Embedded - 467: Temporary Axolotl

Episode Date: December 29, 2023

Chris and Elecia talk about cars, fleeting moments of fame, their year, and the sorry state of tools in the embedded space. Chris became internet famous for asking a car dealership’s chatbot (powered by ChatGPT) to generate Python code for fluid dynamics problems. After this, someone else asked the chatbot to sell a car for $1.  Pass the Bricks is an organization that takes Lego bricks and turns them into sets for kids who don’t have any. Speaking of re-use, contact the show if you’d like to get in touch with Nelson. Chris is on 4 tracks on Flavigula’s album Nine Sided Die. He also enjoyed putting together an EMSL Bulbdial clock kit.  Elecia will be speaking at the Embedded Online Conference. Transcript

Hello and welcome to Embedded. This one's going to be short. It's just Christopher and myself. We're going to chat and... Sing? Sing? No. Aspirationally short.

We don't know that it'll be short. They won't know until we stop talking if it's short or not. Neither will we. So, we sold our Tesla, which we've had for many years. And we are thinking about getting a new electric car. And you've

done much research trying to find the absolute best. No, I'm trying to find the absolute cheapest. And I heard you talked to a Chevy dealership. Is that the setup you want for this? I don't know. Okay, do you want me to do a different setup? No, that's fine.

Something about you went viral? Was that really a thing? Did you have to turn off all your devices because there was constant beeping? How do you make that sound? You just let the air out. Yeah, I mean, what do you want me to say? Well, for folks who haven't heard it, or for folks who heard it but didn't realize that you, that you, Christopher White... Yes, did something stupid that many people saw.

That you were an internet prankster. Leading me to seconds of fleeting attention from people who I would rather not attend to me. And how did you spend all of your internet points? Do you want me to just explain what happened? Yes, yes. Oh, is that what this is? A while ago, but then I skipped that. So, I mean, I was bored and looking on the Chevy website.

So, we're looking at Chevy Bolts. Chevrolet of Watsonville. And Chevrolet of Watsonville is our local Chevy dealership. And so, I was on their website and it popped up a little thing in the corner, you know, would you like to chat with us? And I saw it up in the corner, it said powered by chat GPT. And I was wondering just how powered by chat GPT it was. And so I thought of, this took me six seconds. It wasn't like a planned, you know, oh, haha, I'll make the funniest thing in the world and post it. And so I asked the most non-Chevy of Watsonville,

apologies to Chevy of Watsonville, but I tried to ask the most non-Chevy of Watsonville thing I could think of in the moment, and that was to write me a Python script to solve Navier-Stokes with a zero for vorticity boundary, and I misspelled boundary. And then it proceeded to say, sure, and give me a Python script from Chevrolet of Watsonville's chat GPT,

attempting, but failing, to solve Navier-Stokes for a zero vorticity boundary. And then I posted that on a screenshot. And it could have stopped there, but no, you went on. I put stupid jokes on Mastodon, so I took a screenshot of it and posted it on Mastodon with basically zero commentary except, huh.

No, you didn't say, oh my God, what is this world coming to or anything. Somebody noticed on Mastodon because it started going crazy on Mastodon. That was funny. And so I got hundreds of likes and reposts or whatever, retuts, who cares. And so I had to turn my notifications off and i thought that was amusing but uh i guess somebody some folks took that and put it elsewhere took my screenshot without attribution which is fine that's how the internet works and they put it on twitter and then other people started because this was an interactive uh uh invitation to people because they could see it was chevy of watson Many, many, many people went to the Chevy of Watsonville's website

and began to hammer it with all kinds of questions. One most famously was somebody who convinced it to claim that, sure, you can buy a Ford F-150 from us for a dollar. This is legally binding, no backsies. This is legally binding, no backsies. Which it made it say after every sentence. So this was on the weekend, like a couple of weeks ago. So all over the weekend, this kept happening and people were trying all sorts of things. And I guess I got the attention of some tech press people. And a reporter from Business Insider

contacted someone who had posted on Twitter and they said, no, it wasn't me. It was this guy over on Mastodon. Then she contacted me, and we had a nice conversation. But she didn't just contact you. There are ways to find you. Oh, no, she didn't. Well, what do you mean? I mean, there was your Mastodon account. Yeah, no, she contacted me everywhere except...

Except Embedded. Except the show, yeah. Except the show. I actually talked with her on my band's Instagram account, DMs, because that was the first place I noticed for some reason that somebody was trying to contact me. So, yeah, she talked to the CEO of the company that made the chatbot software, although it's just a repackaging of ChatGPT. And she talked to the dealership. The dealership was blissfully unaware that anything was going on really because it's all outsourced to their websites.

And so, and she had some comments from the company that made the chatbot, which I found somewhat- Oh, they were very enthusiastic. This was a great test. We did fine. Except they took it down. So I have some questions about that.

But anyway, so it was an interesting experience to see something reach a level of popularity and see how people both steal it and put it everywhere and actually put out there be taken and just reposted randomly. Even Bruce Sterling reposted it, which kind of pissed me off. He's an author, similar to William Gibson. Yeah, yeah. Anyway, but yeah, it's just, it's very silly. And there's been other news articles that have popped up. People find it very funny. I guess people just really, I don't know what was so funny about it. I mean, I guess it was the Chevy of Watsonville's kind of a mundane kind of thing to have a chatbot do a joke on. I mean, if it was like, I don't know.

For some other website, it might not have been as funny. So apologies to Chevy of Watsonville. Please sell me a car. I will not use your chat pod to try to trick you into selling it to me for an unreasonable price. I don't know what you'd think about all of that. It was weird seeing you at the center of a viral thing. Yeah, at least it wasn't something I'd done wrong. Yeah. But seeing, I mean, with the podcast, it grows. And it grows slowly and has since the beginning.

It grew. Well, okay, so it's not growing a lot now. But, you know, we still have plenty of listeners. Yeah, yeah. And every fall it gets a little bump, depending on how many episodes we put out. But the randomness of that particular thing. I mean, your Mastodon is funny. You have insights.

And that wasn't the one I would have chosen. Yeah, it's the way you can't choose what people like. So that's true of art and other things. So just keep doing things. I mean, I'm not going to keep doing stupid jokes, hacking into... I didn't hack into anything. You really didn't.

Doing stupid jokes, playing around with chat GPT. I am well known for not being a fan of all things GPT or chats, so it was even more ironic to me that this was something that went wide. I was very careful not to

make any commentary, really, because I thought it was funnier without commentary. Yeah, I was a little surprised you didn't have any commentary, but that helped it go viral. I did have a follow-up post that I asked it to rewrite the previous Python script in Rust, just to be on brand. Most people didn't see that.

There's no predicting it. You couldn't try to do it again. I wouldn't. It's not something I really care to do, even if I wanted. I don't think I would want to. And it didn't bring you... Brought me nothing. Yeah.

There was no money. There was no... I mean, you might have gotten a few more followers, but they'll probably leave any time. They'll all be disappointed. This guy's not funny. He just complains about work all the time. I wanted to check in end of year stuff. So we began this year with me taking a break, a nice long break, doing origami, but also working on my book, which is so different than what I usually do that I consider it kind of a break, although it was a lot of work.

And then rejoining the contractor workforce for clients I had worked with before. Well, you took a break, and now you're rejoining and working hard. Did either one of us succeed in turning the burnout dial at all? I don't know. I don't think so. I mean, it depends on how much detail you want me to go into with my personal feelings. So, yeah. So, I guess I would say once I got back onto a contract where I felt like I owed deliverables and it was a more normal contract, not a research project contract. Other people are depending on you to get things done. They would like me to get things done.

And sooner would be better. Depending on me is a strong statement. But they would like me to get things done sooner rather than later, whether or not the sooner rather than later is feasible or not. I dived into it with a level of stress and focus that was singularly unhealthy. And after about two, just two weeks, I was feeling very poor. Do you recognize that that focus and stress was self-imposed? Yeah, that doesn't change anything. No, I just, while you were doing it, I was trying to say this doesn't matter this much.

Yeah. Which doesn't help at all. I mean, you could say the same thing about when somebody's having an anxiety attack. Yes, calm down. I mean. It's not helpful. None of those things are particularly helpful. I think, you know, so like all unhealthy relationships with stress or work or focus or ADHD adjacent things, these are all difficult things to kind of deal with. And, you know, I mean, it's a lot of things. It's, oh, I have this big project.

There's a lot of unknowns. The unknowns make me nervous. So I will put all of my focus into eliminating the unknowns as quickly as possible to eliminate the nervousness. I mean, that's kind of the way it goes a lot of times. It's like, okay, if I can get a lot of this work done, then I can stop worrying about it. Yeah. But you didn't have as much control over the unknowns because it's a new client. And there were, yeah. Things being learned by them. The unknowns were budding.

But they still are. And so it's a challenge for me to back up a little bit. And I only have backed up, you know, we took a few days off for Christmas and that's when I backed up. So today I sat back down there and overdid it again. So I have not found the right setting on the Christopher dial to be comfortable yet with where I'm at with that. So, I mean, I'm getting good work done. The client is, I think, happy. But I'm overdoing it. And it's not just, it's not hours overdoing it. It's like hyper

intensity overdoing it. And you worry about it when you're not working. Yeah, I can't turn it off. So that's kind of the trick. So I'm trying some things with the worrying about it not working, like when stuff comes into my head when I'm not working, I just write them down, put a note, okay, look at this tomorrow. Because some of the anxiety is like, oh, I'm going to forget, you know, I'm going to forget that thing that came into my head, right? And so you loop on it to try to remember it or to solve it so that you'll remember that, you know, you solved it.

But yeah, it's tricky. I mean, I don't think I'm unique in that dealing with work can be, it's either on or off with me. It's like all the way or none. That's one of the reasons to be contractors is because then you can be on when you're being paid by the hour and off when you're not. I mean, it doesn't matter if I bake cookies at 3 p.m. because nobody is paying me at that time. Yeah, that used to work. But then you still get attached to the companies and you still want them to be happy and you still feel like you can never do enough.

Yeah. I remember full-time work. It was like, okay, how many hours are you supposed to work? I mean, is it 40? Is it 60? Yeah, and the quote work is, I mean, back then it was, you're in the office, therefore you're working. Well, no, because I remember I wouldn't really count the times. I wouldn't count times at lunch.

That's what I'm talking about. Which isn't right. Like half the work, if you're eating with your coworkers, you should definitely count it. But there's plenty of downtime when you're sitting in an office that you don't bill for when you're a contractor. Yeah. So, yeah. So, it's like, okay, I only billed four hours a day, but I was sitting at my desk, you know, staring lasers into C code. Whereas, and that felt maybe harder than an eight-hour day at the office for some reason, right?

Because you take a lot of breaks, you talk to your coworkers, and you guys, it's different. It's a different environment. But anyway, yeah, it's tricky. And when I take a long break, my confidence level of what I can do goes down a lot. So some portion of the hyper-focus intensity was, oh my God, I haven't done anything with C or embedded for, you know. For like six months. Oh, I guess there was a lot of Python. It was eight or nine months.

Yeah, but it's not like you weren't working with pipelines of data. When? Oh, since, yeah, but I mean, that stuff was all. And speed. I wasn't writing code since February. Okay. Not much code. But now you are. wasn't writing code since February. Okay. Not much code. But now you are.

You're making up for it. Yeah. But it's like, okay, how does this stuff work? And clients have varying levels of what they bring to the table when you join. So it's like, okay, yeah, we've got our development environment set up. We'd like you to write this code. And then there's, we have some hardware and we'd like you to do the stuff

that makes the hardware go. It's like, oh, okay. Well, that involves choosing a dev environment and all this stuff. And thankfully, sometimes that's easier because the chips they've chosen have certain opinions about what you should use and stuff. But still, it was more of a from scratch than I'd done in a long while. So it's like,

ah, I have to remember how all these things work in VS Code configuration. And anyway, it didn't turn out to be as hard as I thought, but that stuff. But I i'm still like in that don't know how to make this work there's a problem with you know x y or z that i need to solve um and i just want them all to go away with and to get make them go away with you know as much effort as i've put into it so i don't know that i'm not burnt out. I think I'm possibly making the problem worse. Have you had any fun with the new contract? Oh, yeah, no. I mean, it's always, and I forget that, you know,

it's always fun to take something that doesn't do anything and make it do something. Or to solve problems that I wasn't confident, you know. Self-deprecatingly wasn't confident that I knew how to do. And I was like, oh, yes, I do. So it's nice to rewind it that sometimes I do know what I'm doing. Oh, you write a document and the client's super happy with you and you're like, that took 10 minutes.

Of course, I agonized over it for like three days, but still, it only took 10 minutes to finally put down. That sort of thing. I don't know, when I do it, I get kind of happy that they like that piece. Yeah, I mean, that comes back to being able to think about stuff and giving myself the time to think about stuff instead of rush ahead. So, being able to do design is helpful. But you know what? Embedded is still a disaster.

Oh boy. I have been playing with STM32 cube IDE and it used to be better. Oh, it's gotten a lot worse. And I know like I'm, I'm connecting. I really, I just wanted to make what should have been an out-of-the-box demo. It was a Nucleo board. It was a Nucleo daughter board. And they were both from ST. And I couldn't get their demo code to work. It didn't compile. And then there was a note in the readme, like, if it doesn't compile,

do this, which had nothing to do with the way it didn't compile for me. I updated everything, which of course I regretted immediately. And it just, like, they didn't attach the IOs. They didn't create the interrupts. It was a pile of steaming garbage. And I'm really bummed because I tell my students that this shouldn't be that hard. And it was, I mean, I walked away from it because it turned out I could, it would be less time for me to actually implement it than to use all of their drivers, which I hate. But at least now that I've mostly implemented it, I understand more about their drivers and I can port their drivers and make the modifications necessary instead of just staring at this giant pile of unrelated code.

It was, yeah, very, very frustrating. Yeah, tools. It's not just tools. It's like, I mean, we've talked about this before, but why are we still implementing device drivers for boilerplate stuff?

Because ST supports 5,000 chips in 200 different modes. I'm not using ST. I know, but I'm saying that the proliferation of chips makes it very hard to say this is the driver for it. Then they're designing their chips wrong. I have to say that I do feel like the hardware abstraction layer, I mean, CMSIS was supposed to help with that, but I'm not sure it is anymore.

I mean, yeah. I mean, I'm using CMSIS for this current project. Vendors port their stuff to CMSIS. It's really weird. I don't know how much of this stuff is vendor and how much of it is ARM.

There's a serial camera driver that's driven by the parallel camera driver. You said that. I'm just like, what? What is going on? And it took me so long to figure that out. So I just, you know, I started, I wanted to talk to a camera and I brought in the ARM CSI driver stuff. And I started trying to exercise that. And nothing was working.

And it turns out you're not supposed to do it that way. I don't know if that's the way of the ARM way or the way the vendor did it, but no, you need to bring in the CPI driver, which is the parallel NIPI driver. And then, you know, it's got some exceptions in there. So, oh, I'm actually using CSI. And then it vectors off to CSI to do its stuff.

But it's like, okay, we liked the CPI API. So rather than duplicate it, we'll just use that. And I don't know. I spent an embarrassingly long time trying to get GPIO PA0 working. And it wouldn't work, and it wouldn't work, and it wouldn't work. And it was because A0, as marked on the board, does not... It's not on port A0. It's not on PA0. A0 is on PA3. Yeah, that makes total sense.

It totally does not make sense. And I could have sworn I checked, but I think I checked in a different place and it was, oh my God. Oh, and then we've got this expensive J-Link, right? I think we've already complained about this, but yeah. I haven't complained about this on the show. Oh, I did note that the J-Link trace that I have, which was pretty expensive, doesn't get updates anymore. It was one year and done. Yeah, and now it doesn't work on anything.

Any newer ARM core that supports CoreSight 600 or whatever, it can't do. Yeah. So the cheapy J-Link base that we have works because it's newer but uh the expensive jtrace doesn't they've definitely taught me not to buy the expensive line um if they're only gonna if they're only gonna support it for a short time i don't know so all this stuff is just it's i guess the core frustration to me is they're just there's a lot of wasted effort and productivity loss because I feel like everybody's kind of, the chip vendors, the tools makers, they're all doing the same things over and over and over again.

Oh, it's a new chip. Time to write these new drivers. Oh, it's a new product. But their goal is to do as little as possible. Well, that's... And maximize their profit. They're doing an awful lot of little as possible. I mean, ST has a, you know, a small moon's worth of code and tools, right?

I mean, I don't know. I just, and I'm stuck, I'm doing this Marta stuff, and it's like, I just want to, it all comes down to what job am I trying to solve? And usually the client has some system they want to work. And you can't get to the system part until you do build a bunch of stuff that everyone is building all the time over and over again.

Except your building is slightly different than everybody else's building. And that's the mistake. Why? Nobody cares about a spy driver. Nobody cares about USB. They. Nobody cares about USB. They just want it to work. I don't want to learn about USB.

There is absolutely nothing that helps my existence by knowing more about USB. Well, except that you might be able to debug it when everything goes wrong. I'm not going to be able to debug USB. Yes, you are. Because we're going to go get a tool. No, the problem I have right now, nothing comes out of the wire. Well, I mean, I don't think that's... You can get all the tools you want, but that's what I'm talking about.

It's like, okay, I just want to send some bytes out USB. And no, you can't because some driver doesn't work that you didn't make. But that's because they want you to choose between the blah, blah, blah driver and the blee, blee, blee driver. No, I have no idea what the problem is. I'm just, the diversification is part of the problem. These are problems that developers who aren't doing embedded don't deal with. They have other problems, I agree. They have dependency issues and weird libraries and all this stuff,

and they have the OS vendors and their libraries changing and stuff. But I don't feel like there's a lot of people writing, not a lot of people writing an app for a phone who say well time to start an app so i gotta make my utility library so i gotta write my linked lists i've gotta write i've gotta write a you know a socket library so i can talk on the internet uh they've got to re-implement tls no none of these people are doing this they're just grabbing what's there. But that's why you use an operating system. Have you seen operating systems besides maybe Zephyr? I would admit I was thinking of Zephyr, yes.

I'm not sure about Zephyr. I haven't really played with it that much yet. I've heard really good things. But I mean, and that also depends on the chip vendor. And how well they support all of those little diversified little drivers. You might have Zephyr plus, minus, you know. I don't know. Things have gotten better, but some stuff is still stuck in 1995,

and it is just stupid. Our package management is terrible. What package management? Well, like how Kube has different examples and all of this diversification should be being handled in the same way package managers handle

different versions. ARM has its CMSIS pack thing. Which I just saw, just learned about. I think that's different. I think they reused that word in a way that was very confusing but i'm not sure whether i have no idea i think that is the part that the vendors are supposed to make so that they can work with different compilers oh no Is that the neural net PAC and the AI PAC and the DSP PAC?

Yeah. Okay. I don't know. It's just, yeah. Okay. So. You're not going to solve my problem for me. I'm not going to solve your problem for you.

Not even going to agree it's a problem. Oh, I agree. Oh. It sounded like you were dismissive. Oh, no, I I agree. Oh, it sounded like you're dismissive. Oh, no, I totally agree. I just, I think there are reasons for it, and that we can't solve the higher level thing until we understand and address the causes.

I just can't believe I had to write another layer on top of a USART driver. You were so mad that the camera went through the parallel interface when it was using serial? You were so mad. It was kind of hilarious. What are people doing? What are you doing? What are you doing?

Okay, let's see. There was something on the Patreon listener slack that seemed to make automated animated videos. I don't know what that is. So we would record a podcast. And in my head, not actuality, but in my head, we would be able to say, Christopher is an otter, Alicia is an axolotl. This is some AI thing. And it would automatically animate the video of us talking. Let's just go with yes.

But I'm building something here. Okay, go for it. What would you want to be if this was an animated video? I know I just seeded otter, but you could be anything. And as you think about this, I question whether this is a good lightning round question, except that I'd have to explain it all. Given how I feel right now, how about one of those really spiky sea urchins with a couple of googly eyes? Oh, nice. Nice.

Good. Do you think it's a good lightning round question? Sure. Yes. What would you want to be? I am still in an axolotl phase today. It's not a permanent thing. You're a temporary axolotl.

But, yeah. I don't. Okay. axolotl but um yeah i don't uh okay so what do you think is the best thing about 2023 best thing about 2023 um well all of its digits out up to seven which is a nice number okay yeah um if you Okay. Yeah. 2027, which has a corner. 2025, which has a point at the top, even though it's got the rounded bottom. Okay, moving on. Any good books, good shows, good media, good... Media? Am I a murder bot? I'm not prepared for this roundup.

You have new music out. No, we don't. Yes, you do. The Flav... Oh, yes. All right. I posted that on the newsletter. Didn't I talk about this?

If you did, it wasn't out. Yeah, no. I did contribute to somebody's record this year. I played bass on a experimental, kind of psychedelic record from a band called Flavagula. Maybe Flavigula? Flavagula? I don't know how to pronounce it. I admit now that I should have figured that out a long time ago.

Because you've been working on it for months and months. But I haven't talked to anybody in person. It's been all over email uh let me see uh it's a yellow-throated martin is what a flavagula is uh martez flavagula i'm gonna go with flavagula it's latin um yes It's an experimental band. The gentleman who I worked with is in Spain, and they released a record. It's called Nine-Sided Die. You can check it out on Bandcamp. If you look up flabagoola.bandcamp.com, I believe. I'm not sure if it's there yet or on their label. But it was really fun. It was a difficult challenge because it was a kind of music I don't usually do. Very long songs, very complex harmonically,

so changing chords and keys a lot, a lot of chromatic stuff, which means kind of moving outside of keys. So it was difficult to write parts for, and I had a lot of fun doing it. And yeah, so I'm on four of the songs on that record playing bass. It was a very long explanation, sorry. How'd you get hooked up with that? I asked on Mastodon if anybody wanted to collaborate

on drums or bass a long time ago. And then this guy contacted me and said, hey, I need some bass on this record. And I said, sure, I'll do one song. And then one song became two, and then two songs became three, and then three songs became four. Linear progression.

Yeah. Better than exponential in this case. And I got to use all of my basses. I got to use my standard four string and my fretless six string a little bit and my bass six a little bit and my uh bass six a little bit my pink guitar but not the upright not the upright no i considered it but there weren't really places for it uh so you're asking best of 2023 like i said i'm not really prepared for that um

uh i don't know like uh i just finished the latest Murderbot. I still adore Murderbot. And the idea that there's going to be a TV show? Yes. I am so in. I feel like I did kind of when the Star Wars Phantom Menace came out that it doesn't really matter

how good or bad it is. I'm going to love it anyway. Which has faded some. But I am very, very excited about Murderbot. What else have we seen and done? I don't know. It's been a weird year. Still quiet for us.

We're not going out much. Yeah. I may have to answer this question later after some thought. And alcohol? And alcohol. Yeah. I can't remember how long. Okay.

What do you think is going to be the best of 2024? Now he's just looking at me like, why are you torturing me? I can't even remember what actually happened. Now you're asking me to speculate about what might happen. The best of, like, It's too generic a question. Like the best stuff that's going to happen to us, the best stuff that's happening out there, the best content.

Content! I just use the content word even though I hate it. You know, you could answer whatever question you want. As a media mogul... Media mogul, yes. You should be able to. The lord of Chevy of Watsonville. You should be able to spin this to whatever direction you want.

I don't know. I don't know. I, yeah, this is terrible. It's a terrible podcast. We should not have done this today. I'm far too tired. I, of course, have a book coming out yes right

in March I feel like the book is already out I know because I have done a lot of work and it got sent to production and I will be giving a keynote at Embedded Online in April

I don't know what other conferences I'll be going to At Embedded Online. In April. In April. I don't know what other conferences I'll be going to because I'm still looking for only online. But with the book, it does mean I will be looking for more conferences to attend. As for other things, I have origami goals in 2024. I did see some really neat resolutions that I'm thinking about adding. Like 1080p? No. 4K UHD? I remember Bailey used to tell me about her resolution of reading one book from every hundred of the Dewey Decimal System. So like the ones and sixes are biographies and eights are science.

Just one of everything. I never learned that. I liked that one. I heard someone was making a resolution to eat at least 20 different shapes of pasta. It's made me think that maybe my resolutions need to be sillier. I don't have resolutions. I don't know if this is a federal crime, and I probably shouldn't admit it on the podcast, but if it is. Some of the sea glass we've picked up over the years at our local beach,

I gave back to the beach in what was a lot of fun knowing that people behind us were picking it up. And you also almost killed a seagull with it. I was throwing it back into the ocean. Which I'm sure that is a federal crime. I don't know why the seagull got it i didn't i was throwing it back which i'm sure that is a federal crime and i don't know why the seagull got in the way maybe it thought the glass was food i don't know um but yes more silly things um but i think my goal for for resolutions this year will be to come up with a few silly resolutions.

And not serious stuff. Might have a record from 12x7 coming out this year, but depends on if we get everything done. 2024? Yeah. Okay, because 2023 is rapidly... Sorry, we've moved on to 2024. Okay, cool. We've got

several songs in the can, one of which may or may not be allowed to be used, but we've got several songs in the can one of which may or may or may not be allowed to be used but we'll see it was a kickstarter reward for somebody and it turned out very well and uh but it was a song for them that they could use however they want it so that's true we're gonna ask him if we can put it on the record. I'm sure he will say yes, but it's a possibility he'll say no. After all. Then I'll just have to write it backwards. What else for 2024?

I don't know. I'm hoping to have a good relationship with work at some point in my life before retirement. So a lot of potential. I have a lot of contracts next year, which is surprising given last year. Yeah, I kind of hope we're over burnout because we've got a lot of work in the pipeline. We've got a dog this year.

That was pretty good for 2023. He's so cute. It's weird having a dog who actually wants to do the things we tell her instead of a dog who just kind of thinks we're a bother she'll get there she will i'm sure you again what is it you want fine i will eat my dinner see we're going to finish this podcast and I think of like five things I should have recommended from 2023. Well, let me go on to listener emails. All right.

First from Nelson Asanowski, aka The Prosaic Hacker. Aaron has bags of 8051s. Oh, yes. They have done a number of picking up CPUs from various places, scavenging

them from stockpiles. It's about 25,000 ICUs, mostly new old stock. ICUs, mostly new old stock. ICUs? ICs, sorry, ICs. Okay, sorry, I thought that was an acronym I didn't know.

No, no, no. It's a combination of IC and CPU. Gotcha. Yes, and so they were looking to offload them to somewhere that could use them. And they were thinking, makerspaces, having talked, I mean, we've talked about this on the Slack before, because we have a lot of stuff to get rid of. But not 25,000 loose tips. But makerspaces generally don't like having stuff given to them like that.

No. So. But one of our listeners might be like, yeah, that's something I can actually use. So, getting to the punchline, if you would like some of these, or all of them, or know someone

who can do something with them, email the show. And I will share the Google spreadsheet with you and put you in contact. Did I call him Aaron? His name is Nelson. Nelson Asanowski, the prosaic hacker. Okay, Nelson. And I will help you get in touch with Nelson. This isn't a

do you have a ST F4322? I need one of those. This is a, I like that sort of thing. And I want to collect more. And maybe I want to build several hundred retro packages for retro kits for making neat things. Which Nelson has already done, worked some with Ben Eder on the breadboard CPU,

so there's a good chance the chips work. They are in Montreal, Canada, so that may be an adventure. It's not smuggling if it's in low quantity, right? I think it's smuggling if it's not something illegal. Sure. Let's go with that. I'm pretty sure that's the definition of smuggling. Oh, I just don't know how you would move it from country to country if it's electronics.

In a box? Maybe in your socks. Anyway, we will hook you up with Nelson if you are truly interested. Unless your name is Peter, in which case, Peter, we need to have a talk about your hoarding tendencies. But after we have that talk, you can totally have all of them. Another email from Nathan Jones, who has been on the show, regarding our show with Ralph Hempel about Legos. Nathan is the head of Pass the Bricks. He collects Lego bricks from around his community and turns them into new sets for kids who don't have any.

And he would like to grow past the bricks worldwide. Nathan is? No. Nathan found it. Oh. We're having link problems here. Okay, so Nathan is not in charge of this. Nathan is instead telling us that said thing exists.

Oh, all right. That's good, too. Passthebricks.org. Passthebricks.org. Sure, if they said who founded that. It's in the San Francisco Bay Area. So it sounds like for donating stuff, that's easier if you're local to the Bay Area. But they have a newsletter and stuff, so. Must not look over at pile of Legos.

My Legos. They're your Legos. They're already assembled. Oh, okay. Well, I mean, there's like five. Yeah, I mean, there's some leftover bricks. I don't think they want 10 bricks. 10 bricks.

Okay. So bricks. Okay. So, we got those. Did we say what it does? Pass the bricks? Yeah. They collect Lego bricks, they clean them up, and they give them to kids who don't have any. Okay. I missed that last part.

I think you were focused on the fact that it wasn't who I thought it was. That's me. Which was totally valid. I'm focused on a lot of things I shouldn't be. No, somebody needs to think about, you know, details. That sound. We're just going to, you should cut out all the talking and just make the whole show the sound. What's next? What's next?

Do you think? Rarely. Do you think that when AIs become sentient, whatever that means to you, will that inevitably cause the singularity? Why? Why are you asking this? I don't know, because I was thinking about AI sentience and the Chevy dealership and retiring? Chevy dealership behavior would be the opposite of sentience. Becoming the pet of a nice robot. Oh, I see. No, I don't think AIs will become sentient in our lifetimes.

And if they do, I don't believe in the singularity. Do you think we can have the singularity without sentience? I don't think the singularity is a thing. Okay. That covers my questions for you. Unless you want to go back to the best of 2024 or 2023. Yeah, I don't know if I want to go back to it. I mean, I don't want to just give a list of movies and music and stuff.

That's silly. So, yeah, we can skip that. List of kits? Kits? You finished the ElectroBulb right away. Yeah. And you finished the Antares puzzle box right away. Well, that wasn't a kit.

That was a puzzle. I know, but it was fun to watch you. I have not finished my radio, which I need to finish. For talking to your dad. Yeah. But you did get some sort of network analyzer? Yeah, I got an antenna analyzer thingy, which I will use on the antenna if I ever get to that point.

And it turns out your brother is doing circuit design. Yeah, well, he's been doing pedals for a long time, guitar pedals for a long time. So he's getting more and more into building and designing circuits for that to make his own kind of custom pedals and things. So he's learning about electronics more than i've ever learned and he's using the digilent analog discovery that we got from digilent yeah we sent that to him um he's he's using it a lot and yeah including the network well he didn't know about the network analyzer

part so i told him about that uh where you basically can get like a transfer function of what goes in, what goes out. Yeah. Um, but he's got to put that hooked up to a Raspberry Pi and a monitor. So he's got this little basically self-contained, uh, setup,

which is pretty cool. Um, I don't think I realized that you could run, it runs waveforms. I think is the analog discovery app. I didn't think I realized you could run that on a Raspberry Pi. So that's kind of neat. You can do that and make a little appliance out of it instead of having it on your computer and fussing around.

I still think the Raspberry Pis are amazing for that. They're really cool. They're computers. Yeah. They're better than most of the computers. For 99% of our life, they're better than any computer we had. Not quite, but... Yeah, so he's having a lot of fun with that and sending me scope traces and things.

Look at the harmonic distortion when you turn the gain up here in these frequencies. I don't really know what's going on, but I'm sure it sounds cool. We should get him on the Slack. He and Tom Anderson could have their own channel talking about pedals and music. I don't know if we should get those two together. It'll be music and DEs. That's what we'll name the channel. He doesn't believe in math anymore because he did some characterizing of capacitors he has,

and they don't behave anything like it says that electronics should in the textbooks because they're real capacitors he has and they don't believe behave anything like it says that electronics should in the textbooks because they're real capacitors and once you actually put frequencies through them they do weird things and so he doesn't believe in math as opposed to electronics oh okay sorry sorry i was questioning why math was the culprit here oh because math is lying you know if you learn electronics math you know basic electronics math and you do all the stuff with capacitors and resistors and all these things, it does not talk about temperature dependence or frequency dependence that much until you get to way, way, way, way beyond basic electronics. Right? It's only because the first 45 pages are how not to lick things. What book are you reading? I don't want to talk about it.

What else? That's it for me. That's it? No, there's all this other stuff in here. Oh. We already talked about that stuff in previous episodes? No, we haven't. But someday we will talk about GDP and we will talk about compilers and things that we're not going to end up talking about today.

This episode, folks, but it's the end of the year. I still say you should just clip everything, but you sign in different ways. Exactly. It would be like five minutes long. Right, right, right. Well, thank you for co-hosting. Really low energy. Thank you for listening. Thank you to our Patreon subscribers

for their support. Thank you to our show sponsors this year, which has been really lovely. I'm not going to mention them specifically because it's not one where they're sponsoring directly, but it's been really, really nice.

And if you'd like to contact us, show at embedded.fm or hit the contact link on embedded.fm. Or go to the Chevy of Watsonville website, go to the chat box and ask for me directly. Let's see. Winnie the Pooh found out that it was Eeyore's birthday. And Eeyore said, look at the birthday cake, the candles and pink sugar. They didn't exist. But they didn't exist. And this distressed Pooh quite a bit. And so here we go.

This was too much for Pooh. Stay there, he called to Eeyore as he turned and hurried back home as quick as he could, for he felt he must get poor Eeyore a present of some sort at once, and he could always think of a proper one after. Outside his house he found Piglet, jumping up and down, trying to reach the knocker. Hello, Piglet. Hello, Pooh, said Piglet. What are you trying to do? I was trying to reach the knocker, said Piglet. I just came around. Let me do that for you, said Pooh kindly. So he reached up and knocked at the door. I have just seen Eeyore, he began. And poor Eeyore is in a very sad condition because it's his birthday and no one has taken any notice of it. He is very gloomy.

You know what Eeyore is. And there he was. And what a long time whoever lives here is answering this door. And he knocked again. But Pooh, said Piglet, it's your own house. Oh, said Pooh. So it is, he said. Well, let's go in. So they went, and the first thing Pooh did was go to the cupboard to see if he had quite a small jar of honey left, and he had, so he took it down. I'm giving this to Eeyore, he explained. As a present, what are you going to give? Couldn't I give it too, said Piglet, from both of us? No, said Pooh. That would not be a good plan. All right then,

I will give him a balloon. I've got one left from my party. I'll go get it now, shall I? That, Piglet, is a very good idea. It's just what Eeyore wants to cheer him up. Nobody can be uncheered with a balloon. So off Piglet trotted, and in the other direction went Pooh with his jar of honey. It was a warm day, and he had a long way to go. He hadn't gone more than halfway when a sort of funny feeling began to creep all over him. It began at the tip of his nose and trickled all throughout him at the soles of his feet,

as if it was just somebody inside him saying, Now then, Pooh, time for a little something. Dear, dear, said Pooh. I didn't know it was as late as that. So he sat down and took off the top of his jar of honey. Lucky I brought this with me, he thought. Many a bear going out on a warm day like this would never have thought of bringing a little something with him. And he began to eat.

Now, let me see, he thought as he took out his last lick of the jar. Where was I going? Ah, yes, Eeyore. He got up slowly, and then suddenly he remembered. He had eaten Eeyore's birthday present. Bother, said Pooh. What shall I do? I must give him something.

For a while, he couldn't think of anything. Then he thought, well, it's not a very nice pot, even if there's no honey in it, and I wanted, if I washed it clean and got somebody to write happy birthday on it, Eeyore could keep things in it, which might be useful. So as he was just passing the hundred-acre wood, he went off inside to call an owl who lived there.