Welcome back to the Free Code Camp podcast. I'm Quincy Larson, teacher and founder of Free Code Camp. And today I'm interviewing Jessica Rose. She's a developer and a teacher who's worked on open data projects at Mozilla. And she is launching a free online boot camp this month called the Bad Website Club. Before we hear from her, I just want to give you some quick updates from the community. Free Code Camp just published a new Python course that will teach you how to program your own aerial drone. You don't actually need to own a drone because you're going to be using PI Simverse for simulation and you can practice autonomous flight with that.

You'll learn the basics of drone components, 3D movement, and common computer vision tasks. Then you'll learn about navigation, image capture, hand gesture control, autonomous following, and more. This is a free two-hour course on the Freo Camp YouTube channel. We also published a massive course that will teach you how to program Nvidia's H100 GPUs using CUDA. You'll learn about cutless optimizations, multiGPU scaling, and the primitives that developers use to train large language models. This is a 24hour course, extremely comprehensive, and I wish I could also send each of you a free H100 to use, but free cooking camp just doesn't have the budget for that. If you are able to get a hold of one,

you will be ready. And you'll certainly learn a lot from this course, whether you have access to those expensive GPUs or not. If you've ever wanted to build a video editor or live streaming tool that runs entirely in a browser, this handbook is worth bookmarking. You'll see how web codecs API can give you low-level hardware accelerated control over video processing. You'll learn key concepts like video frames, codecs, containers, and moxing. Check it out. Link in the description. And Kubernetes doesn't have built-in user databases. Instead, it relies on a chain of authenticators. This course will teach you how X509 client certificates work, why they're not ideal for human users in

production, and how to instead deploy your own self-hosted browserbased Open AI Open ID Connect login. That's approximately a 30 minute read over on the free co publication. And this week's song of the week, 1983's Oblivious by Scottish new wave band Aztec Camera. I got a link to it in the description. I love this song's Django Reinhardt style flamco guitars, the mischievous baseline, and the stereo percussion. Believe it or not, the front man, Rody Frame, wrote this song when he was 18 years old, sang it, and played its iconic guitar soul. 18 years old. Um, you're going to love it. I think you'll really enjoy it.

Support for this podcast comes from the 10,338 kind folks who donate to our charity each month. Join them and support our mission at donate.freeccoamp.org. You can also pick up an awesome free coamp shirt and rep the community with pride. Get the shirt for only $20 shipped anywhere in the US. Shop.freeccoamp.org. And now for my interview with Jessica Rose. We're going to talk about how the whole world is hard and how embracing that difficulty rather than avoiding it can make you a better thinker. The Bad Website Club, a free online boot camp

where people learn front-end development together that starts this April and why building silly little things is one of the best things you can do as a learner. Jessica Rose, welcome to the Free Code Camp podcast. Oh gosh, thank you so much for having me. It's a delight. Yeah. Uh you and I have known one another for years at this point, like I don't know, like 10 years or something like that. Uh we've been Time is a round circle like keeps going. Huh. Yeah. Well, I wanted to start off by just asking, you know, you were building websites before CSS was even out, like literally just using the raw HTML. Like, uh, what are some tools that have come out over the

years that have really excited you and gotten your creative juices flowing? Oh gosh. Um, that's such a good question. So, I'm I kind of bandandy myself around as tool agnostic. So, yeah, everybody gets excited when there's a new tool. Um some of the things I've been really delighted to see are sort of um some innovations in um how we handle data on the web. So WOM was really exciting to see. So hey we can do non-JavaScript um programming languages on the web. Um, but really I'm a little bit I've been doing web stuff for a long time and it means that more than being excited about a lot of the tooling, I'm also just a little bit tired. Like there's just been a lot of stuff.

Yeah. And like I guess what was like the point at which you went from being excited with each new like JavaScript framework or something like that to where you started? Oh, the JavaScript framework. Um so actually um in a positive and negative way it was working with uh learners. So working with folks who were coming to learn web skills for the first time that really got me a little bit more a little bit cooler a little bit more suspicious about um yeah about the amount of tooling that gets dumped on people. So back in the day I'm I'm guessing we're roughly the same age. I might be a little bit older. Um, if you wanted to build a website, you had a bunch of

different options for places that would just do all of the hosting and all of the unspooling your website into existence. And really all you needed was either website builder, love it, or a little bit of HTML as you begin to learn and grow and expand. And that's not true anymore. Partially because we have so much power. We can fully fledged um applications. We can do so much more with the web. But as I talk to more and more learners, I get to hear, "Oh, do I need to learn this? Do I need to learn this?" And it gets to be really hard. There's so much. And I hope this isn't unkind. But part of what makes me a little bit tired is the relentless and confident tone of all of the hype.

Oh, are you not using so and so? Everyone's and talking to learners who say, "Oh, do I need to learn this?" and bringing this incredible list of things that I hope no individual person building on the web or no individual developer would actually need. But we keep getting told you need more, you need more, you need more and it's a bit relentless. Yeah. And you're very much a back to basics type person. uh which we're going to get into when we start talking about website club, but like has there been a tool that's come out where it's been kind of like the monkey paw where like I

wish I could do xyz and then one of the fingers on the monkey paw curls and next thing you know like oh no what have I unleashed uh like a good example of that would be react completely kind of redid like of course we have a lot more power but now you have to completely rethink how you write uh you know code for the front end and things like that and you have to put all these kind ugly HTML, you know, um, components into JavaScript functions and things like that. So, actually, I monkeypaw is a really interesting way to say it because I described this as a monkey paw to appear just the other week where um, I was kind of thinking back to website builders on the early web and these were they're not

beautiful, but they're really sort of drag and drop tick. Um, and we still have something really similar with some um like sort of CRM like WordPress where you don't need to have these coding skills to build these. Um, but I was bemoning, oh gosh, I really wish these existed again. Like these kinds of, hey, you're building a semantic HTML website and here's just a picture to help you do it. Um, and instead we've got sort of the LLM driven um I often call um LLM or machine generated text and it feels very much like spicy autocomplete where it's just a very opinionated will do it for you. And that's not quite what I was hoping for, but it does indeed make it easier to do something.

Yeah, absolutely. And uh we'll we'll definitely talk about AI, but uh I get the sense that you're one of the people that is a little bit reluctant to just hand all human thought over to machines. And uh that said, you have been trying to work on projects that help improve AI uh such as uh your crowdsourced open data sets uh that are used by AI toolmakers. Can you talk a little bit about that? Yeah, so I'm I'm not a pure leite. I'm um a selective lite. Uh but for the last three years or so, I was working over at the Misilla Foundation on Gosh, sorry, we've got cats bounding about the table.

Uh I was working over at the Misilla Foundation on Common Voice, which is a set of multilingual data sets that folks use to build models. Yeah. May models uh to make speech technologies work better for speakers of a range of different languages. Um, so I'm not, oh, no, no, no, no AI ever. Um, but for me, these big models and these big tools tend to have widely scraped the web and often in confirmed and lawsuit driven ways, just stolen. Um, so consent in where this data comes from is a really big thing for me. Um, but at the same time, if I want to make something, engaging my critical thinking, like, hey, why am I doing this? How does this work? Building to understand things.

Yeah. Kind of seems like a lot of the whole point, at least for me. Yeah. I mean, if you just sat in math class and just use your calculator every time your teacher tried to teach, oh, I got the answer right here. Uh, yeah. Okay, cool. Whatever. teach, you know, like you wouldn't learn a lot, would you? Further, like if you had the um I'm probably dating myself. Do these still exist? If you had the teachers version of your room of grow hill textbook and it had all the answers in the back. Yeah. Cool. You're not doing that sort of um uh free recall to who uh what year did Mary Todd Lincoln die, which is probably not something we need to know.

Yeah, it's probably something you can look up. with like a mathematical procedure for example like I help my kids with their math homework uh to the extent that I can with my limited recall of high school math. Um, and uh, essentially they are always having to stress over it and there's like the hair pulling and there's the nashing of teeth and then finally the breakthrough that comes with actually grappling with things and you seem to be a big advocate of grappling with things and not just jumping straight to the answer, not just flipping to the or I guess with the teachers version of the book, you don't even need to flip to the back to look at the answers. They're just like in line, there's the answer, you know.

So, I'm really lucky these days. I get to just work on projects I like, like bad at Website Club. Um, and don't have to do too much I don't like, which is very spoiled. I understand how lucky I am. Um, and this gives me enough space to say, hey, why am I doing this? Am I doing this action because I just need it done? Am I doing this action because I want to learn something? Am I doing this because I want to be proud of what I've finished? Yeah. like having a human thumb on the scale. It feels like something I want to do for almost all of my writing and coding. I'm much more ferocious about it for around um art and music as well.

Yeah. Well, you're as you said, you're in a really good position where you can work on things that are very exciting to you. But I want to quote something uh that you said to me before the interview. Uh you said uh I was supposed to work multiple minimum wage jobs around the clock until I died. Taking you back to your teenage years, grow up in uh the US. Um like what was your what can you walk us through like that context in which that major life decision happened? Yeah. Where you kind of became you or you set out on the journey to become you as opposed to somebody me who um Yeah. So, I didn't grow up in a great um great family situation. Um I was the only one of my siblings to make it to 40. Um yeah, it wasn't a

really great situation. And for me, I a large part of this was I just got lucky. Um I got cuz it was easier back then. It was a little bit cheaper but still hard. Um I complet I went to university which was a big deal. Um, and that meant that I qualified for a um, program to work overseas and could qualify for a visa. Um, and just getting out of the situation I was. So, being far from home, having the opportunity to go for professional jobs meant that later on where I said, "Hey, I used to build websites for fun when I was a weird geeky little kid. Maybe I do want to go through and professionalize some of these skills and do this

professionally." um it seemed more possible and it had this sort of stable base that I could do that from. Yeah. And so you were able to go abroad and uh my understanding is you were able to teach in Japan. Can you like walk us through like that period of your life and like how that kind of just very quickly like transitioned into more and more technical roles and leadership roles and things like that. So, um, yeah, got really, really lucky. Wound up, um, there's a government program to teach in Japan and headed out there and wound up in Kyoto and then Osaka, which, um, I'm a big fan of Birmingham, the UK city. And I always say Osaka is sort of the Birmingham of Japan. Just great, great

places to be. Um, but wound up picking up a little bit of extra work after I finished the government contract. So, did some technical writing. Um, and as part of that, I was like, "Oh, cool. Programmers over here seem to get paid. Y'all get paid double what I do. Oh, that's nice. What do you do? You just It's It's just different words, huh? Y'all are just building stuff with different words. It seems difficult, but maybe not impossible. And having that space and being further from a difficult situation meant that I had that space and time, too. Hey, maybe I want to try

this. And it worked out. I feel pretty lucky with the way it turned out. Yeah, absolutely. Well, uh, like so you just went through a series of opportunities, you know, technical writing and then, uh, you've been working with Mozilla on some of their open data sets and then ultimately a few years ago, uh, you had a podcast which was excellent, uh, the pursuit. Oh, good. Yeah. And it was me being selfish. I just wanted to chat to more people. Yeah. Well, uh, you had me on there. You had a lot of people that I listened to a lot of the episodes. Dawal Shaw of course class central creator uh past

free code camp podcast guest um and uh at some point you just decided like hey I actually want to help more people learn programming can you walk us through that decision and uh how we ultimately arrive you know iteration n of the bad I couldn't even tell you which iteration it is um so I'm going to ask everybody to come back with me to 2020 it was a lot right so there or lockdowns. Everybody was afraid of losing their job. A lot of people were out of work or furoughed in different areas, so temporary out of work. Um, but there was weirdly a big tech hiring boom. And for a couple years before 2020, there had been sort of a boot camp boom as well. And some of these were actually I think that there have been

some legal cases that, let me just say, a lot of these were just scams. So, you would have to pay tens of thousands of dollars for not amazing coursework. And I was doing I've been doing these um sort of open onetoone advice calls with folks for a couple years. And during the lockdowns, I was talking to more and more people who said, "Hey, what should I do? I've just done this so- and so boot camp and I owe them 40 grand um and it's crushing me." and looking at the curriculum and talking to these folks, I was infuriated. Um, and you don't give yourself enough credit, but actually it was a conversation with yourself and Dwalt Shaw from Class Central where it's like, look, we've got the curriculum.

Why don't we build a boot camp? So, the extra delivery, the extra support structures around it. Um, and really the success metrics I was looking for was every single time somebody didn't pay for a scammy boot camp, it felt like stealing 10 or 20 or 30 grand away from sort of the worst possible people in tech. It was very spite driven and it was the best thing I've ever done. Yeah. Uh well, free code camp, one of the reasons it's called free code camp is code camp was like a term used just for kind of like cohorted like learning and it predates the boot camp phenomenon. Um but uh I really like that vibe of people coming together

and learning together and building together. Um, and part of it was a reaction to what I thought were, you know, pro promising uh overpromising, underdelivering on technical education. Um, you're kinder than I am. Yeah. But, uh, so you created this program. Uh, why is it called the bad website club? Oh, so it wasn't at first. So um for 2020 2021 and the thing is when we launched this um class central hosted us for the first two cohorts uh first three and they were amazing just and they continue to be amazing. So much nice things to say about them and we were just the free code camp boot camp I think. Um, and later after we broke off and started doing our own

thing, um, and started working with Karma, um, what we started thinking about was what kind of tone we wanted to approach. So, I don't know if you've been outside recently, but the whole world is hard. Like, everything's stressful. Making rent is hard. Your boss is yelling at you all the time. Relationships are hard now. You have to take care of all kinds of personal data on the web that you never had to think about. like everything in the world feels like it's getting harder and these technical skills that we're asking people to learn with us. They're hard, too. So, we wanted to say the bad website club, not because we're teaching you incorrect skills, but because we

just want to take this time when we're learning together and not worry about perfection. Let's make a bad website. Let's make 10 bad websites. Let's have a silly soft little time and experiment. And then after we've made beautiful messy things and learned, we can go ahead and professionalize that and get good and become a good website developer or a good developer later on after we've sort of had an easy time together. Yeah. So essentially just like accepting that the initial process is going to be messy and not having any pretentions of like, oh, I'm going to be making insanely great websites like right off the bat, but just acknowledging

that, you know, before you can be good, you have to be mediocre. And the being mediocre kind of gets you a little bit of flexibility to actually learn. Um, I love tech. I love the things people wear. I especially love the web, but there's a lot of I'm going to teach you to become the best JavaScript developer in the world in 18 weeks or in 60 in six weeks and you just need to follow my system and buy my ebook and no like we're not selling anything and we're going to have a messy time together and it's going to be fine. Yeah. And one of the things that people may not be aware of is this is a completely free program. Uh there's no credit card sign up. uh you basically just join the cohort and you start

learning and can you walk through what people can expect joining that website? Oh yeah. So the other thing I'd also say is like there is an email list to get sort of weekly lessons. You can hang out with us on Discord but there's also zero obligation to sign up for anything. You never have to sign up sign in. So what we do is we take Thank you very much Quincy the free code camp. Um, we're looking at this cohort. We're going to be learning on the responsive web design curriculum and we're working from that curriculum together. And what we've done is we've added a shared schedule. So, hey, here's a shared pace. We're going to keep each other accountable. A community of other learners going through the same thing at

the same time, but also live streamed lessons Monday to Friday. So, hey, during the weeks, we're not going to show you how to do everything. We're not going to do it for you, but we are, hey, here's an especially weird concept or hey, let's talk through what a variable really means. And the thing I really wanted to make sure we had is question answer. So even today, like you still get stuck working on technical stuff, right? Yeah, absolutely. Same. And I always say to learners that the things you can do when you get stuck are you can um look it up, you can ask someone or you can experiment and get weird. So, hey, does this work? And just being able to unblock folks by

letting them ask a question and work through that question in real time and show them what we're working on is really one of my favorite things about it. Yeah. And the experience of asking a human being is very different from just asking an LLM. Like free code camp early on we had read search ask was uh the kind of uh the rubric or curist the sequence. The idea is it's much easier to like read the error, read the, uh, exception, uh, read whatever is in the documentation, then search around and try to find additional supplemental information, which will probably lead you to like Stack Overflow, free codec

camp forum, places like that, and then asking was like, if you do it in that sequence, you're being the most respectful of somebody else's time, but you don't have to respect an LLM's time. In theory, you could just go right in and skip those first two and just go ask your question to get a quick and easy answer, right? Um, but like what is the merit in asking an actual human dev who's who actually understands things as opposed to, you know, um, an LLM that doesn't really understand anything in the sense that like it has never actually, you know, sat in the room and like banged it head against the keyboard trying to get like something to work, right? Uh, it just has the answer

because it's read all the answers and can synthesize an answer from those. Um, what's the merit in asking questions? because you really emphasized there when you were talking about the bad website club that the question answer is a very important part of the experience. Yeah. So in order to talk a little bit about why I enjoy working with um people instead of working with LLMs on learning talking a little bit about how these LLMs work in a very abstracted way might be helpful. So these LLMs are models and they get fed a data set of things they've seen. Usually for these big tools, it means mass scraping the web and getting people's copyrighted material that they didn't consent for.

But if I said, "Hey Quincy, I'm going to go and look at everything on the internet and I'm going to give you back sort of an amalgamation of things from this topic." Yeah. Um, not everything on the internet is right and not everything on the internet is reasonable, right? There's a lot of 4chan mixed in and stuff like that. the fact that this is considered like the there's a lot of training data that I would not necessarily consider great training data that's mixed into these tools. Yeah. Um but like that advice is hoovered up with the same confidence and regurgitated with the same confidence. But it also means that like these LLMs are so confident where they and

they're not always right. And not only they're not always right, but the way they're delivering the learning the way they're delivering information isn't Yeah, we still have quite a lot of um information that's being uh research that's being done right now, but we're not seeing any of the promised gains in AI learning. So, there was just this uh Microsoft and Carnegie Melon um I'll try and find that and chunk it to you in the show notes include a link to the uh study to the study. There's the Microsoft and Carnegie Melon one recently. I think a reasonably small sample size of maybe 400 people um and they found that the folks who were the more AI skeptical who didn't turn to these tools as quickly um

were engaging their critical thinking more that not just having this given to you. Yeah. Well, I want to move on to another question I have about the bad website club and your preference because I imagine you encounter a lot of people who are I'm not sure if this is for me. They're intimidated. uh they don't necessarily want to like join a community of like strangers online who are building their skills and they're going to come up with a lot of excuses to preserve their own personal energy be uh maybe they uh identify as like introverted and they're just like it's really taxing for me to put myself out there and talk with people like can't I just you know be heads down on

on my computer and do this myself and not have to talk to other people like what are some of the excuses that you think people might have in mind that would essentially they'd use to talk themselves out of participating in something like bad website click. Um, so the big thing I'd add is this isn't the best fit for everybody. The whole point behind me being free and there's no sign up is folks can come along and say, "Oh, is this for me? H I don't get a lot of value out of the lessons. Oh, I'm going to come along to just the guest sessions. Oh, that doesn't work for me. Oh, I'm just going to come along to some of the showand tell." So really the ability to say cool, not all of the things that the boot camp

offers is going to work for everybody. And if somebody says, "Oh, I'm not very um outgoing. I don't like to get into these discussions." Folks can go ahead and just take the pieces they want. Um I'm careful about excuses as the term. Um some of the concerns, sorry, I'm a real soft-hearted. Some of the concerns I've heard folks say in the past are uh this is really hard. These technical skills are because a lot of the folks in tech are like r this is hardcore. We're we're the smartest people in the room. So really a lot of it is am I good enough? Am I smart enough? Can I keep up? Um, and the things that I always kind of bring back is the way we've paced this boot camp is we're going to start on April 24th with a little kickoff party. We'll

have a lesson stream Monday to Friday all the way through until July 3rd. But all of the material you're working on is within free code camp. If you go ahead and join and you get sick or you hurt or you're tired and you need to stop, that progress doesn't go away. There's no pass or fail. I mean, the only way to fail is to stop and not come back. But one thing I'm really proud of is a lot of our learners from the first cohort came back and finished on the third. Like the ability to say, "Cool, this is just a really laid-back community. There are deadlines that we're working on this day, but we don't expect everybody to have everything finished on July 3rd. That's a lot of stuff to do. The big thing I

try and say to anybody who's concerned that this might be too hard or too much websites are like the web is hard. There are a lot of really advanced technical skills you can get to in time, but all we're going to do in this boot camp is learn some of the really important foundational technology. We're going to learn some of the core concepts. And what I really like about the web is you're getting the opportunity to sort of paint with light and pixels. We're just going to make some silly little things for the web. Come and see if you like it. And if you don't, not even a big deal.

Yeah, let's paint with light and pixels. Um, let's make some silly little things. Love it. So, you're a volunteer. You're doing this out of the goodness of your heart. I want that to be abundantly clear to everybody here. And you are not alone. You've got a couple other volunteers that work alongside you. Right. You I was just about to interrupt to big them up and I really appreciate it. So, um, since 2021, 2022, she's going to get me. I've forgotten this. I've been working with Carmen on this, and she is a much better programmer than I am. So, she does ton of JavaScript stuff, a ton of um, like Windows applications.

She's been doing this for a little bit of time as well. So, she's just done all of it. And she's fantastic. She's much chiller than I am as well. So, she's been teaching, we've been teaching kids this whole time. Um, and we've got a brand new instructor joining us this year. Um, who's Eda. And Eda was actually a student in our very first cohort in 2020. Um, so she's my favorite. Not my favorite. I don't have favorites. She's a fantastic example of, hey, we're going to just experiment with the basics first and give you a foundation to continue to learn. So she's gone on and she does quite a bit of programming in Rust these days. Um, which is possibly a bit beyond

my own skill set. So she builds really interesting things and she explains them in ways that are very accessible. Um, but she's just gone on to continue to do a ton of open source work, a ton of freelance work as a dev and it's been really delightful to see her grow and to see her come back and join us as an instructor. Yeah, that's so cool that uh the teacher becomes the student. Right. And she's unnecessarily modest. I promised I wouldn't embarrass her, so I will save it for the streams. Yeah. And I meant to say student becomes the teacher.

I think that's way. Yeah. Like all in the kung fu movies and stuff, if you watch a lot of old 70s kung fu movies, that's always like a recurring theme. Um so there's a lot of hype right now around just uh basically having machines write a lot of code for you. How would you justify to somebody that is like are these skills these basic kind of primitives of web development HTML CSS JavaScript like how much more mileage are we going to get out of these? Like does any of this really matter? is I'm sure a lot of people are hearing uh you know the AI hype people who often have you know multi-billion dollar vested interests in people uh just letting go and letting machines do all the work. Uh like what would your

pitch to someone who is maybe on the fence about whether they really even need to learn the stuff anymore because they keep hearing oh you don't need to learn to code anymore you don't even need to bother learning computer science anymore. Uh, I've got a direct quote from Jensen Hong, uh, the founder of Nvidia that I'm gonna read to you and you can react directly to that if you'd like. That sound cool? Yeah. All right. So, the quote is, um, almost everyone says it's vital that your children learn computer science. Everybody should learn how to program. And in fact, it's almost exactly the opposite. That's a direct quote.

So I think it's a very interesting thing that a bunch of folks who work in these tool makers and a bunch of folks that work in that run tech companies um are saying no don't worry about it. development jobs which have traditionally been um we think of them as being especially high paid but really development jobs when we look at sort of the um productivity growth they haven't had that fall off in the 70s that a lot of other jobs have had. So they've sort of continued to grow salary wise with what we would have wanted to see all jobs continue to grow with in the US and Western Europe. Um, I try never to take advice from my ideological opponents. So, here we've got folks who are saying, "Do you know what? For a long time, I've had to pay these

stuckup developers quite a bit of money, but no, no, no. This tooling, this tooling is magical, and this will replace those very, very finicky, very expensive developers." that doesn't seem supported with the code I'm seeing turned out by a lot of the tooling. Yeah. But also these tools are fantastic and props at building small standalone prototypes. You want to build a complex system, you're going to need software engineers, systems thinking folks were telling you don't worry about it. Don't bother building expertise. What a deeply suspect uh piece of advice. Yeah, I mean it does seem like okay there the you know the scouts motto here in the US uh be prepared you know like why not learn how things work. You know life is long is what I like to tell

people. like people if you're only like you know 18 and you're listening to this like it may not feel like life is that long but believe me you're like less than a fourth of the way through your natural life assuming that you're lucky right um you're going to have plenty of time to actually learn these skills and I will say like anecdotally I do a lot of development with AI assistance and like the whole free co team we use AI assistants for building things and it is only because we are software engineers that know what are doing that we can steer these kind of dumb models to like actually do productive things for us and that we can QA it and uh I do think that people who don't understand how or why

of things uh are going to struggle with these tools. So I mean that would be my quick take to kind of reflect on what you said and you've also said something that I'm going to quote directly. Any new novel creative code will still need to be written by people. Why is that? Oh, so the way these work the sort of um collecting a data set, building a model using a tokenizer to spit out the output. Um so data set build the model on the data set. Um and then you ask the model a query and then it's going to sort of use a fancy sort of tokenizer, a weird parser to give you the answer based on what's seen in its data set. it.

I often think of them kind of as hule. So, or um Slim Fast, they're the same thing. Yeah. So, for people who don't know what that is, basically like liquid diet type stuff that in theory gives you like all the nutrients you need um and maybe even has some drugs in it to make you feel full, but you're just drinking it like a shake. So this is the output of these models are just sort of a mixed up very mid slurry of what it's seen in its data set. The ability to make something new make something hasn't seen. The ability to make something creative like if you've got a and admittedly these models are very uh complex.

These are very interesting um bits of text. But what they're doing is they're spitting out copies of what they've seen. There's no Yeah. Yeah. So, by virtue of them just they're not going to recombine a bunch of existing knowledge into something that's truly novel and interesting. Would you like your assertion there? We're there's not the opportunity to combine like you can't blend a bunch of different types of food and wind up with a brand new beautiful creation of a recipe. You're just going to wind up with a different kind of

slurry. And it might work. Yeah. But I think the question here is if you have a boss where they just want this done and the LLM model does this. Yeah, it might be want to be what you do. If you want to build something beautiful or meaningful or silly or technically complex or weird. Yeah, I don't think they do weird especially well. Yeah. Another question that a lot of people are going to have, and this is a question I answer a lot, um, but I want to hear your answer to it. Why should I use any textbooks or any courses or anything like that when the LLM has read all these things? In theory, it's it's represented in the training data. Why don't I just ask a bunch of questions to the LLM? Isn't that faster?

Wow. So, um I just I gave a talk in February. That sounds about right. Um actually with Eda. I was quite proud talking about um AI is forgetting machines. So how to continue to learn and continue to build stuff in an era of just spitting out the answer. And that's really the challenge, isn't it? That there um so I really recommend um Barbara Oakley is a neuroscientist who works in sort of the space of how we learn. She's got some books and courses just fantastic and usually comes to as a guest speaker to our boot camp as well. Um and she talks a lot about the sort of diffuse and focused thinking patterns that you need to sort of build um learning and like journey through the neurons in our brain which is quite

interesting but also talks a bit about the friction we need. So if you were just getting the answers all the time and especially getting the answers all the time from a model that is very confidently and very like they praise a lot, right? Like they're they're very incorrect a lot of the time and they're very incorrect in a very confident way and they'll tell you how smart you are while they're lying to you. So Quincy, here's the incorrect answer. That was such a good question. Good job. um isn't always the best way to learn. One thing that you said earlier, the whole world is hard. Things are getting harder and yet you're telling people, hey, do even harder things, like do things

the hard way, like even as the world's getting harder around them. Why not save your energy and just like kind of like hunker down and you know to survive this storm? And like for a bunch of folks, if you do not have the energy to learn a big new thing, that is completely fine. Like if you're just showing up and getting through the day, absolutely respect that. But for some of the same reasons people like woodworking or pottery, I've just started getting into book binding, and it is a real hard way to wind up with a book. It is much easier to go to the shop and get a little offthe-shelf blank notebook. But the idea of, hey, here's a new set of skills. Yes, these are skills that when you bring them up into higher levels of

expertise, they're still really valuable, but the opportunity to slow down, learn something, and in a really low stakes way. No one's going to yell at you, you're not going to get in trouble. the way to slow down and build something. There was this um sort of software craftsmanship movement a while back and it sounded um at the time to me a bit sort of um it sounded a bit old-fashioned at the time. We're going to slow down. We're going to consider this a craft. Yeah. And in my ceaselessly advancing age, hey, let's slow down. Let's understand this. Let's learn it together. Let's have a low-key silly little time. Yeah. Even if it's just me learning with me, I'm going to have a really nice time.

Yeah. I mean, when I'm trying to learn something, I intentionally slow down and like I want to jump to this conclusion. but let's let's slow down and let's like learn more. And it's hard because you have that extrinsic pressure like, oh, you know, like time is money like the rents do or um you know, people are waiting on me or like I could be doing anything in the world right now and I'm sitting here reading this book in the library. Like what am I doing? You know, like there are all these voices that this entire chorus in your mind that you have to like quiet down if you want to slow down and really focus and be in the moment. um is that your experience and like how do you cope with like the competing priorities and doubt that you feel?

So one thing one sort of big constant in my life is I'm lightly but persistently annoyed when we take systemic problems and put them on an individual say, "Oh, it's your fault. Um it's your fault you're broke. It's your fault you're um having a hard time getting to work in a city with bad transport. It's I think it's not a coincidence that the way we live right now, we're beset by time pressures, notifications. We've got people teaching about, writing about how to make applications more addictive, how to make platforms stickier, which is sort of the tech term for more addictive. Yeah. Yeah, I mean one of the books that was really seinal back in like 2012 2013 when I was like you know building

free co camp uh and kind of thinking about it was called Hooked and it was literally about like kind of like Skinnerbox type uh reward like hijacking people's reward center because that's how you get engagement and that's how you get your multi-billion dollar exit. Yeah. So, with you saying, "I could be doing anything in the world and I'm here in the library reading this book." Um, I often think of this as an incredible type of luxury. You could be doing whatever you want. But for a lot of us, a lot of the time that, and I include myself, that might be looking at my phone. That might be worrying. That might be doom scrolling. Um, I can't tell other folks.

I shouldn't tell other folks how to live. But one thing that I've started slowly introducing into my life is less foam. And please see find me on social media, see me after class. I've got like very concrete structures around this because they're addictive. They're intentionally addictive. But less screen time, less phone when you can and when you can afford it, and more time in this glorious luxury of I'm interested in and it might not be web development. It might be right now I'm really into reading about the sociology of um conspiracy theories and how the underlying factors with which people form groups and start to believe in sometimes weird sometimes harmful things. But that's such luxury to say,

"I'm going to go read some academic papers instead of doing another email." If folks have the time and if something they might want to learn is building websites for and not just for work. I want to build a website to show you pictures of my cat. I want to show build a website for my sister's salon. I want to build a website for recipes. I want to build a website for art. If you Yeah. If there's folks out there who might want to spend their very precious time learning with us, gosh, what an honor. What a joy. Yeah. And one of the things you're hitting on there is like building websites is a form of expression. Uh not every website has to be some utilitarian like let's convert this widget into this

widget or let's uh convey this very specific piece of information or let's get people to whip out their credit card and pay for this thing that they didn't know they needed. You know um there's a lot of exciting weirdness on the web uh and a lot of beauty and you can find a lot of those projects. A lot of them unfortunately have been kind of like centralized, but there was always like centralized like you know geospaces. Yeah. Geio cities. Woo. Uh geocities. What was there? There were like uh Neopets where you could like customize your You used to be able to customize your uh MySpace page.

Uh you could do a lot of that kind of personal expression. You can look for that sort of stuff coming to free co soon. By the way, I love how like as you were saying these things, different generations of people whose back hurts were going Yeah. at different points. One of the reads that I have about Bad Website Club is people are going to get in there and they're going to learn a lot and it's going to be like a creative kind of low pressure environment where they can express themselves and they'll learn the tools and the techniques to further express themselves. And one of the things that like I would observe about these core you know primordial tools that you're using HTML CSS JavaScript like these been around they've been around since the very early

days of the web basically like you learned HTML before CSS was out but CSS has been around for I don't know like 30 35 years at this point I don't know like a long time 30 years probably I don't know CSS first release but um yeah they're not going anywhere almost everything that you when you go to a website now is just using those under the hood even if there's a lot of additional layers of abstraction on top of it. So it sounds like this is something where people are going to learn skills that they can turn around and apply and they're also going to learn strategies and ways of thinking about things that are going to be relevant for a long time coming.

Yeah. And I'm always really cautious. So, we do focus a lot on, hey, a lot of people want these um a lot of people want these skills for getting a job, for getting a different job, for doing better at their job. Um, and it's difficult because secretly in my little goblin heart, I want to push, hey, this doesn't always have to be about making money, about being more financially secure. But in the same way that I wouldn't be teaching this if I was still working two part-time minimum wage jobs forever. Yeah. One thing I'm really conscious of when the boot camp is I'd love to say like you said, "Hey, building for the web is it's art. It's making silly things. It's making beautiful things. It's

expression." But a lot of time in the boot camp, we focus on people want to build these skills in order to get a better job, in order to do better at work. A lot of the incentives are financial and there's a temptation to say, "Oh, no. We're doing this for art. We're doing this for human expression. But being able to pay your bills and being able to eat is a pretty big like in the Maslau's hierarchy of needs, getting closed, housed, fed, and safe has to happen before we've got space to build silly things. Um, I would not be able to build the boot camp and work on it every year if I was still working, yeah, two full-time minimum wage jobs until I died. Um, it would seem really

cherish for me to say, "Oh, now that I'm housed, fed, okay, stable, I demand other people meet me at a level of purity I wouldn't have been able to meet." So these skills are creative, but people can use them for whatever they need them for. Yeah. And that is the practical reality is that we all have bills and we do have to address that. Um you said earlier uh again to go back to your quote, I was supposed to work multiple minimum wage jobs around the clock until I died. like if we could say send some encouraging words or some practical advice back to that young Jessica uh before you'd gone abroad, before you kind of like figured out a way to break out and found a path out of that life, what would you say to yourself that might

make your life a little bit easier? Um or just save you time and thrashing. Yeah. Oh gosh, thrashing is such a good way to say it. Um, I think that's kind of cheating as well because I left in the 2000s and I was younger in the 90s. Um, and things were easier then. Like stuff was still expensive, stuff was still hard, but social mobility was felt a lot easier. Um, what I'd say to me is probably nothing. I haven't always taken the most direct route, but I wouldn't want any spoilers. Um, and do you know what? For folks who might be in a similar position today, I do not have the misplaced confidence to offer any kind of advice. People are coming from a whole bunch of different situations. Everything's hard. I'm not

going to say, "Oh, you just need to." Um, but for me, a big part of why I run the boot camp every year is I guess spite I got really lucky. I had a whole bunch of people help me. And other people deserve this. Other people deserve help. Other people deserve support. Everyone. We can't means test access to happiness. We can't means test. Yeah, giving people skills that might improve things for them. I wouldn't give myself any hints. I would probably just keep bumbling along with my weird, spiteful little desire to make sure that as many people as possible get as many breaks as I did.

Awesome. Well, uh, I want to thank you for all the work you're doing through Bad Website Club. It's my delight. the other work you've been doing uh around the communities helping more people get into tech get these skills it's super helpful uh I want to encourage everybody to consider joining bad website club uh the opening date is u actually I was going to ask if you wanted to come along to oh asking you on the air if you wanted to come along and uh welcome people at the kickoff party which is going to be on the 24th of April so just a couple of weeks, 24th of April, uh I'm going to be there and uh there are going to be a lot of other really cool chill human beings there uh coming together to build creative stuff to help other people

learn how to build creative stuff and uh yeah, that'll be fantastic. I'm hyped about it. Yeah, I'll look forward to them. And then maybe some of you lovely people watching along, I'd be so delighted to learn with you as well. Awesome. Well everybody, thanks for tuning in. Until next week, happy coding.