Embedded - 520: All Sorts of Interesting Facts About Teeth

Episode Date: February 6, 2026

Chris and Elecia apologize, discuss uses and abuses of chatbots, reach out to an uncertain manager, try to help someone out of their professor's draconian rules, and extol the joys of reading.  Chabot Space & Science Center is in Oakland, CA, US. It is wonderful! Some suggestions for UncertainManager: Thanks for the Feedback: The Science and Art of Receiving Feedback Resilient Management Manager's Path Soft Skills Engineering podcast Hang in there! You are probably doing better than you think. Audio books are great! In the US, many libraries have digital libraries with extensive audio collections. There are several apps with different catalogs for the same library Libby, Kanopy, Hoopla, and Palace (check out the California shelf at Palace!). And since you are probably going to ask about the games Elecia doesn't play: Turing Complete shows how logic and logic gates work, building up a processor. Zachtronics' TIS-100 is another logic and processor design game. It is a little ugly in spots (too real world) but it is a really deep dive into learning assembly. It is the precursor to Shenzhen IO but harder to finish. Zachtronics' Shenzhen IO is about circuits and how they work . Human Resource Machine by Tomorrow Corporation is about optimizing resources, it turns out to be a lot like assembly programming. Should you have gotten here because you wanted facts about teeth, Elecia had been enjoying Bite: An Incisive History of Teeth, from Hagfish to Humans. Transcript

and welcome to Embedded. I am Elysio White here with Christopher White. It's going to be just us, and we're going to chat about AI management conferences. My two favorite topics.

My three favorite topics. Before we get really started, though, I do have to apologize the Shibot Science Center. Right, right, right. is not in Berkeley, it's in Oakland, and it's an important distinction. I feel like now I should actually look up to see what the actual name of the Shibou Thingy Bob is. The Shabot Thingy Bob.

Now we're going to get another email. I mean, I've been to the telescopes, I've been to the Science Museum. That's why I chose Science Center. It is the Shabot Space and Science Center, Ashoklyn's Place for Space. It's right in the name, Oakland's Place for Space. Okay. So, yes.

Berkeley has many great attractions, but Shabot is in Oakland. And for those of you in Ghana, the difference is about a mile. Ten miles and two hours of traffic. Between Oakland and Berkeley, it's about three feet. But it is different. It is different. Yes, yes, yes. Oakland doesn't get enough credit for how many cool things they have. That's fair.

Okay, so that said. Got the important stuff out of the way. We have listener emails. We have AI things to talk about. I don't know if he's going to really make it so that you can hear that kind of squeak, sigh, sad thing that he did. That's fine. We can talk about AI. I've been using AI more.

I'm ready to come out with this. I don't use it to code. I do use it to explain hard physics problems to myself when I have hard physics problems because I do a bad job of that. And it will even make me a diagram. And it will make equations and they'll be pretty and that'll make sense. And this is for like IAMU gimbal lock

or given a certain set of conditions or robots balancing kinematics. These are well-known equations. They're written in books. There are a bunch of books that are about this. If I had the right book, I would be happy. Oh, I did a PID anti-wind-up thing that then I didn't believe it, and I asked it where it got its information, and then I went to the book.

So I have been using AI for that. I've been doing it for a while because the physics stuff really bothers me that I can't get it and being able to ask it nine or ten times. You should go get a master's degree in physics because that's what happened to me. I was really bothered and I didn't understand physics. I just went and got a graduate degree, so I don't see what the problem is. That took you a long time and a lot of money. I mean, it didn't take that much money. Not a lot of money, but it was a long time.

And I always wanted, I want the solution to the problem I have in front of me. I don't. I'm just giving you a hard time. Okay, but then, but then, so I've been doing this for long enough that I'm happy with it. Then there were two other AI things I did. First was I heard someone talking about an AI code reviewer.

This isn't an automatic PR system. It isn't an AI widget that writes code and checks it in. This is an AI widget that reads your, PR, the diffs, and comments on it. And I was just like, I don't know. But then I give one of the people I work with far too many comments, and I feel a little bad. And maybe they would like it better if they got comments from somebody who wasn't me. And so I went into it thinking, okay, this is a tool that maybe I could give to another engineer

so that they wouldn't get comments from me. because I know that it irritates them. And then I ran it on my PRs because, you know, you've got to make sure that it does what you think, darn if it didn't find a PID bug that was in saturation, in wind up, and only when the integral term was negative, which I gather you're not supposed to do, but since I think of PID loops as it doesn't matter whether they're positive or negative, it's just math. Anyway, it found an actual bug that would have caused integral windup the thing that I was actually trying to avoid with this code.

It didn't tell to me in a very good way, but that's okay. It also told me things like the title of your PR sucks. I don't remember what it actually said, but. It was right. I had been using the, to get branch name, and I could, you know, an extra 30 seconds of making a title so that people didn't have to unparse whatever I meant to get to what I actually intended was good. It was stupidly useful and it didn't cause a lot of noise. And it was very frustrating that it worked so well. I have never said that it doesn't work. Do you have more to say? I have one more instance I used. So another coworker wrote a very dense piece of code

and then wrote documentation that was somehow even more dense. And I was starting to draw a diagram so that I would understand what the intention was. And instead I popped the description that I was having trouble reading into a clanker and asked it to generate a mermaid live diagram, a code for the diagram. I didn't even specify what kind of diagram I wanted.

I was like, if you want to do UML, you want to do sequencing, you want to do flowchart, I'm not even going to tell you. And the output was 60% useful. It was a really good place for me to start reading this difficult section of comments and sketching out what I would do, what I understood of it with this diagram it gave me and being able to put it all together. Diagram was not perfect. Diagram needed some adjustments,

and if the diagram gets used further, it will get better adjustments. But it was a good place to start. It would have taken me... I want to say, okay, so with this code and this document and understanding this code, if it's really important for me to understand, it's going to take me three passes. The first pass would have been the words to create a diagram. Second pass would have been to look at the diagram and understand what was going on. The third pass would have been finally getting around to the code.

I feel like it cut out the first one. I still have to go through and really understand what's happening, and that's still going to be hard. And I still have to review the code to make sure the code does what the intention was. And even the AI reviewer, I still have to go through the code to make sure that it was intention. The AI reviewer was like one more. It was like a lynch, but with algorithms.

Okay. You may now finish. Your sigh. No, I don't want to do that. I mean, that's fine. Everybody knows what I think. No, you also use... I have tried it this week for some Zephyr stuff,

and I realized at the end of it that I was angry, and that I had spent the time reading the docs, I would have known more about Zephyr and solved the problem in the same amount of time. Okay. Because it was wrong over and over and over again. Was it? Is it because Zephyr changes so fast?

No, it's just... If you don't know, the topic well, and I don't know Zephyr that well right now, having not used it in a while, you can't tell if it's wrong. Yes, yes. And so the point of it, doing stuff for you that you already know how to do escapes me a little bit. And I like learning things. And so I get a funny feeling when it does something that I could have learned because I've learned nothing at the end of those sessions.

when I try it. So, but that's my interaction with it. I feel weird talking to it. I don't like talking to it. I think the whole interaction model is not healthy. Well, then you get some weird interaction models with it. I told you about one of them. Being violent and...

Yeah, I'm not going to go into that. But it's difficult for me to separate that stuff with the political aspects of it, because the people who are making these things, the way they have made them, chat GPT is made by OpenAI, The president of OpenAI just gave $25 million to the GOP. Do I want to support that?

There's definitely a political ramification. Anthropics is all around talking about, oh, we're going to take away all the jobs, but also usher in a beautiful future somehow, which are totally in conflict. Google, I don't know what Google does, but they're off changing their motto away from what it used to be. Anyway, I do not feel clean after using them. And I know they're whizzy and fun. I think they probably do, I'm going to admit,

they save time on certain tasks if you can corral them. I think getting better at coding, and they can do some of that stuff, if you can corral them and if you can review what it's done. But I think they're dangerous. And while I still will attempt to use them from time to time, I, every time I do, I feel like I've done something wrong at the end of it. And that's a personal issue that's surrounded by a whole lot of things like the politics of it, the people who are pushing it.

The fact that it's the only tech product I can remember in history that has to be pushed on us. It is pushed so hard. And I don't like that. Every website I use now from Atlassian to QuickBooks to anything, has pop-ups constantly about have you tried to AI this? Would you like to do some AI? You never use AI in QuickBooks. It's so prone to error naturally. And I know people talk about the agenic stuff and the clod code and all of that that does huge amounts of tasks autonomously and stuff.

That scares me more. But yeah, I mean, I'm not going to win. I admit that right now. This stuff's not going away. I think these companies may go away. I think there's likely to be one or two kind of winners. But even that's going to be a commodity. And I think within 10 years, these are going to be models that you can download from, you know,

you know, random place and run on your phone that'll do stuff. So this stuff's not going away. People are going to continue to use it. I don't think I'm going to win the cultural battle. But, you know, I'm old enough that I only probably have to do this for another few more years and then I don't have to pay attention to computers. Well, the one, the code reviewer, I, that one is pay for. And we have a, you must get reviewed by two people.

we are not taking away that you must be reviewed by two people. So this is actually just adding work. That's fine. And if I hadn't been so freaking confident in my PID what it was telling me, if I had done what it was telling me, it would have all gone horribly wrong.

It was the only reason it was useful was because I was confident enough in what I was doing, to be able to say what the AI was telling me was wrong, but since it was telling me this, I needed to go look at something. I couldn't just follow what it said. I had to have enough strength of conviction

to say, I take your advice, I recognize that means there is something weird, and I will go look for it. Yes, and I do think that after years of using that, that is going to degrade. I think you were going to look. I think not you, but people who regularly use these things are going to degrade their skills. Not just their skills, but their ability to think.

They do have a deep distrust of them still. But that's going to go away. The whole Gemini make me a picture was pretty spiffy. Uh-huh. Partially because I was already in a very lazy, I don't want to draw a diagram for somebody else's coat. But you don't find the, the, the, training stuff compelling. That doesn't bother you. The political parts? Not the political parts. I mean, that's another question, but the how they've trained these. The fact that they used my book,

but they didn't pay me? Amongst other things. I do find that very unhappy making. I find a lot of that unhappy making, and I don't know how to balance it. I just feel like... I'm not sure that not using it makes it better for them to have ripped off my book. Doesn't make it better, but it means you're not supporting it. I'm not paying for most of this stuff. Yes, you are. In some manner. I mean, yes.

Google earns billions of dollars from something. That is true. If you are not paying for the service, you are the service. And in some of these, I am the service, yes. I don't know. We don't need to get into this further. I mean, yes. Like I said, I'm pretty much done with making my points about this.

No, no. I really, I know you think you've talked about it enough. And I definitely am reluctant to do some things. But I also see some engineers who use it all the time. And I don't know that that's a great idea because, as you say, it may make your skills worse. The thing is to me, these things have existed for all of two years, really, two, three years. and everyone acts like, wow, it's like I could never do anything, done anything before. And I've been doing this a long time.

And I don't feel like software quality is improving. I have no evidence the software quality is improving. Like everything I buy. You know in two years? Yeah. Yeah, from big companies like Apple and Google and stuff, everything seems to be regressing. So I don't know. Maybe there's huge performance benefits that are return on investments that we're not seeing.

But, you know, Alassian themselves sent me an ad because I just signed up with a new client for that last year. And it said 96% of companies and are users of AI see no return on investment. Learn how to be the 4% that do. And it's like, that's not a good advertisement. Wow. Okay. So you're just upfront telling me this doesn't really work. Because 4%.

Those are the number of people who can convince themselves of anything. Yeah, I don't know. I am trying very hard to get back into technology and being a productive worker. And yeah, it's still tough. Because everyone's talking about this stuff all the time. It gets a little bit much.

And I'm tired of hearing about success stories because the success stories to me don't matter. this thing could do everything for everyone and the ethics of it makes it tainted as far as I'm concerned. So you would go so far as to really just avoid it entirely because of the... I think if we summoned a demon to hell to drive us around just because we get driven around and it's fun doesn't mean that summoning a demon from hell was a good idea.

How do you balance that with trying it regularly? Because everybody's talking about it and I feel like if I haven't tried it once in a while, I am not equipped to talk about it. Okay. Cool. I know, listeners, you probably have your own opinions, whether you are... But if it went away, I would not care.

The physics thing, I really like the physics thing. The AI reviewer, yeah, that could be nice, but to me, that's basically smart lint. And if they called it smart lint, that would be 100% more chance to... Are you using a linter? I mean, it won't find the PID thing, certainly, but are you using linters for other things?

Oh, yeah, yeah, no. But it caught bigger things. Yeah, yeah. Like your description didn't match your code things, which would be nice. Look, if an LLM comes around that has been trained ethically, and they say,

we've trained on these things that are licensed and freely available, and here is the link to all the things we've trained on, so you can verify it. And, you know, it's trained to have guardrails and say, not generate reprehensible things, then I think that's more viable as a product. If people understand how to use them.

I really think the conversational interface is bad, like bad, bad. But I don't know how to get around that because these are language models. So that's another issue that I think is psychologically problematic. but I don't know how to get around that. The AI reviewer doesn't have, you don't talk to it in language, you ask for it to do a review and that's all you really get to say.

Okay. In fact, I really wanted to argue with it about my PID stuff. And then couldn't, which was fine. I had a test case. I had a unit test. I knew what I was doing. The physics,

I tend to give it, the same amount of languages you would find in the problem set of a terse math book. Again, I'm not criticizing your interaction. No, no, I'm just thinking. I'm thinking through things. The Gemini code to diagram, that one is the weakest of the cases, because what really should have happened there is I should have given it back to the original developer and said, go make me a diagram.

But I was torn between making a diagram myself because I thought that's what needed to happen next and just trying to power through with what I had, which I knew wasn't going to happen. Anyway, yes, varying uses. I'm still not happy with it coding for me. I still disbelieve what it tells me 95% of the time. And the other 5%, I am trying to use it to learn things, not to do my work for me. And I would say, that's fine. consider other ways to learn things for a moment before reaching for it.

Oh, I do. I mean, if I had a book instead that framed the problem the way I needed it framed, I would 100% use that. Anyway, why was this a topic? Did somebody ask? No. Okay. No, it's fine.

I thought maybe somebody asked us to talk about it again. No, I wanted to talk about it. Cool. I'm only two subjects. No, no, no. I thought I had seen somebody ask on Slack. And I couldn't remember if we needed to say who it was. No. We did have one thing that was actually, we didn't bring it up in the survey.

We didn't bring it up in the show about the survey. But it was from the survey. It led to a discussion on our Patreon Slack. And I think the person who initiated the discussion is not there. So it's kind of confusing. Anyway, one of the 2025 survey responses said, I recently became a manager as well as an individual contributor and I am terrible at it. And I think this was in the, do you have anything else who'd like to say? And Svec mentioned that he would offer to talk to that person and see if they want a listening ear or coaching or advice for free. So if that was you that

said that, you can contact him or email us and we'll contact him for you, forward your message. But then that led to a discussion on the Slack when I asked, what would you say to this uncertain manager? What advice would you give to somebody who freely admits I recently became a manager and I'm terrible at it? Well, I mean, it's difficult to have a conversation with a sentence. But I think the first thing, the first thing a new manager needs to do is spend at least two weeks not saying anything. And just listen to people, have one-on-ones, ask for what are your greatest concerns about

the organization, what's keeping you from getting work done, what's impacting your morale, questions like that plus how can I help and that can be both down to people who report to you and upward but then don't do anything just to gather that information over time get a sense for the people you work for and the people you work above you work for both in a certain sense but and then that's a good place to just start because if you don't understand the organization, you can't understand how to make organizational decisions. And the organization, not just structurally, but interpersonally, and the kinds of the way the work is done.

And that's step A, and a lot of people don't do that. Well, a lot of people... A lot of people have ideas. The person also did say that they were an individual. to a contributor, and it's a hard line to walk, manager and I see. Because that sitting back and just listening is much harder when you also are responsible. You do have the advantage of knowing what's going on at a kind of nuts and bolts level that you wouldn't as just a peer manager.

And I think we've both done that every time we've managed almost, right, where we've written code, but also. I mean, I've been a manager and only done management for a while. view, but I did get there by writing code initially. I don't think I ever stopped writing code. I'm not sure. No, I've stopped writing code. I want three, four direct reports and my ability to write code.

Oh, yeah, no, it's terrible, but sometimes somebody needs to jump in. Yeah. And then there's, I mean, I think don't overlook coaching and mentoring, like from people who have managed. And I started with books. Camille Fornier's manager's path takes you from being an engineer to a tech lead to a manager. It starts out with like what should be in a one-on-one, but not what should be in a one-on-one from a manager's perspective, but as a good engineer learning to better manage themselves. And so you just read to the chapter where you are or where you want to be and you don't have to read the whole book to get to what do you do to become.

a CTO. It's very tactical. It's very what you need now. Svec suggested Laura Hogan's resilient management. And then another person suggested the soft skills engineering podcast, since we know you like to listen to podcasts. Philip Johnston said that he thinks it makes sense to feel like you're terrible at doing something new and different. Yes. And there's a lot of potential in that feeling. I think we've all known many truly terrible managers who thought they were great at it,

and everyone else was the problem. So if you feel like you're terrible, you may actually be way, way ahead of where you think you are. At least you're willing to consider becoming better. But yes, and your own evaluation of your own terribleness is, as we all know, difficult to assess the accuracy of that. Well, and understanding that it is different from engineering, it is wildly a different skill set from engineering.

Why do we keep promoting engineers to be managers? Well, part of the reason is, I think, for like a first-line manager, it's difficult, well, okay, yeah, for any manager, it's difficult to communicate to, as you go further up the chart, the people get further and further away from engineering, right? And so you need somebody who can speak that language and explain the process of development to people who maybe don't understand it that well or don't appreciate why it's, you know, difficult to say how long something's going to take. How long will it take you to solve this bug? I don't know. That's not a satisfying answer. So

having people skills and management skills and communication skills, as well as technical skills. Understanding the technical stuff means that you can do the persuasion necessary to protect the people who work for you from having the director, the vice president sit down in their cube and say, so I understand that this is taking you a long time, but we really need to ship this while you're sitting there saying, well, there's something wrong with the laws of physics,

or there's something undocumented about this thing, and we haven't found it yet, or there's a memory leak, and it's one of those things. It takes three weeks or four weeks, and, you know, yeah, you know, you're sort of saying. I do. And experience is so helpful here, both having the hands-on engineering experience to be able to translate the bugs, the difficulties, the schedule parameters into actual things that can happen and things that other people, other parts of the organization can understand.

understand. That bridge, it's not easy. But as engineers, it is mostly the soft people skills that I found the hardest. Yeah, it is the hardest. Yeah, definitely. And there are look up scripts. Then look up management scripts. I'm not saying prescriptions and scripts as in acting scripts. when your engineer comes and says, I'm unhappy. There's actually a list of things you're supposed to do with that point. And you don't have to make them up for yourself.

It's not cheating to go on whatever website has these, Captain Arkward or one of the more professional faced script systems, script bloggers. It was, bloggers to start with who had these. And look up, okay, this person has just started crying in my office.

Do I need to do something? Or do I just accept that this is a lot of emotion and they're handling it? And I just pass over the tears and say, do you want a minute or continue? Do you want a minute or shall we just continue? And that's okay. You don't have to figure out how to deal with all of the, people things. Sometimes you can do it somewhat mechanically as long as you're kind of genuine about it. Yeah, I think that, I think, while it's difficult, because there is a lot of, I'm not going to say

dishonesty. I'm going to say, not manipulation, but like, there's political skills that are necessary. And that's where engineers tend to, I think, at least stereotypically fall down a little but, and getting those political skills is not easy. And being a manager and being a director around that level is very political. And, you know, people are vying for influence. They're teams against teams, even though that's not supposed to be the way things work. That's the way things do work. So there's a lot going on there.

And I remember I wasn't that great at it, the upward management stuff, because my instinct was to say no to a lot of stuff, because a lot of stuff was impossible, or they didn't understand what they were asking. and I wasn't good at kind of, how do I say no in a way that they think I said yes, which is what you've got to be able to do, right? And I would just outright say no. We're not doing that. And that doesn't go over well with the founder and CEO of the company.

I was so protective of my team, even if they didn't eat it. Yeah, yeah. Because you're part of that team, right? It's not just them, it's you too, right? I remember sitting in a group thing and it was a woman who had worked for me many years ago and several other women and the leader of whatever this workshop or meeting was

said, who here has ever gotten a raise that they haven't asked for? And of course, most of the hand stayed down. But of course the person who worked for me, her hand had to go up because I knew she'd gotten raised and asked for her. It was, and that was because I felt like I was part of that team, and I needed to make sure that team felt valued. Yeah, well, it was nice you had the authority to do that because many times I had to go to my people and say, well, you deserve much more than this, but they're giving you this, and there's nothing I can do about it.

Well, I mean, I still had to say that. here is your 1% increase and your four shares of stock thanks for all your hard work yeah no it's tough it's there's a lot going on with management writ large as a discipline that you know and most people are winging it there's no you can read about software development processes and stuff but I don't think people go out of the way to most managers don't go out of the way to learn the discipline of management aside from like MBA kind of of thing, which is not the same thing. And now, one of the things I liked about Fornier's manager's path was that it was specific

to engineering. Yeah. And she came from an engineering background. Yeah. And so the book she wrote was very technical focused. It didn't feel like how to live a happy life. It was more, here are some checkboxes. You don't have to check them all off all.

time, but check off most of them. Or check them all off eventually. Yeah. So Uncertain Manager, if you would like more discussion, feel free to contact us. If you are an uncertain manager and would like more discussion, I guess, the same. But also maybe join the Patreon Slack, which costs a buck, and then you can ask other people because there were a lot of good other suggestions. A lot of people who said, yeah, I feel that way too. Or I felt that way and it's gotten better. I'm still not comfortable

or even a lot of people saying, tell me about it. I can listen. So yeah, it's pretty cool. Oh, right. That led to the other thing. Did you read what I wrote about books recently? Let's just say I didn't. So back to the Slack. Sorry, a lot of stuff happens there. You don't have to join. We'll tell you some of the highlights. But one of the things I've been doing is I had this idea for the software daily devotional,

where every day it kind of reminds you to think about something software-ish. Like, don't repeat yourself or magic numbers are bad, something. And it turns out that I can. can't do this daily because I don't have that much time or that much willingness to write random things. But one of the things that I wrote recently was about books. And I know some people don't read. I love books. I have always been a reader. I look at a book and I just assume that inside there will be adventure and learning and characters and mind-bending things and it will be awesome. Christopher's looking at me like I'm crazy and yet yesterday I told him all sorts of interesting facts about teeth.

No, I'm just trying to think of, I'm trying to think of counter-example books to that because there's a lot of technical books where there's not a lot of adventure characters. That's all. That's all. That was what I was for. But you hand me a book and that's my starting place. Right. I love books in total. Yeah. And that's me. And I wasn't always sure.

I mean, it used to only be fiction books for me. But then I found a nonfiction book while we were on vacation. And I loved learning about ants, something that I really did not need to know about. And that kind of convinced me I code read the technical information if I was. was in the right headspace. But I still, you know, we read data-driven science and engineering last summer, some of us in the book club, and it was hard. And I usually spurn YouTube videos, but I needed them to get through the material,

and I took notes, and it was like a college class. I even, you know, I even got a math tutor. But I, this is from somebody who loves. books. And I know there are people who aren't readers or people who don't read technical books. Maybe they don't like the little letters that deans around or the physical stillness because it is hard to, I mean, I am happy to crash out on the couch with the book, but that's me. And there are many great technical audiobooks, but they're often a little harder than the non-technical audio books, just getting them or putting them in the right place or finding

the diagrams. Anyway, it got me to thinking about, I suggest books for just about everything. You come to me with a problem. Chances I say, have you read, is very high. But that's not a good solution for a lot of people. It's not even a great solution for me sometimes. Because there are books that I have completely bounced off of multiple times. So if someone's not a reader, what do you, how do those people learn?

I don't know. I'm sure they must. I find the highest density way of learning things is from reading to me. But you don't sit down with technical books. very often. Technical books, well, not recently, but I mean, if I have something I'm trying to learn, technically.

The problem is a lot of that information has gone to the web, but that's still reading. So, I mean, there's a distinction between people who don't read and people don't read books that I think is a little subtle

because it's impossible to be an engineer or whatever and not read. But there are a lot of people who get far more information from videos. That I don't understand. Or audio. The only information I prefer getting from video or audio is where somebody's demonstrating a technique, you know, musically or for a craft or for some skill.

But anything thinky? No, I hate it when I have to watch a video about math. Unless it's like a class, but then there's usually a book that goes on. You know, yeah, no, I'm sure there's, I'm sure some people have different skills with learning in different modalities. I don't learn well from video. And books, books I find really high density because if you can read quickly, then you can absorb a lot more information than somebody talking at this pace and conveying information at this pace for an hour. That hour, I mean, if you read the transcript of one of our podcasts, it takes. takes you about five minutes.

Takes me a lot longer than that, but then I'm also correcting them. But if you're reading it quickly and just, you know, five, ten minutes, that's a pretty big speed up. Depends on how fast you read. And people aren't listening to a podcast to read it. Obviously, they're listening to Us Talk, but, and I like podcasts too, but for a different thing, I don't listen to podcasts to learn about, you know, some learning happens. Although, okay, French is different because I'm trying to do immersion.

I'm trying to learn the length. which that's separate. I don't even care what they're talking about. But yeah, no, that's a tough question. So one of our listeners on the Slack said they're dyslexic. And I'm not going to name the names because I didn't ask for permission. But he said to the folks who love learning, he suggests they go out for a run while reading. It's not that he can't read or even doesn't enjoy reading it.

It's just swall. There is no accusation here that everyone should learn from books or reading. That is how I have learned to learn to learn. And having tried some other things, I'm not as good as the other things. And the same person says that they have unusual retention for technical information, that they almost never need to read or listen to material. So they can watch an hour-long video in 30 minutes at 2X and get way more information than 30 minutes of reading.

But there are other people where it's the opposite. I can't do that. 30 minutes reading a good book is way more useful and useful than watching the video. It's important to understand that other people retain information differently. I think we're overstating it too because there's combinations of sources, right? Like school. School is designed to be graduate, university. College.

Thank you. School is, you know, lecture plus homework plus reading. And those things all work together. I think the lecture is actually important part of that because it gives you the gaps between the reading or explain stuff in more depth slower with the reading or gives you repetition. that the reading might not give you. It's very difficult to take, I find it difficult to take a college textbook

without a course associated with it and learn as well as I could from having the course there. Because it dictates a pace. It means you're going to spend weeks on a chapter instead of reading a chapter in a day and thinking, why don't I understand this yet? But if you're capable of absorbing that information differently

from a different source and retaining it, that's great. I've never been able to do that. I have to have lots of repetition. and working with the material. I have to take notes. And I have to take notes. I have to take notes because that's another way

to seal it into long-term memory for me. I have had podcasts where they've gotten so detailed technically that I have gone and taken notes. And then, of course, I remember that. But if I don't, audio will go in one year and out the other. And the thing about notes, I've met people who say, oh, I never take notes. I hate it.

It's useless. I never go read them again. I never go read them again. The thing about notes is, the point is not to go read them again necessarily, although that can be useful when you're reviewing stuff. The point is the process of taking notes does something to your brain while you're learning. But I think that's partially the way we were trained to learn and partially the way we are naturally inclined to learn. Maybe, but since a great preponderance of people have settled on books and lectures and notes as learning, I think there's probably something fundamental there with very, And I have been thinking more about audiobooks.

I've been reading a lot, reading, listening to a lot more fiction audiobooks because my vision sucks. And I can read more things in other areas if I listen to my silly novels instead of using my eyes on them. Christopher is shaking his head like I can't do it audiobooks I realize after I started listening to them this is why I basically stop listening to audiobooks

I've given up I will drift off and a chapter will go by and I will realize I haven't heard a thing they've set in 25 minutes and have to rewind and I just can't do it

I can't the sounds just start to go and then I'm I'm gone. Yeah, I don't have the active... Do you fall asleep? No, I usually do it once doing some other chore or something. Thinking about something else? Maybe if I was lying down with my eyes closed or something.

I don't know if I'm folding clothes or doing some chore around the house, I'll start thinking about something else. But yeah, that's really tough for me. Podcasts, I don't have that problem, but podcasts are light. There's nothing. You know, if I miss a couple of minutes of somebody talking about something and really missed anything narratively, right? The narrative here is pretty loose. And I mention the French. I listen to a lot of French podcasts now, and those I'm actively listening.

Those are easier for me to keep up to not lose the threat of, because if I don't focus, there's somebody speaking something I don't understand, and that's very distracting. And so I have to have a little bit of focus to understand what's going on, and that locks me into it better. Maybe I should listen to audiobooks in French, but since that's rather difficult, it's a difficult language. in written. So anyway, yeah. Well, for people who want to try audiobooks, technical or semi-technical ones, there were a lot of suggestions about using library resources. There are a lot of online library resources like Hoopla and Canopy and Libby. And those all have audiobooks as well as digital books. So if you, I don't know if any of them have the manager's path, but since I keep,

suggesting that, I'm going to keep suggesting it. And for those of you who don't read, don't, there are lots of other options. As somebody who reads a lot, I don't look down on people who don't read. I understand. Some of this came up because I was talking to a manager and I suggested a book for one of the people they work with and the manager was like, I can't recommend a book unless I read it and I hate to read technical books like that. And I was like, I don't want to play technical video games, but I will 100% suggest Shenzhen, I owe to anybody wanting to learn assembly. There are a lot of games

I'm not interested in and yet I keep a list of things that CS majors might like. So even if you don't like books, you don't have to read everything. You can just say you heard it was good and pass it along. Yeah, the only people I will judge are people who don't like learning. Yeah, that's harder for me because I don't connect with that at all. Let's see. We got an email from a listener, Merco, who is studying Lexington. electrical engineering and using the TI MSP 430.

And the professor believes you should never use libraries you haven't written yourself. Cool. Yeah. The question is, how would you go about learning to use a new microcontroller with all of the hardware abstraction layers or with low-level? level drivers. Wait, with those libraries? Well, I think what, what Merko wants me to say is, of course you use the libraries. Of course you use the libraries. Registers are fun, get to know them in college, and then use the libraries. But I'm not going to say that your professor is wrong. Your professor is wrong. The professor, maybe should,

not use the libraries when you are learning to understand how processors work? Sure. That's a different thing. If that's what they said, then fine. Then that's fine. But never use the library. No, use the library. This is like having somebody say you should never eat food you didn't cook yourself. And maybe that's true for them. Maybe there are poisonings over the lifetime really are. You shouldn't drive a car you haven't built yourself. At a certain point, there are actually things that you aren't good at building. in software, even if you're a software engineer that you probably shouldn't build

that you should use other people's that they're experts in doing it. Hardware extraction layers is possibly one of those things because they're usually built by the company that made the processor. But yeah, no, if we want to advance technologically,

not reinventing everything every time you want to use it is one of the big speedups. And yet, if your professor says, We're not using the library in this class. That's fine. That's fine. That's fine. That's fine. That is different. That's different. That is different. If the professor said you should never use a library that you haven't written yourself in those exact words, then I think they might be questionable. But if the intent is we're going to build everything scratch to learn how to build things from scratch and get a better understanding, then I'm 100% behind that. Yeah.

The second part of the question was. Never wear clothes you haven't knitted yourself. You have to grow up too. You can only wear a hair. How do you choose a microprocessor? I don't. This is a similar to a question we got from Chris Gamal. How do you pull a trigger on recommending a different vendor's chip set

and hitting the port code Did we talk about that? We might have. You know, I keep this list and I never delete anything. I feel like we talked about that. We either talked about it on Slack or we talked about it on the show. All right.

So we'll go back to my email where we, how do you choose a processor? It's a really, really hard problem. Comparing flops or other metrics doesn't really work because pure speed is seldom the problem in embedded systems. And DMA capabilities may be the thing that makes it

so that your processing power doesn't matter. So if I need to choose a processor, one of the easiest ways is to look at what everybody else in the space is doing. Are you making a new Fitbit? Well, let's go open a Fitbit. I know they don't exist. It was a good example.

I think they exist. I don't really exist. Do you have primary periphery? for us. BLE is one. That's going to limit your processors. USB, Ethernet. I think you're starting too low level.

I mean, the first question is, what am I making? Yes, yes. I assume that part was defined, although that's a big assumption, one that's too often. What are I making here? What are the requirements? And then that's going to drive you

to a class of processors, and then you can go from that class of processors to manufacturers. Well, then do you have applications? Do you need an MMU? Right. Are you talking about Linux instead of an R-TOS? Yeah, yeah.

And then you will eventually get to what you were saying about, okay, what peripherals do we need and what kind of power profile? Well, peripherals are highest for me because you can't connect a motor with a KEP encoder to directly to something that speaks, I don't know, H-DMI. Right. Right. HTML. Your microcontroller is HTML, you pick the wrong one. That's what I'm saying.

Is there are... Yeah, I know. That's sort of defining who you're talking to. The requirements will get you to the block diagram of your system. Yeah. What are the inputs and outputs?

What is being controlled? That kind of stuff. Yeah. Yeah, I think we're in total agreement. It's just I tend to start a little more top down. I definitely start with the, what do I need to talk to? And what are the most important things I need to talk to?

If it's spy and I square T, that doesn't eliminate anybody. And how fast do I need to talk to them? Yeah. And our operating systems and the cost and the power, availability. How fast does it need to go? Is it 46 megahertz or 8 gigahertz? And that isn't a number I look at. That's not the critical number.

I want an 8 gigahertz quartet. XM. You say that. It surprised me if we get a one gigahertz soon or if it doesn't already exist. I don't know. I think there's pipeline things that you have to do before you can get beyond certain barriers, which I don't think those chips have yet. And then the final thing that Mirko asked about was talking about communities on

discords and in-person socializing events like make affairs and open sauce. But we end up talking about the ones that are in the U.S., and Mirko is not in the U.S. Did not want me to say where he was, but with less than 10 minutes of sleuthing, I came up with six conferences near to him that were mostly in the same country, a European country. And basically, if you look for your country and embedded, start there, that's not going to work for everybody. But...

Or whatever the word in the local languages. Don't assume there's nothing nearby. Because there often is. Hardware IO does go all over the world. The Open Source Summit is in Berlin next year. I think they were in Scotland last year. There are some very interesting conferences that go on in Japan and New Zealand and Australia. but we don't notice those because...

I'm not in Japan, Australia, Newtown. Yeah. And if I was in Japan, I would go to all of the origami conferences, of which there are many. But I don't even go to the U.S. origami conferences. What am I saying?

Boop, boop, boop. Oscar emailed, Hi, sorry for bothering you. And then Oscar goes on to say, really, really nice things about us, about how we're helpful, one of the few beacons in an otherwise dim industry, I think that's a bit much. And that they're happy that we have, that we find good people doing good things. And Oscar was surprised that we have a mission statement, as noted in, in the direct and sensible podcast about the, you know,

and review. But we do. We have the mission statement, and I appreciate it. And Krista has to live with it. While that covers my list, is there anything else you want to talk about? Oh, I just pushed a lot of buttons, and I don't know what they did. We could talk about that, yeah? Looks like it's still recording.

Yeah, I think we probably should. Looks like it's still recording. Do I sound like a squeaky mouse now? I don't think it can do that while recording. Okay, now I have two more. edit. Okay, all right. I think we're good.

Leaving this in. Okay. I've been doing origami. You've been doing French and drums and photography and developing your own film. Yes. But I've mostly been doing origami reading and playing with dog. Any other things you'd like to chat about? No, I'm tired.

All right. And I have a meeting. Because he has an actual client. They expect me to do things. Two, kind of. Two? Oh. Jojo is not my client.

Jojo's our dog, in case you wonder. Thanks, Jojo. All right, well then I'm going to sign off. Thank you for listening. If you'd like to join the Embedded Slack, support us on Patreon or coffee. And you'll get a link. If you would like to email us, show at embedded.fm. Or hit the contact link on Embedded FM.

And now some Winnie the Pooh. We are in the chapter in which Christopher Robin leads an expedition to the North Pole. Winnie the Pooh has just finished singing Sing Ho for the Life of a Bear. He was so pleased with the song that he sang it all the way through to the top of the forest. and if I go on singing it much longer, he thought, it will be time for the little something. And then that last line won't be true. So he turned it into a hum instead. Christopher Robin was sitting outside his door, putting on his big boots.

As soon as he saw the big boots, Pooh knew that an adventure was going to happen, and he brushed the honey off his nose with the back of his paw and spruced himself up as well as he could so he could look ready for anything. Good morning, Christopher Robin, he called out. Hello, Pooh, I can't get this boot on. That's bad, said Pooh. Do you think you could very kindly lean against me? Because I keep pulling so hard that I keep falling over backwards.

Pooh sat down and dug his feet in the ground and pushed hard against Christopher Robin's back. And Christopher Robin pushed hard against his and pulled in his boot until he had got it on. And that's that, said poo. What do we do next?