WeAreDevelopers LIVE - Humans in the Loop: Storytelling in Tech - Pritesh Kiri-এর সম্পূর্ণ প্রতিলিপি

the metro that on the stilts on top of the main on NG Road, right? 00:02:05.359 --> 00:02:09.919 Yeah, and it's gonna be trains that run, it's not like Berlin where it's uh they're always not running. 00:02:10.159 --> 00:02:12.960 So that's the other interesting thing that's have that's happening there. 00:02:13.120 --> 00:02:18.639 Well we've already we've already had the the sort of behind the scenes, haven't we, before the show, you know, because uh Pratech, you were telling us about the venue, right? 00:02:18.719 --> 00:02:20.159 That where we're gonna be hosting it. 00:02:20.240 --> 00:02:25.919 And obviously for us, you know, we we sort of have team members over there that kind of do do all of that and kind of organizing it. 00:02:26.080 --> 00:02:29.520 But for us, it's kind of uh, you know, in DevRel, it we don't know until we get there. 00:02:29.599 --> 00:02:36.800 So it's nice to hear that it's like a great venue and you've been there and you like it, and it's you know, kind of the best place for this kind of event, really. 00:02:37.680 --> 00:02:38.000 Yeah, yeah. 00:02:39.199 --> 00:02:39.840 It is, it is. 00:02:40.240 --> 00:02:42.479 And Bangalore is an interesting place, I have to say. 00:02:42.560 --> 00:02:50.560 I mean, I lived in Mumbai, I was in Delhi a few times, I was in uh uh Goa and these places, and that's why Bangalore was just very approachable. 00:02:50.639 --> 00:02:54.400 So it was a very uh uh place to just hang out and do things. 00:02:54.639 --> 00:02:58.400 It will be interesting how it changed in the last decade, I think. 00:02:58.639 --> 00:03:02.719 But we but it's gonna be a good thing to have, and the lineup is great. 00:03:02.800 --> 00:03:04.319 So you submitted a talk. 00:03:04.400 --> 00:03:05.280 What are you gonna cover? 00:03:05.599 --> 00:03:14.879 It is related to how you uh you write better tool descriptions for your MCP servers to make it more accessible. 00:03:15.280 --> 00:03:23.759 Uh for for people to use your NCP server or for other agents to understand when you say accessible, is that uh available or human accessible? 00:03:24.240 --> 00:03:28.879 It's for agents primarily because we face issues multiple times. 00:03:29.120 --> 00:03:44.000 Uh agents calling the wrong tool and getting the wrong set of information that they pull in and then they eventually give the wrong uh response to the required agent or human, whatever it is, uh in the agent system. 00:03:44.319 --> 00:03:48.319 Uh so to make sure that that doesn't happen, what do we get? 00:03:48.479 --> 00:03:52.560 So that's the major uh cool. 00:03:52.800 --> 00:03:56.240 The honking sounds in the background is that you or is that from Dan? 00:03:56.479 --> 00:03:59.120 Because that sounds like Rickshaws to me. 00:03:59.520 --> 00:04:00.400 It should be me. 00:04:00.639 --> 00:04:03.439 Let me just close my window, just give me a second. 00:04:03.759 --> 00:04:06.560 No, it's good, it actually puts a little puts a bit of effort in. 00:04:06.719 --> 00:04:07.759 I love it, it's really good. 00:04:08.080 --> 00:04:10.080 I'll say it makes exactly, you know. 00:04:10.240 --> 00:04:11.039 Yeah, yeah. 00:04:11.280 --> 00:04:17.920 I mean, the great thing was the MG Road, uh Mahabagandhi Road is like the biggest road in Mumbai, uh, in Bangalore. 00:04:18.000 --> 00:04:25.199 And it's like five lanes, but as there is a lane on demand system in India, like you just pull between two other cars and you make another lane. 00:04:25.360 --> 00:04:31.519 Uh, we we we actually filmed each other back then who's gonna take how how long it will take to cross that road. 00:04:31.759 --> 00:04:41.120 And the idea is basically if you cross a road in India with that many cars, you just walk in one direction with a speed and you don't change the speed, otherwise you're being run over. 00:04:41.439 --> 00:04:53.600 And we basically we had a uh as a British company, of course, we had uh we had a bet going on like who's gonna be the fastest crossing MG Road and who's gonna be the slowest, and it was just hilarious, it was just really, really good. 00:04:54.079 --> 00:04:55.759 Race conditions, yeah. 00:04:56.000 --> 00:04:57.839 Yeah, race conditions, definitely. 00:04:58.000 --> 00:04:58.399 Cool. 00:04:58.560 --> 00:05:10.560 Let's take a look at some of the resources we cover before we can go into detail about what you also you do, that humans of tech stuff that we wanted to chat about in the original thing, especially on master.f, it's basically always front-end stuff. 00:05:10.639 --> 00:05:26.160 But this is basically a really cool article that explains to you uh what generated H AI generated HTML, how it's you sandbox it, but also how you can render it within an iframe to see the issues with it and see if it's something that might be going out. 00:05:26.319 --> 00:05:41.360 So I love the idea of him uh just explaining what he's done and then he found a malicious problem with one of the codes that he's done, and then he just shows you what to do to actually contain HTML generate uh i generated HTML. 00:05:41.519 --> 00:05:43.519 Because I mean, this is nothing new. 00:05:43.600 --> 00:05:51.439 We had XSS before, we had like uh we had like injections before, and now we just generate whole bunches of HTML and hope for the best. 00:05:51.600 --> 00:06:00.879 But this is a very interesting kind of uh approach to make sure that this LLM generated code cannot break out of things and you actually can test it before you apply it. 00:06:03.199 --> 00:06:05.920 Have you encountered anything like that, Friedrich, before? 00:06:06.000 --> 00:06:09.839 Like in in your in I mean, you see working with with servers and stuff like that. 00:06:09.920 --> 00:06:13.920 What is your take on the current, like, oh my god, everything gets hacked by AI? 00:06:16.000 --> 00:06:26.319 Um I believe most of these agent tech tools that I have used so far, the coding agents, do take care of most of these stuff.

00:06:26.480 --> 00:06:57.920 Of course, I personally haven't reviewed it each and every time, but uh uh I did, you know, uh scroll through this article and I do think that it does make sense for the uh tool build like the website building tools which uh helps you build websites, but they don't have checks on uh what is getting injected and if it is relevant enough uh or if it is you know uh secure enough to get published uh uh for your for your business. 00:06:58.399 --> 00:07:08.000 Um uh having a sandbox which does take care of these things is of course uh good security practice, I would say. 00:07:10.639 --> 00:07:14.000 It's interesting that that uh uh I love that master.dev. 00:07:14.240 --> 00:07:24.560 I'm I'm not too happy about the name, but I I think it's uh it's basically where Chris Koya left uh when when uh CSS tricks went all right, and now CSS tricks is back as well, so that's pretty good. 00:07:24.639 --> 00:07:27.600 So we actually have those two blocks to play with again. 00:07:28.959 --> 00:07:37.680 This was one of those interesting bits where somebody uh uh so 3% of the critical software security bugs in all of recorded history have been found in the last month. 00:07:37.839 --> 00:07:45.040 This is what this is on pace to find another hundred thousand bucks this year alone, adding 26% more vulnerabilities to the all-time total. 00:07:45.439 --> 00:08:00.879 Um it feels to me, we talked about biases on the last show, and it feels to me that this is like a prediction that is basically too uh uh uh too scared of what's happening right now. 00:08:01.040 --> 00:08:06.000 Just because we didn't have the CVEs reported back then uh didn't mean it didn't happen. 00:08:06.160 --> 00:08:17.439 I think the the main difference with the with that that's what's happening in security right now is that like models actually uh are asked to hack things and then release those things immediately. 00:08:17.759 --> 00:08:24.800 And uh it's kind of shown as a as a sense of power over LLM, how much you can actually hack things. 00:08:25.040 --> 00:08:32.240 And yeah, if something like PHP has been touched for 20 years and hasn't been maintained in 15 years, I'm not surprised that there was a C V E and there. 00:08:34.639 --> 00:08:50.320 Yeah, I mean, I mean the thing with this is uh uh Chris, are you sort of skeptical of the the prediction, the 26% one, um about being on track for kind of finding the the 26 of 26% of the um sorry, adding 26% more vulnerabilities to the all-time total. 00:08:50.559 --> 00:08:56.480 Um I find this really interesting because it's sort of um, you know, we talked last week about there's this almost like cat and mouse, right? 00:08:56.559 --> 00:09:06.240 That there's we're we're better at kind of recognising and finding the vulnerabilities, but um whether or not we're actually now better and quicker at kind of patching them and fixing them and resolving them, I don't know. 00:09:06.320 --> 00:09:08.559 You know, it's it's sort of something you'd have to sort of tell me. 00:09:08.639 --> 00:09:12.639 But um because you because you know you see a story like this and you think, well, that's good and bad, right? 00:09:12.720 --> 00:09:17.759 That's great and not great, but uh but like you're saying, it's maybe there's maybe a bit more behind the headline, right? 00:09:17.840 --> 00:09:21.360 That if if these vulnerabilities were always here, we just now know about them. 00:09:21.440 --> 00:09:28.960 Um has anything kind of ultimately changed, or or is it just that we're we've got kind of the reporting of the information and we've got the information to hand? 00:09:29.120 --> 00:09:31.440 Um what's your take on it, you know? 00:09:33.120 --> 00:09:40.000 I mean, I I'd I'd rather talk to more people in security what their current take on it is, because I mean we have we've had this before. 00:09:40.080 --> 00:09:50.720 We've basically like, and I mean, the same when people talk about quantum computing right now, and like, oh, everything is insecure because quantum computers are around the corner and they've been around the corner for six years now. 00:09:51.200 --> 00:10:00.080 So uh um the the fixing things for the security of the future is has become much harder because things are much moving much faster. 00:10:00.240 --> 00:10:09.840 But I think there's also a sense of the the the the the false uh uh the false idea of like people things being published and people seeing there's more happening. 00:10:10.000 --> 00:10:14.080 We have the same in uh uh in like uh life here in Berlin, for example. 00:10:14.240 --> 00:10:19.600 People are saying, like, oh, life is so dangerous now with all these crimes happening, with all these crimes being reported. 00:10:19.840 --> 00:10:23.679 The thing is now we have social media where the crime gets filmed and immediately shown. 00:10:23.759 --> 00:10:28.080 If you look at the statistics, crime has been going down for years and years and years. 00:10:28.240 --> 00:10:32.960 It's actually much safer than it was to be than it was in the 90s or in the noughties nowadays. 00:10:33.120 --> 00:10:37.919 But like uh but just because we report these things, they become so much more obvious. 00:10:38.080 --> 00:10:44.399 But of course, it means hardening our software and thinking about that has become a much, much faster race. 00:10:46.960 --> 00:10:57.519 Pritash, do you think when you're kind of um I was thinking then about the fact that you know we're obviously here talking about humans of tech, and yet, you know, the kind of last five minutes we've spoken mostly about kind of agents and and kind of the work they're doing. 00:10:57.679 --> 00:11:09.279 But do you on your side, being on the community side, you know, do you uh get increasingly kind of are the conversations that you're having with developers about agents that have gone awry and uh you know issues that they're having with the agent? 00:11:09.360 --> 00:11:29.360 Is it kind

of the middle that the I say the middle person, you know, as the agent, but um is it still that there's kind of um issues with documentation, or is it that there's that there's agents that are kind of on their side, they've got the problems because they're going awry, and maybe it's not something that you can do much about on your end, you know, um, in terms of improving the documentation so that the agent has got more to go on. 00:11:29.600 --> 00:11:30.960 Um do you see what I mean? 00:11:34.639 --> 00:11:37.360 Um I I don't think so. 00:11:37.440 --> 00:11:53.200 There are a lot of changes that has been because being on the community front, we do have like a variety of discussions that we do as a whether it's a meetup, whether it's uh being part of any conference, but the discussion that we are having right now is something uh we have done in last meetups. 00:11:53.279 --> 00:11:53.600 That's right. 00:11:53.840 --> 00:11:55.120 I'm able to recall that. 00:11:55.279 --> 00:12:07.840 Uh we at Hardest, we are a CICT Tabbox platform, and uh it's like one of the one of the security platforms that helps you make your software deliver your software securely. 00:12:08.080 --> 00:12:31.600 And um the this article mentions that okay, we have there are increasing number of uh uh bugs or vulnerabilities or whatever is is the you know rate in which we are able to identify these issues as increased like anything, but at the same time, uh we are able to remediate them with the help of AI as well. 00:12:31.840 --> 00:13:08.399 Um, it's not manual fixing, it's again if you make it as a pipeline as a part of your CICD pipelines or your uh devs like uh which hasn't been putten out yet, but the most of the companies who do care about the vulnerabilities and to to you know follow the security practices, uh we the discussion majorly revolves around how fast we are able to uh you know fix or remediate the critical most of the critical vulnerabilities before even the you know the software gets shipped to the end users. 00:13:09.840 --> 00:13:10.480 So yeah. 00:13:11.039 --> 00:13:20.080 I think that's the big thing, but with CICD and also rolling out and deployment being so fast these days, like with AI generated things, we deploy the whole platform. 00:13:20.320 --> 00:13:23.759 We don't just basically patch things and upload smaller files. 00:13:23.840 --> 00:13:28.720 So a lot of like old nasties actually cannot get around, uh cannot stay around for long. 00:13:28.879 --> 00:13:46.879 I mean, case in point was actually here that Entropic uh just got was hacked for a few uh hours with a Trojan that actually uh that that actually uh got in through VS to cursor, the VS Code extension, and then basically install uh infected people's computers. 00:13:47.039 --> 00:13:52.559 And uh then 12 hours later they were able to actually roll it back and do set do things differently. 00:13:52.720 --> 00:13:59.919 But around 5,000 developers had actually uh a malware injection through a VS Code extension. 00:14:00.000 --> 00:14:02.879 And I think that's kind of the worrying bit as well to me. 00:14:03.039 --> 00:14:12.559 Like we we we rely on the platform just to do everything for us, but uh attacking the platform will be the interesting thing, interesting for thing for people to do in the near future. 00:14:12.799 --> 00:14:16.879 Like if you cannot get and plugins were always a thing for that. 00:14:18.639 --> 00:14:23.440 Yeah, I mean, I I think there's something interesting as well, Chris, about the fact that you know you said then 5,000 developers. 00:14:23.600 --> 00:14:24.799 Is it developers though, right? 00:14:24.879 --> 00:14:26.320 Is it just people that are building? 00:14:26.399 --> 00:14:31.519 You know, we say building now as the kind of term, and I think that that's that's that's the really worrying bit, right? 00:14:31.600 --> 00:14:44.559 Is that there's there's there's one thing to be said for reading that article and understanding it, and then there's another thing to be said for kind of never seeing the article because you've uh purely kind of vibe coded something with, you know, uh uh with maybe no technical understanding. 00:14:44.639 --> 00:14:46.799 And I think they're the people I would be most concerned about. 00:14:46.879 --> 00:14:54.399 And whilst I know we might say people that use cursor are perhaps more kind of dev you know based anyway, I I think I don't think that's always going to be true. 00:14:54.480 --> 00:15:03.440 Um, and and they're the people that I would worry about the most because then they're not gonna go and uh run the updates even or you know, or kind of remove the plugin or even kind of be on top of this. 00:15:03.600 --> 00:15:05.279 So yeah, so much going on. 00:15:05.519 --> 00:15:09.519 Well, and all the all the other platforms are now doing coding things as well. 00:15:09.679 --> 00:15:22.879 I mean, chat GPT became a coding environment and Slack now released a coding environment that basically anybody in the Slack can talk to a developer that is an agent and actually build software for your company in Slack. 00:15:23.200 --> 00:15:44.879 And uh I mean that sounds like a great thing in terms of democratizing and letting people do their own thing, but it also worries me that these environments that uh that we're getting into a world where everything wants to be super app and do everything rather than just and keep you in it, rather than just allowing people to have one app for developing, one app for coding, right for for communication. 00:15:45.120 --> 00:15:52.320 I think that like uh the the kitchen sync of making everything into one tool uh tool platform never worked for me. 00:15:52.480 --> 00:15:54.799 It it felt it felt it feels weird. 00:15:54.960 --> 00:16:05.360 Like I'm uh dangerous to to for people to generate random agents and let them run from Slack, and you basically you always wanted to make that not that not vulnerable. 00:16:05.600 -->

00:16:12.879 Then again, Discord just announced that they're now uh actually making open servers indexable by Google and search engines. 00:16:13.120 --> 00:16:20.080 So that's another thing that is that is fascinating to me that like okay, everybody wants to be in that play in that playground at the moment. 00:16:23.200 --> 00:16:38.960 And the last thing that was uh some that kind of gelled with me lately because I'm actually redesigning my uh uh my blog and I'm looking at like blog posts from 2005 to now, and there's so many uh old uh iframes and injection points that I had. 00:16:39.039 --> 00:16:45.679 Like I had all these YouTube embeds that were just like object IDs and embeds because there was with there were still flash movies. 00:16:45.919 --> 00:16:57.039 I had like the whole uh uh like lots of services that allowed you to host your your uh your uh your slides, and all of those had these little players that with JavaScript you had installed. 00:16:57.279 --> 00:17:02.559 And uh yeah, a lot of them are defunct now, and I basically ran a ran a link scanner over it. 00:17:02.799 --> 00:17:13.039 But uh this was interesting that basically there is uh there's just a tag uh controller, much like the these uh share for social media bits that we used to have. 00:17:13.119 --> 00:17:19.759 Remember those sharing buttons had like 6,000 services that all are dead now, like stumble upon and these kind of things? 00:17:19.920 --> 00:17:21.200 And this was something like that. 00:17:21.359 --> 00:17:31.519 So basically, that that service is now bought by another company and basically managed to kill your iPhone uh just with an old JavaScript include in your pages. 00:17:32.559 --> 00:17:46.960 And I think that's kind of worrying, but uh it's it's up to us as maintainers of products to actually get rid of dead things, but it's really hard to actually report uh to find them because you have to be on the fly. 00:17:47.039 --> 00:17:52.000 And it's uh it's kind of sad when you have like startups that that offer you cool things. 00:17:52.160 --> 00:17:56.559 I've got about 80 blog posts where I had like different people that did video transcription. 00:17:56.720 --> 00:17:57.920 None of them exist anymore. 00:17:58.160 --> 00:18:03.599 Most of them are 404, some of them are three or two to another service, one of them went to a gambling site. 00:18:03.839 --> 00:18:07.839 You really, really have to be on your toes with content that you have on the web. 00:18:09.440 --> 00:18:10.319 It's tough though as well. 00:18:10.400 --> 00:18:13.359 Like Pratash, I was just thinking then, you know, you you produce a lot of content. 00:18:13.440 --> 00:18:15.680 Um, and what where was it you kind of started though? 00:18:15.839 --> 00:18:16.160 Did you start? 00:18:16.319 --> 00:18:22.640 Because I mean, Chris, you know, you talked about kind of you know radio journalism and then and blogging and writing and it being, you know, what which medium was it you started in? 00:18:22.720 --> 00:18:29.759 Because now you do so much video that it kind of got me curious about whether that was kind of your your your first kind of medium or not. 00:18:31.039 --> 00:18:36.319 Um, so just to give you my background, I am a mechanical engineer by degree. 00:18:36.559 --> 00:18:36.880 Okay. 00:18:37.440 --> 00:18:44.400 And then I did sales and marketing for one year right after my degree, and then I got into tech. 00:18:44.640 --> 00:18:55.519 And the way I got into tech was very unnatural, and uh, it was not like I had to work for like one year just to prove myself that okay, I'm good enough you enough have a different degree. 00:18:55.680 --> 00:19:03.119 So when I started my journey as as a you know uh to get into tech, I started documenting everything on Twitter. 00:19:03.440 --> 00:19:10.880 Okay, so back then we used to uh we didn't have chat GB to create to create harcode posters and write content for us. 00:19:11.039 --> 00:19:14.400 So I had to write everything manually, have to frame nice. 00:19:14.880 --> 00:19:17.519 So back then we had we used to have the like threads, right? 00:19:17.599 --> 00:19:32.400 Like we have threads right now as well, but previously we had to think a lot because we have like 280 characters, just 280 characters in which we have to write a hook, write a value, and we have to make it so good so that people would read it, right? 00:19:32.640 --> 00:19:35.440 So we have to think a lot before what we write and how we write. 00:19:35.759 --> 00:19:39.200 Uh and that made me a good writer, is what I feel right now. 00:19:39.359 --> 00:19:45.920 Even everyone is generating content with AI, but that you know, skill of creating good content, the written content. 00:19:46.000 --> 00:19:52.640 Uh I got it from from my days when I when I started learning about tech and documenting my journey on Twitter. 00:19:52.880 --> 00:19:53.519 So yeah. 00:19:53.839 --> 00:19:56.480 That was my first medium of content creation. 00:19:57.039 --> 00:20:12.720 Yeah, because I couldn't remember where I'd kind of first seen you because it's like, you know, we we've kind of had a few people in DevRel on in the last few weeks, and it's um it's always nice that everyone's got their kind of slight angle, you know, the thing that they do a little bit more of, but but yeah, you've always kind of you know, uh or at least in the last year or so, you've been sort of more so in on video. 00:20:12.799 --> 00:20:13.440 So that's cool. 00:20:13.519 --> 00:20:15.839 That's interesting to hear where you started, you know. 00:20:16.079 --> 00:20:24.000 Yeah, well, I mean that's the worry as well that you basically you use Twitter as as the thing to get out there, and yeah, then it became X. 00:20:24.400 --> 00:20:26.240 And then how do you archive it? 00:20:26.319 --> 00:20:27.680 How do you get the content back? 00:20:27.839 --> 00:20:30.559 There were so many things that we did in closed environments. 00:20:30.720 -->

00:20:38.160 I mean, Medium was another big one that people do, and then you have like uh uh um Stack Overflow, these kind of things. 00:20:38.240 --> 00:20:44.799 So people put a lot of content in other people's uh uh environments rather than owning where they actually have it. 00:20:44.960 --> 00:20:52.240 Have an archive somewhere that you control, and I think that's incredibly important because companies get bought, get changed, uh, these kinds of things. 00:20:52.400 --> 00:20:59.279 It's really uh fascinating how fast how fast people make money with your content and not give you anything for it. 00:20:59.519 --> 00:21:04.960 I mean, the best story was that x.com domain, that was basically the the Twitter X account. 00:21:05.119 --> 00:21:14.880 There was a person that was actually you have having that X account and um on Twitter, and then he basically had like 40,000 followers and was greatly successful. 00:21:15.119 --> 00:21:32.319 And when uh some person bought Twitter and turned it into X, they just took his uh took his uh uh name and changed his uh account to X1234567 or something like that, and basically offered him to uh come come to the office and get an X t-shirt instead of just getting something else. 00:21:32.480 --> 00:21:43.759 But before he was he was offered$40,000 to sell that that the X dot com uh Twitter name because those short Twitter names were really necessary for a lot of companies as well. 00:21:44.000 --> 00:21:49.759 So it's just there's people out there that run big things that have no qualms about ripping you off. 00:21:49.920 --> 00:21:56.079 So make sure that you actually own your stuff and you can actually um uh get it out there as well. 00:21:57.359 --> 00:21:59.759 Now, this was interesting uh um that. 00:22:00.240 --> 00:22:10.000 Uh the in 1954 IBM said Fortran would eliminate coding and debugging, what the numbers say about AI and the work and guess it developers Tuesday in 2020. 00:22:10.880 --> 00:22:13.839 So we always say, like, when is it unnecessary to code? 00:22:13.920 --> 00:22:24.880 And it was quite fun that actually in 1981, 1977, these are all the links where basically you say you don't code anymore, you just use other things. 00:22:25.119 --> 00:22:26.480 It will be interesting to see. 00:22:26.559 --> 00:22:31.279 We had uh I just had a coffee with developers with Bernard Vogels and asked him about that as well. 00:22:31.440 --> 00:22:38.160 Like, uh, is it really that we just are gonna be people talking to agents that code for us, or will code stay? 00:22:38.319 --> 00:22:50.000 And I'm I'm I I'm not sure, but I've like it really that it's really that long ago that they said, like, okay, this is the end of coding, you don't need to worry anymore, the machines are gonna do it for you. 00:22:51.519 --> 00:22:57.200 I I wonder, I mean, you know, Pratasha, we we'd love to get people on the show and kind of know what they what they feel about this, right? 00:22:57.359 --> 00:23:01.599 But when

you hear these kinds of predictions, I mean, like Chris said, you know, who knows ultimately? 00:23:01.759 --> 00:23:02.640 We're kind of not sure. 00:23:02.799 --> 00:23:05.279 But what does your what does your kind of gut instinct say? 00:23:05.359 --> 00:23:12.880 You know, is is code on the way out because it maybe won't be important, everything will be so abstracted away that it's just it's almost irrelevant. 00:23:13.119 --> 00:23:18.000 Or is there something to be said for maybe that we don't want it to go away, so it won't go away? 00:23:18.079 --> 00:23:18.960 Because we do like it. 00:23:19.119 --> 00:23:24.960 We do, you know, what it can be uh you know uh the bane of our life writing code, but we do love it, right? 00:23:25.039 --> 00:23:30.960 And I wonder if uh maybe that will be enough to actually keep it around and for us not to abstract things away. 00:23:31.119 --> 00:23:32.000 Um, I don't know though. 00:23:32.079 --> 00:23:32.720 What do you think? 00:23:34.240 --> 00:23:50.799 I think uh uh we the older generation, right, who used to write code by their hands know the uh dopamine head that we used to get when we see a successful render if you are a front-end developer or successful API call, okay. 00:23:50.880 --> 00:23:53.519 It it happened successfully, yay, we did it, right? 00:23:53.839 --> 00:23:56.240 Uh, but that's not the case now, right? 00:23:56.400 --> 00:24:04.000 It's like everything is done by AI, and we don't even the newer generation doesn't know that happiness of a successful code. 00:24:04.240 --> 00:24:15.680 Uh you know, how how does it look like, like, because everything is done by AIJ, just direct them to do it, and they don't even know what the code does if if you talk about the extremely newer generation. 00:24:16.079 --> 00:24:33.200 Um, so that that changes something uh it's it's hard to depict, and I think the more software we generate with more code, uh it's it would be difficult to uh uh review it for the newer generation. 00:24:33.359 --> 00:24:44.799 Of course, the generation in which we all used to write code, understand that what is wrong, what is right, but uh uh and how the product looks like, how the leadership wants the code to be like, right? 00:24:45.039 --> 00:24:55.119 But the newer generation is something exp especially the student one, uh they they are very uh stuck in the loop on what should they learn about. 00:24:55.359 --> 00:24:59.039 They are stuck in a loop that should they learn coding or not? 00:24:59.359 --> 00:25:08.559 Uh because everything I was in a conference and a Google Device conference uh yesterday, and the most asked question was should we learn coding or not? 00:25:08.640 --> 00:25:10.400 Because everything is done by AI. 00:25:10.880 --> 00:25:20.559 Uh it's again, uh it's very hard for the newer generation to you know take a call that what should they do in this in this space right now. 00:25:21.759 --> 00:25:28.240 Yeah, I mean that's kind of like should we learn to drive a car when everything is self-driving, uh and it isn't yet. 00:25:29.920 --> 00:25:43.440 It's it's uh it's an i it I mean that that that always fashion that always fascinated me that the craft from the from the work, like knowing the stuff, how to do it, and uh then you actually use tooling to make it for you. 00:25:43.680 --> 00:25:49.839 But I think learning the craft is still a very important thing because it gives you an independence of the tooling, and the tooling might go away. 00:25:49.920 --> 00:25:51.519 You might be out of tokens. 00:25:51.839 --> 00:25:57.599 The the company that gave that free cool uh uh model might actually think next week it's too expensive. 00:25:57.759 --> 00:26:06.720 I mean, yeah, again, X just announced that the Crockbot was like super expensive to run and impossible to actually maintain, and all of them will come up with that. 00:26:06.880 --> 00:26:14.319 So I think the the answer for that one will be uh will be still uh yeah, depending on how secure you want your job to be. 00:26:14.480 --> 00:26:23.599 If you want to be somebody that goes in quickly, releases something, becomes a serial entrepreneur, and releases one app after another, do that right now, make the money and run. 00:26:23.839 --> 00:26:29.680 But uh, it if you want to have a career and actually stay longer in the market, that might not be the right way of doing it. 00:26:29.839 --> 00:26:34.319 But uh it will be interesting what's what's happening in the next few months about this. 00:26:35.680 --> 00:26:36.799 I just have a question. 00:26:36.960 --> 00:26:37.680 Can I ask? 00:26:38.079 --> 00:26:41.920 Yeah, just a follow-up question to what Chris Chris mentioned. 00:26:42.079 --> 00:26:58.079 So uh uh you mentioned that um we should learn coding at Spect if we want uh uh you know a secure job and future because the coding agents might go away, or you know, that can be a limit on the tokens that businesses give to the developers. 00:26:58.480 --> 00:27:16.799 But at the other end, if you see um the the way companies are hiring, uh they have stopped asking those uh coding related questions, they have started asking more on uh uh how they use AI in in you know making their developer job more productive. 00:27:17.039 --> 00:27:18.720 Uh so what do you think? 00:27:18.799 --> 00:27:20.720 Uh what's your take on this? 00:27:21.440 --> 00:27:22.640 Well, I think that makes sense. 00:27:22.799 --> 00:27:24.880 I mean, I hated whiteboard interviews. 00:27:25.039 --> 00:27:26.240 I think that makes no sense. 00:27:26.400 --> 00:27:39.839 I mean, having uh having a world that we're in right now where we do software collaboratively, um, I never had to actually look at the code or ask somebody in a job interview to show me to write code from scratch. 00:27:40.000 --> 00:27:43.680 Because I looked at stuff they've done before and I know what that person can do. 00:27:43.839 --> 00:27:52.079 I looked at people's GitHub repos and I looked

how they reacted to pull requests, how they reacted to comments, and that way I already knew what that person was doing. 00:27:52.240 --> 00:28:11.279 So the clever uh uh the clever questions that we had in the past, like oh, how do you optimize this uh uh this like uh uh leaf structure or this tree-shaking algorithm and these kind of things, I always found them gatekeeping and kind of pointless because uh algorithmic questions in interviews, you never use it in your job, anyways. 00:28:11.359 --> 00:28:13.200 You didn't use it when we didn't have AI. 00:28:13.359 --> 00:28:18.240 You just basically, it was just a write of passage to do you know that thing. 00:28:18.480 --> 00:28:26.480 And uh, I mean, I remember at one job interview they asked me to write a sudoku validator, and I'm like, I didn't know the rules of Sudoku, so I asked them that first. 00:28:26.960 --> 00:28:38.799 And like uh an uh British uh BBC once made me write XSLT on a whiteboard at like nine in the evening, and I'm like, okay, I'm out, I'm I'm not staying here, this is not a fun interview. 00:28:39.039 --> 00:28:53.359 But the uh the way I guess a lot there's a lot of news lately that people say, like, oh, we don't do coding interviews at all, anything, and it's people like Entropic that actually bring that out, and everybody's like, oh my god, they're so ahead, we have to do the same thing. 00:28:53.599 --> 00:29:07.440 I think it's still important for people to take a look at code and not only what they're using AI for, but also uh, I think the review cycle, as you said before, younger generations don't even know what bad code is. 00:29:07.680 --> 00:29:18.880 I think I always was more excited uh showing people code in in job interviews and ask them what's wrong with it and see them how they would improve it and what they can do, because that's a skill that we still need to have. 00:29:18.960 --> 00:29:24.240 And if your answer is like I just asked Claude to do it, then I might as well hire Claude, not you. 00:29:25.519 --> 00:29:28.240 I think I think just to jump on that as well, I think there's something really interesting. 00:29:28.319 --> 00:29:34.640 I was just thinking about from like I was working in education in as a teacher, and uh it was uh the year that ChatGPT came out. 00:29:34.720 --> 00:29:38.000 I want to say, what was it, 2022, something like that, but uh 23 maybe. 00:29:38.079 --> 00:29:43.039 But what when it came out, it immediately meant we had to change the degree, the course that we were teaching. 00:29:43.119 --> 00:29:54.480 Because you know, before that we had some form of essays, you know, very sort of uh only a few, but um we instantly it meant that they just weren't that important because it was something that we weren't testing the right thing, really. 00:29:54.559 --> 00:30:03.680 You know, we were if you see that as an examination, it was uh their ability to, I suppose, use AI to uh put the information in a way that sounded like it was them. 00:30:04.000 --> 00:30:07.839 Now that's not useful, that's not really, you know, that's that's pointless, really. 00:30:08.000 --> 00:30:23.519 So that would be the equivalent if you mapped it back on to a kind of coding exercise that they've submitted a load of code that we don't know if it's written by them or not, and uh they're not at any point asked to try and explain it or you know, uh explain even their kind of architect architectural decisions, let's say. 00:30:23.759 --> 00:30:32.079 So I I think that it's kind of the same thing that we're just at a point where we're trying to figure out what we're trying to test and what which skills we actually need to test. 00:30:32.240 --> 00:30:34.400 What skills does a developer actually need? 00:30:34.640 --> 00:30:40.319 And I think just to uh sum up that, Chris, you mentioned Werner Vogels uh from um uh Amazon, uh the CTO. 00:30:40.480 --> 00:30:45.599 And I I I've got this quote in my head that he said this week, you know, which was people are asking me, would my job go away? 00:30:45.759 --> 00:30:56.880 And he said, um, maybe if you don't evolve, and he said you need to start developing the skills that are less about kind of writing code and syntax, but they're higher-level architectural um understandings, you know. 00:30:57.039 --> 00:31:07.200 And I think there's something to be said for that there, that writing an essay, writing a sentence, writing syntax, writing one line of code is is not as important as the the big picture and having an understanding of that. 00:31:09.920 --> 00:31:10.480 Cool. 00:31:10.720 --> 00:31:12.240 Uh, this is also an interesting one. 00:31:12.319 --> 00:31:16.079 If you build a website, that's a website's back what uh what a website should do. 00:31:16.240 --> 00:31:17.759 So, and I found it really interesting. 00:31:17.839 --> 00:31:25.599 That's a really interesting uh tool to actually uh take a look at like what you do out there because websites aren't dead yet. 00:31:25.680 --> 00:31:32.559 If anything, I think it's more important to publish them properly so agents and people can index them and find it for you. 00:31:32.720 --> 00:31:36.799 So uh interesting little uh little website, uh app specification. 00:31:37.680 --> 00:31:38.240 Specification. 00:31:38.960 --> 00:31:40.960 Also exists if you actually write these things. 00:31:41.119 --> 00:31:44.799 Sometimes you forget the one thing that might make it successful or not. 00:31:45.200 --> 00:31:48.480 Now, um, Dan, this is down your alley, isn't it? 00:31:48.559 --> 00:31:50.960 Like uh I love it. 00:31:51.200 --> 00:32:03.920 The the biggest problem that we have nowadays is if you get like uh uh flyers for events and these kind of things, they're all these horrible AI-generated chat GPT things that I just cannot see anymore. 00:32:04.079 --> 00:32:15.039 And this is a library of like handwritten, really badly looking stuff, but at the same time human-done stuff that you can actually uh start

using and actually editing for you. 00:32:15.119 --> 00:32:21.200 And you've been been in bands, you've been promoting bands, your first websites were for bands as fan projects. 00:32:21.279 --> 00:32:22.400 So, what do you think of that? 00:32:22.720 --> 00:32:23.279 Oh, I love it. 00:32:23.359 --> 00:32:33.200 I mean, I mean, I I think I think it's because, yeah, aside from the nostalgia of of kind of being, you know, like a teenager and even trying to, you know, recreate something that you've seen, like an art style. 00:32:33.279 --> 00:32:41.440 You know, I used to try and sort of make the website look like the kind of band artwork used to and things like this, you know, with CD covers and all sorts, really showing my age. 00:32:41.599 --> 00:32:47.279 Um, but yeah, we we talked before about, you know, it's it we don't want to sound too much like a kind of art school person, right? 00:32:47.359 --> 00:32:53.440 But there's something to be said for um something that's maybe not great, but feels really human. 00:32:53.680 --> 00:32:57.119 And you you know that it's a human and there's something DIY about it. 00:32:57.200 --> 00:33:00.559 I mean, okay, I this takes me back seeing that, Chris, the gaffer tape. 00:33:00.799 --> 00:33:12.079 Because there was a there was a real there was a real trend in the kind of music and and rock world, you know, of bands wanting things to look like they were, whether they were gaffer tape or Polaroid photos or kind of real world objects, you know, laid down. 00:33:12.240 --> 00:33:20.160 Um but yeah, there's something to be said for that, that you know, that like you're saying, about whether it be capturing your content and making it so that you've you've you've got it forever. 00:33:20.240 --> 00:33:22.000 Uh some of these things I don't want to go away. 00:33:22.160 --> 00:33:28.640 Some of these kind of slightly rubbish DIY, very human, um, things that, you know, the that there's no one getting paid for it. 00:33:28.799 --> 00:33:34.160 There's no, you know, it's the kind of local c stand-up comedy show or whatever, but there's just something about that. 00:33:34.240 --> 00:33:37.680 And I love the fact that somebody's taken the time to kind of keep hold of them. 00:33:37.759 --> 00:33:49.039 Um because I I think that that that that's the risk, right, with AI that everything becomes uh very cookie-cutter, everything's the same, loses uh the character, loses the human kind of edge. 00:33:49.440 --> 00:33:54.960 Uh even if even if from a design point of view, uh the human one is rubbish. 00:33:55.839 --> 00:33:57.119 It's still got character. 00:33:57.920 --> 00:33:58.240 Yeah. 00:33:59.519 --> 00:34:06.079 And the other one that was uh also obviously a vibe coded thing, I think, was that word art generator, which cracked me up. 00:34:06.720 --> 00:34:16.239 I love that they're basically you can use all the word art that you had to have in Word and now in HTML and get it generated as C D as CSS to actually use later on. 00:34:16.559 --> 00:34:17.039 That's cool. 00:34:17.920 --> 00:34:19.760 Yeah, there's a few of those that are on there. 00:34:19.840 --> 00:34:26.079 But I I like the fact they've done this, yeah, the actual um UI of the of Windows, you know, and pet and paint. 00:34:26.239 --> 00:34:27.679 Um because a lot of them they don't really. 00:34:28.079 --> 00:34:30.079 Or when they do, it's too modern. 00:34:30.239 --> 00:34:33.039 So this is what's what why why which one is that though? 00:34:33.199 --> 00:34:40.800 That's that is it just the UI now of of of paint, or you know, or or is that kind of a specific is that Windows 95 or something? 00:34:40.880 --> 00:34:42.159 I can't I can't picture it. 00:34:42.239 --> 00:34:44.079 It's nostalgia, but it's vague. 00:34:44.480 --> 00:34:47.119 I I think yeah, it's not 95. 00:34:47.280 --> 00:34:52.719 I think it was a mixture of something else, but yeah, it's just fun that these things are possible nowadays. 00:34:53.280 --> 00:34:53.760 Cool. 00:34:53.920 --> 00:35:00.880 Um let's talk about two quick tools before we go into Vega News and then talk more about what uh uh what Pretesh is doing. 00:35:01.119 --> 00:35:16.159 This came out from um from Clear Left in England, and I just love that it's uh it's a uh hue um accessibility tester on color, a color tester for uh hue colours and pass criteria and all these hits. 00:35:16.320 --> 00:35:23.039 And I just love that the the title, this usual pass, is just uh having a dad joke in the title of a tool is a good thing. 00:35:25.360 --> 00:35:26.719 I like the look of it as well. 00:35:28.719 --> 00:35:44.480 So yeah, fun to do if you have if you have colours and you want to do uh see if all of them check and you don't want to use the um the developer tools or use uh uh build things built into Figma, this actually quite works and it's quite fun to see. 00:35:45.119 --> 00:35:48.239 And uh, what do you use when you when you're kind of doing that? 00:35:48.320 --> 00:35:51.280 Because that because as in if you're kind of picking a colour palette, you know. 00:35:51.360 --> 00:35:56.559 I mean, obviously I used to have a kind of website that had lots of colour palettes on it, so I've always been interested in how kind of people come to that decision. 00:35:56.639 --> 00:36:06.719 But if you're doing that, if you're picking a colour palette, I know obviously now people might just say, give me something that looks professional or whatever, but um, you know, what was the what was the kind of uh uh approach you would take? 00:36:06.880 --> 00:36:17.599 Because I usually would use tools like this where I need something visual that's a little bit like Photoshop, but I don't want to open the software and I want an eyedropper so I could grab something from somewhere else as a starting point, you know? 00:36:17.679 --> 00:36:20.639 Um, or do you do this in in the kind of developer tools? 00:36:20.800 --> 00:36:24.960 Because I I personally I've done it

maybe once or twice in the developer tools, but not not usually. 00:36:25.280 --> 00:36:36.159 I I normally just randomly put numbers, uh put colors in in the developer tools and in VS Code and then see if it fits, and that's why my stuff looks like shit. 00:36:36.320 --> 00:36:39.039 But it's actually uh it's actually mine, so that's I don't care. 00:36:39.199 --> 00:36:46.000 I mean, I like that the uh that the color picker that you have in the developer tools actually has this accessibility check built in. 00:36:46.159 --> 00:37:01.280 So if you actually go to somewhere and um like for example, you go to this one here, it actually will tell you the color, and you can click that here, and you can see if the contrast ratio is good enough. 00:37:01.599 --> 00:37:11.760 So basically, those two lines tell you if the contrast of background and foreground is good enough, and you can pick and if it's not good enough, you just pick that one and just copy it, and then you're done. 00:37:11.920 --> 00:37:15.360 So that's something, and that's actually in this in VS code as well by now. 00:37:15.440 --> 00:37:20.400 So basically, when I when I do that stuff, for example, main. 00:37:22.159 --> 00:37:29.920 I love that every in C S code, in VS Code that every color that you use has a well, it's ironic now that there's no color in there. 00:37:31.920 --> 00:37:33.840 Uh has this patch in there as well. 00:37:33.920 --> 00:37:42.000 It doesn't have the accessibility check, but you can just play around in the NVS code itself, so you don't even need to go to a developer tool to do it. 00:37:42.239 --> 00:38:00.159 So I found that actually quite uh quite it does the job for me, but I'm also not a professional designer, and it's basically, I guess, doing a uh uh uh doing a like I mean you built a a color picker, didn't you, that gives you like uh uh uh colours that match the other one around them. 00:38:00.239 --> 00:38:02.480 So you probably know more about that than I do. 00:38:02.800 --> 00:38:10.719 I mean it's just lots of there's lots of these little, I mean I kind of started doing it myself, and then you find there's lots of these little libraries, and they're actually very simple in terms of finding the the colour contrast. 00:38:10.880 --> 00:38:23.920 Um but like you're saying with that there, you know, having the the kind of double A and and uh triple A passing, like that that should be the basic, I think, that people are at least kind of checking that because I I see it quite often, and I've even been called out for my for it myself, you know. 00:38:24.000 --> 00:38:30.800 I think we had a past guest, Emma, who who once kind of sent me a message and she said, by the way, I've just done an audit of the accessibility of the colours on your site. 00:38:30.880 --> 00:38:33.760 And you know, you're dangerously close to them passing, but not quite. 00:38:33.840 --> 00:38:56.719 So like you've either you've either uh checked the contrast and then actually changed it because you wanted it to be slightly nicer looking in your mind, but less accessible, you know, and it kind of was good to be called out for actually um being guilty of making a decision sometimes because you think it looks better, but actually forgetting that there are standards for a reason and that you you as the designer are the least important one in that equation. 00:38:56.880 --> 00:38:59.679 Um, it's about the users, but yeah, yeah. 00:39:00.000 --> 00:39:14.079 I I also use this uh the accessibility insights for web that's a Microsoft uh tool that's also used internally, and it's an extension for the browser that actually runs a quick check over everything, and then you get these warning signs here that tell you what the problems are. 00:39:14.320 --> 00:39:16.639 And uh you can even file an issue directly. 00:39:16.719 --> 00:39:20.800 So in Microsoft itself, we had it with the uh with the bug filing tool. 00:39:21.039 --> 00:39:28.400 So basically, we went through that and tested it in our browser, and it's just a browser extension for Chrome or for Edge or any other browser out there. 00:39:28.639 --> 00:39:34.000 That's something that I use for that color checking as well, that also simulates the grayscale kind of thing. 00:39:35.360 --> 00:39:45.360 Another tool that I I I will have a blog post about, and I'm almost done with it, uh, which I've done uh um which I've used for years is Name Changer. 00:39:45.440 --> 00:39:50.480 And that's something that's uh so simple that it it annoys me that it's not part of the operating system. 00:39:50.719 --> 00:40:13.199 Like if you're in Finder in uh um if you're in Finder on Mac, you can not many people know that you can actually highlight more than one file and say like rename, and then you get a chance to replace the text in each of them, then you get a chance to add text to each of them or to format text accordingly, like name and index and these kind of things, and that renames all these files. 00:40:13.440 --> 00:40:24.320 The problem that I have with it is that it only shows you the one of them and it doesn't show you all of them, it doesn't show you if there's issues when you rename them, if you overwrite files that way, or if you lose information. 00:40:24.480 --> 00:40:31.199 Uh so uh you don't have wildcards and you don't have regular expressions, and regular expressions is something I live and breathe, and it's a wonderful thing to have. 00:40:31.360 --> 00:40:40.800 And that's why I use uh if you have that problem that you have to rename a lot of files, there's this tool that's free, it's been out for around for decades, I think. 00:40:40.960 --> 00:40:45.599 It's got name changer, and that one you drag all the things in there, and then you can do a lot more. 00:40:45.679 --> 00:40:57.440 You can replace the first occurrence of something, the last occurrence of something, all occurrences, you can do a wildcard, replace, append, prepend, remove characters, insert

characters, change case, sequence, and regular expression. 00:40:57.519 --> 00:41:09.599 And regular expression, of course, is the thing that I love the most, where I say, like, okay, everything that is.com uh followed by a dash, remove that one and it shows you live how the changes are coming in as well. 00:41:09.760 --> 00:41:12.719 And if there's any issues with it, then it would show them in red. 00:41:12.800 --> 00:41:16.239 So basically, there's no way you can actually do these things wrong. 00:41:16.400 --> 00:41:29.599 So if you actually need to change a lot of names of files, which is something that when you, for example, when I shifted now my 25 uh two my 21 years of blog posts, I had to do a lot of search and replacing in file names. 00:41:29.760 --> 00:41:32.000 Um, that tool is a really, really fun thing. 00:41:32.400 --> 00:41:37.760 It also has an image browser, so you've got images, you can get images directly from the XF data in the images. 00:41:37.920 --> 00:41:40.880 So, and you can also get data from the file itself. 00:41:41.039 --> 00:41:50.960 So, if you're uh if you want, for example, to include the date when it was generated, you can get that from the file metadata, and that's something that uh that finder doesn't do for you. 00:41:51.199 --> 00:41:52.880 So that's just name changer. 00:41:52.960 --> 00:42:00.400 If uh if you ever need to edit a lot of files, uh a lot of files on your machine, it's a tool that saved me a lot of times. 00:42:01.360 --> 00:42:01.920 Cool. 00:42:02.719 --> 00:42:03.039 Nice. 00:42:03.360 --> 00:42:03.840 All right. 00:42:04.159 --> 00:42:08.480 Print dash, you know, uh probably we have a thing called uh fake on news. 00:42:08.719 --> 00:42:15.039 So uh we're gonna read your headlines, uh, and you have to tell us if I made them up or if they're real. 00:42:15.360 --> 00:42:20.960 And we have uh we have uh um uh a top list of people around the show before. 00:42:21.039 --> 00:42:22.800 Uh we will see if that works out. 00:42:22.960 --> 00:42:26.000 And uh yeah, let's play fake on news and see how you fare. 00:42:26.320 --> 00:42:30.079 Tesla's newest feature lets cars drive themselves to charge at any time. 00:42:30.320 --> 00:42:31.360 Looks like a news. 00:42:31.599 --> 00:42:33.280 It's possible, it's possible. 00:42:33.920 --> 00:42:34.559 It's fake. 00:42:35.920 --> 00:42:37.440 100 video calls per day. 00:42:37.599 --> 00:42:40.079 Models are applying to be the face of our i scams. 00:42:40.480 --> 00:42:42.079 100 video calls per day. 00:42:42.320 --> 00:42:48.320 Maybe if China's doing it, then it's possible to stick a news, whatever. 00:42:49.679 --> 00:42:51.039 It was it was news, yeah. 00:42:51.280 --> 00:42:55.599 Quantum internet achieves first successful intercontinental connection. 00:42:55.840 --> 00:42:57.920 No, I don't know what quantum internet is. 00:42:58.159 --> 00:43:01.760 I I feel like there's a lot of words there just trying to get us to believe it's news. 00:43:01.840 --> 00:43:02.639 Do you think it's fake? 00:43:02.800 --> 00:43:06.239 As far as I I've gone wrong in all, so you take the call this time. 00:43:06.800 --> 00:43:09.119 I'll go f I'll go for fake, but apologies. 00:43:09.360 --> 00:43:09.840 Yes. 00:43:10.800 --> 00:43:12.079 It's it it was fake. 00:43:12.239 --> 00:43:12.480 Okay. 00:43:12.719 --> 00:43:18.880 Pentagon treating military-based freezer malfunctions as suspected cyber attack, officials say. 00:43:19.360 --> 00:43:21.039 Uh let's make it a news. 00:43:22.400 --> 00:43:22.960 Yes. 00:43:23.440 --> 00:43:24.239 And it is news. 00:43:24.880 --> 00:43:27.599 Maxing is dictionary.com's word of the year. 00:43:27.679 --> 00:43:29.360 It's cooler than saying you're doing more. 00:43:29.599 --> 00:43:30.400 It has to be news. 00:43:30.480 --> 00:43:31.599 I've seen that word everywhere. 00:43:31.679 --> 00:43:32.559 It has to be word of the year. 00:43:32.719 --> 00:43:35.360 Yeah, say a lot of Gen Zoo there. 00:43:35.599 --> 00:43:36.000 Yeah. 00:43:36.159 --> 00:43:38.079 I mean, I even see it on the news now. 00:43:38.320 --> 00:43:39.280 Yep, it's news. 00:43:39.519 --> 00:43:43.280 Beavers chew down sign warning walkers about beavers in Poland. 00:43:43.679 --> 00:43:44.320 That feels real. 00:43:44.400 --> 00:43:45.360 It's so bonkers. 00:43:46.000 --> 00:43:46.719 No idea. 00:43:46.800 --> 00:43:48.400 Or or do you've done it? 00:43:48.960 --> 00:43:53.760 I I I feel like it's real because that if if you can make that up, you've got a brilliant imagination, you know. 00:43:53.840 --> 00:43:55.920 Um, but also the irony is is amazing. 00:43:56.079 --> 00:43:57.199 I I would go with news. 00:43:57.440 --> 00:43:58.400 It is news. 00:43:59.360 --> 00:43:59.519 Meta. 00:44:00.159 --> 00:44:03.679 Reportedly built a humanoid robot prototype that can do household chores. 00:44:03.920 --> 00:44:04.880 Yeah, it's new. 00:44:05.920 --> 00:44:07.360 Yeah, I I I'm wearing it. 00:44:10.480 --> 00:44:11.760 You're getting it, Pitash. 00:44:11.840 --> 00:44:12.880 Yes, the technicalities. 00:44:14.559 --> 00:44:15.039 What do you think? 00:44:15.199 --> 00:44:20.800 I know Nemo, Nemo, there was some some uh Robo name, but I don't know if it's on meta. 00:44:24.639 --> 00:44:26.639 It is it is meta, it was real. 00:44:26.880 --> 00:44:27.360 It's meta. 00:44:27.679 --> 00:44:28.480 Uh yeah. 00:44:28.719 --> 00:44:34.000 Bank of America kindly asks you that you don't share your banking password with your AI assistant. 00:44:34.239 --> 00:44:35.840 Yeah, that feels real, yeah. 00:44:36.239 --> 00:44:37.440 Ideally it should be real. 00:44:37.599 --> 00:44:41.119 Yeah, the advice is so simple, but I yeah, lots of people need to hear that. 00:44:41.840 --> 00:44:42.320 Yep. 00:44:42.480 --> 00:44:42.800 Yeah. 00:44:43.199 --> 00:44:48.320 Attackers impersonate AI bots to scan organizations for keys and credentials. 00:44:48.960 --> 00:44:50.079 This also sounds real. 00:44:50.239 -->

00:44:53.519 Yeah, I mean we we know it's true at least, it's whether or not it's real. 00:44:53.840 --> 00:44:54.559 Yeah, yeah. 00:44:55.360 --> 00:45:00.239 Highway thieves are spoofing GPS to steal millions in data center bound cargo. 00:45:00.559 --> 00:45:02.880 It's tricky to say it, yeah. 00:45:03.280 --> 00:45:03.840 It's real. 00:45:04.000 --> 00:45:04.320 Sorry. 00:45:04.880 --> 00:45:09.599 AI agent swarms unleashed in Minecraft to build a thriving civilization. 00:45:09.760 --> 00:45:14.000 50 agent teams fail after failing into paralyzing bureaucracy. 00:45:22.800 --> 00:45:23.920 It is news. 00:45:24.159 --> 00:45:30.000 French MPs vote for a 0055 tax on every gigabyte sent to users. 00:45:31.920 --> 00:45:32.960 Yeah, strange amount. 00:45:33.039 --> 00:45:35.519 Well, then maybe that's the reason it's real, because it's so specific. 00:45:35.840 --> 00:45:36.480 I don't think so. 00:45:36.559 --> 00:45:41.599 Then it should be real and should be heart, and uh I think it's real. 00:45:41.920 --> 00:45:42.159 Cool. 00:45:42.239 --> 00:45:45.039 What does that mean for the hell for for the for the leaderboard? 00:45:45.199 --> 00:45:46.480 You are on place four. 00:45:46.639 --> 00:45:47.199 Look at that. 00:45:47.360 --> 00:45:48.960 Oh, not bad. 00:45:49.440 --> 00:45:55.360 You've got yourself a bronze medal because you are actually joined with Roxy Rodriguez Becker, who's just above. 00:45:55.440 --> 00:45:57.760 So you've actually got yourself a bronze medal, you're joint third. 00:45:58.000 --> 00:45:58.639 Congrats. 00:45:58.880 --> 00:45:59.280 Cool. 00:45:59.679 --> 00:46:00.480 Nice, excellent. 00:46:00.639 --> 00:46:01.440 That's not easy. 00:46:01.840 --> 00:46:04.480 All right, then that was fake on news for today. 00:46:04.559 --> 00:46:05.920 So let's chat about you. 00:46:06.079 --> 00:46:08.400 Um, you're doing this humans of tech thing. 00:46:08.480 --> 00:46:09.280 What is that? 00:46:10.239 --> 00:46:13.519 Yeah, um, should I start with how it started? 00:46:13.679 --> 00:46:15.039 Like, yeah, yeah. 00:46:15.360 --> 00:46:21.280 So in 2020, uh, so I got into tech into in the year 2022. 00:46:21.519 --> 00:46:42.559 Uh, and um I was there is this page called Times of uh sorry, uh humans of New York, and in India we also have humans of uh humans of Bombay, so it's a very popular page in India, and then I I was researching around and um I I didn't found something like humans of tech. 00:46:42.800 --> 00:46:50.719 And back then we used to have a very active Twitter community, and we used to call it Tech Twitter, uh, and then that was Tech Twitter India. 00:46:50.880 --> 00:47:00.559 Back then we used to do a lot of spaces on Twitter, you know, it was very active uh before Elon Musk bought the company, uh, and the algorithm was different. 00:47:01.039 --> 00:47:08.880 Uh, and during that time, I the first thing I searched was the domain, and it was available for some reason, I don't know how. 00:47:09.119 --> 00:47:12.800 And I just bought it, I didn't know what to do out of humans of tech. 00:47:12.960 --> 00:47:16.639 I just was like, let me buy the domain first, and I got the domain, I bought it. 00:47:16.719 --> 00:47:25.360 I star I all I secured all my socials of Instagram, Twitter, and LinkedIn, and then I I started with tutor spaces. 00:47:25.519 --> 00:47:53.280 I so what I thought process was uh I want to onboard people uh who are in tech, uh, but they we want to share a human story uh apart from the tech story that they have, and the people are doing good and doing some inspiring work in tech, and just want to get a human out of them and want to share that human story uh uh to the people. 00:47:53.519 --> 00:47:58.320 Um, and we I started with hosting Twitter spaces. 00:47:58.400 --> 00:48:05.119 So if you see the initial few uh podcasts, initial like 10-15 podcasts, I did it with my friends on Twitter. 00:48:05.280 --> 00:48:12.239 Uh Hemandiga was also one of them who will be your next uh uh next uh guest uh on 11th of November. 00:48:12.400 --> 00:48:14.400 Uh and a lot of other people. 00:48:14.559 --> 00:48:22.239 We we Francesco, I wanted to interview, but after that, I just discontinued that, and we have Toby Loba and everyone else. 00:48:22.400 --> 00:48:24.000 Uh a lot of good people in tech. 00:48:24.079 --> 00:48:30.000 I I did got a chance to, you know, uh share the human story of theirs, and I got a really good traction. 00:48:30.400 --> 00:48:32.880 People actually wanted to listen to them live. 00:48:33.039 --> 00:48:36.880 Um, and then I started video podcasts similar to what we are doing right now. 00:48:37.119 --> 00:48:45.920 Um, and after a point of time, I I started, I found another gap, which was uh conference blogging. 00:48:46.079 --> 00:48:46.800 Um, right. 00:48:46.960 --> 00:48:48.320 Have you heard of conference blogging? 00:48:48.400 --> 00:48:52.719 Have you heard of blogging people to do travel vlogs and everything, right? 00:48:53.199 --> 00:48:58.000 Um, so my thought was what what what if I do a conference vlogging? 00:48:58.159 --> 00:49:02.239 What if I uh show the world how does a conference looks like? 00:49:02.400 --> 00:49:06.719 Because for me it's privileged to be part of any conference because I live in Bangalore. 00:49:06.880 --> 00:49:22.719 But for a lot of students and what what a lot of working professionals who are in tier two and tier one, tier three cities in India, they might not get a chance to you know experience that how does a conference look like, or what is the conference at the first place, all right? 00:49:22.960 --> 00:49:37.119 Um so I just want to this this this thing I thought, and then I thought, like, let me let me create an experience, let me create a video, let me create a blog where I can share that okay, this is how a conference looks like, and this is how you why you should be part of. 00:49:37.280 --> 00:49:37.519 Right. 00:49:37.599 --> 00:49:41.679 I started whenever

I go to a conference, I record it nicely. 00:49:41.760 --> 00:49:53.920 Uh, and I also take a like mini interviews with people, sharing them, sharing about what they're doing, and also uh how how was the conference, that feedback, and uh some some good questions, some funny questions. 00:49:54.159 --> 00:50:10.880 And I the first one was why I did was the GitHub constellation that happened in India, uh, which was after a very long time, and uh I got really good response again, and then after that I did a few more a Bix conference, and then uh we also did DevilCon India as well. 00:50:11.119 --> 00:50:19.679 Um, recently when I went to Korea for an MCP dev summit, there also I did the same thing, and I I feel like I I feel joy while doing this. 00:50:19.840 --> 00:50:21.920 I feel like okay, uh it's fun, right? 00:50:22.000 --> 00:50:24.960 It's also an icebreaker for me to chat with people, okay. 00:50:25.280 --> 00:50:29.440 If I want to chat with someone, if I go it directly, I I feel conscious. 00:50:29.519 --> 00:50:34.559 But if I go that, hey, I want to take a small interview of yours, could could you do you have two minutes for it? 00:50:34.719 --> 00:50:42.800 And it's an icebreaker for me to start a conversation with someone, especially with the organizers or some good people who are from the good company. 00:50:43.039 --> 00:50:48.239 Uh and yeah, the this was the idea, and this is what I've been doing with this uh page. 00:50:49.440 --> 00:50:55.519 What made you pick Twitter Spaces for that or X Spaces or rather than just going on YouTube? 00:50:55.840 --> 00:51:04.960 And is it just because it was easier or uh because I mean I'm I I'm more over like where where is it gonna be later on? 00:51:05.039 --> 00:51:16.079 Like the we we always say like you gotta be in one platform to reach people, but it feels to me like we're doing this to record it and then publish on our own platform, but you also publish to all the others as well. 00:51:16.320 --> 00:51:21.760 Like uh getting people into one platform feels uh feels kind of uh limited, I think. 00:51:21.920 --> 00:51:25.119 But was it just a choice that you made because of reach? 00:51:25.760 --> 00:51:37.280 Yeah, so back then I had like decent reach on my Twitter, like I had like six, five, six thousand people uh followers, and the people to whom I reached out, they also had a decent following, right? 00:51:37.519 --> 00:51:48.400 So later on, I was anyways going to record the space and put it on YouTube, but the live thing will always happen on Twitter space, and uh and I was not confident to be on camera back then. 00:51:48.480 --> 00:52:15.119 I I was very new, I was just out of college, like one year, two years, and Twitter space felt like an easy option and good for me to distribute to a good set of people who are uh in my my following and also the following of the guests, and most of them had good following, so it's it was like exchange of the audience, uh uh and it did help me get some good followers for humans of tech as well, because they were also one of the hosts. 00:52:15.360 --> 00:52:22.639 So the initial uh you know the initial uh reach that I I wanted for my page, I I got it from that. 00:52:24.639 --> 00:52:31.119 What were the interesting human stories that you found that you were that surprised you that you didn't think of before? 00:52:31.199 --> 00:52:34.159 What was the most beautiful thing you heard from somebody? 00:52:34.639 --> 00:52:35.199 Yeah. 00:52:35.440 --> 00:52:37.039 Um so it has been long. 00:52:37.119 --> 00:52:44.880 I have done inter like the uh uh interviews on humans of tech, and lately I have been doing this for this vlogging only. 00:52:45.039 --> 00:52:53.840 But uh I remember I I um on I had a I had a chat with this uh uh this girl called Shruti. 00:52:54.480 --> 00:52:59.599 She she she started her tech journey when she was 16. 00:52:59.760 --> 00:53:03.039 She she was very it started very early. 00:53:04.239 --> 00:53:19.840 The way she navigated from one place to other with the help of uh networking with the right people, um and you know, contributing to the value uh that she can provide to the businesses. 00:53:20.000 --> 00:53:23.679 She started as a freelancer, uh, she's still doing well, I believe. 00:53:23.760 --> 00:53:26.079 Uh it's been a while I have been touched with her. 00:53:26.320 --> 00:53:46.719 But uh the way she narrated her own story, uh starting from when when she she started when you she was 14, 15, 16, and then how uh she got her first project uh as a freelance website building, and she started coding and she was at least like I don't even know how to use a computer when I was 15 or 16. 00:53:47.039 --> 00:53:51.119 I just I was just playing that road rash game or a Wake Op 2 or something else. 00:53:51.599 --> 00:53:56.800 And these people are uh doing crazy stuff, and this is before AI thing, right? 00:53:56.880 --> 00:54:12.559 It's not like AI is support for supporting her, but the way she she lived, she's living, she started living the tech journey, and the way she shared about it was very inspiring, and that made me feel that there's a long way to go and there's a lot more to do. 00:54:12.800 --> 00:54:13.519 So, yeah. 00:54:14.239 --> 00:54:19.119 I think it's also nice to see that you know you it's very clear that you know that obviously there's added benefits, right? 00:54:19.199 --> 00:54:24.400 When you when you kind of do something like humans uh um humans of tech, you get the added benefits of, like you say, sharing audiences. 00:54:24.559 --> 00:54:30.159 But um, it seems like there's still quite a clear thing for you about giving other people the opportunities, right? 00:54:30.239 --> 00:54:48.880 Like that you're saying you feel very lucky to actually go to conferences, and I think that that's that's great because we don't, you know, we can take these things for granted, these opportunities we get, but that there's you wanting to share it

with other people and bring other people into tech because I think the thing I've seen and the thing that I've always been drawn to in other people in tech is people that are not gatekeepers. 00:54:49.039 --> 00:54:53.679 And if someone is a gatekeeper, they sometimes can do it for for a kind of legitimate reason, right? 00:54:53.760 --> 00:54:57.039 Like they might be saying, don't just dive into something if you don't understand security. 00:54:57.199 --> 00:54:58.800 It's like sensible gatekeeping. 00:54:58.960 --> 00:55:06.159 But I've certainly seen from you that you're like the opposite of a gatekeeper, that you're trying to use the word on board, but like bring people into tech. 00:55:06.320 --> 00:55:20.480 And uh sometimes I think that the way doing it, like the way you're doing it with storytelling, is more interesting because if you don't work in tech, you probably don't understand or really enjoy or love to hear the kind of nitty-gritty tech detail. 00:55:20.639 --> 00:55:26.159 You just want to know about a person and then be inspired by what they've done because then you think, well, I could do the same. 00:55:26.400 --> 00:55:36.880 Uh so and I can speak as somebody who's seen these kinds of things around, you know, and and kind of uh I could see for myself, like, oh, there's people who have made the career switch in the way that I was looking to. 00:55:37.039 --> 00:55:45.920 So I think that's something that I think you're probably being quite modest, but I really appreciate in what you do, Pratesh, is that you're definitely trying to kind of like lend a hand to other people and bring them into tech. 00:55:46.000 --> 00:55:51.840 And uh and I love I love seeing that because it's not all computers and agents, there's people behind it all, you know. 00:55:52.239 --> 00:55:54.719 Yeah, true, true, true, very true Daniel. 00:55:55.679 --> 00:55:55.840 Yeah. 00:55:56.559 --> 00:56:00.159 So, how does the conference vlogging work with your company? 00:56:00.320 --> 00:56:06.000 Do you do it in your own free time or is this something that your company appreciates you doing as well? 00:56:07.119 --> 00:56:10.639 Uh I I have been doing this long back when I was a developer itself. 00:56:10.719 --> 00:56:20.159 So nothing to do with my full-time thing, but they do appreciate uh the way I do it because I've done some um whenever an internal event happens. 00:56:20.239 --> 00:56:26.239 Uh so I'm very I believe like I've learned to film uh events really well. 00:56:26.639 --> 00:56:42.000 How do we start it, how do we record it, uh uh you know, um holding a camera to editing it well, to adding a good music and make it like make it a lively story out of an event is something that I've learned in a few years. 00:56:42.159 --> 00:56:45.519 Uh and I and the good part is I do enjoy doing that. 00:56:45.599 --> 00:56:48.880 Um so whenever I and it's also a memory for me as well. 00:56:49.039 --> 00:56:59.360 So whenever I see this, like let's say 20

years afterwards, and it's on YouTube, and I see they I was part of this conference, I've met X number of people over there and stuff like that. 00:56:59.599 --> 00:57:03.920 So uh yeah, I I just enjoy doing that. 00:57:04.880 --> 00:57:08.320 Yeah, the networking that you do over the years is super important. 00:57:08.559 --> 00:57:25.519 I mean, I just had uh I mean I just went to the list of the conferences that I was in the last 20 years, and that the amount of people that you know from really old conferences or people that you uh conference or people that become conference organizers as well and community organizers, it's it's it's really gold. 00:57:25.599 --> 00:57:34.320 It's so much worth so much worth your career as well to have these people to fall back on and to actually talk to in the near future and in the later future as well. 00:57:34.559 --> 00:57:49.679 So um it's going to a conference is not just there for the swag or so some talks, but actually to uh to to appreciate what's happening in the uh in the water cooler moments in between and in uh uh from crock to talk. 00:57:49.760 --> 00:58:01.280 And we keep saying that about our event, like we have a lot of talks, and you can it can be overwhelming that if you go to our event that you miss more than you see, but it's totally okay because all of them are filmed and all of them are gonna be recorded. 00:58:01.360 --> 00:58:07.760 So actually, uh fear of missing out can be a joy of missing out in our event because that's not the thing that you actually need to. 00:58:07.920 --> 00:58:11.360 That said, it's depressing for a speaker to be in an empty room. 00:58:11.519 --> 00:58:17.599 So it's from time to time it's a good thing to appreciate them as well, to go in there and see them coming. 00:58:17.840 --> 00:58:25.679 So uh I guess you you're coming to our event and you do your vlogging there as well, or is it uh are you gonna be freaking out because of your presentation? 00:58:26.159 --> 00:58:44.880 Uh yeah, I I I skip that part since I have a presentation, and sometimes I do feel like in in the sense of recording a conference, I do miss out some important uh conversations or you know, uh my mind goes in like what I should shoot, what should I do, just that, and everything else. 00:58:44.960 --> 00:58:55.599 And other and other and I miss out the part of networking sometimes, but most of the time when I'm speaker at some conference, I I'd I'd rather be a speaker and and I skip logging part. 00:58:55.760 --> 00:59:01.440 But I do create a short uh one, two minutes video of every conference I go to on my Instagram. 00:59:01.599 --> 00:59:05.840 Uh that's uh yeah. 00:59:06.000 --> 00:59:10.960 Well, you'll see me walking around with a camera, so no doubt I'll um I'll grab you for two minutes for an interview. 00:59:11.039 --> 00:59:12.960 So we'll we'll get something between us protesting. 00:59:13.119 --> 00:59:13.920 Don't you worry about that. 00:59:15.519 --> 00:59:19.440 And yeah, if you said like recording and dismissing it now, try organizing an event. 00:59:19.519 --> 00:59:29.599 I mean, the amount of stuff that I missed that happening, and it's great to see people like you doing the interviews and then hearing what people have been doing at the event that I had no time to even look at. 00:59:29.840 --> 00:59:37.519 I mean, for the last three events, I didn't even look at the uh at the booth of our partners because uh I just saw people walking around with like backpacks and stuff. 00:59:37.679 --> 00:59:39.199 Oh, where's that from? 00:59:39.440 --> 00:59:45.199 But uh I have had to moderate the main stage, so I had nothing, no time to do the any of these at all. 00:59:45.840 --> 00:59:54.239 So um, yeah, it's good to have people that actually record these things and bring them out there, and it's not just to show off, but actually just to share the experience of it. 00:59:54.400 --> 00:59:58.480 Yeah, thanks very much, and we see you next week and next Wednesday. 00:59:58.719 --> 01:00:02.079 And uh, we actually gonna see you in India very, very soon. 01:00:02.159 --> 01:00:04.079 So we have to get started packing. 01:00:04.400 --> 01:00:05.039 See you soon. 01:00:05.119 --> 01:00:05.760 Bye bye. 01:00:06.320 --> 01:00:07.760 Bye bye, take care.