2025-10-17_friday_brainforge_demos_retro_f0e89fcf

Meeting Title: Friday Brainforge Demos & Retro Date: 2025-10-17 Meeting participants: Samuel Roberts, Mustafa Raja, Rico Rejoso, Casie Aviles, Amber Lin, Henry Zhao, Ryan Brosas, Raymund Verzosa, Awaish Kumar, Uttam Kumaran, Demilade Agboola, Robert Tseng

---

WEBVTT

1 00:00:26.260 ⇒ 00:00:27.310 Samuel Roberts: Hey.

2 00:00:29.250 ⇒ 00:00:30.450 Mustafa Raja: Hey, how are you?

3 00:00:31.060 ⇒ 00:00:33.690 Samuel Roberts: Doing alright, trying to get that Heroku thing working.

4 00:00:35.640 ⇒ 00:00:36.920 Samuel Roberts: Oh yeah, the script?

5 00:00:38.010 ⇒ 00:00:43.280 Samuel Roberts: Yeah, so I somehow was able to… it’s a little weird, because you can have the environment variables.

6 00:00:43.750 ⇒ 00:00:45.019 Mustafa Raja: In an ad.

7 00:00:46.160 ⇒ 00:00:51.280 Samuel Roberts: But then there’s also the pipeline, and that’s what supplies them to the review apps, right?

8 00:00:51.280 ⇒ 00:00:52.100 Mustafa Raja: lists…

9 00:00:52.310 ⇒ 00:00:58.649 Samuel Roberts: But their API and their CLI don’t really let you access the pipeline as easily.

10 00:01:00.220 ⇒ 00:01:07.749 Samuel Roberts: So I did it sometime before, because that’s how I got them all up there. I didn’t paste them one by one, but I gotta put it in a script, and I’ll get that.

11 00:01:07.750 ⇒ 00:01:08.269 Mustafa Raja: Oh, internet.

12 00:01:08.270 ⇒ 00:01:12.220 Samuel Roberts: That’ll be very helpful, because then that can just be our source of truth, and we can update it in Heroku.

13 00:01:12.510 ⇒ 00:01:15.819 Mustafa Raja: Yeah, yeah, that’ll make a lot, things a lot easier.

14 00:01:15.820 ⇒ 00:01:20.390 Samuel Roberts: Yeah, there’s definitely other tools we can use eventually, but if we’re just doing it here, and there’s only a few of us on it, I think it’s…

15 00:01:21.650 ⇒ 00:01:22.569 Samuel Roberts: Here we go.

16 00:01:22.570 ⇒ 00:01:25.029 Mustafa Raja: At least for the platform, it’ll be smooth a lot.

17 00:01:25.260 ⇒ 00:01:28.859 Samuel Roberts: Definitely, definitely. For the platform, I think it makes the most sense.

18 00:01:31.870 ⇒ 00:01:32.829 Rico Rejoso: Hey, guys.

19 00:01:33.320 ⇒ 00:01:33.900 Mustafa Raja: Hank.

20 00:01:34.220 ⇒ 00:01:35.169 Samuel Roberts: How are you?

21 00:01:36.840 ⇒ 00:01:39.230 Rico Rejoso: good, so far?

22 00:01:40.270 ⇒ 00:01:40.940 Samuel Roberts: Good.

23 00:01:43.420 ⇒ 00:01:45.019 Rico Rejoso: Yeah, let’s be 20 others.

24 00:01:45.300 ⇒ 00:01:46.660 Samuel Roberts: Of course, of course.

25 00:02:18.000 ⇒ 00:02:19.870 Samuel Roberts: Henry, the Amber.

26 00:02:20.840 ⇒ 00:02:21.960 Henry Zhao: Hey, guys.

27 00:02:23.340 ⇒ 00:02:24.830 Henry Zhao: Amber, how was your time off?

28 00:02:25.420 ⇒ 00:02:30.200 Samuel Roberts: Yeah, she’s still connecting to audio, hopefully something’s… Not a problem there.

29 00:02:48.440 ⇒ 00:03:00.989 Amber Lin: Hello, I think Udam is in a sales call. Let’s wait a little bit for everybody, and then I think we can get started. You might join a few minutes later.

30 00:03:02.000 ⇒ 00:03:02.650 Samuel Roberts: Okay.

31 00:03:19.380 ⇒ 00:03:22.559 Samuel Roberts: How was everyone? Amber, how was your… how was your time off?

32 00:03:23.150 ⇒ 00:03:24.070 Samuel Roberts: They’re here.

33 00:03:24.480 ⇒ 00:03:28.420 Amber Lin: It was very nice. My parents just left yesterday night.

34 00:03:28.420 ⇒ 00:03:29.540 Samuel Roberts: Oh, boy.

35 00:03:29.540 ⇒ 00:03:33.860 Amber Lin: Yeah, so, I also got to see my brother, which is very, very interesting.

36 00:03:33.860 ⇒ 00:03:34.200 Samuel Roberts: That’s.

37 00:03:34.200 ⇒ 00:03:34.820 Amber Lin: new.

38 00:03:35.540 ⇒ 00:03:36.460 Samuel Roberts: Good kid, Ken.

39 00:04:02.270 ⇒ 00:04:03.090 Amber Lin: Okay.

40 00:04:03.920 ⇒ 00:04:10.740 Amber Lin: Rico, are you sharing screen, or is Sam sharing screen? I think we can get started, then Utem will probably join.

41 00:04:12.510 ⇒ 00:04:13.430 Samuel Roberts: Oh, yeah.

42 00:04:13.740 ⇒ 00:04:14.470 Samuel Roberts: Okay.

43 00:04:15.850 ⇒ 00:04:18.849 Samuel Roberts: I can go ahead and share mine if that’s how you want to do it.

44 00:04:19.690 ⇒ 00:04:21.279 Samuel Roberts: Recap, that works.

45 00:04:21.700 ⇒ 00:04:23.059 Amber Lin: Yeah, sounds good.

46 00:04:23.060 ⇒ 00:04:23.900 Rico Rejoso: Definitely enjoyable.

47 00:04:23.900 ⇒ 00:04:27.780 Samuel Roberts: I can… Okay, hold on. Slide…

48 00:04:28.330 ⇒ 00:04:32.330 Samuel Roberts: What’s the command to get into the… The view.

49 00:04:32.710 ⇒ 00:04:34.889 Samuel Roberts: Oh, okay, I’m Eastern, okay.

50 00:04:35.550 ⇒ 00:04:38.580 Samuel Roberts: Where? Oh, okay, hold on.

51 00:04:39.360 ⇒ 00:04:43.369 Samuel Roberts: My computer’s freaking out now, there we go. Still freaking out. Okay.

52 00:04:44.620 ⇒ 00:04:46.650 Samuel Roberts: Let us share…

53 00:04:52.510 ⇒ 00:04:54.969 Samuel Roberts: Alright, can everyone see what I got up here now?

54 00:04:58.320 ⇒ 00:04:58.860 Amber Lin: Yep.

55 00:04:59.440 ⇒ 00:05:01.569 Samuel Roberts: We are good. Okay, cool.

56 00:05:01.710 ⇒ 00:05:10.579 Samuel Roberts: So, I don’t know if I’m walking through all this, but I, I, I am… with the agenda, I got the icebreaker and the live chat today, so,

57 00:05:10.980 ⇒ 00:05:15.930 Samuel Roberts: I guess we need to start with that. Is that the best way to go through this? That’s the next thing.

58 00:05:16.890 ⇒ 00:05:19.850 Samuel Roberts: In fact, I probably should have just, oops.

59 00:05:20.680 ⇒ 00:05:36.430 Samuel Roberts: Jumped over to here. Okay, so, my lab chair and my icebreaker are going to be kind of related. So, I’m curious, what is everyone’s… I’m probably dating myself a little bit here, but I’m curious what people’s first memory of the… of using a computer or using the internet was.

60 00:05:36.770 ⇒ 00:05:38.930 Samuel Roberts: So, I’m…

61 00:05:39.670 ⇒ 00:05:50.120 Samuel Roberts: I remember using a computer as a kid a little bit. I played some games and stuff. We had a computer a few different times in different phases, but I remember the first time getting online, it was in the basement.

62 00:05:50.170 ⇒ 00:06:01.649 Samuel Roberts: of my house, because that’s where the connection came in, and all the dial-up stuff. I’m that old. But I also remember the first time using it in school. I remember discovering Google for the first time in fifth grade, so that really dates me a bit more.

63 00:06:01.650 ⇒ 00:06:14.859 Samuel Roberts: But I’m curious what other people’s first memories of, like, the internet, or the computer in general, or whatever it is that you guys have. So, I guess I’ll, hold on, I don’t have it… I don’t have the list in front of me of who all is here.

64 00:06:14.860 ⇒ 00:06:15.450 Henry Zhao: I’ll go next.

65 00:06:15.450 ⇒ 00:06:18.689 Samuel Roberts: Perfect, thank you. I was gonna toss, but I’m seeing the presentation.

66 00:06:18.690 ⇒ 00:06:29.589 Henry Zhao: I actually… yeah, I actually remember when my computer, like, started up on, like, the command prompt, like, Windows doesn’t even start up automatically, you have to do, like, runwin.exe to even start.

67 00:06:29.730 ⇒ 00:06:33.579 Henry Zhao: And all there was was, like, a file explorer, minesweeper, and, like.

68 00:06:34.680 ⇒ 00:06:36.569 Henry Zhao: Word, I think, maybe. I don’t even remember.

69 00:06:36.780 ⇒ 00:06:37.580 Samuel Roberts: Yeah.

70 00:06:38.380 ⇒ 00:06:41.340 Samuel Roberts: Yeah, I have some strong memories of that, too.

71 00:06:41.450 ⇒ 00:06:42.560 Samuel Roberts: That’s really cool.

72 00:06:43.150 ⇒ 00:06:46.169 Samuel Roberts: Yeah, and then in school, I would play, like, word munchers and…

73 00:06:46.170 ⇒ 00:06:48.880 Henry Zhao: math munchers, those kind of things. It’s, like, very basic.

74 00:06:48.880 ⇒ 00:06:49.620 Samuel Roberts: Nice.

75 00:06:49.810 ⇒ 00:06:54.820 Samuel Roberts: That’s cool, yeah. I remember, Chip’s Challenge. Does anyone remember that game? It was like a…

76 00:06:55.600 ⇒ 00:07:04.129 Samuel Roberts: an old game I played. It was old for me, too, and I was already, you know, probably a little older than everyone else, but… yeah, alright. Cool. Yeah, anyone you want to toss it to, then?

77 00:07:04.710 ⇒ 00:07:07.240 Henry Zhao: Toss it to… Amber.

78 00:07:09.220 ⇒ 00:07:11.680 Amber Lin: It’s just sending in a channel, so I…

79 00:07:11.680 ⇒ 00:07:12.340 Samuel Roberts: Oh, fair.

80 00:07:12.340 ⇒ 00:07:30.939 Amber Lin: I was playing, this game in the summer when my parents were not home with my cousin, when we were supposed to do homework, and then we would play this until they come home, and we would turn the computer off in one minute, and then blow the fan on the computer to cool it off, so they didn’t.

81 00:07:30.940 ⇒ 00:07:31.569 Samuel Roberts: Oh my goodness.

82 00:07:31.570 ⇒ 00:07:32.640 Amber Lin: playing games.

83 00:07:33.490 ⇒ 00:07:34.450 Samuel Roberts: That’s so fun!

84 00:07:36.650 ⇒ 00:07:37.330 Samuel Roberts: I don’t know.

85 00:07:39.100 ⇒ 00:07:41.909 Samuel Roberts: Cool, cool. Anyone else wanna… who you wanna toss it to?

86 00:07:42.160 ⇒ 00:07:44.370 Henry Zhao: We both definitely had Asian parents, Amber.

87 00:07:45.120 ⇒ 00:07:48.519 Amber Lin: Yes. That is how they are.

88 00:07:48.740 ⇒ 00:07:50.110 Samuel Roberts: Yep, yep

89 00:07:53.200 ⇒ 00:08:02.830 Samuel Roberts: Okay, so let’s see, we just mentioned playing Mario, there you go, that’s a good one. Ryan mentioned Windows XP, going down to yours revenge. Oh, Windows XP was such a good one.

90 00:08:03.330 ⇒ 00:08:06.800 Samuel Roberts: I kept that way longer than I should have on certain computers.

91 00:08:08.170 ⇒ 00:08:11.120 Samuel Roberts: anyone else wanna chime in?

92 00:08:14.510 ⇒ 00:08:16.250 Samuel Roberts: Come on, it’s an icebreaker, guys.

93 00:08:16.870 ⇒ 00:08:18.539 Awaish Kumar: Gotta break the ice some more.

94 00:08:20.750 ⇒ 00:08:21.760 Samuel Roberts: Alright, well…

95 00:08:21.760 ⇒ 00:08:23.440 Awaish Kumar: Internet, I think I…

96 00:08:23.610 ⇒ 00:08:23.960 Samuel Roberts: Yeah.

97 00:08:23.960 ⇒ 00:08:26.820 Awaish Kumar: I remember using, like, wireless cards.

98 00:08:27.530 ⇒ 00:08:30.399 Awaish Kumar: On top of my internet package that I have…

99 00:08:31.270 ⇒ 00:08:35.020 Awaish Kumar: Like, 100 dollar, 100 rupees, like, God.

100 00:08:35.159 ⇒ 00:08:41.469 Awaish Kumar: Would you want to give you some… Access to internet, like, an hour or something?

101 00:08:41.470 ⇒ 00:08:43.770 Samuel Roberts: Yeah, yeah, I remember, yeah.

102 00:08:43.960 ⇒ 00:08:45.459 Samuel Roberts: Yeah, it’s crazy how, like.

103 00:08:45.600 ⇒ 00:08:50.119 Samuel Roberts: different things used to be, and that’s kind of what I wanted to get into with this. So yeah, I’m curious… yeah, that’s great.

104 00:08:51.230 ⇒ 00:08:53.189 Awaish Kumar: And that’s when I started to, like…

105 00:08:53.440 ⇒ 00:08:56.540 Awaish Kumar: learn, like, what Gmail is, and…

106 00:08:57.540 ⇒ 00:09:01.360 Awaish Kumar: Well, like, at the time, Gmail also had, like, chat box.

107 00:09:01.630 ⇒ 00:09:02.260 Awaish Kumar: Prevent.

108 00:09:02.260 ⇒ 00:09:02.660 Samuel Roberts: Yeah.

109 00:09:03.090 ⇒ 00:09:05.570 Awaish Kumar: Instead of regressingly.

110 00:09:06.480 ⇒ 00:09:08.690 Awaish Kumar: So Okay. Yeah.

111 00:09:08.690 ⇒ 00:09:09.560 Samuel Roberts: Oh, thankful.

112 00:09:10.610 ⇒ 00:09:15.839 Samuel Roberts: Well, I have lots of things to talk about with this, because it really… what’s… I read a book a number of years ago called,

113 00:09:16.150 ⇒ 00:09:17.590 Samuel Roberts: because internet?

114 00:09:17.980 ⇒ 00:09:21.600 Samuel Roberts: Which is an interesting book about, kind of, the,

115 00:09:21.750 ⇒ 00:09:34.409 Samuel Roberts: linguistics of the internet, and, like, emojis, and, you know, elite speak, and all this other stuff that, like, as we’ve seen because of the internet. And one of the really interesting things was, like, when people joined the internet at different times, so, like, your internet

116 00:09:34.490 ⇒ 00:09:50.200 Samuel Roberts: generation is one thing it talked about. Like, if you join the internet now, you know, you’re joining it, and TikTok’s the big thing, but if you joined it a few years back, it was more Instagram, it was more Facebook, it was more MySpace, it was more, you know, all these other things that were the big social things on the internet.

117 00:09:50.200 ⇒ 00:09:58.099 Samuel Roberts: Which I find very interesting, and would love to do more, talk more about, but this one I’m gonna keep, because I’ve done something like this before, so I wanted to talk through

118 00:09:58.130 ⇒ 00:10:06.239 Samuel Roberts: This is actually just, how the internet works. So I know we have varying degrees of, tech knowledge.

119 00:10:06.260 ⇒ 00:10:19.170 Samuel Roberts: But, I wanted to just give everyone a baseline of, like, what is going on with the internet and computers in general. Last time I did this was when I was running a hair care company, so the audience was very different.

120 00:10:19.170 ⇒ 00:10:30.080 Samuel Roberts: So please, if anything is, like, too in the weeds, or not in the weeds enough, ask questions, feel free to interrupt me. I’m just gonna kind of walk through, like, very basically, like, how does what we’re doing right now even work?

121 00:10:30.210 ⇒ 00:10:36.610 Samuel Roberts: So, I wonder if we can do, like, a quick, like, show of hands or something? I don’t know how best to do that on…

122 00:10:36.720 ⇒ 00:10:40.490 Samuel Roberts: on here, especially now I can’t see everything.

123 00:10:40.490 ⇒ 00:10:41.479 Henry Zhao: It’s gonna feature them.

124 00:10:41.480 ⇒ 00:10:47.480 Samuel Roberts: Yeah, yeah, I just… yeah, so people… who has a pretty… who thinks they have a pretty good sense of, like, how the internet works, and…

125 00:10:47.670 ⇒ 00:10:49.999 Samuel Roberts: Like, understands…

126 00:10:50.960 ⇒ 00:10:54.910 Samuel Roberts: Any amount of, like, what’s going on right now when we’re in a video call, for example.

127 00:10:55.360 ⇒ 00:10:58.630 Samuel Roberts: I don’t know where I see that now that I’m sharing my screen and everything, but…

128 00:10:59.080 ⇒ 00:11:03.860 Samuel Roberts: Who doesn’t have a great sense? Maybe let’s try that, too, because I don’t… There we go.

129 00:11:06.700 ⇒ 00:11:10.869 Samuel Roberts: Is there a… is there any way to do this? I don’t even see… what am I looking for here?

130 00:11:13.100 ⇒ 00:11:15.449 Samuel Roberts: I don’t see the hands thing, does anyone see the hands thing?

131 00:11:16.610 ⇒ 00:11:18.330 Henry Zhao: I saw Amber’s red X…

132 00:11:19.380 ⇒ 00:11:23.329 Amber Lin: There’s a raised hand on the bottom.

133 00:11:24.530 ⇒ 00:11:28.809 Samuel Roberts: Yeah, I don’t see that, because I… since I’m sharing my screen, I just see, like, all the sharing stuff.

134 00:11:30.080 ⇒ 00:11:36.510 Amber Lin: But I don’t see anyone… Oh, that should be, if someone raised their hand, It…

135 00:11:36.510 ⇒ 00:11:37.590 Samuel Roberts: There we go, now I see it on the.

136 00:11:37.590 ⇒ 00:11:38.899 Amber Lin: It would happen.

137 00:11:38.900 ⇒ 00:11:40.010 Samuel Roberts: I see it on the chat, okay.

138 00:11:40.010 ⇒ 00:11:42.159 Amber Lin: But it’s not me, I don’t know it.

139 00:11:42.350 ⇒ 00:11:52.209 Samuel Roberts: That’s fine, okay, I see a green check. Okay, cool. So who doesn’t, then? Let’s buy a show of any kind of thing. Who doesn’t have a great sense? Because I’m kind of just wondering who I’m talking to today.

140 00:11:52.430 ⇒ 00:11:54.980 Samuel Roberts: Alright, cool.

141 00:11:55.350 ⇒ 00:12:03.140 Samuel Roberts: That’s good, okay, because sometimes I… I can go really deep into the… oh, let’s go back. I can get really deep into this, I can go not… I also don’t have a perfect understanding, I…

142 00:12:03.190 ⇒ 00:12:21.930 Samuel Roberts: was trained as a mechanical engineer in college. I got into web development after college, and I did not understand all of this until I was, you know, maybe a couple years in, and I finally understood, like, oh, that’s what that’s doing in here. And, like, there’s lots of different levels here. So let’s… let’s jump in. First level is binary. I think most people probably

143 00:12:22.570 ⇒ 00:12:40.379 Samuel Roberts: have a good sense of what this is, if not, like, how it works, but basically, everything digital is ones and zeros. On, off, you know, light on, light off, whatever you want to think of it as, it’s like switches, and that’s what builds up everything that we’re doing right now, including the video, including everything displayed, the audio.

144 00:12:40.380 ⇒ 00:12:51.009 Samuel Roberts: everything that we’re processing here is digital. So, for example, letters are all represented in certain different kinds of codes. There’s ASCII, there’s Unicode, there’s different ways of representing this digital data, and then computers do

145 00:12:51.150 ⇒ 00:12:53.899 Samuel Roberts: All kinds of different math on it to do what we’re doing now.

146 00:12:55.980 ⇒ 00:13:02.959 Samuel Roberts: Basically, then, what happens is, when you want to then get to, like… you have a computer that can process all this stuff, do all the things we’re used to.

147 00:13:02.990 ⇒ 00:13:22.320 Samuel Roberts: But then, as this starts to travel across networks, which we’re gonna get into, you need a way to send these signals across different computers. So if you imagine two computers sitting next to each other, there’s no way for them to communicate, because they’re two separate boxes. And so we need to send this binary data through different means. It can be wireless, like radio waves, like Wi-Fi or Bluetooth.

148 00:13:22.320 ⇒ 00:13:33.000 Samuel Roberts: We can use light, we can use electronic signals, we can use a whole bunch of stuff. And that physical layer that combines these different computers is what kind of creates the base layer of a network.

149 00:13:33.030 ⇒ 00:13:33.810 Samuel Roberts: Okay?

150 00:13:34.190 ⇒ 00:13:45.150 Samuel Roberts: So here’s a graph I pulled from the internet that kind of shows how this works. So you can imagine there’s different layers of complexity on how we do things. So if you imagine, you wanted to build Zoom from the ground up.

151 00:13:45.340 ⇒ 00:14:00.070 Samuel Roberts: and there was no computers in the world. You need to build a physical layer, right? You need to have things that actually do the work of taking the video signal, pulling that in, sending that info around. But then, on top of that, you have these different layers of abstractions that are,

152 00:14:00.420 ⇒ 00:14:09.000 Samuel Roberts: simplifying some things and giving you some other benefits. So there’s the data link layer, but this is what actually sends the data. So the transmitting raw bits is not that helpful unless you can interpret them.

153 00:14:09.150 ⇒ 00:14:25.769 Samuel Roberts: Then there’s the network layer, which figures out where things need to go once you have a more complex network, which we’ll get to. Transport layer, I don’t know if you guys have heard of TCP, UDP, these are different kinds of protocols that are built on top of this network layer that allow you to send data in different ways with different pros and cons.

154 00:14:25.990 ⇒ 00:14:33.280 Samuel Roberts: The rest of these are kind of a little more detailed than we need to get, but I’m going to jump into building a network. So, if you imagine back in the day.

155 00:14:33.410 ⇒ 00:14:40.399 Samuel Roberts: We’re talking, like, the 60s, there were computers, big computers, they called them

156 00:14:40.660 ⇒ 00:14:46.749 Samuel Roberts: minicomputers, they were the size of, like, not quite the size of whole rooms, but they wanted to talk to each other across

157 00:14:46.830 ⇒ 00:14:58.140 Samuel Roberts: the country, at least here in the US and probably around the world, but the one that became big was called ARPNET, and it was basically across telephone lines, these, like, colleges and other research centers were communicating.

158 00:14:58.140 ⇒ 00:15:10.010 Samuel Roberts: And so, from that became the internet. So you can imagine, like, big groups of computers talking to each other. And so you need certain addresses, that’s what an IP address is, if you’ve heard that before. It’s your internet protocol address. It’s basically…

159 00:15:10.110 ⇒ 00:15:28.170 Samuel Roberts: your street address for your machine, or for your, depending on how different ISPs do it for your area. Everyone has a router that actually is how their internal network at their house speaks to the external network, which is basically what we call the internet. So it’s essentially a network of networks, which, looks kind of like this.

160 00:15:28.770 ⇒ 00:15:40.549 Samuel Roberts: So the internet is a massive web, which, not to be confused with the World Wide Web, which I will get into and why that’s different. But basically, if you imagine, like, you’re one of these nodes on your computer, it’s talking to a router.

161 00:15:40.550 ⇒ 00:16:00.309 Samuel Roberts: that router is also, in some sense, a computer. It’s talking to tons of other things on the internet. That’s what your ISPs do, that’s what all of the cloud services do. And then when you want to send, for example, my video to everyone out there, and everyone’s video to me, it’s being routed through all kinds of different paths here. And you can do things that are more direct, you can do, like,

162 00:16:00.660 ⇒ 00:16:13.400 Samuel Roberts: like, virtual private networks, like VPNs, that kind of say, like, okay, we’re going here to here, let’s find that route, and let’s encrypt it and do all this stuff. You can do other things that are just, like, you know, sending out a message and sending an email, for example, and that has to travel through a bunch of different routes, and you can

163 00:16:13.400 ⇒ 00:16:28.430 Samuel Roberts: find all that information, and it’s interesting, and not interesting, and however deep you want to go. But that’s the idea of the internet, it is a collection of networks. So then, on top of that, we have these protocols. So imagine these bits going down this wire, whether it’s light, whether it’s radio waves, whether it’s electricity,

164 00:16:28.600 ⇒ 00:16:48.539 Samuel Roberts: that all has to be communicated, and there’s a very interesting field about, you know, error correction and things like that, because if you imagine, this is all real-world stuff. We’re not working in a perfect environment anymore, so bits get dropped, things have to be corrected, you know, you can imagine blurry videos coming in, or messages, you know, like, think back to, like.

165 00:16:48.540 ⇒ 00:17:06.010 Samuel Roberts: an image of the 90s, where someone’s, like, loading an image and it’s, like, going line by line. Like, that’s because it’s all coming in and it’s having to be corrected and stuff. TCP is one that’s very widely used, we’re using it probably right now in some form, that makes sure that all the data’s there. So there’s a back and forth, which means it’s a little bit slower than some others, but it means all the data got through.

166 00:17:06.089 ⇒ 00:17:23.279 Samuel Roberts: Other things that are important here, DNS, which I’m not going to get too into, is basically how you know that those IP addresses mean Google.com, or brainforce.ai, or whatever it is. So there’s, like, a lookup table that basically says, like, alright, this means that, and there’s a whole system that I’m not going to get into here, because it’s much more in-depth about how all that works.

167 00:17:23.410 ⇒ 00:17:35.719 Samuel Roberts: So, coming back to what we just talked about here, this is similar to what we just looked at for the layers of the internet, but this is a little more of, like, the protocols that are used. So if you imagine this link layer, Ethernet, Wi-Fi.

168 00:17:35.850 ⇒ 00:17:51.879 Samuel Roberts: That’s just like connecting machines, right? This IPv4 and v6, so if you’ve ever seen a number.number.number.number, that’s an IPv4 address. IPv6 has tons more, because we’re actually running out of the IPv4, which is a whole fascinating thing, as we’ve had more

169 00:17:51.880 ⇒ 00:17:59.270 Samuel Roberts: Internet of Things and phones and, like, not what people initially thought was going to happen when they first invented it. But that’s how we then communicate between the machines.

170 00:17:59.270 ⇒ 00:18:15.529 Samuel Roberts: The TCP UDP is how we send that information, and then on top of it, the things you might have seen before, HTTP, like, in front of every URL, that’s telling you that it’s using the hypertext transfer protocol, to send this information across. And that then takes us to… we’re gonna skip through this.

171 00:18:15.550 ⇒ 00:18:21.130 Samuel Roberts: We’re gonna skip through some of this, because I want to get to the World Wide Web. So.

172 00:18:22.240 ⇒ 00:18:26.370 Samuel Roberts: Very basically, the internet and the World Wide Web are two distinct things.

173 00:18:26.660 ⇒ 00:18:28.910 Samuel Roberts: So the internet is this network.

174 00:18:29.350 ⇒ 00:18:33.800 Samuel Roberts: The World Wide Web is a kind of subset of this that is what we view in the browser, basically.

175 00:18:34.110 ⇒ 00:18:50.110 Samuel Roberts: So, HTML, CSS, JavaScript, all the browsers, they’re all on the web, effectively. But if you’ve heard of things like, I don’t know, like the dark web, or even just, like, cloud stuff when we’re hitting APIs and things, we’re working on sometimes different protocols, you can send information different ways.

176 00:18:50.190 ⇒ 00:18:54.229 Samuel Roberts: This was an… this is… literally was invented by Tim Berners-Lee.

177 00:18:54.230 ⇒ 00:19:12.920 Samuel Roberts: in a specific time period at CERN, because before this, there were a bunch of different ways you could do it. I was just watching a video recently about one called Gopher that’s just text-based. It’s just text and files, and there’s no video, there’s no images, there’s no links, things like that. So what was great about this and why it kind of blew up was because the ability to link things together across this network.

178 00:19:14.390 ⇒ 00:19:22.379 Samuel Roberts: So the World Wide Web kind of sits on top of the internet, which sits on top of all these other protocols, which sits on top of just a physical network of machines.

179 00:19:22.540 ⇒ 00:19:40.319 Samuel Roberts: Brief history of what I just kind of talked through for kind of reference. I wanted to throw this in. I mentioned ARPNet, ARPNet, sorry, goes back to the 60s. Email was invented in 1971, so if you imagine, like, people weren’t browsing the web before that, but you could still send messages to each other over this protocol called email.

180 00:19:40.340 ⇒ 00:19:54.020 Samuel Roberts: That’s where some of these things come for, like, the at symbol and stuff, which I think predates that a little bit, but that’s what it was used for, because you’re identifying what server someone was on and all this. The web itself didn’t come around until, like I said, 1990. It was invented in 89, got all

181 00:19:54.210 ⇒ 00:19:55.540 Samuel Roberts: you know, used…

182 00:19:55.690 ⇒ 00:20:09.789 Samuel Roberts: a mosaic came out, a whole bunch of things, and then that’s kind of where everything took off from there. So you can see there’s a pretty big gap of time here, there’s a pretty big gap of time here, but if you imagine, like, everything that happened on the internet exponentially, like, this is where all the stuff we’re used to has kind of come about.

183 00:20:11.100 ⇒ 00:20:17.169 Samuel Roberts: This is just kind of modern stuff that we don’t really need to get into, recap, and I think that’s all I got.

184 00:20:17.170 ⇒ 00:20:35.859 Samuel Roberts: So, I kind of breezed through that, because I didn’t want to take too much time, and I can go way more in-depth on certain things, and there are plenty of things I don’t have the knowledge to even discuss, but that’s a very high level of how, like, your machine sends ones and zeros to other machines, to other machines, all bouncing around, and then a video gets streamed to you through that.

185 00:20:37.380 ⇒ 00:20:45.940 Samuel Roberts: I find it very interesting. I’m not sure if everyone does, but this is at least now you have a little bit of a sense, if you didn’t know, about the layers that it goes through. I’m happy to…

186 00:20:46.300 ⇒ 00:20:51.960 Samuel Roberts: talk more about this if anyone has any questions or anything, but that’s my lab share, for today.

187 00:20:51.960 ⇒ 00:20:52.380 Uttam Kumaran: Yeah.

188 00:20:52.380 ⇒ 00:20:53.119 Samuel Roberts: Sorry if I went a little.

189 00:20:53.120 ⇒ 00:20:59.260 Uttam Kumaran: Can I ask questions about, updates in, like, protocols?

190 00:20:59.260 ⇒ 00:20:59.610 Samuel Roberts: Yeah.

191 00:20:59.610 ⇒ 00:21:04.030 Uttam Kumaran: Did you see the update? Like, what was the most recent update to, like,

192 00:21:04.430 ⇒ 00:21:08.690 Uttam Kumaran: like, one of the IP protocols, where there’s, like, V5 to V7, like, what are.

193 00:21:08.690 ⇒ 00:21:09.190 Samuel Roberts: Yeah.

194 00:21:09.190 ⇒ 00:21:11.290 Uttam Kumaran: what are the updates that are happening, and like…

195 00:21:11.500 ⇒ 00:21:15.230 Uttam Kumaran: are these, like, rela… like, yeah, V4 to V6, like.

196 00:21:15.230 ⇒ 00:21:15.920 Samuel Roberts: Yeah, that’s.

197 00:21:15.920 ⇒ 00:21:23.380 Uttam Kumaran: What is in these updates? And, like, how do they get permeated? Like, who… Like, does it, and like…

198 00:21:23.660 ⇒ 00:21:39.210 Uttam Kumaran: is it, like, a… something on my machine? I assume there is, like, a… I assume I have to have the right card to use that. I guess, and for everybody on the call, maybe you could… if you could somehow equate it to, like, USB-C versus B…

199 00:21:39.680 ⇒ 00:21:43.490 Samuel Roberts: Totally, yeah. I don’t know, yeah. Yeah, yeah. So if you imagine,

200 00:21:44.010 ⇒ 00:21:53.299 Samuel Roberts: This is actually something I realized when I was living in London, and I realized the phone plugs were all different. Because when phones were invented, the world wasn’t quite as…

201 00:21:53.520 ⇒ 00:22:08.529 Samuel Roberts: flat, and, like, they had different phone plugs than we had here, whereas Ethernet came around in a much flatter world, and that’s pretty standard everywhere, right? So that’s… that’s helpful, is that, like, we now have these international organizations, that

202 00:22:09.380 ⇒ 00:22:16.410 Samuel Roberts: kind of have all of these processes and committees and voting procedures to update these protocols, because if you imagine, like.

203 00:22:16.470 ⇒ 00:22:31.450 Samuel Roberts: when they invented IPv4, or IP whatever, probably previous to that, but the one that we’ve been using for years is IPv4, it’s… it’s four, what is it? Octets of numbers, so it’s, like, 123.456.789.

204 00:22:31.580 ⇒ 00:22:33.620 Samuel Roberts: 000, whatever.

205 00:22:33.880 ⇒ 00:22:38.839 Samuel Roberts: there’s only so many of those, right? So when something needs to update to this kind of V6,

206 00:22:39.040 ⇒ 00:22:52.129 Samuel Roberts: Those numbers are wildly different, but there’s way more of them. But in order to do that, they have to agree on how we’re going to communicate. Some things are going to be hardware level, so not everyone’s going to be able to get it initially. IPv6 has been around for…

207 00:22:52.970 ⇒ 00:22:57.869 Samuel Roberts: way longer than you would realize, considering the fact that we still use a lot of IPv4.

208 00:22:58.700 ⇒ 00:23:04.920 Samuel Roberts: Basically, it’s a slow process of people suggesting these things, like…

209 00:23:05.850 ⇒ 00:23:22.969 Samuel Roberts: proposing them, there’s all kinds of proposals, and then those make it through a committee level, and those make it through, like, a standardization until everyone kind of agrees on it. And while that’s happening, people are also then building hardware, including it into the hardware they sell, because they want to future-proof it a little bit more. Think of it like, VHS,

210 00:23:22.970 ⇒ 00:23:25.970 Samuel Roberts: Betamax, DVDs,

211 00:23:27.300 ⇒ 00:23:37.569 Samuel Roberts: I figure out what was DVDs at the same time, but then, like, Blu-ray and HD DVD. Like, they’re different protocols, and we want to kind of avoid that with the internet as much as possible, because you want…

212 00:23:37.740 ⇒ 00:23:45.460 Samuel Roberts: people to be speaking the same language across the internet so things communicate well. So there’s some hardware that’s gonna have to get updated, there’s some firmware that’s gonna have to be updated.

213 00:23:45.460 ⇒ 00:23:48.899 Uttam Kumaran: And then what are the ben… like, what are the core benefits, though? Like, if you’re.

214 00:23:48.900 ⇒ 00:24:08.230 Samuel Roberts: Oh, sure, yeah. So, part of it is, like, speed limitations. If you’ve heard of, like… like, Ethernet isn’t a catch-all, but there’s different types of cabling, for example. And so, like, Cat5 or Cat6, if anyone’s heard that before, are different types of cables that are just essentially the same cable, but just, like, capable of speeds that are a little bit better. And so if those speeds are improved…

215 00:24:08.230 ⇒ 00:24:11.419 Uttam Kumaran: SBC, for example, you can transmit power and data.

216 00:24:11.420 ⇒ 00:24:18.759 Samuel Roberts: Exactly, so yeah, right, from the USB-C standpoint, you can do a lot more. That’s hardware, that’s software, that’s,

217 00:24:18.760 ⇒ 00:24:33.569 Samuel Roberts: you know, error correcting, all these algorithms that people are improving. All of those things go together to say, like, oh, we can force more through the same cable using either faster hardware, you know, better algorithms, better…

218 00:24:33.570 ⇒ 00:24:41.839 Samuel Roberts: like, experience, like the plug-in stuff, like, that makes a difference. How many… there’s more wires in a USB-C cable than any USB cable, so there’s just more throughput in general.

219 00:24:41.840 ⇒ 00:24:50.710 Samuel Roberts: for things like the internet, for IPv6, they’re actually, like I said, they’re running out of addresses. So if you imagine, there’s only a set number of IPv4 addresses that can exist.

220 00:24:50.880 ⇒ 00:24:52.370 Samuel Roberts: Because it’s…

221 00:24:52.550 ⇒ 00:25:07.669 Samuel Roberts: it’s a readable number, basically, and the IPv6 is, like, just this long string that is very hard to human read, but you can have… I don’t want to even try to say, tens of thousands, millions more addresses, and when we live in a world today versus, you know, whatever I said that was,

222 00:25:08.730 ⇒ 00:25:25.510 Samuel Roberts: like, 1983, it’s a very different need right now. Like, the speeds can increase, and you can use the same kind of IP format to pass data, and you can improve that, but to then connect millions more machines, you need a different addressing system, which is a whole other piece of this process.

223 00:25:25.510 ⇒ 00:25:32.320 Samuel Roberts: So yeah, USB, the USB-C is kind of, like, there’s a lot of different pieces. There’s hardware, there’s software, there’s firmware.

224 00:25:32.400 ⇒ 00:25:48.629 Samuel Roberts: You know, there’s also, like, length of cables. Like, you can only get a USB-C cable that’s a certain length long, and they’re rated different ways, and so the hardware, meaning, like, the cabling is also different. There’s lots of different things we don’t think about at, like, kind of the microscale of when we’re just using our computers, but if you imagine scaling that up to the world.

225 00:25:48.950 ⇒ 00:25:54.319 Samuel Roberts: all of that stuff becomes more applicable. So I don’t know if that really covered the question very well,

226 00:25:54.320 ⇒ 00:25:55.100 Uttam Kumaran: Okay.

227 00:25:55.100 ⇒ 00:26:06.339 Samuel Roberts: But yeah, if you imagine, like, the speeds they were getting back then versus now, versus the volume back then versus now, a lot of these things have to change to support that massive infrastructure that’s… or the massive influx into that infrastructure, so…

228 00:26:07.190 ⇒ 00:26:14.580 Samuel Roberts: I’m not sure if that was a helpful explanation. I don’t necessarily know the details of USB, well enough, but to compare it is basically that kind of thing, where.

229 00:26:14.580 ⇒ 00:26:17.339 Uttam Kumaran: I mean, tell me about post-2020s, like.

230 00:26:17.660 ⇒ 00:26:18.000 Samuel Roberts: Yeah.

231 00:26:18.000 ⇒ 00:26:22.490 Uttam Kumaran: A couple things that are on my mind about internet. One is you have…

232 00:26:22.640 ⇒ 00:26:26.509 Uttam Kumaran: Now that we have, sort of satellite mesh.

233 00:26:26.690 ⇒ 00:26:30.969 Samuel Roberts: It’s really cool about how many parts of the world can now.

234 00:26:30.990 ⇒ 00:26:32.450 Uttam Kumaran: Get access to internet.

235 00:26:32.460 ⇒ 00:26:40.899 Uttam Kumaran: at not only just, like, basic, but at, like, extremely fast speeds. So there is gonna be some disruption of the traditional internet providers, right? Like.

236 00:26:40.910 ⇒ 00:26:54.890 Uttam Kumaran: And that’ll kind of transfer to companies like Starlink, the Amazon that are doing these big satellite launches, so I think that’s one thing. I guess I’m also curious about, like, there’s a lot of fiber, so, like, my house has fiber.

237 00:26:55.890 ⇒ 00:27:03.609 Uttam Kumaran: Right, that’s another thing. I think, but, you know, the downside is, like, everybody now is streaming 1080p, 4K.

238 00:27:03.610 ⇒ 00:27:05.710 Samuel Roberts: That’s what happened. That’s exactly what I was gonna say.

239 00:27:05.710 ⇒ 00:27:21.740 Uttam Kumaran: Like, and we’re all, like… like, I’m on a 1080p 4K zoom right now with all these people across the world, so think about how much of that… so I guess I’m interested in, like, what you think is gonna be for the future of the internet, if there are gonna be these, like,

240 00:27:21.830 ⇒ 00:27:27.719 Uttam Kumaran: There is gonna be another, like, clear delineation of, like, a next phase, you know, or not.

241 00:27:27.900 ⇒ 00:27:39.880 Samuel Roberts: Yeah, I think there definitely will be. I mean, you hit the first thing I was gonna say, which is the fact that, like, as speeds increase, like, volume… it’s like the, induced demand kind of thing with traffic. Like, you build another road, or build another lane, you’re gonna.

242 00:27:39.880 ⇒ 00:27:40.450 Uttam Kumaran: Yes. More traffic.

243 00:27:40.450 ⇒ 00:27:43.959 Samuel Roberts: And if you’re gonna just create another traffic jam at a higher volume of traffic.

244 00:27:43.960 ⇒ 00:27:45.679 Uttam Kumaran: That’s a problem here in Texas, by the way.

245 00:27:45.680 ⇒ 00:27:46.400 Samuel Roberts: Yeah, it’.

246 00:27:46.400 ⇒ 00:28:04.049 Uttam Kumaran: they’re very stupid, and so what they do is they just add another lane, thinking that solves the problem, when in fact, like, it doesn’t. And so there’s a lot of traffic theory, you know, that, like, applies, where if you increase the amount of availability, like, the demand will increase.

247 00:28:04.490 ⇒ 00:28:13.530 Samuel Roberts: Yeah, exactly. And I think the internet is… either follows that in some way, but also, like, as we are able to process 4K video more locally, because

248 00:28:13.530 ⇒ 00:28:25.479 Samuel Roberts: you know, chips have gotten better, graphics have gotten better, we’re then able to send that, and we need the capacity to do that, and as we build the capacity to do that, the hardware’s getting better, and it’s just kind of back and forth between, you know, using up all that bandwidth.

249 00:28:25.500 ⇒ 00:28:40.199 Samuel Roberts: So I think there’s definitely some of that where, you know, things just… the overall numbers just get bigger. But where I think that’s interesting is, like you said, like, as more… like, I forget exactly the numbers of, like, people in the world that are online, and that obviously has room to grow a ton more with more…

250 00:28:40.320 ⇒ 00:28:46.579 Samuel Roberts: Satellites and mesh technologies and other things, but that’s one thing that I think is really interesting from

251 00:28:46.630 ⇒ 00:29:03.630 Samuel Roberts: from my perspective, which is maybe a little bit different than what you were asking, but some of this stuff comes from, like, the really, like, the open internet. Like, a lot of this stuff was open source, and we’ve gone into such a closed source world, partially. And a lot of people are working in different ways to open that up, but the hardware level, the network level, is very hard to do

252 00:29:04.170 ⇒ 00:29:15.639 Samuel Roberts: Mesh is something that may be a little bit better at that, getting people online, you know, through, you know… Right now, you have to go through an ISP, but you could theoretically, like, not use any closed-source stuff, and you could do a whole bunch of other things, and use a different internet.

253 00:29:16.000 ⇒ 00:29:16.320 Uttam Kumaran: Yeah.

254 00:29:16.320 ⇒ 00:29:21.170 Samuel Roberts: All these different things. Mesh networks really open that up to, like, more people getting online, and not necessarily being…

255 00:29:21.170 ⇒ 00:29:21.620 Uttam Kumaran: Helios.

256 00:29:22.710 ⇒ 00:29:28.080 Samuel Roberts: Yeah, yeah, yeah, exactly, exactly. Some of that stuff is good, like, you know, Web2 is very, like…

257 00:29:28.810 ⇒ 00:29:37.379 Samuel Roberts: how we ended up with Google and Facebook and Amazon, like, there are good things underneath it that were, like, the open internet that people kind of were promised.

258 00:29:37.520 ⇒ 00:29:38.580 Samuel Roberts: And then

259 00:29:38.790 ⇒ 00:29:52.090 Samuel Roberts: but there was no, like, monetization built in, which is a big thing. And so, like, now we’ve kind of… I don’t know if we’ve turned the corner and started, people are thinking about that in a way that’s, like, as more people get online, are they going to be going through all these gatekeepers? Are they going to be going through…

260 00:29:52.090 ⇒ 00:30:00.050 Samuel Roberts: these ISPs, how much control does the ISP have? There’s a whole bunch of net neutrality stuff here as well. There’s a lot of really interesting things that I don’t necessarily have a good

261 00:30:00.240 ⇒ 00:30:03.280 Samuel Roberts: Sense of where it’s going.

262 00:30:03.420 ⇒ 00:30:08.120 Samuel Roberts: I mean, IPv6 is one of those things that’s been on the horizon forever, because we need it, because…

263 00:30:08.590 ⇒ 00:30:13.970 Samuel Roberts: Everyone’s gonna be online eventually, and that’s billions of people, and there’s not billions of addresses, but,

264 00:30:14.360 ⇒ 00:30:18.740 Samuel Roberts: you know, and also you’re online with multiple machines. You have your phone, you have your laptop, people have

265 00:30:19.070 ⇒ 00:30:21.179 Samuel Roberts: People have, you know, all of these different things, and so there’s…

266 00:30:21.440 ⇒ 00:30:28.339 Samuel Roberts: Internet of Things is a whole other… other side of this, where, like, everything’s connected, everything needs an address, but I don’t have a good sense, yeah, go ahead.

267 00:30:28.340 ⇒ 00:30:43.060 Uttam Kumaran: Yeah, we do a lot of work around data privacy stuff, so, like, affected a lot in, like, even when I was in industry was, like, when CCPA came out, when GDPR came out, there’s a lot of restrictions, but what you’re finding now is.

268 00:30:43.130 ⇒ 00:30:56.479 Uttam Kumaran: Although those rules exist and the fines can be big, adhering to them are very… is very difficult. And this is where the regulatory body, was, like, extremely dumb, in that they built these rules in order to just, like.

269 00:30:56.480 ⇒ 00:31:05.549 Uttam Kumaran: have something there, but the EU in particular has, like, no real great way of enforcing, and they’re just pro- they’re just mainly using it as a way to, like.

270 00:31:05.880 ⇒ 00:31:25.580 Uttam Kumaran: in some ways, curb free speech, in some ways just make money off of, like, charging Google and Apple, and they’ve shifted these regulations, so that’s really, really tough. There was a whole push for net neutrality and things like that. I think the other piece, you know, and this is where, like, I want to ask, like, some of our friends that are in Asia. When I go to India, for example, they skipped

271 00:31:25.750 ⇒ 00:31:26.989 Uttam Kumaran: Like, Wi-Fi.

272 00:31:27.480 ⇒ 00:31:29.940 Uttam Kumaran: So they went straight from, like.

273 00:31:31.040 ⇒ 00:31:34.059 Uttam Kumaran: I mean, I guess, like, not much to just cellular, right?

274 00:31:34.440 ⇒ 00:31:41.200 Uttam Kumaran: and I’m sure this is the case in Pakistan, in Asia, you can get Sims with pretty hefty

275 00:31:41.330 ⇒ 00:31:44.569 Uttam Kumaran: Like, you know, amount of gigs, like…

276 00:31:44.610 ⇒ 00:32:01.680 Uttam Kumaran: for pretty cheap. In the US, in particular, if you come here to the US, you see there’s a huge mafia around that, like, you have to pay, like, hundreds of dollars a month for unlimited, where in India, like, dude, you go get, like, a 100GB SIM for, like, $0, you get 5G everywhere, and, like, there’s nobody’s on Wi-Fi.

277 00:32:02.380 ⇒ 00:32:06.230 Uttam Kumaran: So, I don’t know, I feel like the US has a long way to go in some of these.

278 00:32:06.230 ⇒ 00:32:06.700 Samuel Roberts: Oh, definitely.

279 00:32:06.700 ⇒ 00:32:24.239 Uttam Kumaran: Mainly because what we do have positive is, like, in capitalism, is, like, we have some more optionality, but then you end up with industries like internet or cellular, where there are… there is, like, regulatory capture, and there is only a couple of companies, and then they just jack the prices.

280 00:32:24.240 ⇒ 00:32:24.900 Samuel Roberts: Comment.

281 00:32:25.220 ⇒ 00:32:28.719 Uttam Kumaran: So, yeah, that’s the… That’s the tough part.

282 00:32:28.990 ⇒ 00:32:39.870 Samuel Roberts: Totally, totally. I mean, there’s a lot of that where, you know, fiber has taken so long for us to get here in the US, and other developed countries have had better speeds for longer. Places in the US still don’t even have.

283 00:32:39.870 ⇒ 00:32:40.710 Uttam Kumaran: Alex, should we even…

284 00:32:40.710 ⇒ 00:32:41.199 Samuel Roberts: and continue.

285 00:32:41.200 ⇒ 00:32:49.019 Uttam Kumaran: We need to invest in… on-the-ground stuff when, like, Wi-Fi and satellite is getting so fast, like.

286 00:32:49.130 ⇒ 00:32:50.010 Samuel Roberts: Why? Yeah.

287 00:32:50.010 ⇒ 00:32:53.369 Uttam Kumaran: like, why… why do you need fiber, right? So this is, like…

288 00:32:53.520 ⇒ 00:33:01.860 Uttam Kumaran: Where the U.S, unfortunately, like, the regulatory regime is just really stuck in the Stone Age on a lot of this, and

289 00:33:02.420 ⇒ 00:33:05.509 Uttam Kumaran: I don’t know, I feel like in other countries, because they’re more, like.

290 00:33:06.000 ⇒ 00:33:11.769 Uttam Kumaran: the government has control, can make faster decisions. Like, India, for example, rolled out,

291 00:33:12.190 ⇒ 00:33:24.109 Uttam Kumaran: like, UPI, like, Universal Payments, like, interface, and in other countries, too. Like, I don’t know, Mustafa or Wish, like, is there a UPI equivalent or a payment equivalent in Pakistan?

292 00:33:25.120 ⇒ 00:33:30.970 Mustafa Raja: There are wallets, but I don’t think there’s an alternative for UPI.

293 00:33:33.480 ⇒ 00:33:36.639 Uttam Kumaran: Are you guys particularly cash, or is there… is there a.

294 00:33:36.640 ⇒ 00:33:44.090 Mustafa Raja: I don’t know. We do have digital payment solutions. We have wallets.

295 00:33:44.340 ⇒ 00:33:48.159 Mustafa Raja: That are regulated by State Bank of Pakistan.

296 00:33:48.160 ⇒ 00:33:50.450 Uttam Kumaran: Yeah. But they’re not really banks.

297 00:33:51.410 ⇒ 00:33:52.889 Mustafa Raja: They’re just wallets.

298 00:33:52.890 ⇒ 00:33:59.479 Uttam Kumaran: like, for example, in India, every place has a UPI, like, QR code.

299 00:33:59.480 ⇒ 00:33:59.930 Samuel Roberts: Yeah.

300 00:33:59.930 ⇒ 00:34:03.920 Uttam Kumaran: as a business, that’s how you collect… that’s how you collect, and everybody’s phone…

301 00:34:04.070 ⇒ 00:34:10.129 Uttam Kumaran: you… when you register, you have to then sign up for that so you can pay. So, inbuilt in the infrastructure around.

302 00:34:10.139 ⇒ 00:34:10.569 Samuel Roberts: you, you know.

303 00:34:10.570 ⇒ 00:34:28.690 Uttam Kumaran: and your phones are this payment interface, where here we have… we have… you have to do Apple Pay, or Touch2Pay, or Tap, or swipe, and my dad works a lot in payments, so I’ve done… so I’ve learned a lot about that world. There’s a lot of regulatory capture, because certain people want to have the rails, you have Visa and MasterCard that own those rails.

304 00:34:28.830 ⇒ 00:34:44.729 Uttam Kumaran: this is… the way this is going now is you’ve… if people have read a lot of… if anyone who’s in crypto knows about stablecoins and, like, USDC, and what stablecoins is allowing to do is you… you can transact faster and have a record of

305 00:34:44.840 ⇒ 00:35:04.289 Uttam Kumaran: like, money movement, and have a delimiter in U.S. dollars versus, like, one person has to store it, transfer, and there’s all these brokers. And so what’s… what’s… what is most likely going to happen is… is you’re… we’re going to be able to enable a lot faster digital payments, like real-time settlement. Like, a problem for us at our company, like, these are a perfect problem.

306 00:35:04.290 ⇒ 00:35:20.969 Uttam Kumaran: I have money. I’m trying to pay our team. The money takes sometimes a few days, sometimes, like, more than a week to land. That is something, as a business owner, I have no control over, and I have no alternative. Like, I can’t… the settlement time is actually impacted by those

307 00:35:20.970 ⇒ 00:35:32.360 Uttam Kumaran: each of these people in the chain, like, if I have to pay through RAMP, then RAMP takes a little cut, then RAMP hands it to someone else, they take a cut, then they go to your bank, they take a cut. Instead, like, if we could have real-time settlements.

308 00:35:32.360 ⇒ 00:35:43.839 Uttam Kumaran: in crypto, in USDC, it would be very easy, because I could just transfer it to you. And so, a company that’s… that was doing a lot in this is called, Circle. Circle,

309 00:35:45.200 ⇒ 00:35:48.729 Uttam Kumaran: Let me find it… USDC…

310 00:35:50.830 ⇒ 00:35:56.050 Uttam Kumaran: So these guys were sort of the pioneers in a lot of,

311 00:35:56.680 ⇒ 00:36:00.359 Uttam Kumaran: USDC, and even a friend of mine was like, maybe you should consider

312 00:36:00.710 ⇒ 00:36:11.720 Uttam Kumaran: you know, having folks on your team go on circle, and they were… they’re able to accept payments there because it improves the real-time settlement, but these are all things that even affect us as a business, right?

313 00:36:11.860 ⇒ 00:36:12.770 Samuel Roberts: Yeah.

314 00:36:12.770 ⇒ 00:36:13.700 Uttam Kumaran: So…

315 00:36:14.410 ⇒ 00:36:24.420 Samuel Roberts: This is something I discovered when I was in London, was how much faster the banking system is there. And I think that has to do with the fact that the U.S. has been so, yeah, captured by so many

316 00:36:24.420 ⇒ 00:36:34.849 Samuel Roberts: like, big companies, they haven’t needed to change, whereas the EU, when they finally got all these countries together, had to figure out a newer system, and so, like, money moves immediately, like…

317 00:36:34.850 ⇒ 00:36:39.560 Samuel Roberts: in those banking systems, even pre, you know, whatever crypto stuff. Like, whereas in the US, it takes

318 00:36:39.560 ⇒ 00:36:47.950 Samuel Roberts: days, it’s ACH, it’s wire, it’s all this other stuff, because no one had to step in and figure a new system out to make all these things talk better together and faster.

319 00:36:48.570 ⇒ 00:36:59.130 Demilade Agboola: Just… just to add to that, like, in Nigeria, because I went to Nigeria recently, in Nigeria, you can actually send money to, like, POSs. So instead of, like, tap-to-pay, you just, like.

320 00:36:59.130 ⇒ 00:36:59.910 Samuel Roberts: Oh, yeah.

321 00:37:00.150 ⇒ 00:37:03.510 Demilade Agboola: And then it enters, like, as soon as you see the successful, like.

322 00:37:03.900 ⇒ 00:37:08.950 Demilade Agboola: share receipts, like, as soon as it pops up successful, you have the POS ding. Like, it starts.

323 00:37:08.950 ⇒ 00:37:09.980 Samuel Roberts: Mmm!

324 00:37:09.980 ⇒ 00:37:18.479 Demilade Agboola: That’s cool. Yeah, it’s across different banks and everything. It’s really cool, and, like, transferring across banks is, like, basically instantaneous.

325 00:37:18.480 ⇒ 00:37:20.270 Samuel Roberts: Yeah, that’s what we don’t have here, it’s crazy.

326 00:37:20.270 ⇒ 00:37:37.579 Demilade Agboola: We don’t really… we don’t really do… because people want their money, like, ASAP, but there’s also a trust factor as well, because there’s a worry that if the person goes out of sight, and you haven’t gotten hold of your money, that’s the money gone forever. So, like, things need to happen really quickly, and it’s pretty cool, to be honest.

327 00:37:37.910 ⇒ 00:37:55.229 Samuel Roberts: Yeah. Yeah, when I was in YC, there were a bunch of companies doing, like, phone payment stuff back, when was this, in 2019, I think? A lot of, like, Latin America and Africa founders were doing all kinds of stuff like that, and I haven’t kept up with that, so I don’t know where we’re at with that now, but that’s really cool to just hear the thing ding right then.

328 00:37:56.370 ⇒ 00:37:57.130 Samuel Roberts: Cool.

329 00:37:57.490 ⇒ 00:38:01.430 Samuel Roberts: Any other thoughts about the internet in general, or anything?

330 00:38:02.510 ⇒ 00:38:03.500 Samuel Roberts: Questions?

331 00:38:05.460 ⇒ 00:38:06.710 Uttam Kumaran: This is a great topic.

332 00:38:07.630 ⇒ 00:38:25.149 Samuel Roberts: Yeah, I was thrown into this, when I did the hair care company, and we started doing this, and I had to be the first one, because I was the, like, guinea pig, and I was like, no one here… we’re all, like, haircare people, no one knows, and I was like, this is the perfect topic. And then as I did it, I was like, I don’t even know all this stuff by detail, you know, that well. Like, it’s… there’s so many moving pieces.

333 00:38:25.900 ⇒ 00:38:26.550 Uttam Kumaran: Yeah.

334 00:38:27.120 ⇒ 00:38:29.750 Samuel Roberts: And it’s, it’s… it’s that, like,

335 00:38:30.190 ⇒ 00:38:33.580 Samuel Roberts: You know, no one knows how to build a mouse, or something like that, you know, that…

336 00:38:33.780 ⇒ 00:38:44.460 Samuel Roberts: forget exactly what… like, a thought experiment, but, like, how would you construct that? Like, you need the plastic, you need all the engineering that goes into the board. Like, there’s so many pieces and moving parts that, like, you can’t…

337 00:38:44.750 ⇒ 00:38:50.670 Samuel Roberts: do it yourself, and that’s what it feels like learning about the internet, and, like, even computer engineering all the way down, it’s like, we’ve…

338 00:38:51.080 ⇒ 00:38:52.829 Samuel Roberts: Built so much up that we’re.

339 00:38:52.830 ⇒ 00:38:58.590 Uttam Kumaran: Yeah, dude, I have a great article for you. This guy, he basically

340 00:38:59.290 ⇒ 00:39:06.350 Uttam Kumaran: He has a YouTube channel, but he basically built every machine that you would need to rebuild civilization.

341 00:39:06.350 ⇒ 00:39:08.379 Samuel Roberts: Yes, I’ve seen stuff like this, yes, exactly, it’s crazy.

342 00:39:08.380 ⇒ 00:39:24.089 Uttam Kumaran: And he was like, okay, well, we need… there’s, like, 20 machine types that you… that if you had them, you could basically build everything, like, mills and, like, tractors and stuff like that. And he’s completely off the grid, and he literally built, like, all this machinery himself.

343 00:39:24.270 ⇒ 00:39:26.360 Uttam Kumaran: And it’s very, very interesting.

344 00:39:26.800 ⇒ 00:39:28.939 Samuel Roberts: Very cool. Yeah, I’ll try to…

345 00:39:29.040 ⇒ 00:39:34.960 Uttam Kumaran: I’ll try to find it, let me see… do rebuild…

346 00:39:35.120 ⇒ 00:39:37.380 Uttam Kumaran: Let me see if I could just…

347 00:39:39.830 ⇒ 00:39:43.680 Uttam Kumaran: Yeah, I don’t know, I’ll have to… I’ll have to find it, but

348 00:39:43.930 ⇒ 00:39:49.350 Uttam Kumaran: Yeah, I guess I don’t… I also had a little bit of a lab share, if I could go through that. This is a…

349 00:39:49.820 ⇒ 00:40:01.229 Uttam Kumaran: different topic, but close to my heart. Not really, but, maybe I’ll give everyone just, like, 2 minutes to read through this article. I may have sent it before,

350 00:40:01.390 ⇒ 00:40:03.840 Uttam Kumaran: but I think about this…

351 00:40:04.880 ⇒ 00:40:07.910 Samuel Roberts: A lot, and oh, yeah, yeah.

352 00:40:07.910 ⇒ 00:40:12.390 Uttam Kumaran: We’re iterating on this, sort of on a week-to-week…

353 00:40:12.830 ⇒ 00:40:19.409 Uttam Kumaran: basis here at Brainforge, but I’d be curious if everyone wants to just give this article, a short…

354 00:40:19.790 ⇒ 00:40:21.929 Uttam Kumaran: Read, it’s really short.

355 00:40:22.030 ⇒ 00:40:25.319 Uttam Kumaran: Maybe I just give everyone, like, 2 minutes to read through this.

356 00:40:25.650 ⇒ 00:40:27.350 Uttam Kumaran: And then I would just, like.

357 00:40:27.560 ⇒ 00:40:30.839 Uttam Kumaran: maybe we can have a close chat, because I really…

358 00:40:31.270 ⇒ 00:40:35.180 Uttam Kumaran: I think this is a, this is a cool one.

359 00:41:55.530 ⇒ 00:41:57.189 Henry Zhao: Yeah, this is super interesting.

360 00:41:58.050 ⇒ 00:41:59.699 Uttam Kumaran: Yeah, does everyone have a chance to…

361 00:42:00.670 ⇒ 00:42:01.790 Uttam Kumaran: To give it a read.

362 00:42:05.960 ⇒ 00:42:08.279 Uttam Kumaran: So, kind of, like.

363 00:42:10.020 ⇒ 00:42:15.309 Uttam Kumaran: The gist of this article is, like, what is the future of, like, org structures gonna be?

364 00:42:15.380 ⇒ 00:42:30.190 Uttam Kumaran: And this is something that I think Robert and I think a lot about, especially because I think our advantage as a small company is to adopt these new paradigms when a technology as revolutionary as AI comes in. So.

365 00:42:30.490 ⇒ 00:42:31.979 Uttam Kumaran: the baseline…

366 00:42:32.240 ⇒ 00:42:39.089 Uttam Kumaran: for this conversation, a couple of things have to be true. You have to bet that AI is…

367 00:42:39.270 ⇒ 00:42:55.259 Uttam Kumaran: here, and that it is important and effective in our work, right? So, this is where some people in your life, or some people in my life sometimes come to me being like, oh, AI is all fluff, it doesn’t really work. I sort of kill the conversation there, like, I’m not willing to… I don’t, like, really

368 00:42:55.740 ⇒ 00:43:07.060 Uttam Kumaran: engage in those types of conversations anymore. I think if you’re at this company, you have seen how effective it can be, and so I hope that everybody here is on that same page, where if someone in your life is like.

369 00:43:07.060 ⇒ 00:43:16.000 Uttam Kumaran: well, I don’t really get this AI thing, it doesn’t really work. You can just say, like, okay, maybe you’re off the… you’re off the train. So, given that assumption, I think one thing that

370 00:43:16.090 ⇒ 00:43:25.270 Uttam Kumaran: you know, I think a lot about is, like, what is our org structure? And so, one way of thinking about this is in, like, ratios between

371 00:43:25.330 ⇒ 00:43:31.079 Uttam Kumaran: Leaders, middle management, and sort of, like, employees.

372 00:43:31.120 ⇒ 00:43:48.810 Uttam Kumaran: And I think one thing that, you know, we’re a small team, so in some sense, everybody here is playing both player, coach, is managing at least their time, and also executing. We also have people that are lean more management, and of course, you have Robert and I that are, like, doing quite a bit of management work.

373 00:43:48.820 ⇒ 00:43:51.330 Uttam Kumaran: And there’s also different ways of the way people have

374 00:43:51.570 ⇒ 00:43:57.170 Uttam Kumaran: rose through those ranks. Some people are very meritocratic, some people are more,

375 00:43:57.600 ⇒ 00:44:11.879 Uttam Kumaran: like, this is more seniority or tenure-based. But one thing that this article talks about is, like, okay, what is the first configuration, the typical configuration, which is short pyramid,

376 00:44:12.090 ⇒ 00:44:14.749 Uttam Kumaran: You have one person, and then…

377 00:44:14.880 ⇒ 00:44:29.429 Uttam Kumaran: You have… so this is the… this is the first thing. This is a normal world. One leader, several managers, bunch of individual contributors. In the next world, you have one leader, and you have seven, like, AI agents, and then…

378 00:44:29.590 ⇒ 00:44:38.289 Uttam Kumaran: those AI agents go do work. You also have, like, Rocket Ship, which is, like, more of one person…

379 00:44:38.450 ⇒ 00:44:51.330 Uttam Kumaran: sort of 7 agents, and then you have, like, people under them. And then, kind of, like, one thing that it’s talking about is, like, how can we move things from…

380 00:44:52.080 ⇒ 00:45:07.600 Uttam Kumaran: like, these traditional pyramids into things that are more short, where the gap between people are shorter, and you have, like, some sort of AI coming in and augmenting some of the work.

381 00:45:07.670 ⇒ 00:45:15.850 Uttam Kumaran: So I think that this is a really interesting way of thinking about this, where for us, like, the way we think about AI is I think about it, like.

382 00:45:15.940 ⇒ 00:45:19.970 Uttam Kumaran: taking over roles, right? But I don’t think about it in the world of, like.

383 00:45:20.420 ⇒ 00:45:30.149 Uttam Kumaran: for example, a company like Amazon, for example, they already have hired hundreds of thousands of people. They are actively finding ways to replace those people with AI.

384 00:45:30.150 ⇒ 00:45:43.109 Uttam Kumaran: In our company, we have way more work than we have the people for, and so there is not a really situation where we’re, like, replacing people with AI. In fact, most of the AI efforts we do are either speeding up.

385 00:45:43.110 ⇒ 00:45:50.919 Uttam Kumaran: team’s ability to execute, or taking on work that is not being done. For example, generating startup… generating stand-up notes.

386 00:45:50.920 ⇒ 00:46:05.440 Uttam Kumaran: that’s something that, like, we just don’t have the staff to do that for every single stand-up, every single day. That is something we can offload to AI. What does it help us as a business? Helps us run leaner, helps us run with less people. And so one of the, also the things that, you know, I got…

387 00:46:05.580 ⇒ 00:46:07.670 Uttam Kumaran: shared,

388 00:46:07.950 ⇒ 00:46:21.910 Uttam Kumaran: is this concept of, like, like, another node, let me just, I don’t know what the exact graphic is, but there’s, like, another node in network, in, business.

389 00:46:22.150 ⇒ 00:46:33.630 Uttam Kumaran: communication. So basically, what… I have to find the article, but a friend of mine was like, every time you add a node in, business communications, and this is gonna be more about

390 00:46:33.690 ⇒ 00:46:49.290 Uttam Kumaran: network communications, but basically, if you think about, in a business, you have several nodes, right? You have, like, me talking to Amber, talking to someone on the team, talking to a client. Every time you add another person, you add another node of communication, another game of telephone.

391 00:46:49.320 ⇒ 00:47:00.770 Uttam Kumaran: And so one of the things that I think a lot about is, like, the benefits of a small team is that I can go direct to Demolade on a team and ask for help with something, instead of going through

392 00:47:00.770 ⇒ 00:47:14.850 Uttam Kumaran: Manager 1, Manager 2, Product Manager 1, blah blah blah. Like, I don’t like that. And that shortens the pyramid, right, and keeps us closer. But we still… there are benefits to those middle management roles, and so what… one thing we’re trying to decide on is, like.

393 00:47:14.850 ⇒ 00:47:20.969 Uttam Kumaran: How do we create Sort of this, like, lean team with very little hierarchy, but…

394 00:47:21.240 ⇒ 00:47:37.150 Uttam Kumaran: the hierarchy typically comes because you have to have these separations of roles, and have to have people that are guiding a system. And so how can I turn… for me, I think the point of this article is you have to make… you kind of have to make a choice. You have to decide, whether…

395 00:47:37.210 ⇒ 00:47:46.020 Uttam Kumaran: we turn into something where there’s a lot of leaders, and then there’s… every one of them gets a bunch of AI agents.

396 00:47:46.020 ⇒ 00:47:57.489 Uttam Kumaran: or there’s leaders, there’s, like, managers, and those people manage AI agents, or it’s something in the middle. And so this is where I think we’re sort of thinking about, okay, how does our platform

397 00:47:57.600 ⇒ 00:48:16.470 Uttam Kumaran: sort of help support this idea, whereas a… for example, I’ve been project managing a lot this week. One of the things that I need to do a lot of is go back to meetings, copy stuff into ChatGPT to help me articulate, like, follow-ups, or tell me my to-do list. Those are things… before, as a project manager, I would have to write on a piece of paper every single to-do and go back.

398 00:48:16.480 ⇒ 00:48:19.879 Uttam Kumaran: Now, that task that maybe would take 2-3 hours.

399 00:48:19.940 ⇒ 00:48:34.560 Uttam Kumaran: like, I could do in, like, 10 minutes. And so that is a helpful thing where an AI is assisting me in an existing process. Well, you can take that to the limit, and you can say, well, AI actually… maybe an AI agent can sit in our meetings and proactively do those tasks for us.

400 00:48:34.560 ⇒ 00:48:40.669 Uttam Kumaran: And then I become more of a manager of that person, right? And so, I think this is just an interesting article.

401 00:48:41.250 ⇒ 00:48:56.609 Uttam Kumaran: talking about, you know, how we see… how we’re thinking about the world with AI and a consultancy, but, like, curious if anyone else on the call has, like, thoughts about this, or has started thinking about this, given, like, some of the stuff we’ve been doing internally.

402 00:48:57.070 ⇒ 00:48:57.920 Uttam Kumaran: Yeah.

403 00:49:03.040 ⇒ 00:49:14.639 Samuel Roberts: I like the point that not every department will be the same structure. Like, different departments can utilize things different ways, or need different things. So, like, as an org, it might be one whole thing, but then each

404 00:49:14.880 ⇒ 00:49:19.530 Samuel Roberts: Use of different tools and needs, can be different.

405 00:49:20.020 ⇒ 00:49:20.660 Uttam Kumaran: Yeah.

406 00:49:21.430 ⇒ 00:49:30.030 Samuel Roberts: Because it is… it can be very… yeah. Again, it’s not one-size-fits-all, like it says. I think that’s… even on the micro scale of, like, within teams and stuff, it’s good to know.

407 00:49:30.630 ⇒ 00:49:31.240 Uttam Kumaran: Yeah.

408 00:49:36.980 ⇒ 00:49:38.800 Uttam Kumaran: Anyone else have any thoughts?

409 00:49:40.450 ⇒ 00:49:42.719 Awaish Kumar: Yeah, like, one of the things…

410 00:49:43.160 ⇒ 00:49:51.529 Awaish Kumar: Is that, like, we can utilize… like, one of the areas which you said we are lacking is communications with clients.

411 00:49:51.960 ⇒ 00:49:58.999 Awaish Kumar: And that part, like, we can maybe try to offload some of it to AI agents to…

412 00:49:59.340 ⇒ 00:50:05.170 Awaish Kumar: Like, generate weekly summaries of I think…

413 00:50:05.740 ⇒ 00:50:12.669 Awaish Kumar: of the leader, and meetings, I mean, like… like, the document we are working on in Notion, maybe we have some…

414 00:50:12.970 ⇒ 00:50:20.359 Awaish Kumar: Something ready. Just review it and edit it and send it out.

415 00:50:23.130 ⇒ 00:50:23.700 Uttam Kumaran: Yeah.

416 00:50:31.340 ⇒ 00:50:38.219 Awaish Kumar: And one more thing, which I have been, like, talking about since I came in regarding like… The client…

417 00:50:38.570 ⇒ 00:50:41.750 Awaish Kumar: Normally, like, if we say we build a pipeline, like, that’s…

418 00:50:41.880 ⇒ 00:50:47.180 Awaish Kumar: not going to catch their attention. Instead, if you say, okay, your revenue is,

419 00:50:47.470 ⇒ 00:50:49.589 Awaish Kumar: 20% less than last week.

420 00:50:50.120 ⇒ 00:51:02.919 Awaish Kumar: So… and that’s why, like, there are the reasons, like, four reasons, and that… that’s why it has been down, and that might catch their attention. And we can have some AI agents to build those

421 00:51:03.030 ⇒ 00:51:05.099 Awaish Kumar: Reports from data for us.

422 00:51:05.710 ⇒ 00:51:06.490 Uttam Kumaran: Yeah.

423 00:51:06.490 ⇒ 00:51:07.030 Demilade Agboola: Oh.

424 00:51:08.000 ⇒ 00:51:29.899 Demilade Agboola: Just to add, something that came to my mind that I’d seen an article on is that there’s some people who have, like, AI agents to query data, so you just, like, will talk in, like, regular English, and then it will create the SQL query that queries that. And sometimes I think about, like, some of the hidden tasks, and, like, being able to have that set up for us, so that, like.

425 00:51:30.390 ⇒ 00:51:46.079 Demilade Agboola: I mean, obviously, it will generate the SQL so we can look at it and just confirm that the SQL is right and does make sense, but instead of us constantly having to, like, redo, like, simple one-off requests, that could be really helpful.

426 00:51:47.090 ⇒ 00:51:47.660 Uttam Kumaran: Yeah.

427 00:51:50.420 ⇒ 00:51:53.570 Uttam Kumaran: Yeah, this is where I think, you know, I’ve talked to Sam a lot about, like.

428 00:51:53.790 ⇒ 00:52:03.889 Uttam Kumaran: I think we’re sitting on this stuff right now, and I think, you know, I think a lot of… about AI as, like, sort of the, like, self-driving.

429 00:52:04.030 ⇒ 00:52:10.559 Uttam Kumaran: levels of automation, right? So if I Google, like, self-driving automation levels.

430 00:52:11.020 ⇒ 00:52:21.260 Uttam Kumaran: The NHSDA, the Natural Highway Transportation Safety Authority, they basically have described,

431 00:52:21.410 ⇒ 00:52:39.099 Uttam Kumaran: level 0, 1, 2, 3, 4, 5 for autonomous vehicles. And I think that we have a… I think about our business in that way, and every function has 0, 1, 2, 3, 4, 5. And I think on data, we’re probably just cracking, like, level 0, level 1, where all of us are using cursor now.

432 00:52:39.100 ⇒ 00:52:51.950 Uttam Kumaran: Right? So we’ve improved our process, but there is another layer to, hey, can every person, every client get a text-to-SQL interface for their client’s database? Right? And so now simple questions can get answered faster, and then

433 00:52:52.120 ⇒ 00:53:11.249 Uttam Kumaran: The level 3 could be, hey, why don’t we have an AI agent every morning go through and find an insight? Okay, that’s level 3. And I think this is where we want to go. I… I’ve always said that, like, I think engineering is going to be the hardest part to impact. You’re seeing that we’re already doing this on sales, marketing, and project management.

434 00:53:11.250 ⇒ 00:53:13.639 Uttam Kumaran: I think there, we’re almost, like, level 2.

435 00:53:13.640 ⇒ 00:53:25.270 Uttam Kumaran: Level 3 is where we need to start having, like, agents, autonomous agents, ambient agents that are working in things. But on the data side, totally, like, I think, Demolata, if you’re like, hey, for Eden.

436 00:53:25.770 ⇒ 00:53:35.520 Uttam Kumaran: by next week, I want to… I need something to query our BigQuery with natural language. I think that’s a great task for our team to work on, to help support you, and if you can make the case

437 00:53:35.520 ⇒ 00:53:47.250 Uttam Kumaran: And this is where I think, for folks that are on clients, if you can make the case that, hey, this will get us another 5-10 points in margin, it totally makes sense for our internal team to work on that solution for you.

438 00:53:47.430 ⇒ 00:53:56.769 Uttam Kumaran: Right? And that’s what I’m sort of looking forward to hearing from everybody, is like, in your day-to-day work from clients, if you find that, hey, I’m doing something with ChatGPT,

439 00:53:56.770 ⇒ 00:54:10.090 Uttam Kumaran: or I’m finding myself doing something with the cursor, and I need a little bit of a better solution, and I think if I had this, I could be this much more efficient. With the same quality or better of the deliverable, it is a no-brainer for us to work on that, you know?

440 00:54:16.700 ⇒ 00:54:23.280 Uttam Kumaran: Cool. I’ll send a couple more articles about, like, AI and services. I just dug a couple of them up.

441 00:54:23.460 ⇒ 00:54:30.530 Uttam Kumaran: Which, you know, people, can take a look at and read, and I’m just, like, curious about everybody’s thoughts. I did a lot of reading about this, sort of, like.

442 00:54:30.840 ⇒ 00:54:32.600 Uttam Kumaran: Last year, around, like.

443 00:54:33.070 ⇒ 00:54:47.250 Uttam Kumaran: March, April of last year, and a lot of it has stayed the same. I think the only thing that has changed is actually it’s not moving fast enough. Like, I think the technologies in AI are moving quite fast, like the language models and things like that, but the adoption is

444 00:54:47.300 ⇒ 00:54:53.149 Uttam Kumaran: abysmally slow, and I think we’re even seeing firsthand, like, how hard it is to

445 00:54:53.450 ⇒ 00:55:01.689 Uttam Kumaran: for even our crew of people to understand where AI can impact, but I think we’re also pretty far ahead on the curve for a company our size, so…

446 00:55:01.980 ⇒ 00:55:11.520 Uttam Kumaran: Yeah, maybe the last thing that I wanted to share while I have everybody on the call is just some platform improvements, and I’ll probably steal

447 00:55:11.620 ⇒ 00:55:22.570 Uttam Kumaran: AI team’s thunder here, but we made a lot of really nice improvements to stuff, and I just want to even just make sure I have up, our project reviews.

448 00:55:22.710 ⇒ 00:55:26.829 Uttam Kumaran: To share this, and I shipped… Some features this week.

449 00:55:26.990 ⇒ 00:55:29.170 Uttam Kumaran: And I am not a full-stack engineer.

450 00:55:29.490 ⇒ 00:55:38.290 Uttam Kumaran: And I’m not as smart as these guys, yet, you know, I still was able to do it, so I’m very, very happy. A couple of things,

451 00:55:39.180 ⇒ 00:55:49.409 Uttam Kumaran: Okay, I’ll just make sure I’m, like, at a good zoom level. So one, we ship departments, so you can see admin, delivery, engineering, marketing, and sales here.

452 00:55:49.770 ⇒ 00:55:51.530 Uttam Kumaran: So you should be able to see…

453 00:55:51.870 ⇒ 00:55:58.880 Uttam Kumaran: like, just similarly, we have meetings categorized by clients. You can now see them categorized by… for our internal meetings.

454 00:55:59.770 ⇒ 00:56:05.509 Uttam Kumaran: My ask is if you see something that’s miscategorized, just let our team know, and we can, learn from that and improve.

455 00:56:05.540 ⇒ 00:56:20.729 Uttam Kumaran: But let’s say you want to change that yourself, you can actually go in here and just change the department it’s assigned to, or change the team it’s assigned to. So those are kind of some nice features. Another thing you may have noticed is we made improvements to the date range.

456 00:56:20.750 ⇒ 00:56:28.660 Uttam Kumaran: This is something that was… has been really on my list to do, and so we improved, like, your date picker, so you can select two dates here, hit apply.

457 00:56:28.770 ⇒ 00:56:39.469 Uttam Kumaran: Things will refresh. You can also quickly go here, say last 30 days, last 7 days, today. That’s really nice. As usual, you can go through and select your participants as well.

458 00:56:39.680 ⇒ 00:56:41.450 Uttam Kumaran: So that’s great.

459 00:56:42.900 ⇒ 00:56:58.429 Uttam Kumaran: Also, you can now multi-select meetings, so let’s say I want to chat over all of these three Urban STEMS meetings, or let’s just say these two, I can add this to AI Context and say, tell me what happened in each of these.

460 00:56:58.500 ⇒ 00:57:08.100 Uttam Kumaran: And so, previously, a user like me would have to copy the transcript from each into ChatGPT, then ask this question. You can now ask this here.

461 00:57:08.670 ⇒ 00:57:09.759 Samuel Roberts: Are you still in Notion?

462 00:57:11.500 ⇒ 00:57:12.840 Samuel Roberts: You’re showing Notion right now, I think.

463 00:57:12.840 ⇒ 00:57:14.609 Uttam Kumaran: Oh, really? Oh, shoot, sorry.

464 00:57:14.870 ⇒ 00:57:16.320 Samuel Roberts: I was like, what?

465 00:57:16.320 ⇒ 00:57:21.190 Uttam Kumaran: Oh, oh, sorry, sorry, sorry. Can you see this now?

466 00:57:21.440 ⇒ 00:57:22.290 Samuel Roberts: Yes.

467 00:57:22.290 ⇒ 00:57:30.480 Uttam Kumaran: Okay, cool, sorry. So I was just walking through, like, this demo. So you can click on a date range here, select these things, apply.

468 00:57:30.620 ⇒ 00:57:33.149 Uttam Kumaran: You can go into a meeting, see, like.

469 00:57:33.600 ⇒ 00:57:40.000 Uttam Kumaran: different Meleep meetings here, you can go in and… Change departments…

470 00:57:40.380 ⇒ 00:57:47.820 Uttam Kumaran: I was showing multi-select, so you can click on multiple meetings, add the transcripts, the AI context, ask a question about it.

471 00:57:48.230 ⇒ 00:57:58.719 Uttam Kumaran: Another thing for, like, sort of anybody that’s creating tickets is I shipped a little bit of a better UI for this tickets workspace, where you can type in

472 00:57:58.920 ⇒ 00:58:02.750 Uttam Kumaran: I don’t know, generate a few sample tickets, I don’t know if this is gonna work.

473 00:58:02.910 ⇒ 00:58:04.219 Uttam Kumaran: But…

474 00:58:04.380 ⇒ 00:58:22.340 Uttam Kumaran: what this is gonna do is, now instead of, like, seeing it on the blow, you can see it on the right here. There’s some quick, like, QA checklists, and I think what it should generate, maybe, is several here. This is probably something that we can move to Mastra and make it a little bit quicker.

475 00:58:22.340 ⇒ 00:58:26.330 Samuel Roberts: Yeah, I don’t know what’s gonna do with very little contact there, too, so there we go, okay.

476 00:58:26.330 ⇒ 00:58:37.219 Uttam Kumaran: Yeah, so you can see here that these are tickets I want to create, and so, I can edit the contents of the tickets, change the team, quickly approve and push them.

477 00:58:37.320 ⇒ 00:58:39.400 Uttam Kumaran: And so just something that…

478 00:58:39.740 ⇒ 00:58:59.530 Uttam Kumaran: one thing that is a huge workflow for PMs is, like, this creating ticket workflow. Another thing that you can do now is, when you go into a meeting, and you go ahead and click copy transcript, the transcripts are actually very small now. So, as you can see here… oh, actually, maybe it’s,

479 00:59:00.350 ⇒ 00:59:09.489 Uttam Kumaran: I don’t know where the best place is to paste this, but the transcript now, instead of being, like, this long thing, is now very short, so…

480 00:59:09.670 ⇒ 00:59:26.019 Uttam Kumaran: it’s what we ran… we have this little compression algorithm that, like, compresses the text by… removes a lot of the stuff that’s not needed, compresses it, so you can shove more into your context for AI. This is something that I do often, which is hit copy transcript, put it into AI for stuff.

481 00:59:26.180 ⇒ 00:59:38.219 Uttam Kumaran: Another thing is if you go to Marketing Assets, and Robert, you will find this helpful, is you can quickly filter to, like, the stuff we need. Hey, I’m looking for, like, an e-com one-pager, we don’t have any.

482 00:59:38.220 ⇒ 00:59:47.900 Uttam Kumaran: Or we may have some for other industries. But either way, it’s very… it’s very fast to filter now. You can be like, oh, where’s that video? Where’s that one diagram?

483 00:59:48.100 ⇒ 00:59:50.209 Uttam Kumaran: Where’s everything we did for one client?

484 00:59:50.400 ⇒ 00:59:55.009 Uttam Kumaran: everything is tag-based, and I took… we took the filters out of the modal, so…

485 00:59:55.710 ⇒ 01:00:01.659 Uttam Kumaran: We’re gonna… this is gonna start to double and triple in size, and so we want to make it easy for users to find.

486 01:00:02.230 ⇒ 01:00:03.140 Uttam Kumaran: the appropriate.

487 01:00:03.140 ⇒ 01:00:04.130 Robert Tseng: Amazing.

488 01:00:05.160 ⇒ 01:00:09.240 Uttam Kumaran: You can also create assets directly in here, so Hannah is using this.

489 01:00:09.240 ⇒ 01:00:10.300 Robert Tseng: So she can go ahead and…

490 01:00:10.620 ⇒ 01:00:13.009 Uttam Kumaran: You can go here, add, add assets, yeah.

491 01:00:13.010 ⇒ 01:00:29.970 Samuel Roberts: You can’t yet upload directly to it, but if it’s a link to, like, a video or a Notion doc or something, it can go here. The plan is to eventually just, like, drop something in here and see this get created with it uploading to an S3 bucket and not GitHub and stuff, so… yeah. Excited, but don’t, don’t, don’t… yeah, it’s not quite there, but it will be. It will be.

492 01:00:29.970 ⇒ 01:00:34.029 Uttam Kumaran: It’s there. It’s there. Go ahead and use it.

493 01:00:34.030 ⇒ 01:00:36.220 Samuel Roberts: But it’s so… you can’t upload to it, is my point.

494 01:00:36.220 ⇒ 01:00:52.140 Uttam Kumaran: I know. Soon. But this… now, Hannah could go in here, add assets, so these may be YouTube videos, external sites, whatever, so we’re starting to centralize all our assets. This is a replacement for a typical CMS, content management system.

495 01:00:52.220 ⇒ 01:00:56.480 Uttam Kumaran: Which we don’t need one of those clunky pieces of software.

496 01:00:56.950 ⇒ 01:01:01.050 Uttam Kumaran: And then also, you can bulk edit, so you can go in here and edit a bunch of them.

497 01:01:01.280 ⇒ 01:01:03.859 Uttam Kumaran: Change tags and things like that.

498 01:01:03.990 ⇒ 01:01:11.159 Uttam Kumaran: Again, this is a lot of the stuff that design handles. The last thing is the nav is a little bit better, so…

499 01:01:11.330 ⇒ 01:01:16.269 Uttam Kumaran: it’s not, like, clunky and over the thing. You can actually go in here and, like, yeah, so…

500 01:01:16.430 ⇒ 01:01:26.109 Uttam Kumaran: Couple things next. We have… we’re gonna implement role-based asset control, we’re gonna hook it into Google Auth, so that when a user gets created in Google.

501 01:01:26.150 ⇒ 01:01:39.569 Uttam Kumaran: You’ll be able to… a user immediately gets created here, and, like, it’s all sort of, like, one flow through Google. We’re gonna add some more helpful features for our sales and project mark… project management team directly in here.

502 01:01:40.230 ⇒ 01:01:43.109 Uttam Kumaran: And then they have the ability for operations to create clients directly.

503 01:01:43.240 ⇒ 01:01:44.940 Uttam Kumaran: Stuff like that, and there’s some behind-the-scenes…

504 01:01:45.540 ⇒ 01:01:47.490 Samuel Roberts: Yes, we’re a little bit over, but…

505 01:01:47.790 ⇒ 01:01:52.810 Uttam Kumaran: Lot of great updates this week on platform… Forge platform, so…

506 01:01:55.770 ⇒ 01:01:58.969 Uttam Kumaran: Yes. And now you’re seeing… this is what… how I used to…

507 01:01:59.230 ⇒ 01:02:03.470 Uttam Kumaran: lead products, so they’re seeing the product manager in me come out, so…

508 01:02:04.010 ⇒ 01:02:18.090 Uttam Kumaran: I’m pumped, this is getting… it’s getting better, and so the last thing I’ll mention is we’re gonna make… we’re making it easier for anybody in the company to ship features. One of my big directives for the AI team was the AI team is not gonna get to the size where they can ship everything.

509 01:02:18.090 ⇒ 01:02:31.189 Uttam Kumaran: it has… you have to… we have to make it easier for folks like me, folks like Hannah, anybody on this call, to actually make impacts onto the Forge. And so, one of the things is, if you have a feature that you want to add here, and you have the idea for it in your head.

510 01:02:31.370 ⇒ 01:02:45.220 Uttam Kumaran: I could show you how to go from that idea to design, to in the platform in less than an hour, I promise you. And so, if you have an idea that you’ve been wanting to add here, please ping in Slack.

511 01:02:45.220 ⇒ 01:02:58.879 Uttam Kumaran: And I will show you how to get this… get your design into, like, a view that you like it, chat with it, get it to the perfect spot, get it approved, and then how to actually implement this with Cursor, get a PR created, and then the AI team will approve it and merge it.

512 01:02:58.950 ⇒ 01:03:03.070 Uttam Kumaran: So, if you are… if you have a feature in mind, and you want to do this.

513 01:03:03.250 ⇒ 01:03:05.579 Uttam Kumaran: I’m available all rest of the day.

514 01:03:05.900 ⇒ 01:03:07.400 Uttam Kumaran: And I work here.

515 01:03:07.770 ⇒ 01:03:13.059 Uttam Kumaran: 12, 15 hours a day, so I’m here all the time to help your ship off onto the forge.

516 01:03:14.640 ⇒ 01:03:29.169 Uttam Kumaran: I challenge you to think about a feature that we can’t ship end-to-end in less than an hour, because I did 6 of these features in 2 hours at 2 AM the other day, and so it’s not… I’m telling you, you could do it too. And yes.

517 01:03:29.440 ⇒ 01:03:35.790 Uttam Kumaran: If you want to put this on your resume that you’re a full-stack engineer and you ship features in production, this is a great way to…

518 01:03:35.940 ⇒ 01:03:39.939 Uttam Kumaran: Just keep that. Just keep that.

519 01:03:42.260 ⇒ 01:03:43.000 Uttam Kumaran: Cool.

520 01:03:43.560 ⇒ 01:03:46.550 Uttam Kumaran: Okay, that’s it. Anything else?

521 01:03:49.240 ⇒ 01:03:54.850 Uttam Kumaran: And thanks, everyone. I know we’ve changed a lot of processes this week, and there’s been, like, people in and out of the company.

522 01:03:54.910 ⇒ 01:04:10.580 Uttam Kumaran: I… I think we’re… we’re in a… of course, we iterate every week, and I think this new stand-up process is… is really great, but I want to thank Rico in particular for a lot of help this past week on helping me manage, but also, I know we change a lot of meetings around, so… I love seeing everybody every morning, and

523 01:04:10.850 ⇒ 01:04:21.619 Uttam Kumaran: I’m, like, just dashing through tickets, and I think we are really aligned. I think in another week or two, we’re gonna be in a really good spot getting updates to clients every day.

524 01:04:22.480 ⇒ 01:04:29.100 Uttam Kumaran: So yeah, I feel good. Been a very busy week, though. I have not slept much, so…

525 01:04:32.210 ⇒ 01:04:32.940 Uttam Kumaran: Cool.

526 01:04:33.430 ⇒ 01:04:34.080 Samuel Roberts: Alright.

527 01:04:34.560 ⇒ 01:04:35.520 Robert Tseng: Alright, thanks.

528 01:04:35.520 ⇒ 01:04:36.390 Uttam Kumaran: Thanks, everyone.

529 01:04:36.390 ⇒ 01:04:36.789 Henry Zhao: I get it.

530 01:04:36.950 ⇒ 01:04:38.159 Samuel Roberts: Talk to you soon.

531 01:04:39.350 ⇒ 01:04:40.630 Demilade Agboola: Thank you, bye.