Latest / Coffee with Developers / WeAreDevelopers LIVE - Read More, Prompt Less

Transcript

WeAreDevelopers LIVE - Read More, Prompt Less · Coffee with Developers

• 0:08
  SPEAKER_01: Hello and welcome to another We Are Developers Live.
• 0:10
  SPEAKER_01: We're here uh Wednesday again.
• 0:12
  SPEAKER_01: I'm technically on vacation, but I came in to help you out anyway.
• 0:16
  SPEAKER_01: So uh because it's fun to actually talk about some new things.
• 0:19
  SPEAKER_01: And we've got a guest who we've been who had before, who's always fun to have around.
• 0:24
  SPEAKER_01: So, Luis, how are you doing?
• 0:26
  SPEAKER_03: Doing fine.
• 0:28
  SPEAKER_03: Hi, you're welcome.
• 0:31
  SPEAKER_01: Stephen65 princess.com.
• 0:34
  SPEAKER_01: Like, yeah.
• 0:35
  SPEAKER_01: My you used you used everything in that name thing down there.
• 0:38
  SPEAKER_01: That's quite impressive.
• 0:39
  SPEAKER_03: So yeah, I mean, normally people just we have to beg them to put their job title in and things, and yeah, so it's yeah, I realized today that there was just like a character limit because I wanted to purchase like hashtag troublemaker or something like that, because official job title is chief troublemaker.
• 0:55
  SPEAKER_03: I I feel like that's accurate.
• 0:57
  SPEAKER_02: I remember you saying that when you like I didn't I didn't know that was actually official, right?
• 1:02
  SPEAKER_03: Like there's a few of those at Microsoft, like the global I mean, I am not at Microsoft, so whatever I feel like is official is official because I am the one who's officially saying that about myself with authority, you know.
• 1:16
  SPEAKER_01: Yeah.
• 1:17
  SPEAKER_01: I mean Mozilla had that.
• 1:18
  SPEAKER_01: We had like a fire starter kind of guy, like, and uh it's yeah, it's but I mean uh when I give advice to people about their career, it's always fun that startups give you these cool titles, but they mean nothing in an HR level sense.
• 1:32
  SPEAKER_01: That's the problem about that.
• 1:34
  SPEAKER_01: Like if you later on join another company and you want to negotiate a good salary as the first thing in your interview, you have to have a job title that's an official job title that's accustomed uh connected to an HR salary level.
• 1:48
  SPEAKER_01: So um it's always quite fun when you get a chance to call yourself the chief code monkey or whatever, but it's actually it can bite you in the back if you want to negotiate later on and actually have somebody else take in uh take it instead of you because they have their official title title on the uh on the CV that the AI scanner found immediately, kind of thing.
• 2:08
  SPEAKER_03: Yeah, so kind of the privilege is when you work as an independent and aim to do this for a lot longer, I guess.
• 2:16
  SPEAKER_01: Yeah.
• 2:17
  SPEAKER_01: So yeah, I was thinking about like having Luise here and basically uh uh being uh uh being German, another native German speaker.
• 2:26
  SPEAKER_01: I I thought about in uh uh throwing Dan uh a curveball by like by like bringing up a new topic here.
• 2:32
  SPEAKER_01: Uh we're gonna we're gonna have a new section called N versus German because uh we we we have uh this being an Austrian company, uh most people speak something resembling German.
• 2:44
  SPEAKER_01: Uh like uh and Dan is always confused when we have meetings, and then he comes in and I switch half halfway through the sentence from German to English.
• 2:53
  SPEAKER_01: And uh let's find a term every single time that we actually give Dan, because he was confused or he was amazed that German has these one sentence one word that actually mean whole sentences and very complex terms.
• 3:06
  SPEAKER_01: What was the one that you found out the other day about Dan?
• 3:10
  SPEAKER_02: Um well I heard about the person that can only use the toilet at home, which is a specific to German thing, I apparently that somebody uh is it uh I don't want to say the word on the stream, but um, if you know German, you can fill that one in.
• 3:22
  SPEAKER_02: The one I did like was is it um uh Uber Video Wachten, something like that, which I see it on like the uh the the trains, and that is a great example because it's like three sort of seem you know familiar words to me as an English speaker, like it's like saying super video watching, and to me it like perfectly sums up where I go, Oh yeah, that's like a sort of wonky translation of like this filming happening everywhere.
• 3:45
  SPEAKER_02: Um, but I love some of the the the really niche ones, Chris.
• 3:48
  SPEAKER_02: You can tell me some of those, but uh yeah, there's some clever ones.
• 3:50
  SPEAKER_01: Yeah, the one I got for you today is called Kumersbeck, and that actually means that's the weight that you gain when you're when life is not going well.
• 4:00
  SPEAKER_01: Uh so the literal translation would be grief bacon.
• 4:06
  SPEAKER_02: That's amazing! That's so see, that's the best bit, the literal translation.
• 4:10
  SPEAKER_02: Because that if you put that into into like a you know, imagine being in Britain and saying to somebody, oh, carrying a little bit of the grief bacon, you'd be like, come on, that is really harsh.
• 4:19
  SPEAKER_02: But I love that in German, it sounds sounds completely innocuous.
• 4:23
  SPEAKER_02: That's great.
• 4:25
  SPEAKER_03: Yes, I I love that a lot.
• 4:27
  SPEAKER_03: That's a good one.
• 4:29
  SPEAKER_01: Oh, brilliant.
• 4:30
  SPEAKER_01: Next week.
• 4:31
  SPEAKER_01: So we found one we found one each week and see what we can do with that.
• 4:35
  SPEAKER_01: I mean, I love that with English as well.
• 4:37
  SPEAKER_01: There's like uh there's uh a lady who's actually part of countdown of the TV show, and she writes books on uh on the meaning of words and the the uh where words come from.
• 4:48
  SPEAKER_01: And ironically, I forgot the word what words come from, but uh etymology.
• 4:52
  SPEAKER_01: And uh she actually uh she actually has words that are not available anymore, not used anymore, but are beautiful.
• 4:58
  SPEAKER_01: So she has all these like Georgian and Victorian English words that are absolutely wonderful to read as well.
• 5:04
  SPEAKER_01: So she's a great follower on X uh if you want to learn something about the English language from time to time.
• 5:10
  SPEAKER_01: So that's good too.
• 5:11
  SPEAKER_01: Yeah, we all hear about the education and about writing and stuff.
• 5:14
  SPEAKER_01: So I mean we're gonna come back to that later.
• 5:16
  SPEAKER_01: Louis Lise, you just wrote a book, right?
• 5:19
  SPEAKER_03: I did, in fact, yes.
• 5:21
  SPEAKER_03: Why would anyone do that?
• 5:27
  SPEAKER_03: Yeah, so um why did I write a book?
• 5:30
  SPEAKER_03: So um here's the book, it's uh from Prototypes to Production, it's a 90-day playbook for shipping enterprise AI.
• 5:38
  SPEAKER_03: Um and I wrote this because I saw a pattern um across all of my projects, and that were we have a lot of prototypes, a lot of proof of concepts, a gazillion of those, even, and nothing ever runs in process.
• 5:52
  SPEAKER_03: It is everything is an experiment, and we just need to try again and try again and try again, and nothing ever reaches real users, and that is because we don't have we don't have clean data, so the data is a mess, uh, the processes are a mess, and there is no owner and no budget and no funding and blaming culture everywhere, and frankly, a lot of fear as well.
• 6:17
  SPEAKER_03: And so they opt for another proof of concept and another proof of concept and another proof of concept, but they do not have any access strategy to either scale or pivot or stop that, and this way they cannot materialize any ROI because if you do not ship, you can't materialize that.
• 6:38
  SPEAKER_03: So um I developed a 90-day program for my customers, and then I wrote it down so because I don't scale, but the book could.
• 6:48
  SPEAKER_01: I mean, I think the biggest problem with that one is that it the fundamental change that people have, that that software is finished, which is something that I told people for decades as well.
• 6:57
  SPEAKER_01: But to me, the iteration was like building something, testing something, improving something, and then basically, when it's running, do the next iteration, do A-B testing, do more of security testing on it to find things out.
• 7:10
  SPEAKER_01: But now we're actually in a world where we ask an agent to build something, and if it's not good enough, just build a new one from scratch, and we don't even do any of these uh uh UI testing, user testing at all.
• 7:22
  SPEAKER_01: We we generate agents that simulate users that actually uh uh go through the thing, but it's uh the proper being sitting behind a mirror and looking at people failing at using your product is so incredibly insightful.
• 7:36
  SPEAKER_01: But people just seem to say it's too expensive, we don't do it any longer, we might as well just ship a new one and see if that one sticks.
• 7:43
  SPEAKER_03: Yeah, and I feel like this is just like so stupid, and I call this organizational amnesia because we established just like so many great patterns on how to make and ship good software over the course of decades, even, and now we just like collectively forgot about them and we're not um sticking to those anymore, and we do not measure anything anymore.
• 8:09
  SPEAKER_03: And I feel like, yes, use your agents if that makes you better by all means.
• 8:15
  SPEAKER_03: But what does better mean?
• 8:16
  SPEAKER_03: And we have metrics for that.
• 8:18
  SPEAKER_03: So, for example, in my project, I try to stick to the Dara framework and we establish a baseline and see so how are you doing, just like in all of the key metrics, for example, just like mean time to recovery or anything like that, um, just like without agents, and then we just look how is that just like with agents?
• 8:37
  SPEAKER_03: And if your metrics do not improve, then you held yourself just like very, very busy with some tech that you don't know how to use correctly.
• 8:48
  SPEAKER_03: Uh, because the metrics they don't care about how your code came to life.
• 8:54
  SPEAKER_03: It's not interested in that, it just like measures very, very observable things.
• 9:00
  SPEAKER_03: And if you get better with that, yeah, by then by all means, just like stick to the process.
• 9:06
  SPEAKER_03: We established that over years we know how to ship software in general.
• 9:12
  SPEAKER_03: That means yes, iteration, iteration, iteration, because software is done in the same way as the lawn is mode.
• 9:19
  SPEAKER_03: I did not say that.
• 9:21
  SPEAKER_03: Um by some dude.
• 9:24
  SPEAKER_03: Um but yeah, just like over and over and over again.
• 9:29
  SPEAKER_01: So now you published a book that people will download as a pirated EPUB and then ask an agent to summarize it for them instead of reading it.
• 9:37
  SPEAKER_03: Yep.
• 9:38
  SPEAKER_03: Let's do that.
• 9:41
  SPEAKER_03: Let's see how successful you will then be.
• 9:45
  SPEAKER_01: It is interesting though.
• 9:46
  SPEAKER_01: Like, I mean, when I wrote my books, it was basically for my parents, so they have something physical in their hand and know I have a job.
• 9:52
  SPEAKER_01: Uh and and uh I mean, at that time it was still not as common, like uh, and it fascinates me just how the quality of publishing books has gone down as well.
• 10:03
  SPEAKER_01: The amount of books that you see that have prompts in them that were obviously AI generated.
• 10:08
  SPEAKER_01: I mean, there's a few influences out there that release a book every two weeks, and you're like, no, it takes about seven months to write a book.
• 10:15
  SPEAKER_01: It doesn't like, I mean, the with all the reviewing, with all the text layouting, with all the things that you need to do.
• 10:22
  SPEAKER_01: And I hated my tech editor with a passion after the end of my pro uh book, but it made it a better book because those people are about to challenge you to say, like, explain this easier, make it easier, make it better, make it make it simpler, these kind of things.
• 10:35
  SPEAKER_03: Make it easier, make it simple, can you condense this?
• 10:39
  SPEAKER_03: I don't know, six thousand-word paragraph because you feel like you need to explain the entire world, just like in this book, you don't.
• 10:48
  SPEAKER_03: So please condense this, and then you just like sharpen your axe and sharpen your axe and sharpen your axe over and over and over again.
• 10:55
  SPEAKER_03: And yes, it takes time, and it also takes time.
• 10:59
  SPEAKER_03: So I think by writing and I think by talking.
• 11:02
  SPEAKER_03: So it's not I first think something and then I write it down, but just like the thoughts and the ideas develop as I'm writing them down, and that means I start a sentence and then I will delete a few words again, so x, x-base, and then just like write just like similar and then just like delete again, and then this way I just like iterate myself over, and then I read it again, and just like, oh fuck, that was bullshit.
• 11:27
  SPEAKER_03: So I will just like delete again and then just like rewrite it again, and then I get closer to what I actually wanted to say.
• 11:35
  SPEAKER_03: And of course, I experimented just like in blog posts, which is like, what if I just like gave a copilot, oh, insert your your favorite product here, just like some thoughts, and just like tell it, hey, make a blog post out of this.
• 11:49
  SPEAKER_03: This is my style, you go ahead.
• 11:52
  SPEAKER_03: And it wasn't me because I needed the writing experience to know how I feel about things and what I wanted to express.
• 12:00
  SPEAKER_03: Um, so that is just like kind of a journey to oneself as well, if you just are able to think through writing or drawing or talking or whatever is your mode of expression.
• 12:14
  SPEAKER_03: And that applies even to coding.
• 12:16
  SPEAKER_03: So, and of course, we all know coding agents can help with some tasks, but I actually enjoy writing the code.
• 12:24
  SPEAKER_03: So don't take this away from me.
• 12:26
  SPEAKER_03: I like to solve the puzzle.
• 12:28
  SPEAKER_03: I like to think, oh, I'm presented with this problem, and now I know five ways on how to approach this problem, and I know why four of those approaches will lead to later problems.
• 12:38
  SPEAKER_03: So I will stick to solution C.
• 12:40
  SPEAKER_03: Something like that.
• 12:42
  SPEAKER_03: And please don't take this away from me.
• 12:45
  SPEAKER_02: There's so much in there I just want to jump on, actually, because I thought there's so many interesting things.
• 12:48
  SPEAKER_02: I think um the first thing was that I don't know if it was like Stephen King or somebody that said like the best writers are actually the best editors, and it's that idea of like the first thing that you write is actually kind of irrelevant because it's what you said, it's about well, maybe I'm just putting something in a bullet point form because I'm gonna think it through in detail later, or I'm just trying to get the skeleton or whatever.
• 13:06
  SPEAKER_02: Um, but then when you put all of that back on to software, you know, what we're saying is that kind of most people really have forgotten that and they think that the generation part is actually the the editing part, all rolled into one.
• 13:17
  SPEAKER_02: And that, you know, almost by I was thinking some of the stuff you're saying then about user testing, even going right back to that and how that links, that this idea that we think that because the LLMs have got access to like all the information that's on online and it's the kind of average of human intelligence, that as long as the what we get, whether it be the book that they produce with just a prompt or the software that they produce that maybe is inaccessible, but the UI looks okay, like this idea that the average is sort of okay for most people, and that's all they're striving for, and that we know that there's major problems with that, right?
• 13:49
  SPEAKER_02: But but do you see what I mean?
• 13:51
  SPEAKER_02: This idea that for some people it's just about the like speed and the fact the thing exists when it didn't before, the quality is kind of this slopified world, right?
• 14:00
  SPEAKER_02: Um, that's slopify, not Shopify, but yeah, it's just really interesting.
• 14:04
  SPEAKER_02: Just good you're picking the opposite.
• 14:07
  SPEAKER_02: Stopify, yeah.
• 14:08
  SPEAKER_02: I think that's a great word.
• 14:09
  SPEAKER_02: I need to create, I'm gonna get that domain name now if I don't have it.
• 14:11
  SPEAKER_02: But yeah, I think this is something to be said for for people that are doing what you're doing, where you can use AI to help to be your sort of co-pilot, really, your collaborator.
• 14:19
  SPEAKER_02: Um, but then it's knowing where the limits are for it, isn't it?
• 14:22
  SPEAKER_02: That it's good that you're that you're kind of you know really pushing that.
• 14:25
  unknown: Yeah.
• 14:26
  SPEAKER_01: I mean, the the average of human intelligence sounds like a very scary thing to me.
• 14:30
  SPEAKER_01: I mean, that there's so many people that drag that down that average level that it's just incredible.
• 14:35
  SPEAKER_01: I mean, we we we also talk with people and they say, like, why'd you why don't you just talk to your computer and let it actually write the things for you?
• 14:42
  SPEAKER_01: And I I like the concept of writing because I think about while I'm writing it.
• 14:47
  SPEAKER_01: I say a lot of things, I don't think about them.
• 14:49
  SPEAKER_01: This live show is a great example.
• 14:51
  SPEAKER_01: Uh, but then actually, when I write something, I have I'm in a different frame of mind.
• 14:55
  SPEAKER_01: I'm actually writing and um thinking about it more because it is an extra effort to do it, much like people need some people need to note things down or scribble things down to remember them.
• 15:06
  SPEAKER_01: They need that physical extra thing, like uh uh haptic learners that actually have that kind of extra level.
• 15:12
  SPEAKER_01: And I think taking all of that away and going immediately to the final product, much like your book says as well, like we're getting the production, uh getting the thing into production, is something that makes sense to give yourself a bit of sense of uh uh sense of think it first, then write it down, kind of thing.
• 15:30
  SPEAKER_01: And I love seeing words appear when I write them.
• 15:33
  SPEAKER_01: It's just it feels like a very geeky thing to do, but it feels it feels uh wonderful to see a sentence.
• 15:38
  SPEAKER_01: I'm like, I can shorten that, I can do that, I can do that.
• 15:41
  SPEAKER_01: That grammar is a wonderful thing, and I think more people should do that.
• 15:45
  SPEAKER_01: And uh, but generally when we talk about book sales and stuff as well, it's pretty brutal to see how the uh the average reading level across every country on this planet is plummeting, and that's mostly because parents don't read to their kids any longer, which is super sad.
• 16:05
  SPEAKER_03: So I'm so my kids are grown up and nearly grown up, so my daughters are 16 and 21.
• 16:12
  SPEAKER_03: So um that and they are both readers because I read to them when they were small kids and not able to read for themselves, and I always encourage them to read and both read books and just like paper books, not even just like on the Kindle or anything like that, just like they read paper books every single week, not every day, unless it's for school or university, but they read, and literacy is just like such a powerful instrument, and we need to get back to this day where it's just like normal to read to kids, and it's normal just like to read with kids and uh for every adult as well.
• 16:55
  SPEAKER_03: And I feel like it it just like got almost that absurd that people read that we have this term, at least in Germany, Germany, it's just like yeah, it's just like performative reading, just like in public.
• 17:07
  SPEAKER_03: No, it is normal to read, but we just like feel this is just like so special because no one reads in public that everyone who does this and just like enjoys a little bit of fresh air on a park bench and is reading a book and enjoying themselves and minding their own business and working on themselves is just like now performative reader.
• 17:25
  SPEAKER_03: And it's just like, no, this is wrong.
• 17:27
  SPEAKER_03: This is just like a person who reads a book.
• 17:30
  SPEAKER_03: There is no performance in there.
• 17:33
  SPEAKER_02: Maybe it's the fact we like the kind of idea of the high-techness, you know, and that we all we automatically think that things are like less high-tech, old-fashioned, less sort of interesting in whatever way you want to think of it.
• 17:43
  SPEAKER_02: And that maybe it's just that that people are like, especially some younger generations perhaps will kind of have this thinking of like, oh, well, the information in the book must be out of date compared to what I can get online, you know, rather than it's probably more credible because of the process it's had to go through to get printed, rather than just anybody, you know, publishing whatever they want.
• 18:00
  SPEAKER_02: Um, and I think there's there's something to be said for that as well, right?
• 18:03
  SPEAKER_02: That maybe we're in like a transition point, a bit like the kind of um MP3 vinyl kind of cycle that it would go back to these old technologies because the thing that separates MP3 from vinyl is the quality, right?
• 18:16
  SPEAKER_02: It's the bitrate of the audio.
• 18:17
  SPEAKER_02: So you can hear things on a vinyl recording that you can't hear on an MP3, they're just not there.
• 18:22
  SPEAKER_02: And so there I think there might be something there with you know, this idea that uh programming and coding and just building things, uh for some people, some of the love of it will be will have been lost.
• 18:31
  SPEAKER_02: And the same for writing, like the love of writing or reading a book, and that maybe we will cycle back to some of these older ways of doing things, even if it's like a niche, uh, but just not performative.
• 18:41
  SPEAKER_03: But also it's just like it's a it's a cultural technique to just like be able to listen to vinyl, just like at home and just like putting that on the turn table and really just it's kind of like guarding the fire, you know.
• 18:54
  SPEAKER_03: So of course I have a Spotify account and just I I have just like a gazillion of songs on my phone and la la la la la.
• 19:00
  SPEAKER_03: But still, it is just like it is different when I put on a vinyl and just like listen to that because then I really listen to this, and when I'm writing, I'm really writing this, and to the books who just like carry all the old knowledge.
• 19:14
  SPEAKER_03: How did we get to the conclusion that up to date is always just like better?
• 19:19
  SPEAKER_03: Because we always say, just like, yeah, history repeats itself, and there's just like so much knowledge that the generation of our parents had, and the the generation before of that had, and even just like the one before them.
• 19:35
  SPEAKER_03: And all of these generations, they wrote things down because they had this cultural technique of just writing things down so that others could benefit from that.
• 19:45
  SPEAKER_03: And I feel we could learn a lot more if we just read more and write more and yeah, exposed ourselves more to just like what's already there in knowledge.
• 19:57
  SPEAKER_01: It goes back to that slow and fast thinking model that basically that like yeah, uh that that just read that one up.
• 20:04
  SPEAKER_01: That's an interesting one.
• 20:05
  SPEAKER_01: I think the other bit about vinyl is also uh that we listen to the whole album and we actually try to we found the weird songs that we didn't know about.
• 20:14
  SPEAKER_01: Whereas like with streaming, you listen to singles, you listen to two simple to one song at a time and make yourself a playlist of like random stuff.
• 20:22
  SPEAKER_01: I mean, Dan was in the car with me.
• 20:23
  SPEAKER_01: I've got this wonderful playlist of like 960 songs that are completely random, and it's uh uh and but at the same time, it's just I remember uh discovering uh albums that are basically making sense.
• 20:35
  SPEAKER_01: I mean, that there's like uh concept albums, the Warren song leads to another kind of thing that may be uh a bit too much, but it's interesting to see that as a band, you just basically it's much harder to release these days because you have to be a one-hit wonder and then become a second hit wander if you're lucky.
• 20:53
  SPEAKER_01: Uh, but your your audience you get through live concerts, you don't get them through streaming because nobody has any connection with you through streaming.
• 21:01
  SPEAKER_01: And I think it's the same about a book.
• 21:03
  SPEAKER_01: Buying a book is already a um uh dedication.
• 21:07
  SPEAKER_01: It's like you you actually say, like, okay, I want to learn this now.
• 21:09
  SPEAKER_01: No, I buy this book that only has this one thing as a topic.
• 21:13
  SPEAKER_01: I could read 10,000 blog posts, I could ask uh ChatGPT to summarize five books for me, but I still think the the extra level of learning will not will not go away.
• 21:24
  SPEAKER_01: And that's good because people uh should get money for writing, and uh it may it makes sense to actually publish those things still.
• 21:31
  SPEAKER_01: I find it harsh to see that some uh publishers just don't give a crap about quality anymore.
• 21:37
  SPEAKER_01: Like, I mean, when you see like prompts inside the text being printed just because they want to release something that week that was a cool feature last week, you cannot write a book that fast.
• 21:47
  SPEAKER_01: That makes no sense.
• 21:48
  SPEAKER_01: Like, you know.
• 21:50
  SPEAKER_03: Yeah, I know, I know.
• 21:51
  SPEAKER_03: But I I feel like it's um it's about time that we get back to quality and to to to to value people who.
• 22:00
  SPEAKER_03: That and yes, listening to an entire album is art because also there are tracks on the album that will never make it to Spotify or Apple Music.
• 22:10
  SPEAKER_03: Um, this is what my daughters recently only discover that usually on a vinyl there's tracks that are not will be found uh on streaming services.
• 22:20
  SPEAKER_03: So um we heavily invest in vinyl.
• 22:23
  SPEAKER_03: And um this goes out to all of the parents, just like make sure that your kids have access to vinyl antibiotic.
• 22:30
  SPEAKER_01: That would be CDs also had hidden CDs also had hidden tracks at what uh at some time as well.
• 22:35
  SPEAKER_01: So that could get you could find that was really cool.
• 22:38
  SPEAKER_01: And um, I mean that I mean, of course, you then get the people that overdo it.
• 22:42
  SPEAKER_01: I mean, remember like I've got audiophile friends that basically they warm up the cables to their turntable before they listen to the song, and you're like, yeah, okay.
• 22:50
  SPEAKER_01: I mean, it's that was especially hilarious when HTMI came out and people were like, Oh, I need a gold-plated HTMI cable, and you're like, it's digital.
• 22:58
  SPEAKER_01: It doesn't really matter what the quality of the HTMI cable is, it's either one or zero.
• 23:03
  SPEAKER_01: There's nothing in between.
• 23:06
  SPEAKER_01: For now, I can't do it.
• 23:08
  SPEAKER_01: Quantum computing.
• 23:09
  SPEAKER_01: Yeah, yeah, that's it.
• 23:12
  SPEAKER_03: I love this a lot.
• 23:14
  SPEAKER_01: Cool.
• 23:15
  SPEAKER_01: Let's take a look at some resources and stuff.
• 23:17
  SPEAKER_01: Uh, I mean, uh, let's skip them a bit because we already had too much time that we wasted on other bits.
• 23:22
  SPEAKER_01: And uh, yeah, cool.
• 23:24
  SPEAKER_01: So uh yeah, the first one is something that we were worried about earlier.
• 23:28
  SPEAKER_01: If we've got to talk about it too much, we cannot monetize this.
• 23:31
  SPEAKER_01: But a certain person has basically said that artificial intelligence is not to be used anymore.
• 23:36
  SPEAKER_01: It's now called super intelligence, and anyone else is the enemy of a certain country if they do doing that.
• 23:41
  SPEAKER_01: And now it's just like uh what happened?
• 23:44
  SPEAKER_01: What just happened?
• 23:45
  SPEAKER_01: Is it uh is like the 1990s comedy shows from England like brass eye?
• 23:50
  SPEAKER_01: Is that becoming our reality at the moment?
• 23:53
  SPEAKER_01: This is just bonkers.
• 23:55
  SPEAKER_01: Sean I have a I have a problem with uh super intelligence because it sounds just so lame, it just sounds like uh a badly translated manga, like you know, super intelligence force.
• 24:09
  SPEAKER_01: I'm not sure, you know.
• 24:12
  SPEAKER_01: Let's see too much.
• 24:14
  SPEAKER_01: Yeah, yeah.
• 24:15
  SPEAKER_03: I feel like we can we can just like scroll past this because it is yet another thing that yet another dude said, and we can just like skip past that.
• 24:26
  SPEAKER_01: Yeah, so Margaret Hamilton is dead, uh, and uh that did the rounds.
• 24:31
  SPEAKER_01: I mean, everybody who is something saw the picture of her with the printout of the code from the Apollo moon landing and stuff, and she was just an incredible, incredible person.
• 24:42
  SPEAKER_01: And if you haven't uh read up on her yet, it's so so good.
• 24:46
  SPEAKER_01: At the same time, it was very uh uh disappointing to hear that the Grace Hopper uh um uh like parties and uh and events and stuff are not happening this year because of funding being cut in this area.
• 25:00
  SPEAKER_01: But women in tech is a super important uh uh topic, and I think she's just a person that if you have uh kids that actually want to look up to somebody, that's a person to read up on.
• 25:12
  SPEAKER_01: I guess Louise, you you hammered that one into your daughter's heads already, right?
• 25:16
  SPEAKER_03: Uh I I I I did.
• 25:18
  SPEAKER_03: I I even have um Margaret Hamilton as a legal model, just like with all the code.
• 25:25
  SPEAKER_01: I love that.
• 25:25
  SPEAKER_01: Yeah, the minifig here, yeah.
• 25:27
  SPEAKER_03: Yeah.
• 25:28
  SPEAKER_04: Exactly that.
• 25:32
  SPEAKER_01: It's just really, really cool.
• 25:35
  SPEAKER_02: Um sorry, carry on.
• 25:39
  SPEAKER_03: It is super cool.
• 25:42
  SPEAKER_01: Also, in that vein, if you haven't seen Hidden Figures yet, uh the movie, uh that should be a mandatory viewing in every in every school, I think.
• 25:51
  SPEAKER_01: That's about the uh uh the the black mathematicians that actually helped NASA women uh mathematicians that helped NASA land on the moon.
• 25:59
  SPEAKER_01: And they listened to one of her since till 1997 because she was better than computers and more uh uh correct in the things that she was she was talking about, and it showed the discrepancy of like being uh an expert in math to help the uh the country reach the moon at the same time not being allowed to use the same toilet as the other people because of your skin color.
• 26:19
  SPEAKER_01: So it's uh it's an incredible, incredible movie, and I think it's something that we should advocate more for.
• 26:26
  unknown: Yeah.
• 26:27
  SPEAKER_01: Now, the big thing also that went around uh was this one uh where somebody used uh AI to reverse engineer basically the whole of the Adobe uh suite as open source and free uh tools and called them something something craft.
• 26:44
  SPEAKER_01: He's now working on the same for the for the um for the office suite.
• 26:49
  SPEAKER_01: And there's also a reverse engineering uh task uh tool that actually allows you to reverse engineer almost anything down to the level of uh down to the level of um binary code, finding out what's going on there and comparing it to others to other bits.
• 27:09
  SPEAKER_01: And I think this is cool.
• 27:10
  SPEAKER_01: I'm I think uh uh I I don't like monopolies.
• 27:15
  SPEAKER_01: I I think there was at one time where a certain company like Adobe uh ruled the market and could do whatever.
• 27:21
  SPEAKER_01: And this one shows that the uh that nowadays you can copy anything, but I think the I have opinions.
• 27:30
  SPEAKER_03: I I think you have opinions.
• 27:34
  SPEAKER_01: Wow.
• 27:35
  SPEAKER_03: Yeah, yeah, surprise, surprise Lisa has opinions.
• 27:37
  SPEAKER_03: So there's um so there's two things I need to address here.
• 27:41
  SPEAKER_03: So first is um, yes, at first glance, that looks of course impressive, like any prototype looks impressive.
• 27:51
  SPEAKER_03: That is just like the entire purpose of that looking impressive.
• 27:55
  SPEAKER_03: If you talk to actual designers and UX people who use the Adobe switches like on a daily basis, they will tell you that even the most basic features will not work at all.
• 28:11
  SPEAKER_03: Um and of course you could now say, yeah, well, then it's open source raising issue, just like someone will fix that power of open source community, yada yada yada.
• 28:20
  SPEAKER_03: But um we all know that this won't be happening at that scale because the suite is too big for that.
• 28:29
  SPEAKER_03: The second thing is speaking of the big suite, and that both applies to um to the entire Adobe suite as well as to Office or M365.
• 28:42
  SPEAKER_03: It is not about the app level.
• 28:45
  SPEAKER_03: So, yes, of course, you can just like create an Excel or Word or PowerPoint or Outlook clone and just like and self-host this just like so that you're able to um to save your files.
• 28:59
  SPEAKER_03: Of course, you could do that, or you could use LibreOffice or whatever.
• 29:03
  SPEAKER_03: But we're not interested in the UI per se, we're interested in the platform layer, and the platform will then just like make it work for everyone, because of course, as much as I need to criticize the hyperscalers for just like being monopolists, and I see this as like politically very, very critical.
• 29:28
  SPEAKER_03: I dislike this, but there is no alternative right now to that because no one has a platform, everyone focuses on this on this app level where you do not really have a benefit.
• 29:41
  SPEAKER_03: So, yes, if it's just like you or you and your family, or you and the three mates that you're just like founding a startup with, yes, then it's fine.
• 29:53
  SPEAKER_03: But if you need to power an enterprise with that, or even just like a mid-sized company with 210 employees, you are screwed with um something like that.
• 30:04
  SPEAKER_03: Because then you need governance, you need compliance, you need policies, you need all of the things that just like come with an integrated platform that we do not have in just like pure source products.
• 30:16
  SPEAKER_01: And that applies to I think the other side of it is like just copying the code and making the code work is like uh what you're not doing is actually giving uh having done all the research that went into it.
• 30:29
  SPEAKER_01: I mean, there's reasons why Photoshop has certain features, there's reasons why Word has certain features.
• 30:35
  SPEAKER_01: Sadly enough, we could never take it out again.
• 30:37
  SPEAKER_01: That's why when you release things in an enterprise company, like when I work for Microsoft, for example, it's like every feature that you put in has to be 100% sure that you really want to put it in there.
• 30:48
  SPEAKER_01: People always want new features, but you can as an enterprise, you cannot take out a feature from a product because people will go, uh, there will be like torches and things in front of your house.
• 30:58
  SPEAKER_01: Like it's uh people are very excited about their features that they got used to.
• 31:03
  SPEAKER_01: So uh of course, it's it's easy it's easy to copy.
• 31:06
  SPEAKER_01: Uh well, it's not easy, but it's now the tools allow us to make it easy to copy these things uh bit by bit.
• 31:12
  SPEAKER_01: But the research and the uh the knowledge of the people who build it is not part of this.
• 31:18
  SPEAKER_01: So uh that's why I always found it weird when people do a clone one-on-one of something and then call it like, oh, that's that's as good.
• 31:24
  SPEAKER_01: And you're like, no, actually, you should have done something different with it.
• 31:28
  SPEAKER_01: I mean, the other thing I tried to reach out to him, but he didn't have time uh so far.
• 31:32
  SPEAKER_01: I mean, there's not photos, photo p.
• 31:37
  SPEAKER_01: So photo p is another one that I use instead of Adobe Photoshop because I'm not uh I'm paying for it$50 a month, uh$50 a year, and it actually is uh uh it has uh it is ad-driven if you want to use it for free as well.
• 31:50
  SPEAKER_01: And it's basically Photoshop in the browser and it does everything that you want it to do.
• 31:55
  SPEAKER_01: Now that person made about three million a year with that tool so far, and now he's actually being attacked because not attacked, but he actually feel uh sees that there are thousands of clones showing up from Photope where people just reverse engineered his clone of Adobe, basically, and uh and he uh he complained to the companies hosting those things, and basically it's not even that they don't take them down, he gets like pull requests from clones on his main repository, like uh fix that thing that's broken in that pirated version of your tool.
• 32:32
  SPEAKER_01: So I think this is an interesting one that like SAS, uh a software as a service, is facing a very big threat of uh of being able to uh to just clone things and to reverse engineer things.
• 32:44
  SPEAKER_01: That has always been the case in the game uh development uh world, but now it's coming for like uh for like proper suites as well.
• 32:51
  SPEAKER_01: It's interesting to see.
• 32:52
  SPEAKER_01: Yeah.
• 32:52
  SPEAKER_03: And now that you say that, there is a fourth issue.
• 32:55
  SPEAKER_03: So I raised two, you added another one, and now I have another one as well, with with which is uh the fourth issue with that.
• 33:02
  SPEAKER_03: So if we're just like all the time, yeah, but big AI is just like stealing our data and stealing our ingenuity and our books and our code, and now they just like sell this back to us for uh$20 uh subscription, then how ethical is it now to just like steal the work of just like thousands of developers who created the Adobe Suite, who just like created the M365 Suite and any other um application to then just like reverse engineer it, which basically means I will steal the work and then I will just like claim I did this, and even if I open source this, who gave you permission to be just like modern Robin Hood for tech?
• 33:45
  SPEAKER_03: It's just like so.
• 33:46
  SPEAKER_03: This is this is also an ethical problem now.
• 33:50
  SPEAKER_01: I think just uh it gets ironic when people start monetizing the clones.
• 33:55
  SPEAKER_01: That's I've seen as well.
• 33:57
  SPEAKER_01: Like basically, like it's fine to say, like, I'm gonna pirate it and uh and open source it.
• 34:02
  SPEAKER_01: I mean, that's what we've done in the piracy scene, uh, or what people have done in the piracy scene who would dabble in these kind of illegal things uh uh for years.
• 34:10
  SPEAKER_01: But basically the the but then actually going back and making money with it or selling it to other people, this is where it gets really frustrating and really weird because like, yeah, you've you're not you're not Robin Hood in this case, you're just a a thief.
• 34:25
  SPEAKER_02: Yeah.
• 34:26
  SPEAKER_02: I wonder as well if there's something to be said for the fact that a lot of these features, you know, not only are we stealing from the developers, right?
• 34:32
  SPEAKER_02: But we're actually kind of uh I think we don't um ex uh understand how complicated some of the seemingly simple features are to actually get into this software.
• 34:40
  SPEAKER_02: And I've got a story of once I'd reached out, I think I just posted about um Adobe XD, which is their kind of like um UI, you know, it's the Figma equivalent within the Adobe um kind of suite.
• 34:51
  SPEAKER_02: And there was a time where I was just sort of using it because I thought, well, it makes sense to kind of have it all within that suite because you can share files between the different software more easily.
• 34:59
  SPEAKER_02: So I was just seeing if it was like a viable alternative to Figma.
• 35:02
  SPEAKER_02: And I noticed when I was on um um on Adobe XD, two really simple or seemingly simple features were missing.
• 35:08
  SPEAKER_02: One was the ability to add rulers, so I couldn't actually pick like check if anything was pixel perfect or where it was positioned.
• 35:15
  SPEAKER_02: Um, and the second one was um color spaces or like color profiles.
• 35:19
  SPEAKER_02: Uh so it meant that if I exported something from there, all the greens, let's say, would be like really vibrant or really over the top, and it and I couldn't change the color space or the color profile.
• 35:28
  SPEAKER_02: And so I posted about it on X saying, how can we be in like 2024, let's say, at that point, and we don't have these features, uh, they're really simple features or something like that.
• 35:37
  SPEAKER_02: And one of the product managers actually responded with like very politely, but they just said something like the features themselves are simple to use and they might not seem to do a lot, but getting them in there is a nightmare.
• 35:47
  SPEAKER_02: Like building these things, you know.
• 35:49
  SPEAKER_02: And um, so I think there's just something to be said for we're not just stealing the F the actual work of those developers, but it's like the uh the the the number of features that will just be actually be missing if we just clone something that isn't perfect.
• 36:02
  SPEAKER_02: Uh but also we're just skipping over the fact that they've probably put a huge amount of work into some things that we just overlook.
• 36:16
  SPEAKER_01: Because yeah, you cannot you cannot do messy things and whole people don't find them.
• 36:21
  SPEAKER_01: Like you're basically there's a lot of like undocumented code in closed source that basically would never survive any pull request in the open web.
• 36:30
  SPEAKER_01: So that's the kind of thing that that is the the tricky bit there.
• 36:33
  SPEAKER_01: I mean, is it the end of closed source?
• 36:36
  SPEAKER_01: That's one thing, of course, that's another problem.
• 36:38
  SPEAKER_01: But then you have the issues with Proto P, for example, that if you're open source, that people just clone your stuff anyway, and then you get even more requests.
• 36:46
  SPEAKER_01: Or when we had uh Daniel Stenberg here on the show, when he says like he closed all resources because he gets like 6,000 pull requests for curl a day that are AI generated nowadays.
• 36:57
  SPEAKER_01: So that's like moving uh uh dark matter uh code into the open source world.
• 37:04
  SPEAKER_01: This is force brute forcing it, but for a company, it also means a cultural change, and that is not gonna happen in the Adobe's and the Microsoft anytime soon when it comes to these products, I think.
• 37:15
  SPEAKER_03: Yeah, I great.
• 37:19
  SPEAKER_01: Cool.
• 37:20
  SPEAKER_01: Um okay, let's go to some weird stuff.
• 37:22
  SPEAKER_01: I love this one.
• 37:23
  SPEAKER_01: Uh, the DNS Museum.
• 37:25
  SPEAKER_01: It actually shows you all the things that people have done with DNS over the past, and it's just mental.
• 37:33
  SPEAKER_04: Oh yeah.
• 37:34
  SPEAKER_01: I mean, that's right down your alley, isn't it?
• 37:37
  SPEAKER_02: Like, I mean, yeah, I just it's the same thing I say every time, you know, in that it's just like I see that and I say to myself, developers, you know, because it it's I just imagine that the moment when you know that thing of like when you sort of uh that the bubble is popped and you suddenly realize that you're also as be as well as being a nerdy developer doing your nerdy developer things, you're like a normal person in the real world.
• 37:57
  SPEAKER_02: And somebody says, What is it you're up to, or what have you been working on?
• 37:59
  SPEAKER_02: And then you show them this, and they have that mixed reaction of like, I don't really understand it, but I'm sort of impressed, but I don't know if I'm supposed to understand it.
• 38:08
  SPEAKER_02: And I think a project like this, I wouldn't even have a have a clue how to explain to somebody what I've been doing with my time, but I love it anyway.
• 38:17
  SPEAKER_03: Exactly that.
• 38:17
  SPEAKER_03: It's just like normal people wouldn't even understand.
• 38:22
  SPEAKER_03: It's just and also and we can't really explain, but yeah, it's just like it it it I like those websites a lot.
• 38:31
  SPEAKER_02: I feel it's a little bit like when um if if you said to like a young child, like, why did you put your finger in the in the electrical socket?
• 38:37
  SPEAKER_02: And they would say, like, because I could, and it's kind of that, you know, it's just like this is what we're doing.
• 38:42
  SPEAKER_03: The best reason because I can't.
• 38:44
  SPEAKER_03: It's just like why did she do it because I can't.
• 38:47
  SPEAKER_01: This is this is great.
• 38:49
  SPEAKER_01: It's the lyrics of the bad horse song from uh Dr.
• 38:52
  SPEAKER_01: Horrible's sing-along uh blog as t as a DNS lookup as a trace route.
• 38:57
  SPEAKER_01: It's just so good.
• 39:00
  SPEAKER_01: Anyways, people do that, and uh uh I mean I just I was just offline for two days uh for a day with my christianheiman.com because uh my um my SSL uh uh certificate uh that I bought could not be installed with the hoster that I have at the moment.
• 39:17
  SPEAKER_01: And yeah, I was like two and a half hours on the phone with people in two different countries, and now it's fixed more or less.
• 39:24
  SPEAKER_01: Uh, but it just it just drives me nuts how some some of these things that are very, very cool and geeky, but others are just basically unnecessary any longer.
• 39:34
  SPEAKER_01: Like nowadays, buying or sell buying a certificate to SSA uh to HTTPS your website and installing it by hand, that's something that Cloudflare and Let's Encrypt and others have been have been solving for the last 10 years.
• 39:47
  SPEAKER_01: Yeah, that's just basically making money nowadays by hosters that don't make uh that that are not there for the users, but they're just there to actually sell you more things.
• 39:56
  SPEAKER_01: And you realize that when you try to debug it, it's a situation literally.
• 40:03
  SPEAKER_01: It was brutal.
• 40:03
  SPEAKER_01: It was just like I was just thinking this is really disappointing.
• 40:06
  SPEAKER_01: And I I'm gonna write about this.
• 40:09
  SPEAKER_01: There were so many good hosters in the past that were geeks for geeks, like Dreamhost and uh uh Media Temple, and a lot of people that I know in WebDE back then in Germany, like even one-on-one, like people like very obvious uh uh providers, zen.co.uk, my first provider in England, they were basically nerds that helped me out to do my nerdy stuff and not just like people who sold you hosting packages.
• 40:32
  SPEAKER_01: And that kind of market has really, really gone down a hill.
• 40:39
  SPEAKER_01: It's been going around for a while.
• 40:40
  SPEAKER_01: Uh uh the uh destroy any website, which basically allows you to shoot up websites and if it loads, I don't know what's going on there.
• 40:50
  SPEAKER_01: But yeah, it's like basically a little slide where you can shoot the website and it's actually quite fun.
• 40:56
  SPEAKER_01: Oh, it's great.
• 40:57
  SPEAKER_01: So you can do Wikipedia.
• 41:00
  SPEAKER_01: Yeah, it's it's it's quite uh it's so much fun there.
• 41:07
  SPEAKER_02: It doesn't work here.
• 41:08
  SPEAKER_02: If it does work, change change to the second weapon.
• 41:11
  SPEAKER_03: Yeah, but can we please protect Wikipedia and open knowledge?
• 41:14
  SPEAKER_03: Thank you.
• 41:14
  SPEAKER_02: Yeah, that's true.
• 41:15
  SPEAKER_02: Yeah, it's a bit of an innocent sort of bystander, isn't it?
• 41:18
  SPEAKER_02: And we're just beating up the Stickman Wikipedia page.
• 41:20
  SPEAKER_01: Yeah.
• 41:22
  SPEAKER_01: But I mean it's it's fun to see, and I think that's basically where where this like uh Canvas in HTML came in as well, because that one would be much harder to do in the past, but now we have that kind of stuff.
• 41:34
  SPEAKER_01: This is also great, then that's probably down your alley as well, that English to Claudish translator.
• 41:39
  SPEAKER_02: Yeah, yeah, it's actually um really good.
• 41:41
  SPEAKER_02: If you go to the demos page, I don't know if you've uh got that there, we're quite behind, I know, with the stream.
• 41:46
  SPEAKER_02: There we are.
• 41:46
  SPEAKER_02: Um because at first I thought, you know, I couldn't tell quite what it was.
• 41:50
  SPEAKER_02: It's one of those things where it's sort of like, is this just something that's going to translate what I'm saying into uh kind of clawed, um, like a clawed-ish type response?
• 41:58
  SPEAKER_02: But actually, some of these examples I thought were quite where are we?
• 42:01
  SPEAKER_02: Can we see them?
• 42:02
  SPEAKER_02: I don't know if they can't load for us, nothing's loading there.
• 42:04
  SPEAKER_01: Yeah, the release can go out after Alice approves the final report is English, and Alice final report approval is the hard gate here.
• 42:11
  SPEAKER_01: The release can land after that, uh, is basically the claudish.
• 42:16
  SPEAKER_02: It's how I was I mean, I was just on the line of knowing uh um it's like a we've said this before, there's a few things you're never quite sure are satire, and then you kind of keep going and you you're not quite sure if it's got like a real useful, you know, um, because it's sort of like at first I thought, okay, maybe it's some sort of token reduction type thing, and it's just gonna sort of but but it it walks that line where it did take me a while to be 100% sure if it was satire or not, really.
• 42:46
  SPEAKER_01: Well, it's the it's the the kind of show-off y language that you get a lot from a lot of tech people that especially when they get into public speaking and they think they have to put like lots of everything in like uh uh um like passive voice, first of all, and then like semicolons in there somewhere, and it basically feels like uh this this looks to me like a comparison of a for loop and a while loop.
• 43:09
  SPEAKER_01: A while loop is really easily readable by every human person.
• 43:13
  SPEAKER_01: A for loop needs three variables to take part uh to take into consideration to understand what's going on in the for loop, and that's the same thing there that the upper one is a while, the lower one is the four.
• 43:24
  SPEAKER_01: At the same time, we always start with a four when we teach people coding.
• 43:29
  SPEAKER_03: That is a lovely comparison.
• 43:31
  SPEAKER_03: I will steal that from your I will credit you though, but just like this is this is amazing.
• 43:36
  SPEAKER_03: I like this.
• 43:39
  SPEAKER_01: And we got Doom in the database, of course we do.
• 43:42
  SPEAKER_01: Uh I mean Doom is in everything now.
• 43:44
  SPEAKER_01: I'm I don't have time at the moment.
• 43:50
  SPEAKER_01: I don't have time at the moment, but I wanted to use Claw to actually create wood.
• 43:54
  SPEAKER_01: You know, like if you put the Doom logo around, it says wood, so it would be just basically a 3D environment run.
• 44:00
  SPEAKER_01: Around a forest and cutting down trees with a chainsaw.
• 44:03
  SPEAKER_01: So that would be a fun project to build and show and show to John Romero at the next Weird Developers World Congress, which is coming back to Berlin and is coming to India in a few weeks.
• 44:15
  SPEAKER_01: So hopefully we will have something in there as well.
• 44:18
  SPEAKER_01: Now, in terms of tooling, I don't have any pre-prepared this week, but I found this one interesting.
• 44:23
  SPEAKER_01: I mean, Dan, you probably would be the person to say more about this.
• 44:26
  SPEAKER_01: So this is Shotcraft.
• 44:28
  SPEAKER_01: Uh it's a video scope for Cloud Code and Codex that actually allows you to do videos that uh with remotion that are as good as thing people that we normally pay for, like Dan.
• 44:41
  SPEAKER_01: What do you think of that thing?
• 44:43
  SPEAKER_02: Do you know what?
• 44:44
  SPEAKER_02: It's it's it's kind of timed actually, well timed, because I've actually just finished writing an article about how you can do a much lower kind of tech version of this, really.
• 44:52
  SPEAKER_02: And the reasons that I've kind of written that article is that I'd noticed that if you um if you were to build things in like crawled code, it will often over-engineer.
• 44:59
  SPEAKER_02: And the first thing it reaches for is remotion.
• 45:01
  SPEAKER_02: It reaches for uh next as a framework, it gives you remotion, it installs loads of MP uh node node modules that you don't really need.
• 45:09
  SPEAKER_02: And uh all of that really is in service of you seeing what video you're making.
• 45:14
  SPEAKER_02: That's probably going to be a description of like a prompt, you know, or a prompt that you're giving it in terms of what you want the video to do.
• 45:20
  SPEAKER_02: So I think for this, it's kind of largely like pre pre-made things that remotion, like uh templates that remotion can use, and then it's got variables in there like numbers and kind of you know, insert image here.
• 45:32
  SPEAKER_02: And so you put the kind of images in the correct place, like you would a website, really, and then it will kind of uh uh pick them all up and put them into the into the template.
• 45:39
  SPEAKER_02: Um, my thinking actually is that I found that you can do some of this in a kind of not very sexy way.
• 45:46
  SPEAKER_02: And what I mean is um when the early chatbots were kind of starting to start to do image generation, one of the big things I found for kind of getting consistent styling is that when you're deciding on that sort of design style, whatever you want to call it, design library, you know, whatever, um doing it in JSON format was like a really, really good way of me making sure that I would get this standardized styling, use of fonts, spacing, you know, conning on the on the lettering, all sorts.
• 46:11
  SPEAKER_02: Um, and then actually that is way, way uh less heavy in terms of your token usage and actually the production of it, because I'm saying to it, here's your JSON, here is the uh um another JSON file that's or no, it's markdown, sorry, saying here's the things I want you to produce, the assets.
• 46:28
  SPEAKER_02: And then it's essentially a few a few really simple steps that I'm going through.
• 46:32
  SPEAKER_02: Now the difference is um when I do it, it just makes it with HTML because we all know that Claude has that tendency to just like build whatever you've asked or spoken to it about with HTML straight away.
• 46:43
  SPEAKER_02: So I kind of cut out the whole need of like remotion because I thought, well, just use HTML and CSS and then export it with FFmpeg.
• 46:50
  SPEAKER_02: Now, this is uh putting all that in context, this is a really, really good library, but just be aware that there's a hell of a lot of stuff in there that you probably won't need.
• 47:00
  SPEAKER_02: And I would say that you can do some of this yourself, um, and then at least you're gonna have a little bit more of that control over it yourself.
• 47:08
  SPEAKER_02: Um, and I don't even know if you need to use remotion.
• 47:11
  SPEAKER_02: So I would say this will be really good for like high-end product display videos, really polished.
• 47:17
  SPEAKER_02: It's got a very cinematic look to it.
• 47:19
  SPEAKER_02: Um, but for most things that we'll need it for, what I was needing it for, where I wanted to make some overlays that animated, and I could put them over a video, it it it was much, much simpler to take a kind of um no library approach, to be honest.
• 47:32
  SPEAKER_02: But it's cool.
• 47:33
  SPEAKER_01: No, that would be that would be that would be our call now to reverse engineer that thing and actually turn it into a lightweight one that only does HTML and CSS.
• 47:42
  SPEAKER_02: Interesting.
• 47:43
  SPEAKER_02: Uh that's my week sorted.
• 47:46
  SPEAKER_01: We don't even need to steal it because it's Apache 2 license, so that's good.
• 47:49
  SPEAKER_01: You cannot steal something that's free, so that's kind of cool.
• 47:52
  SPEAKER_01: But yeah, it's uh it's interesting though that like I'm also worried about like when you get these things that they will get overused and you will get the same looking intro sequences uh all over the shop.
• 48:04
  SPEAKER_01: I mean, there's already an AI uh um design language that we're getting tired of.
• 48:10
  SPEAKER_01: Uh so it might be that this is happening in the video space soon as well.
• 48:15
  SPEAKER_03: Yeah.
• 48:16
  SPEAKER_01: Cool.
• 48:17
  SPEAKER_01: So let's get uh let's challenge Louise a bit and actually play a game called uh fake on news where you have to tell us uh we show you headlines and you have to tell us which ones we made up and which ones are actually uh real headlines that happened that we happened to stumble upon when we looked up things for the Dev Digest.
• 48:35
  SPEAKER_03: Okay.
• 48:51
  SPEAKER_02: Fantastic.
• 48:56
  SPEAKER_02: Fair enough.
• 48:59
  SPEAKER_01: Cool.
• 49:02
  SPEAKER_01: Poetry is the new AI security thread as poet Poe LLM malware infects 3K plus servers.
• 49:11
  SPEAKER_03: I hope this is fake.
• 49:14
  SPEAKER_02: Yeah, I would hope so.
• 49:15
  SPEAKER_02: It's a good name actually.
• 49:16
  SPEAKER_02: Poe LLM, Poe L L M.
• 49:18
  SPEAKER_02: Not bad.
• 49:19
  SPEAKER_03: I I hope this is fake.
• 49:20
  SPEAKER_02: Okay, we'll go for fake.
• 49:24
  SPEAKER_01: It's actually news.
• 49:25
  SPEAKER_01: I mean, um it is my username, like with an 8 at the end, and the amount of people who never read that properly is just fascinating as well.
• 49:34
  unknown: I see.
• 49:36
  SPEAKER_01: OpenAI's revenue is reportedly 20 billion less than previously projected.
• 49:43
  SPEAKER_03: I can imagine that this is um actually true.
• 49:46
  SPEAKER_03: So this is yeah news.
• 49:48
  SPEAKER_01: I think so.
• 49:54
  SPEAKER_01: 20 billion.
• 49:55
  SPEAKER_01: That's quite a that's quite a number to try to hide.
• 49:57
  SPEAKER_01: It's quite amazing.
• 49:59
  SPEAKER_01: Builders are using reverse engineering to fix window security issues before Microsoft.
• 50:04
  SPEAKER_03: Totally true.
• 50:05
  SPEAKER_03: I I guess that's true, unfortunately.
• 50:08
  SPEAKER_02: I think so.
• 50:09
  SPEAKER_02: There's a lot of reverse engineering on today's episode, so I think I think so.
• 50:15
  SPEAKER_01: Which is why I put it in as a fake.
• 50:17
  SPEAKER_01: Very good.
• 50:18
  SPEAKER_01: Red Herring.
• 50:21
  SPEAKER_04: Yes.
• 50:23
  SPEAKER_01: After renaming artificial intelligence to superintelligence, the White House plans to replace cloud with sky computing as the future is satellites.
• 50:33
  SPEAKER_02: Ah, I think she's on to you, Chris.
• 50:36
  SPEAKER_02: I think she knows that I I feel that's fake as well, and I think that's another uh red herring.
• 50:41
  SPEAKER_02: It's fake nicely done.
• 50:45
  SPEAKER_01: Number five.
• 50:50
  SPEAKER_01: Anthropic band's cruel behavior against his clothed AI.
• 50:57
  SPEAKER_01: Protector.
• 51:06
  SPEAKER_00: Tesla to rename self-driving to aid at driving in EU to comply with new law.
• 51:15
  SPEAKER_03: This could be true.
• 51:16
  SPEAKER_03: So this could be news.
• 51:19
  SPEAKER_01: Yeah, I haven't heard about it.
• 51:23
  SPEAKER_01: No, they have other things they have to do for Europe, uh, but they I think they don't call it self-driving at all anymore.
• 51:31
  SPEAKER_00: Number seven transmit data using high-latency molecular vibration technology.
• 51:40
  SPEAKER_03: Yes.
• 51:42
  SPEAKER_02: That's one of these ones that you know it's a bit too early for me to even figure out what that means.
• 51:46
  SPEAKER_01: Uh yeah, it's news, it's actually vibration of uh of materials to send data to other things.
• 51:54
  SPEAKER_01: You've seen it in the past where people saw like uh videos of like potted plants and how they moved and actually realized what people said around them.
• 52:02
  SPEAKER_01: An application in Lisp to grow, you grow by talking to it.
• 52:07
  SPEAKER_03: News.
• 52:10
  SPEAKER_02: It feels like it's sort of too perfectly written, the fact that it's like in Lisp as well, but then uh yeah, I'm gonna say news.
• 52:17
  SPEAKER_02: Let's have a look.
• 52:20
  SPEAKER_01: It is news.
• 52:21
  SPEAKER_01: Feels weird that uh that they talk to something called Lisp, right?
• 52:24
  SPEAKER_01: That's just really wrong.
• 52:25
  SPEAKER_01: But at the same time, whoever put that S in Lisp is a really terrible person as well.
• 52:30
  SPEAKER_01: Yeah, yeah, true.
• 52:32
  SPEAKER_01: Number nine.
• 52:32
  SPEAKER_01: Yes.
• 52:33
  SPEAKER_03: We can't be friends.
• 52:35
  SPEAKER_01: Taiwanese company unveils batteries that charge themselves with device movement.
• 52:41
  SPEAKER_01: Surely we would have heard of that.
• 52:43
  SPEAKER_03: Yeah, that yeah, works.
• 52:45
  SPEAKER_03: I mean it works for a watch as well, so yes, news.
• 52:57
  SPEAKER_02: We're doing so well.
• 52:58
  SPEAKER_01: Number 10.
• 53:01
  SPEAKER_01: Scoop, the real reason AI companies pursue medical model conscience is to get tax exemption.
• 53:10
  SPEAKER_02: Um I I feel that it is.
• 53:16
  SPEAKER_02: I think the also the the scoop thing, I don't know, it feels like it's uh trying to get us to think it's from a real story.
• 53:22
  SPEAKER_04: Yes, yes.
• 53:25
  SPEAKER_01: It is fake.
• 53:26
  SPEAKER_01: But I have a lot of scoop headlines lately.
• 53:28
  SPEAKER_01: That seems to be a new thing a lot of outlets are doing.
• 53:31
  SPEAKER_03: Okay.
• 53:32
  SPEAKER_01: Interesting.
• 53:32
  SPEAKER_01: Number 11.
• 53:35
  SPEAKER_01: Superintelligence AI rebrand causes unprecedented demand for Slovenian website names because the domain is.si.
• 53:44
  SPEAKER_03: Yeah.
• 53:45
  SPEAKER_01: I've I've I've heard this in the news as well.
• 53:47
  SPEAKER_02: Have you seen it, Luis?
• 53:49
  SPEAKER_03: Uh I did not because I was not being I I I can't consume news on a daily basis.
• 53:54
  SPEAKER_03: Uh it's too depressing for me.
• 53:56
  SPEAKER_03: But I I feel that could be true.
• 53:58
  SPEAKER_01: Yeah.
• 53:59
  SPEAKER_01: Mistake that it is true.
• 54:02
  SPEAKER_01: Makes you wonder if uh if the wife of a certain person who is from Slovenia has a has a gain in that.
• 54:10
  SPEAKER_01: Interesting.
• 54:11
  SPEAKER_01: That would be a conspiracy.
• 54:12
  SPEAKER_01: Who knows?
• 54:12
  SPEAKER_01: Number 12.
• 54:13
  SPEAKER_01: Uh there you go.
• 54:15
  SPEAKER_01: Retailers are pulling back from self-checkout after years of theft errors and questionable time savings.
• 54:21
  SPEAKER_03: Fake.
• 54:22
  SPEAKER_01: Yeah, I don't see them doing that.
• 54:23
  SPEAKER_01: Have you?
• 54:24
  SPEAKER_03: So there is a lot of errors, but they will not be uh pulling back.
• 54:29
  SPEAKER_02: Nicely done.
• 54:30
  SPEAKER_02: Let's have a look.
• 54:36
  SPEAKER_01: Actually, they do.
• 54:37
  SPEAKER_01: That a big retailer, I think, in the Netherlands, has done it.
• 54:41
  SPEAKER_01: I mean, it's weird, like every supermarket you see people being confused about these things.
• 54:45
  SPEAKER_01: Yeah, so how we stand.
• 54:47
  SPEAKER_01: Let's have a look on the leaderboard.
• 54:51
  SPEAKER_02: Oh, fourth, joint third, though, actually.
• 54:54
  SPEAKER_02: In fact, nope, if you follow all the way up.
• 54:56
  SPEAKER_02: Oh no, no, it is fourth because uh Brahmas has got a higher streak, but still fourth.
• 55:01
  SPEAKER_03: That's not bad.
• 55:02
  SPEAKER_03: That's fine.
• 55:02
  SPEAKER_03: That's fine.
• 55:04
  SPEAKER_02: Good show.
• 55:05
  SPEAKER_01: Not bad.
• 55:06
  SPEAKER_01: It's a weird world, isn't it?
• 55:08
  SPEAKER_01: Like news are not uh really interesting to see these days.
• 55:12
  SPEAKER_01: Cool.
• 55:13
  SPEAKER_01: So yeah, uh, tell us about what you how what fame and fortune is now awaiting you with your book sale.
• 55:19
  SPEAKER_01: Are you going for like poetry readings?
• 55:21
  SPEAKER_01: Are you are you getting more people involved?
• 55:24
  SPEAKER_01: Like, what was the first feedback?
• 55:25
  SPEAKER_01: I mean, it's your first book, right?
• 55:27
  SPEAKER_03: It is, it is my first book.
• 55:29
  SPEAKER_03: Uh second book is in the making and needs to be at the editor on Friday.
• 55:34
  SPEAKER_03: So I have deadlines um once again.
• 55:38
  SPEAKER_03: Um yeah, um, it's just like um everyone is super interested.
• 55:42
  SPEAKER_03: Um, I already got great feedback about the book.
• 55:45
  SPEAKER_03: Um customers love it.
• 55:48
  SPEAKER_03: Um, let's see.
• 55:50
  SPEAKER_01: I mean, the good thing is you don't make money with books, but you make money with workshops around them.
• 55:55
  SPEAKER_01: And you're a freelancer, so you basically you can do that.
• 55:58
  SPEAKER_01: In my case, I do I don't didn't make much money with my books because I always had a job.
• 56:02
  SPEAKER_01: But the the turning it into uh into uh into workshops and turning it into presentations is where the real money lies.
• 56:09
  SPEAKER_01: So that's the really interesting thing there.
• 56:11
  SPEAKER_03: So also it's more like a credibility and an authority thing.
• 56:15
  SPEAKER_03: Uh I'm just like, yeah, I wrote a book about this, so I know what I'm talking about.
• 56:19
  SPEAKER_03: And which is like I I don't really have problems to get new gigs uh because people are yeah, there is a lot of um there's a lot of good recommendations uh from my customers too, then the new customers.
• 56:32
  SPEAKER_03: Um, but yeah, I I love this so far and uh we'll continue writing.
• 56:37
  SPEAKER_01: I mean it's mostly because we were part of We Are Developers World Congress a few times, you were on the show a few times.
• 56:43
  SPEAKER_01: That's where most of the fame comes from.
• 56:45
  SPEAKER_03: And yes, so this is where the real fame comes from, and still know when I met Daniel for the very first time at We're Developers.
• 56:55
  SPEAKER_02: Yeah.
• 56:56
  SPEAKER_02: Yeah.
• 56:57
  SPEAKER_03: So um yeah.
• 56:58
  SPEAKER_02: We've talked about it a few times actually, as you know, that was a important conversation, wasn't it?
• 57:02
  SPEAKER_02: Because I was about to moderate the the stage one, I think it was, yeah.
• 57:06
  SPEAKER_02: And I hadn't actually gone up on stage and I had this little moment of nerves where I went, What am I doing here?
• 57:10
  SPEAKER_02: I didn't even, you know, I was still a teacher at that time.
• 57:13
  SPEAKER_02: I I I hadn't moved across or anything, I was kind of uh doing it as a moderator.
• 57:17
  SPEAKER_02: And then yeah, and then you said, No, I've seen the things you've built.
• 57:19
  SPEAKER_02: And I was like, Yeah, yeah, yeah.
• 57:21
  SPEAKER_02: And then you went, so you develop things, and I said, Yeah, and you went, So you're a developer, shut up, just go on stage.
• 57:25
  SPEAKER_02: And I was like, Oh, great, that's the brilliant German no-nonsense nuss that I needed.
• 57:30
  SPEAKER_02: That's exactly what I needed in that moment.
• 57:31
  SPEAKER_02: And then you were like, Nobody cares, just your job title, who cares?
• 57:34
  SPEAKER_02: Just go on stage, you can do and it was great.
• 57:36
  SPEAKER_02: So, yeah, I always uh I'll always say thanks.
• 57:38
  SPEAKER_03: And now look at you, and just like this is really, really cool.
• 57:42
  SPEAKER_03: Um, so yes, the real fame is uh obviously yet we are developers, but uh no for real.
• 57:48
  SPEAKER_03: So this um conference helps me grow tremendously.
• 57:51
  SPEAKER_03: There's always just a great crowd and people to learn from, but oh always just like so much good business that also comes from that conference.
• 57:57
  SPEAKER_03: So I love to be there.
• 58:00
  SPEAKER_01: Love hearing that.
• 58:01
  SPEAKER_01: And talking of which, where we're we're currently in the making and planning and doing all kinds of bits and bobs.
• 58:07
  SPEAKER_01: So if you actually take a look at what's happening, uh we're coming back to Berlin 14th or 16th of July 2027.
• 58:14
  SPEAKER_01: You're you're gonna be moderating again and probably gonna be speaking, right?
• 58:18
  SPEAKER_04: Yes.
• 58:20
  SPEAKER_01: Good.
• 58:20
  SPEAKER_01: And we're coming to India on the 25th and 26th of November, and we're currently getting our vaccinations, our uh our passports stamped and all kind of stuff.
• 58:31
  SPEAKER_01: And uh yeah, hopefully it's gonna be an interesting one.
• 58:34
  SPEAKER_01: It's a bit smaller than the others, but it's gonna be a trial uh to actually do it in India for the first time.
• 58:39
  SPEAKER_01: And uh it's gonna be fun to be back in India, and then you're gonna be doing a lot of more stuff than there than you've done before.
• 58:47
  SPEAKER_01: So that's gonna be interesting to see.
• 58:48
  SPEAKER_01: And we invite a lot of people for the podcast and for uh for this show as well from India.
• 58:55
  SPEAKER_01: So we're gonna be doing late night recordings soon of these sessions.
• 58:59
  SPEAKER_01: So it's gonna be interesting to see.
• 59:03
  SPEAKER_02: Just one thing to add to that, Chris.
• 59:04
  SPEAKER_02: Um, on the We Are Developers homepage now.
• 59:07
  SPEAKER_02: For that for anyone that hasn't seen it.
• 59:08
  SPEAKER_02: I mean, it's just like a huge um kind of upload, if you like, that you can take a look at.
• 59:13
  SPEAKER_02: Um, if you go on to the video section, you've got all of the talks now uh publicly available, or for anyone that's a member.
• 59:18
  SPEAKER_02: Um, on uh the video section, you can see all of the talks from North America.
• 59:23
  SPEAKER_02: So um, I mean, we already had like thousands in there, and so then there's another 300, you know, that have just been uploaded um in full.
• 59:29
  SPEAKER_02: And we've got Kelsey Hightowers uh Zero Token Architecture going out on Friday in the newsletter.
• 59:36
  SPEAKER_02: And I had loads of people actually at the Congress coming up to me after that talk in particular and being like using the words like it was refreshing, um, he didn't hold back, he put us in our place, things like this, because he kind of told off a lot of developers for things that we are doing that we shouldn't be doing and gave us tips on how to get around it and spend less tokens.
• 59:54
  SPEAKER_02: So, yeah, check that out as well.
• 59:57
  SPEAKER_01: Cool.
• 59:58
  SPEAKER_01: And I guess we're at the end of the end of the hour, so it was great to have you again, Luis.
• 1:00:03
  SPEAKER_01: So I'm now going back uh to my vacation mode.
• 1:00:07
  SPEAKER_01: So we're gonna go to an uh to a lake here in Berlin, which is actually a uh a dog uh facility.
• 1:00:12
  SPEAKER_01: So there's about 200 dogs for all shapes and sizes, none of them on lead, having fun with each other.
• 1:00:17
  SPEAKER_01: So that's gonna be great.
• 1:00:18
  SPEAKER_01: Yeah, we're going dog stalking there because we don't have ours at the moment.
• 1:00:22
  SPEAKER_01: So basically that's what we do.
• 1:00:24
  SPEAKER_01: But uh yeah, it's it's great to see, and it's also a fancy part of Berlin.
• 1:00:28
  SPEAKER_01: So you see people in like 300 euro valleys walking around and stuff.
• 1:00:32
  SPEAKER_01: So that's the other little part there that's interesting about that.
• 1:00:36
  SPEAKER_03: So the zip code starts with one four.
• 1:00:39
  SPEAKER_01: Yeah, it's it's an interesting one.
• 1:00:42
  SPEAKER_01: Yeah, I'm hoping it's fun about it.
• 1:00:45
  SPEAKER_01: Yeah, yeah.
• 1:00:47
  SPEAKER_01: All right, thanks, Dan.
• 1:00:49
  SPEAKER_01: Thanks, Luis.
• 1:00:49
  SPEAKER_01: This was the live show for today.
• 1:00:51
  SPEAKER_01: Um, we see you next Wednesday when I'm back in action.
• 1:00:54
  SPEAKER_01: And Dan is be freaking out about all the things we have to do for the India event, but it's gotta be so much fun to go back there.
• 1:01:00
  SPEAKER_01: Yeah, and I'm already eating a lot of Indian food to get sort of accustomed to it, but yeah, that's great.
• 1:01:06
  SPEAKER_01: Cool, see you then, bye bye.