran: I think we crushed the first part, which is just getting to that outline.
317 00:29:53.720 ⇒ 00:29:59.419 Uttam Kumaran: It would be great for us to start thinking about, okay, like, what else could the designers need at that stage that maybe, like.
318 00:29:59.750 ⇒ 00:30:01.120 Uttam Kumaran: But they would…
319 00:30:01.230 ⇒ 00:30:08.590 Uttam Kumaran: they may butcher, they just may not be able to fill the gaps about, like, explain this like I’m a 12-year-old, like, what does this company do?
320 00:30:08.700 ⇒ 00:30:22.399 Uttam Kumaran: Right? So helping them, like, rough out some of those basic things that, like, if they were to get wrong, it would impact, like, the whole deck, you know? So, putting a lot of that into the notion first, because even on our team, sometimes the designers, like.
321 00:30:22.410 ⇒ 00:30:30.550 Uttam Kumaran: they never… they didn’t grasp, like, we’re building case studies for people. I’m like, just get a basic sense of, like, what we’re… what this company does, or what we’re doing for them.
322 00:30:30.590 ⇒ 00:30:36.499 Uttam Kumaran: You know, and so some of that we can fill out in Notion as well.
323 00:30:37.490 ⇒ 00:30:38.180 Matthew Good: Yeah.
324 00:30:38.470 ⇒ 00:30:39.190 Uttam Kumaran: Yeah.
325 00:30:39.190 ⇒ 00:30:44.979 Matthew Good: It might be use- I mean, I don’t know, I’m just, like, thinking out loud, it might be useful even on one of these Friday things to have Sylvia.
326 00:30:45.180 ⇒ 00:30:45.580 Uttam Kumaran: Yes.
327 00:30:45.580 ⇒ 00:30:47.139 Matthew Good: pop in and just be like…
328 00:30:47.200 ⇒ 00:30:50.400 Uttam Kumaran: I was gonna ask that.
329 00:30:50.400 ⇒ 00:30:55.320 Matthew Good: she should just join, because one of the things she sent this morning was, like, the handoff between, like.
330 00:30:55.820 ⇒ 00:31:02.740 Matthew Good: the narrative, and then, like, how I might be envisioning this slide going versus what she is, because she’s also context switching between
331 00:31:03.030 ⇒ 00:31:07.100 Matthew Good: even this week of, like, a laser company that’s, like, pitching to Sequoia on Sunday.
332 00:31:07.100 ⇒ 00:31:13.069 Uttam Kumaran: No, it’s tough. And it’s even hard for us, right? It’s hard for you, and imagine if you’re just…
333 00:31:13.790 ⇒ 00:31:18.280 Uttam Kumaran: Like, if you… if you’re not in the wor- if you’ve not been in this world, Yeah.
334 00:31:18.280 ⇒ 00:31:34.560 Matthew Good: I mean, I have feel for her, and I’m, like, giving her a bunch of different narratives, and then she’s like, well… but, yeah, basically, like, closing the gap on that pickup time, where she’s like, okay, I understand in, like, five senses what’s going on here, so then I can use that to be, like, to just, like, sanity check.
335 00:31:34.990 ⇒ 00:31:40.920 Matthew Good: what we’re outputting, you know? Yeah. So, yeah, I’ll… I can have her join next week, or whenever.
336 00:31:40.920 ⇒ 00:31:43.579 Uttam Kumaran: Okay Okay. That’d be great. Yeah, let’s do that.
337 00:31:43.580 ⇒ 00:31:45.000 Matthew Good: Cool, okay, I got the invite.
338 00:31:45.320 ⇒ 00:31:47.869 Samuel Roberts: The other thing I would say, and I think Utah mentioned this, like, where do you…
339 00:31:48.040 ⇒ 00:32:01.290 Samuel Roberts: like, where is the, like, more important work for the human to be doing? Like, the design… like, what are things that… that would be something we have to learn from her, but you might even have, like, places you would rather her focus her time than others, and I don’t know exactly
340 00:32:01.290 ⇒ 00:32:12.950 Samuel Roberts: the details of… of that, but I think thinking about those things… obviously, like, we had this… you were already using cloud, so it was kind of easy to see, like, okay, this is a good spot for that, but just thinking about, oh, I spent a lot of time doing
341 00:32:13.000 ⇒ 00:32:16.800 Samuel Roberts: this kind of thing, or wrapping my head around this sort of stuff. Like, that’s a good…
342 00:32:17.020 ⇒ 00:32:17.520 Matthew Good: No.
343 00:32:17.520 ⇒ 00:32:23.560 Samuel Roberts: If outputting something that is just like, here’s the lowdown for the designer, Based on…
344 00:32:23.960 ⇒ 00:32:27.110 Samuel Roberts: that kind of thing. Like, whatever it is, like, whether that’s text or whether that’s, like.
345 00:32:27.300 ⇒ 00:32:32.229 Samuel Roberts: Simple lo-fi mock-ups, or, you know, just ideas of layouts, or whatever it is that might be.
346 00:32:32.230 ⇒ 00:32:32.570 Matthew Good: Yeah.
347 00:32:32.570 ⇒ 00:32:49.120 Samuel Roberts: that could either not just speed up her time, but let her focus on the things that, like, the human should be doing, and the AI could do a few, like, passes of something and then choose kind of thing. I think that’s… I would say think about that a little bit, like, where… and obviously we can find out from her where she thinks her time is, but obviously your
348 00:32:49.430 ⇒ 00:32:52.880 Samuel Roberts: What is most important for the human to be doing is great, too.
349 00:32:53.230 ⇒ 00:33:01.200 Matthew Good: Yeah, yeah, yeah, absolutely. Okay, I’ll add her to that call next week. I know we’re a little bit over time, sorry guys, but I’ll add her into that call next week, and then I’ll send you over that Dropbox, today as well.
350 00:33:01.200 ⇒ 00:33:02.219 Samuel Roberts: Perfect, thank you.
351 00:33:02.220 ⇒ 00:33:02.970 Uttam Kumaran: Okay, perfect.
352 00:33:02.970 ⇒ 00:33:03.580 Matthew Good: Gents.
353 00:33:03.770 ⇒ 00:33:05.010 Uttam Kumaran: Thank you. Yeah.
354 00:33:05.010 ⇒ 00:33:05.430 Samuel Roberts: a good one.