Embedded - 464: Please Make This Monster Look Scary
Episode Date: November 16, 2023Chris and Elecia talk about their favorite processors, their breakfast preferences, large language model ethics, presents, and Eeyore's birthday. Elecia’s new edition of her book Making Emb...edded Systems is finished! (Except for a couple months of tech reviews, updating, copyediting, and drawings.) It will be out in March. All of the back issues of Byte Magazine Chris’ radio kit that he mentioned but didn’t name is the QRP Labs QCX+ 5W CW Transceiver. Transcript Nordic Semiconductor empowers wireless innovation, by providing hardware, software, tools and services that allow developers to create the IoT products of tomorrow. Learn more about Nordic Semiconductor at nordicsemi.com, check out the DevAcademy at academy.nordicsemi.com and interact with the Nordic Devzone community at devzone.nordicsemi.com.
Transcript
Discussion (0)
Welcome to Impedded.
I am Alicia White, alongside Christopher White.
It's just us this week, and we have many things to talk about.
And we have a new co-host I'd like to introduce.
Say hello, Jojo.
Jojo is our new little dog. Who doesn't say much. Who doesn't
say much. This is not going to be a dog cast. Yes. Like the podcats. I promise. You posted on
LinkedIn that you were looking for a job and I have some questions about that. A contract. I have
a job. I am fully employed by Logical Elegance Incorporated.
And in fact, I'm an officer of said corporation.
So I'm not looking for a job.
I'm looking for clients of Logical Elegance.
Who which, to which I might write code.
At which.
For which.
Got it.
Sandwich.
Do you want a job with Can?
I would accept a job with can.
Do you want a job with lan?
I will not do a lan.
What about ham?
Wait, what?
As in green eggs and ham.
I was going for a whole poem here.
I heard han, and I thought that was something else I didn't know.
It says in the show notes, Chris is looking for a job in poem form.
I thought that we were headed to Dr. Seuss.
I just didn't know what hand was.
I heard hand.
I thought you were talking about a Fourier transform window. I will not process anything sold, sell anything bought or processed, process anything bought or sold.
Yeah, I think we're...
Okay.
Yes, I posted on LinkedIn.
I have had many responses.
Thank you to those responders.
And so you're probably not looking anymore?
I'm in talks with several entities.
So?
But if more entities would like to talk, I can talk with them.
Perhaps they would like to talk to you.
I'm pretty booked until January.
Perhaps they'd like to talk to Jojo.
Jojo would like any work that involves...
Hiding.
Hiding, yes.
Which, you know, to be fair, I share that.
I think JoJo would be a really good manager
as long as the answer is always
nothing.
Don't move.
Yeah.
Yes, so I'm on the contract train.
I expect to be
having way too much to do
in the near future.
That's okay.
I guess.
But yeah, it's probably time for me to start doing things.
The holidays are coming up.
So, you know, people won't be expecting quite as much.
I was going to intro this show telling you that I wasn't ready for summer to end.
And then you were going to tell me that fall is about to end
and that I'm an entire season behind.
I think maybe you're approaching two.
Of course, I don't really understand because as far as I can tell, Halloween was last week.
It sure feels like it.
Okay.
All right.
The other thing that needed to be talked about was the Nordic giveaway.
Right.
We gave away the Power Profiler Kit 2 yes to me wait what because i wanted it
and because i control the giveaways i'm pretty sure there's laws against this sort of thing
wow actually raul uh won the contest but the contest will be going on and I'm sure I can win next time.
Okay. So the way the contest is going to go on is we will have a new drawing every month for five more months.
Right.
So including November. And everybody who didn't win in the previous round
will roll over into the next round?
Yes. And the question is the same.
Okay.
What features are you most interested in?
Although after the third month,
I think we're switching to a different question.
Okay, so we'll probably clean the slate.
Clean the slate after two more tries.
Okay.
And we'll clean the slate and ask a new question.
Which means you'll have to re-enter.
Right. Okay. But until then, you you'll have to re-enter. Right.
Okay.
But until then, you don't have to re-enter.
We'll let everybody know.
And if you haven't entered, you feel free to enter for this month and next month.
Right.
Showitembedded.fm is the best way to enter.
Or hit the contact link on Embedded FM.
I'm almost ready to say it's the only way to enter, but, you know.
I mean, the entries through Squarespace comments or YouTube comments.
YouTube comments is pushing it.
Okay, but it's a little easier for us if you just hit email and then I can put you all into a folder.
I'm not.
And then I shout through the house, Christopher, give me a number between one and however many people have entered.
I don't really check the YouTube comments all that often.
I do get emails.
So it worked out this time, but email is so much easier.
Yes.
Okay.
Well, thank you for entering, everyone who entered, and we will have more.
And there'll be an ad later in the show, but we'll do a separate ad.
Perhaps right now.
The show is sponsored by Nordic Semiconductor this week.
Nordic is a market leader in IoT connectivity, providing hardware, software, tools a wide technology portfolio such as Bluetooth Low Energy, Low Energy Audio, Bluetooth Mesh, Thread, Matter, Cellular IoT, and Wi-Fi.
They have thousands of customers worldwide, with more than a 40% market share in Bluetooth Low Energy, nearly 2 million systems on a chip produced every day. I can see that
getting into the Nordic developer platform can be a little intimidating, but Nordic has their
Developer Academy, which is an online learning platform that equips developers to understand how
to build their IoT products with the Nordic solutions. They also have the DevZone, which
brings together tech support
and a community of customers to provide assistance, troubleshooting, and all kinds of information on
technical topics. Visit nordicsemi.com for more information. Visit academy.nordicsemi.com
and devzone.nordicsemi.com for more information about their learning platforms.
Okay, so now because people have given us money, we need to find ways to get rid of it.
I question your upbringing.
Okay, sure, sure.
Okay.
Not, you know, we can afford to pay for JoJo.
No, sure.
Ways to get rid of it.
Ways to get rid of the money.
Drop computer equipment on the ground.
No, no, no, no. I'm sorry.
Okay. Drop computer equipment on the ground. No, no, no, I'm sorry.
Okay, so I saw that somebody made an incredibly beautiful, colorful, adorable ruler out of PCB with red and green and white and light green and copper.
It's colored silkscreen, right?
Is that how they do it?
Well, I think they only had two colors of silkscreen and they used the board and the masking as the other colors.
All right.
Which was, I mean, it was really nice.
And so I was thinking, well, maybe we should do one
because we also got a request to do a badge after the super con badges um which is weird
because we don't go anywhere so why would we need anybody to have badges i don't think they're real
badges anymore i think i think once they're like a foot wide with a tft 15 inch display
and stereo speakers it ceases to be a badge. It's more of a plate armor.
And so between these two things,
I have been contemplating trying to make an embedded FM board.
An embedded FM board.
I polled the Patreon listener Slack group. Uh-huh.
And the highest priority, and possibly I did influence this some with my description, was that it be pretty.
And then after that, the next criteria was that it be a radio or that it be used for learning embedded.
And those were tied as criteria.
A radio?
And I looked up simple FM radio circuits, and that's pretty straightforward.
That's like 10 components. I looked up one of the STM32 disco boards, which is one I use in my classroom. And that was a lot of components,
in part because there's the whole programming subsystem.
I wondered about that as a, at least that would be useful.
I don't want to really make something that isn't useful.
I mean, a radio isn't useful.
I know, that's the problem.
Yeah, I mean, we could make a board that has an FRAM radio chip on it,
and you could use that as one of your things to program with.
I don't know. Yeah, program with. I don't know.
Yeah, I mean, I don't know.
Just another dead board is kind of weird.
I mean, if I could use it for teaching and stuff, it might be nice.
But the disco board I was looking at retails for $20.
Yeah.
And there's no way we would be able to make it for less than $40 because we wouldn't be making as many.
And once I made it pretty, we would be able to make it for less than 40 because we wouldn't be making as many and once I made it pretty it would be bigger ramify?
how do you use that as a verb?
I'm sure that it works here
the complexities ramify throughout?
sure, yes
one decision ramifies into further downstream decisions.
It's a very Neil Stephenson word.
Yeah.
Well, 40, did you actually go and kind of price a bomb out or?
No, I used it.
If it retails for 20 and the bomb in 10 KQ probably is five then, which means that in hundreds, it would probably be four to five times that.
I mean, we could mostly break even on it if it was just for fun, but there's a lot of work involved.
You're not just going to slap a board together.
Well, and I'm not going to be the one to be able to put together this board.
I mean, even with schematics from,
like, the disco boards available,
if I want to make it pretty,
I'm going to go talk to somebody who's done that a few times.
So, you know, I started to be excited about this,
and then I started to think about
another T-shirt design.
What about something smaller?
Like a ruler?
I mean, I think the dev board idea is more useful, but it's also...
Why?
Way more work and why.
So I think we should continue to think about it,
but maybe think about something that's more show-related,
maybe has some programmability,
but is like something that could run CircuitPython.
That would not be
as challenging, I think.
As a just
a, you know, STM32
with a bunch of stuff on it.
CircuitPython with some pretty lights
and some fun stuff, the logo.
Put the
FM radio on it and have it
do something with it. I don't know.
Let's put some more thought into it or have people,
if they have suggestions of what they'd like to see.
Because it'd be a fun little artifact from the show.
People like that stuff.
Yeah.
Circuit Python might be the way to go.
Then we don't have the programming subsystem.
But we do have USB.
And one of the things, I started talking to Carrie at Alpenglow.
Yeah.
And she pointed out that if you're making these, you're going to have to test them.
And there aren't going to be enough to make a testing jig.
Yeah.
So we're going to be programming these ourselves.
And now the amount of work starts to scale up.
I'm certainly not programming a bare-paddle disco board with anything.
Or testing it. So if you want to do CircuitPython,
that's pretty extractable. Even that, you still have to...
Anyway, it's a thought.
A ruler is still an option. That would be very simple. It would still be pretty.
But kind of useless.
But then the Patreon folks
did say that they wouldn't mind another ruler.
The other option
to take some of the work away from us
is to still do a pretty board,
still have something that does something,
but just ship it as a kit.
So we'd have to build one
to make sure the kit works,
maybe two, program it but otherwise
it's a soldering you know make sure we pick parts that are solderable and stuff but make a soldering
kit thing educational i mean that wouldn't work for probably your class stuff because you don't
want to make students actually build their own boards but um those could be separate option but
that would be something that could be like you could take that later and say, okay, send
this to an assembly house for class purposes.
I don't know.
That would take some of the labor out of it.
And, you know, if it doesn't work, then we'd have to interact with people and say, okay,
well, maybe you need a replacement, this, that, or the other thing.
But then there's also the idea of just doing another t-shirt design.
Yeah.
Which I would not do through Teespring again because they were not great.
Yeah.
Sorry.
So, yeah.
I don't want it to be like a teensy where we are in the board manufacturing.
Not unless for some weird reason we're suddenly super popular and we can just outsource the whole thing and live in the Bahamas as board mavens.
Of course, Sergio mentioned it had been a while since they last got an embedded FM sticker.
I asked if those were still a thing and should we use generative AI to design the next one.
Let me get that first question out of the way.
They are still a thing.
I meant to send some to Supercon for people to give out.
And then I didn't because I just slacked on it.
I have a box of stickers, and I should send them out.
I guess send me your mailing address, and I will make sticker Christmas cards.
Maybe that's the way to go.
Holiday cards.
Holiday cards. Winter cards. New Year's cards.
Solstice cards.
I mean, for the most part, I only have New Year's cards.
The second part of your question, should we use generative AI for this?
The answer to your question is no.
I'm going to go get a cup of coffee.
No, no, I'm done.
Oh, okay. The answer was no. is no. I'm going to go get a cup of coffee. No, no, I'm done. Oh, okay.
The answer was no.
Oh, just that.
I mean, I like our current design, so I don't know if we need a new one.
Sure, yeah.
But if we did, the answer would still be no?
Yeah, the answer would be no.
Unless you want to convince me.
Osau asked, how do you make an ethical choice of
LLM slash AI? Not the ethics of using
LLM? Large language model.
Not the ethics of using one, but
assuming you are going to use one, are there
ethical choices or ethical pathways you should follow in order to choose which one?
I'm not an expert on this, but I do know some things.
I think, so primarily what you're asking about is how the, it was two parts, but the first major part is what it has been trained on.
I think that's the major ethical issue,
just for general use,
that people might be concerned about,
apart from some downstream things.
And that's a bit of a challenge.
So the big ones have been trained on the wide internet.
So I think from OpenAI, ChatGPT
has been trained on
whatever is publicly available on the internet, whether it's
copywritten or not.
So Reddit,
publicly available images, stock images,
news images, people's
personal images, Twitter, whatever.
All that stuff. But also
Google has a bazillion
books. Bazillion books,
possibly.
Many of which are copyright.
Possibly.
Artwork,
you know,
DeviantArt is a big,
big independent art site,
and I think they've all trolled through there.
So,
it's difficult.
So the big ones,
I think BARD is Google's.
I think that is similarly trained on whatever Google can see.
ChatGPT
is obviously trained on
a lot. There are some smaller ones, and
I don't have them off the top of my
head that are supposed to be
ethically trained,
where they have
either found
things that are obviously public domain
to train on, or gotten people's permission for large sets.
I think Adobe has been trying to do that.
So I think the stuff that's built into Photoshop, I think Firefly is what it's called.
Don't quote me on that.
I think that has been trained on things that are licensable so that there are fewer copyright issues.
And I know there's a couple of smaller open source kind of chat GPT things,
large language models that are claiming to be trained on things that are permissible.
You're just going to have to look and see what details you have
about the particular model that you're looking at using.
And it's different between ChatGPT and, say, the image models, although those are starting to cross-pollinate.
So ChatGPT and DALI are now kind of merged, so you can do image stuff back and forth with text stuff and have conversations.
So that's doubly confusing as to what to do there.
So it's not really a great scene.
I think as time goes on,
more things that will be clear about their training
will emerge,
and it will be easier to make a choice
that you can feel better about
on the upstream side, on the training side.
On the downstream side,
there's still the issues about what are you displacing?
Are you really solving a problem
or are you playing with a fun toy?
There was something else about...
Yeah.
So there's a lot of difficult kind of things
that go into the choices right now.
I use some of this stuff.
I've been using OpenAI has a model called Whisper, which is for transcription,
which I've been experimenting with for the podcast.
It's not quite there yet.
Mostly because it doesn't recognize different speakers. So it just gives you this wall of text of everybody speaking.
And there are some apps that you can drop in all the separate
audio files and it'll transcribe them separately, but it's not quite working right yet.
That's very useful. But that's separate than the transcripts we pay for
and are available, and those are all clean and
speakers are identified.
Yeah. What do you think about this question? and are available and those are all clean and, and speakers are identified.
Yeah. Yeah. Yeah. What, what do you think about this question?
Well, I mean,
it's kind of fun because you get pretty irritated anytime anybody brings this up.
So there's some, I can talk reasonably about it. I get irritated when,
I get irritated when people don't think about the usage and are just evangelists for it.
I wanted to write a joke with it. that your posts to LinkedIn be in Seuss form or be in the form of a dating profile where you were looking for a device to date,
not date, to work on.
Uh-huh.
And these were jokes that I just wanted to think about.
I didn't want to fully finish.
But you've kind of convinced me
that letting it write my jokes is probably not that good for me or for the world. Um, and that
part of that is because I saw, uh, from three brown, one blue or three blue, one brown. I never remember which way that goes, but the YouTube people with the math. The drilling is really important and that mathematicians and scientists and physicists
through history, the people we really like look up to, Newton and folks, drilled throughout their lives. They did problems. They did problems
they knew the answer to, and they practiced. And that practice ends up building your intuition.
When we think about solving problems with analytics and intuition,
you don't get that without all the drills.
Or if you do, you're super lucky.
But I need to see a problem a few times before I can really, you know,
ace it off the bat.
And I have to admit, looking at students in class,
I feel like I'm constantly being quizzed on things.
Oh, yeah, yeah.
But I think that makes me better because I am constantly being drilled on embedded topics.
Like, why would you do this versus that?
And why do the bit masks work this way?
And spy versus I2C.
I am never going to get those two confused i
know some people will but i will always know when to use those i squared c is the crappy one
that is not true but go ahead um they have their purposes no no it's fine i just never get it right
the first time whereas with spy i almost always get it right the second time at least. me to write my jokes, I kind of wonder if I do fall into the habit of letting it do
that, if I get out of practice myself.
I think that's certainly something I haven't necessarily...
And it's something I enjoy, except when I kind of just half-ass it like I did with
you should do that in a Seuss poem.
That's something that's getting on the edges of something that bothers me about the LLMs.
And probably would bother me about any AI that purports to converse and think, or not think, but, you know, that we can have conversations with, that we can treat as a sidekick.
And has knowledge enough to do many, many tasks.
One of the things is what you're saying, like getting out of practice of mundane stuff.
Now, plenty of technology gets rid of mundane stuff for us.
Calculators.
Right, right.
And I'm not going to argue that necessarily being everyone needs to be super fast at arithmetic in a world with calculators.
That's definitely a benefit.
But it sure helps for estimation.
But I wonder about communication and writing.
See, that's the thing.
I was hoping somebody would write a thank you letter.
And I'm like, okay, chat GPT is the way I would go with this.
But maybe that's because that's not a natural skill that I am good at.
Polite talk.
Yeah, yeah.
So one of the things I said on Mastodon a couple days ago that I've been thinking about,
and it was triggered by somebody who was doing some writing,
and they threw their writing at chatGPT to have it proofread
and check on some things and stuff.
I'll just say what I wrote.
I have trouble articulating a lot of my qualms about large language models
and where they're headed, but one of them is that it's going to be so easy
to just grab one and have it work with you on writing or brainstorming
or getting a second opinion on something,
rather than reaching out to someone and having a chat.
Especially for people who, the latter,
reaching out to someone takes a bit of activation energy,
interrooted people.
Some people will probably say that's good,
that you have this thing to bounce ideas off of
or to check you at a moment's notice.
You're not wasting anybody else's time.
Okay, so first of all, I'm interested in why you think that would be a waste of somebody's
time.
If somebody asked me to proofread something, I would say yes pretty much immediately.
Wouldn't you?
Sweet.
I have many things for you.
Do you understand what I'm saying, though?
I do, but...
I mean, writing is a human process.
Communication is a human process.
And I feel like if we're...
I just worry about things.
I worry that reducing opportunities for human interaction
and improvement of communication and going,
replacing most of that with communicating with something that is the average of the internet
is not a net benefit because everything will be dulled down. It's like, you know, if you had
Clippy, right? And Clippy was real good and you let it just
you know and people would react negatively that because clippy was bad but
uh how much would you let microsoft word write for you if you knew it was microsoft word really
i mean for cover letters no things that i don't care about? Fine. You know, I'm not the writing police, but certainly boilerplate legal kind of documents are fine, whatever.
My book?
No.
Those have been templatized forever, right?
You could have gone before ChatGPT existed to Nolo Press or wherever and found, give me a template for a business letter complaining about this.
That's fine.
But for a piece of professional writing or a piece of fiction or your book,
you don't think your book would be worsened by running it through this thing?
Yes, it would be. Because, I mean, from the sidebar on how we still use modems to the detailed factual information that has been checked by experts, no, means they're going to be worse writers because one of the ways to learn to write is to read other people's writing and find out what's wrong with it because it's hard to do that for your own work.
And people outsourcing other stuff are going to get out of practice of the mundane writing, but the mundane writing is what keeps you sharp. There's a lot of emails we write and stuff.
I'm getting emails, I'm sure, written by a chat GPT now
because they just have that kind of fragrance.
Well, but those are spam and junk emails.
Some of them, but some of them I think are on LinkedIn.
LinkedIn now, I just saw while I was there,
has a button you can press to do all of your outreach stuff using an LLM.
It'll write all your...
Yes, but those totally feel like this person hasn't bothered to read my profile.
For now.
Anyway, I don't want to go on too long about this because, you know, I get off track and people know how I feel.
But, I mean, having said all this, a few people seem to think I'm anti-AI.
First of all, I hate the term AI because I think it's completely whitewashing all of this.
I prefer the term ML because that covers a lot more things that aren't necessarily neural
networks. It covers types of neural networks that do very useful things in limited scoped ways.
Anyway, I do not hate AI and ML. I, in fact, have one and possibly two clients
where I'm working on ML stuff. So I am familiar with it. I know how it works.
And my opinions are probably verging on philosophical at this point and not technical.
Okay, I have some lightning round questions. Are you ready?
It's a little late in the show for that, but okay.
Simon, finish one gig or start a dozen?
At this rate, I'm starting a dozen.
I currently have four open tasks.
Between teaching class, finishing the book,
and two clients,
I have a lot.
I'm going from
zero or maybe.25 clients
to possibly three.
Svek would like to know
pancakes or waffles?
Ooh.
Well, wait, wait.
Which kind of waffles?
I believe you can choose your own.
All right.
I'm going to admit something here.
I eat a lot of toaster waffles.
It doesn't necessarily mean I like them better than pancakes,
but they're easier to make.
The problem with pancakes,
I'm having an intense sense of deja vu.
Did we not have a very long discussion about waffles and pancakes
on this very show at some point in the last five years? I don't know. Somebody find that out. The problem
with pancakes is they're delicious for three bites. And then suddenly, for me at least,
it feels like I've eaten that spray foam that you use to seal your house with. And then it comes out
real small, but then it expands to 18 times its size so so i have trouble with pancakes from a from a from a from a mass mass to uh to enjoyment ratio standpoint and your
toaster waffles don't expand at all because they have like negative calories oh they've got calories
they don't have very many calories like you know normal but if okay belgian waffle final answer You know, normal. Okay, Belgian waffle.
Final answer.
I would say pancakes, whether you are talking about the food or the style of holding hands.
The what?
Moving on.
Excuse me?
The style of holding hands?
Yes.
If you hold hands with somebody and you hold hands and your palms are together and your fingers, your fingers touch your fingers, that's pancake hand style.
If you do waffles, you interlace them.
And that's waffle style of holding hands.
I've never in my life heard this.
And waffle style is okay unless your partner's hands are much bigger than yours in which
case waffle styles can be uncomfortable but pancake is worse if you're sweaty handed i'm learning so
much from this podcast i really i really think you know people people really we're really helping
people okay one more silly one and then we'll actually do some embedded stuff. Butter. Salted or
not? I mean,
for just eating,
salted. What are you, a lunatic?
Straight out of the carton.
For baking, unsalted.
Didn't we...
I saw Paul Hollywood
just answer this question.
For unsalted, if you're baking,
you can control the amount of salt in the recipe
if you use unsalted.
Otherwise, you don't know how much salt you're adding.
Okay, as the person who bakes,
we always use salted because it lasts so much longer.
Well, there's that too.
And then you just taste to figure out
how much salt you need to add.
But I know that that's not the correct way of baking,
but it's so much easier. Yeah, well, I'm not going to taste raw cake better. I know, and that's not the correct way of baking, but it's so much easier.
Yeah, well, I'm not going to taste raw cake better.
I know, and that's much to your detriment.
No, it isn't.
Okay, I promised.
I knew it.
If you could only use one microcontroller for the rest of your life, Pedro would like to know which one and why.
The basic stamp. That's not true. I know, one and why. The basic stamp.
That's not true.
No, I just thought I'd be contrary.
One microcontroller for the rest of my life.
How long am I going to live?
Let's just get the parameters defined here.
Let's say at least another 10 years.
All right. I mean, you know,
nobody ever got fired for using an STM32F4 something.
Yeah, I have an STM32F4
and an L4.
I know those are kind of different,
but they're not that different.
I mean, you can do an awful lot with one of those.
And the ecosystem's there and it's not going to go away.
And it's well documented and everybody's used it.
I mean, that's kind of the boring mainstream.
And I've seen some M55s and some M85s coming out,
but I'm not very familiar with the specs on those
other than they go faster.
Now I'm thinking, I don't know.
I don't know.
The problem is I don't have any particular,
everybody wants me to use something different
when I show up at a client,
so I don't really form attachments.
I know the ones that I enjoy using.
Yeah, you know, having said that, I'm like,
but you have the Nordic Thingy 52 on your desk
and you're playing with it and having a good time
and you don't even remember what chip is inside there half the time.
And then...
So, okay.
So, yeah, I'm going to switch it around.
I'm going to say the easiest thing to use
for a wide variety of applications,
and that's probably not an STM,
because you have to have a lot of ancillary stuff with a lot of those.
So something Nordic that's got Wi-Fi and Bluetooth.
It would depend on what I wanted to work on.
But that's the thing.
I mean, if I have to choose one microcontroller to use for the rest of my life,
something's gone wrong,
and therefore, I don't think I want to work on
embedded systems anymore. I'm probably digging a moat.
Okay, yes.
Yeah, and
I mean,
I don't know that I
would use...
I think about projects I would do.
I wouldn't use either one of those to do a big robotics project.
I would actually use something Linux-based.
If I had to do a small robotics project with lots of motor control,
I would go to the TI Piccolo line.
If I had to do Bluetooth, I would go to the TI Piccolo line.
If I had to do Bluetooth,
I would use Nordic unless somebody really made me use TI.
Not that TI is bad,
just Nordic's a lot easier.
And that includes Bluetooth thread matter.
Well, you don't have to do...
Zigbee, whatever.
They're difficult.
Nice.
Okay, split the difference. Zigbee whatever's. They're difficult. Nice. Okay.
Split the difference.
A bag of 80-50 ones.
Final answer.
Thank you.
I'm sure that is your final answer.
Is there a contract or job that you didn't take that you wish you had and why?
Oh, gosh.
I kind of wish I had gotten hired by Fitbit
when there were only a couple of engineers there
before Heiko was the manager.
Not that that would have been bad,
just that I could have been hired then
when there were only a couple of engineers
and then I would have gotten lots of stock options.
Oh, I see.
I don't think there's any jobs I gave up that were financial screw-ups.
There's definitely jobs I took that were financial screw-ups.
Let me think.
There were too many jobs that I didn't end up taking.
Contracts, no. too many jobs that I didn't end up taking contracts. No, every contract I've turned down
has been the correct thing to do in hindsight. I mean, they usually go off and you never hear
from them again. Right. But sometimes I check back or hear things from people who's too,
you know, sometimes I, you know, was in not a formal group of people, but I knew people who also were contracting at those places and heard later about things that happened.
No, I've been pretty lucky or good about choosing, so there's no ones that got away, I don't think.
No. Yeah, I agree. yeah yeah i agree certainly been things that might have been lucrative that people had expressed
interest but never got to the point of like me interviewing or anything because i just didn't
want to do it so there's a lot that i nipped in the bud before it even got to the stage of
interviews or anything yeah because there's a lot of things you don't want to work on
exactly because you have the luxury of not having to work
on them yeah okay let's see uh the next question was from ambad ambad at embedded uh who asked the
favorite jack of all trades micro or a favorite specific task micro i feel like we've kind of
covered that yeah uh there's a follow-up question from same person is the arduino uno now obsolete
oh my god it was obsolete when it came out i mean it depends on your your your definition of
obsolete i mean if you've got one it can do things there's plenty you can do still uh so
but even when it came out it was quite expensive for what it could do. Yeah, that's the thing.
I mean, it's a really old processor.
There's a lot you can get for the same or less price now that can do a lot more.
Do they still sell them?
Oh, yeah, they still sell them.
Oh, okay.
But on the other hand, they are the easiest, lowest common denominator.
Almost everyone can get a compiler
to run on their system.
That's the thing about this stuff. That's why I'm saying
it's hard to define obsolete.
Computers get obsolete
because the software stops
being supported.
A laptop from 2005
may not even be able to surf the web
anymore because it can't handle
it doesn't have an you can't upgrade it to an operating system that can even do HTTPS or something, right?
That's not the case with embedded stuff.
Like, that's why 8051s are still being used in modern things, right?
Or derivatives of 8051s on FPGAs or things.
So I don't think it's obsolete as long as it does what you want it to. or derivatives of 8051s on FPGAs or things.
So I don't think it's obsolete as long as it does what you want it to.
But again, there's other things I would pay more for unless you had a really good reason to pay.
There's other things I would, for equivalent price,
would buy instead of a Nuno for hobbyist stuff before I'd reach for a Nuno.
Circuit Python would be much higher on my list.
Yeah, Circuit Python or, you know, and that doesn't even dictate what board it is, right?
You could have a Raspberry Pi 2040 or, you know, teen CDs.
Aderfruit has a Circuit Python little itty-bitty chip on a ruler that cracks me up.
I have one on my desk.
Yeah, I mean, but if it works, then you've got them.
So talking about historical,
I went and somebody linked to the Byte magazine back issues.
Oh, neat.
And I looked at the first one from 1975
and went through a couple of them
kind of thinking it would be fun to do a live
I didn't tweet, live mastodoning. That doesn't sound right, no.
Live tooting sounds even worse.
But reading those and commenting, you know, trying to read a magazine
a day or a week and commenting.
And it was kind of amusing.
I did it a little bit for the Patreon Slack group.
And just, you know, kind of making fun of the idea of needing to repurpose keyboards because they weren't standard yet
and so you'd have to change them around.
But then there were other articles
that were like how to write your own assembly
with the idea of how to understand assembly
as you write it, like what is the piece.
That article could be published today
and would be just as useful.
That's because embedded is still quite... Stuck in 1975.
Still somewhat backwards. But now we have new keyboards,
so it's okay.
You don't even have a big clacky one.
No. That's what all the
coffee achiever people type on.
So you can hear them from 10
miles away.
Mechanical keyboard people,
you're fine. Sorry, I'm not really mad at you.
It's just that he had to share
an office with me for so long that he knows
that if anything makes noise,
it's my opportunity to turn around
and glare at it. Yeah, I had to type on a pillow.
Little letters that I sharpied on.
Keyboard was inside there.
It mostly lined up.
Okay.
So let's see what else we have next.
I recall an episode where you talked about installing an antenna to talk to your dad.
How is that going?
You know what?
It's going.
So we always knew it was going to take us a while because we
both had to learn Morse code and we both had to get radio stuff and learn how that works. So
I have gotten Morse code to the point where I'm bad at it at about 10 words per minute,
which is better than when I started, which was I wasn't able to do anything. So I can send okay.
Receiving is still challenging, so I'm working on that.
I think my dad has gotten somewhat more proficient than me.
He has a radio and he's been testing it.
He's set up his antenna and he's receiving stuff.
I have a radio kit. He has the same radio I do, but he's receiving stuff. I have a radio kit.
He has the same radio I do, but he got it assembled.
I got the kit.
And I started it this week and I have wound a grand total of one toroid.
How many do you have to wind?
I think there's three or four, but this is a hard one because it's got multiple winds.
So it's got 38 winds with two ends and then five and then five and then five.
So it's teensy.
Like it's the size of a dime, so it's kind of fiddly with all the magnum.
I'm just nodding.
Anyway, what?
Do you know what teensy means?
No, no, it's not the teensy part that I'm nodding.
It's the winding and the fives and the 38.
Five wines.
Five wines. It's just winding. You count them. That's a lot of wine. Where's the winding and the fives and the 38 lines and the lines it's just winding you count them
that's a lot of wine where's the cheese anyway so that's going to take me at least a week or two
to finish and then um yeah i'm gonna set it up and see if i can hear anything and we'll proceed
from there but so yeah i don't know you have the wire the antenna, but we haven't set it up yet.
No, I have.
Yeah, I have the antenna.
And I'm going to probably put it in the front yard.
Oh, I thought.
Oh, good. That's better than my idea of getting a crossbow and firing it into one of the taller trees.
You know my track record with projectile weapons.
I was going to get to fire the crossbow.
Projectile weapons and technology and trees.
Yeah, the last time I used a projectile weapon
in the hopes of doing something technology-related,
I ended up in the ER.
Yeah, so, yeah, the front yard, because he's east.
Okay.
So I need to have it, and if I put it back here, the house, I'd have to get it pretty high up to clear the house.
So if I do it in the front yard, I don't have to clear the house.
So it'll be oriented this way because it radiates.
We should put it along the path where the squirrels are.
Maybe we can talk to them.
Yes, I'm certain they know Morse code too.
We can build a squirrel army. So yeah,
the project is going, I'm hoping in the next month or two to try it out. The nice thing about,
you know, trying to talk to someone that you have an out-of-bound communication with is we can kind
of pick a frequency while texting to each other or even on the phone and say, okay, did you hear, did you hear my CQ?
Um, so we'll know if it's, it's working. We're not going to just kind of
randomly try at random times and see if anybody's listening. So, um, yeah, the hard part about Morse, uh, the hard part about Morse that I've talked about before is I thought I got to the
point where I could understand the letters, but once real words were coming past it was harder to harder to copy what was being being said and the just the the
patterns of letters when it's random they tend not to uh have a lot of the quick letters
together as often as as real words with you know because e-I-E-I-O sounds a lot like S-T-Q.
Right. No, no Q.
Not W.
Yeah, exactly.
And common in English, common English words have the common letters frequently, right?
And the common frequent letters are very short in Morse code for obvious reasons.
But if you're not trained to hear them go by real quick, you get lost.
So there's that problem.
And then there's all these...
I almost said something that would get ham radio people mad at me.
But I can't hear them because I don't have a radio.
Salacious. All of these salacious.
I can't hear them because I don't have a radio.
There's all these abbreviations for stuff.
Like CQ?
Well, there's that.
I mean, there's the standard HAM abbreviations,
the Q codes and stuff.
CW.
Those are all, there's a few of those
and most of them are fine.
But there's all these.
Never mind, go on.
I don't know any of them.
I'm just randomly now putting letters together.
There's a mode in the app I'm using to learn which is called qso bot and it's a little
it's not it's not a chat bot but it's a it's a bot where you morse code at it and it morse
codes back at you in kind of a simulation of a normal code conversation and i i got one back
from it and it's just all random letters.
I'm like, this is supposed to be English.
And in the app is a little glossary, it turns out, of common Morse code abbreviations of just normal words.
They're not like specific ham stuff.
Oh. So it was like, you know, R for some word, you know, FR for from, and, you know, so all these common words
are just shrunk to, and I don't know any of them. So I'm sitting here copying this going,
what is this thing saying to me? And it's like, oh, how's the weather out there? And it was like,
you know, H, WX, RR, you know, so. And I don't think we have to use that with my dad,
but, you know, if I am going to talk to anybody else,
I don't really know how to prepare for that.
I don't, it's hard enough to learn Morse.
I don't want to learn somebody's...
Language.
...stenography, you know, so.
I don't know how often that stuff is actually used,
probably pretty commonly because it's, you know,
Morse is extremely slow.
So obviously abbreviating things makes sense,
but I don't know any of those.
So I may just have to print out a thing and a big poster.
But thank you for asking.
It is continuing slowly, as most of my projects do.
Let's see.
I should mention the book,
because I told O'Reilly that I had a podcast with listeners who might be all been super nice and very helpful and have found things.
And have often said, maybe you should show them work here instead of just jumping from here's what it is, here's the equation, here's the answer.
Like actually show it out.
So thank you to the tech reviewers.
When the book comes out, we will be having some giveaways.
There will be some discounts for patrons or newsletter subscribers or people who ask for them.
I don't know.
But that's all in the March timeframe.
It's still on early release on the O'Reilly Learning System.
I can get a 30-day subscription for that for anybody who wants it.
But again, it'd be better if you're actually interested in my book instead of just the O'Reilly Learning System.
You might want to wait until March when it goes from having diagrams
that I have scribbled out to professional diagrams.
Oh, scribble is the cool part.
There is the one that says, please make this monster look scary.
And the monster I drew is just not scary.
It's not even attempting to be scary.
So yeah, that's closer to done. And the tech review stage is hard because
I know that these people are helping me and they're really nice people. And they've,
I mean, they're going above and beyond. And yet there's still always the temptation to argue.
You know, like, I totally disagree with you because I wrote it that way for a reason. And that's not really true. I just wrote it that way
because I didn't think of what they were saying. It's like all ego in code reviews. You kind of
have to get around it and realize that it doesn't... They're not attacking me. They're trying to help me make it better.
And they are succeeding when I let them. And then class part is doing their asynchronous cohort,
which has turned out to be kind of weird and not much less work than a regular cohort so that part wasn't successful um i don't know
what's gonna happen there i don't know if we're gonna do another one um stay tuned i guess stay
tuned i guess but don't hold your breath uh for the book i am starting to look for conferences to present at.
Of course, not very many in person or mostly local to the Bay Area if it is in person.
And being paid is really nice and keynote positions are really spiffy and all that.
But I am working on a couple of pretty cool topics.
You should go on some podcasts.
I should.
Actually, actually.
I hear people like podcasts.
Did you do that on purpose?
Of course I did.
Do you know what podcast I'm on?
Yeah, the Building Something, the Builders Quadrangle.
The Builders Circle.
I was close.
I did a podcast with Circle. I was close. I was on a podcast devoted to people with startups and how to make their startups work.
And she's a mechanical engineer and she's talked to a lot of electrical engineers, but she didn't have a lot of firmware experience or really understand
the whole process. So it's like I talked for an hour, hour and a half solid very quickly
in order to like download everything. So yes, if you would like to hear more of me talking about
embedded systems, which nominally is what this show is supposed to be about.
Oh, we didn't talk about any of it today.
Please check out some other show.
Please check out The Builder's Circle.
And there'll be a link in the show notes, of course.
Okay, cool.
The episode doesn't come out until the 28th.
Yeah, I think that's right.
Yeah.
What else is going on?
There's other stuff.
There must be other stuff.
Does there have to be other stuff?
Yeah, there should be other stuff.
Because then if there isn't other stuff, I have to go do some work.
Oh, I see.
Or exercise.
Actually, given the time, we both need exercise.
Okay.
Well.
And since we got a dog who is actually a lump of coal, we can't take her for a walk because she's too afraid. Instead, we will have to exercise boring ways.
I'm taking a picture of the dog in the podcast zone here.
I see.
All right. Well, I'm sorry to everyone.
Thank you to Chrissy for producing and co-hosting.
What did you call me?
Thank you to Christopher.
Oh, okay. I heard Christer.
Thank you to Christopher for producing and co-hosting.
Thank you to Nordic for their sponsorship.
Thank you to our Patreon listener Slack group for their support and for their questions.
You can enter the contest
or talk to us at showatembedded.fm
or hit the contact link on Embedded FM.
You can do a comment,
but we don't respond as quickly.
We might miss them.
We might miss them.
Especially on YouTube.
If you need pictures of our dog, let us know.
If you would like to imagine what a picture is, all you have to do is go outside,
look at the sun for about one second with just one eye,
and then wherever you look after that, you will see our dog.
It's kind of like a void in space.
Okay.
And now for some Win the poo let's see let's see let's see we had the the nothing poo bear nothing we can't all and some of us don't that's all there is to it
and then uh poo hummed and sang Coddle Stone Pie.
All right, here we are.
That's right, said Eeyore, saying umpty tiddly umpty too.
Here we go, gathering nuts and may.
Enjoy yourself.
I am, said Pooh.
Some can, said Eeyore.
Why, what's the matter? Is anything the matter?
You seem so sad, Eeyore. Sad? Why should I be sad? It's my birthday, the happiest day of the year. Your birthday, said Pooh in great surprise.
Of course it is. Can't you see?
Look at all the presents I have had.
He waved a foot from side to side.
Look at the birthday cake, candles and pink sugar.
Pooh looked, first to the right and then to the left.
Presents, said Pooh. Birthday cake, said Pooh looked, first to the right and then to the left. Presents, said Pooh. Birthday cake,
said Pooh. Where? Can't you see them? No, said Pooh. Neither can I, said Eeyore.
Joke, he explained. Ha ha. Pooh scratched his head, being a little puzzled by this all.
"'But is it really your birthday?' he asked.
"'It is.' "'Oh, well, many happy returns of the day, Eeyore.
"'And many happy returns to you, Poo Bear.
"'But it isn't my birthday.'
"'No, it's mine.
But you said many happy returns.
Well, why not?
You don't always want to be miserable on my birthday, do you?
Oh, I see, said Pooh.
It's bad enough, said Eeyore, almost breaking down.
"'Being miserable myself, what with no presents and no cake and no candles "'and no proper notice of me at all.
"'But if everyone is going to be miserable, too.'
"'This was too much for Pooh.
"'Stay there,' he called to Eeyore as he turned and hurried back home,
"'as quickly as he could, for he felt he must
get Eeyore present of some sort at once, and he could always think of a proper one after.