Chris on AI, autonomous swarming, home automation and Rust!

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Jerod (00:04):
Welcome to the Practical AI podcast, where we break down
the real world applications ofartificial intelligence and how
it's shaping the way we live,work, and create. Our goal is to
help make AI technologypractical, productive, and
accessible to everyone. Whetheryou're a developer, business
leader, or just curious aboutthe tech behind the buzz, you're

(00:24):
in the right place. Be sure toconnect with us on LinkedIn, X,
or Blue Sky to stay up to datewith episode drops, behind the
scenes content, and AI insights.You can learn more at
practicalai.fm.
Now onto the show.

Chris (00:48):
This is Chris Benson, co host of the Practical AI Pod
cast. I was recently invited bymy good friends, Jared Santo and
Adam Stokoviak to be their gueston the Chainplog Podcast, which
is one of the most popular opensource and software development
pod in the world. Thoughindependent now, Practical AI
used to be part of the Changelogfamily of podcasts and we remain

(01:10):
very close. We have reproducedthat Changelog episode to be
this Practical AI episode. Andin this episode, we're gonna
cover a wide variety of topics.
AI, drones, robotics, swarming,home automation, and the Rust
programming language. I hope youenjoy our conversation as much
as I did.

Jerod (01:31):
Today, have Chris Benson, Practical AI co host and
longtime friend. Welcome to theshow, Chris.

Chris (01:38):
Hey. Thanks a lot. It's great to be on the on the guest
side of the equation here.

Jerod (01:42):
Yeah. You've been interviewing folks for a long
time, but now you, sir, arebeing interviewed, so to speak.

Chris (01:47):
Mhmm. Indeed.

Jerod (01:48):
Does that make you nervous?

Chris (01:50):
Well, I got it. You know, you guys taught me everything I
know. So like, yeah, a littlebit. It's kinda like,

Jerod (01:55):
We have back and forth tricks. We're gonna unleash them
on you on this shit.

Chris (01:58):
Oh my God. Okay. But yeah, like, you know, you guys
were the OG originals. Daniel,Whitenack and I learned
everything we know from youguys.

Jerod (02:09):
Well, you guys are good at what you do. So I'll take
that as a compliment. Yeah.

Chris (02:13):
Well, thank you.

Adam (02:14):
What's funny is how back well, how far back we go. I
think there's some context togive here. Jerry, just for an
exercise here, I went andsearched the name Benson because
Chris's last name is BensonCorrect. In my calendar just to
see if the history was there.And literally, April 3 at
10:30AM, Chris Benson on Skype.

(02:37):
That's how far that's how

Jerod (02:38):
far What year?

Adam (02:39):
What's the year? 2018. Did I say the year? My bad.

Jerod (02:42):
No, you didn't.

Adam (02:42):
04/03/2018 Chris Benson 10:30AM Skype. That's what
you're doing.

Chris (02:47):
That was way back when we use Skype, you know.

Jerod (02:50):
That's right. We had to.

Adam (02:51):
That was so wild.

Jerod (02:52):
It was our only option.

Adam (02:53):
And that's what the that was the original conversation
that started the host co hostPractical AI. I think it was a
data show back then even. I'mnot even sure if it had a name.

Jerod (03:02):
It didn't have a name yet.

Adam (03:04):
The beginnings of Practical AI and and this long
history of relationship.

Chris (03:09):
It was funny because I know I had reached out to you
guys. Then so you guys had gotime, and there was this
changelog family that wasalready there. I wasn't part of
it yet, but Daniel Whitenack andI were both the data AI people

(03:30):
in the Go community at the time.And so I was thinking, I was
listening to Changelog andstuff, and thinking, boy, maybe
these guys need to start an AIfocused podcast or something.
And I was thinking, but I'd liketo do that.
But I was thinking, but I needsomebody to do it with. And I

(03:51):
was thinking, I got to reach outto Daniel. He's the other AI
data. So I reached out to Danieland he's like, Oh, by the way, I
just started talking to Jaredand Adam about this. I was like,
Perfect.
I just sent them a message. Sothe timing, yeah, it all just
came together. The timing wasperfect.

Jerod (04:07):
You guys were so far ahead of the curve.

Chris (04:09):
Yeah, well, was very clear. If you were really
plugged into the AI world atthat point, it was very clear
that this was going. Where itwas going, things change all the
time. But it was very clear bythat time that the gas pedal was
on and sky was the limit andthere was some kind of journey

(04:30):
ahead. And at that point, Danieland I, we wanted to be steering
that journey for everybody.
And that was how you know, andyou guys were awesome in terms
of saying, this would befantastic. And we'd love to do
it. And that was back in 2018.And here we are in 2025, late
twenty twenty five.

Jerod (04:48):
Yeah, things have changed, but have stayed the
same as well. Here's a funnystory that you might not know,
Chris. I've given you credit forthis before, but I don't think I
ever told you this, which is atsome point, the four of us were
on a call and this is like postlaunching Practical AI, but pre
ChatGPT moment. And you werelamenting that we like missed

(05:09):
NVIDIA or something like you youwe were talking about the run
up. I think NVIDIA had just hada huge run up with regards to
first it was gaming, but thenalso, you know, machine learning
was kind of starting to takeoff.
And you were like, man, I can'tbelieve like, look at Nvidia,
it's crazy the hockey stickgrowth on that stock. You're
like, but we're too late now,we're too late. And this is

(05:30):
2019. I was

Chris (05:32):
so wrong.

Jerod (05:32):
Yeah, here's

Adam (05:33):
the funny

Jerod (05:34):
part Chris, I thought to myself, are we though? I said, I
was like, are we? And I actuallyleft that call and I went and I
bought a little bit of Nvidiastock thinking, if Chris thinks
we're too late, this guy'salways ahead of everything. So I
think he's ahead. So I have tothank you for a stock tip that
has paid off nicely.

Chris (05:50):
You're welcome, you buy contrary to my advice, that's
probably probably the

Jerod (05:57):
Right, so I need to talk to you more often and kind of do
the opposite thing.

Chris (06:00):
Yeah, there you go.

Jerod (06:01):
So yeah, thanks for that. That was cool. Unfortunately, I
didn't buy enough to like justquit everything else and retire.
You know, I'm happy that youthought we missed it.

Chris (06:11):
I'm glad I was wrong on that. They've done amazing
things. I think it's kind offunny, just in AI in general, AI
has been around at some level.Even the modern form of AI has
been around for decades. It'snot a recent thing, because I
got introduced to it by myparents who were actually

(06:33):
technical people, Georgia Techand Lockheed and things like
that.
And they were doing stuff backin the late '80s and early '90s
and stuff. And my dad introducedme to neural networks, which is
still the basis of all thisstuff, in 1992. And I think it
was funny. The tie in here toNVIDIA is we went through

(06:56):
another AI winter. There's beena series of where everyone gave
up on AI for a little while andthen circled back around.
They're called AI winters. Andso the last AI winter kind of
happened at the end of the '90sgoing into the 2000s there for a
few years before the modern era,if you will, picked up. But I
think the difference is that thenotion of modeling and the

(07:19):
software basis of AI was there,and there were a lot of great
ideas, and a lot of the stuffwe're doing today originated
back then conceptually, but wedidn't have the hardware. We
couldn't actually do the thing.We didn't have these GPUs and
now other types of chips thatenabled all this to happen.
And so it was really like thehardware side of things had to

(07:42):
catch up so that the softwareAnd when people say, Well, why
did we have an AI winter? And Ithink to a large degree, it
wasn't the lack of amazingbrainpower to solve these
problems and create the models.It was the fact that you didn't
have the hardware infrastructureto do the things that people
were envisioning were possible.And it wasn't until NVIDIA came
along and became really the AIhardware company. I know they do

(08:05):
a lot of software stuff, butthat made the difference.
And Google came along eventuallywith TPUs and lots of other
players jumped in, but bothsides had to be there. So a
little journey down memory lanethere. It's the benefit of being
old.

Jerod (08:23):
You've seen it all, Chris. You have seen it all.
I've been around. I'm old asdirt. From your purview, this is
not stock advice, but from yourpurview, here at the 2025, and
you have Nvidia, you have AMD,you have Google, you have Meta,
you have these large playersmaking huge investments.
OpenAI, of course. I mean, thelist goes on and on and on.

(08:46):
Which single entity do you thinkis best positioned to like
succeed over the next ten years?If you had to pick one of the
top contenders like, is itGoogle? They seem like they've
really turned the corner, butI'm not sure if their capital
investment on their owninfrastructure is going to be
the big win that some people aresaying that I don't know.

Adam (09:03):
What do you think?

Chris (09:04):
I think there are so I'm going cheat a little bit. I
don't really have a one. For along time, people would say
OpenAI. And before that, weresaying Google. There is a top
group, and they are certainlydoing well.
And I think at the risk ofgetting slightly in terms of

(09:25):
social issues, there's growinginequality between those group
of haves and a lot of othersthat are have nots in that way.
But I know one of the things Ithink is that I really think
that open models are becomingincreasingly important because
the difference If you go back afew years and it wasn't coming

(09:48):
out of OpenAI, there was a bigperformance difference in what
you were able to do. And if youlook at the closing of the gap
between what's possible I mean,are millions of open models out
there, and there are hundreds ofthem that are kind of like they
are nipping at the heels of theleading ones. That gap between

(10:12):
the latest, greatest thing fromone of these big name companies
and what's possible in the openworld has narrowed dramatically.
And what that's really doing ispushing model creation into
something of a commodity area.
And I think you've seen that interms of what some of these big
companies, they've builtservices, and they're building

(10:34):
separate businesses, and they'regoing into verticals and things
like that. But that's becausejust the model generation is not
going to be the profitable thingfor years and years going
forward. And so they're turningfrom being AI providers
explicitly into AI serviceproviders now that are specific

(10:55):
to different types ofbusinesses. And I think they'll
do quite well. I don't know.
I'm afraid, especially afterpointing out my horrendous- Oh,

Jerod (11:04):
you drilled it last time.

Chris (11:05):
Yeah, I was gonna say after my horrendous NVIDIA
prediction, the last thing I'mgonna go do is pick a winner
here. But yeah, I mean, they'remaking a lot of money, by by
pivoting, you know, within thescope of what they do, and they
have that the expertise. And Imean, like meta or as we're
talking now, meta is just like,just purely buying the AI

(11:26):
talent. I don't care whatGoogle's gonna pay you. I'm
gonna pay you 10 times more, andthere's no way you're gonna go
any place but us.
And trying to kind of catch upto that OpenAI, which is still,
as we speak, probably still ofthe gold standard there. But
with a few others such asGoogle,

Adam (11:45):
as

Chris (11:45):
you mentioned, and several others that are kind of
nipping at the heels there. Soit's interesting times.

Adam (11:51):
So a long winded answer is OpenAI? Is that what you're
saying?

Jerod (11:56):
You have to go back and analyze what Chris said and
tease out the truth of it.

Chris (12:00):
Oh, I tried to escape that. Adam, that was not fair. I
worked really hard for fiveminutes to kind of square my way
out of your question there. Yes.So very close.
When you say OpenAI, very close.You got the word open, right?
How's that?

Adam (12:13):
Okay. All right.

Jerod (12:13):
Chris's answer is the open models will catch will
commoditize the frontier model,so to speak. And these people
that are just buying all theGPUs and just training,
training, training, and then ofcourse inference as well.

Chris (12:27):
I mean, what you can do, it's requiring you, we're seeing
this progression where we'rebuilding out frontier models is
costing less money. There's aton of money in some of them,
but the efficiencies that arenow built into training from
some of the latest research hasmade it to where you can build
some amazing stuff with notquite as much as you might have

(12:47):
expected a year or two ago interms of relative performance
against the hardware that

Jerod (12:52):
you need

Chris (12:52):
to support that. So it might be, who knows, where the
research is taking.

Jerod (12:57):
Is there such thing as peak parameter? I mean, think I
read that XAI's next modelcoming out whenever is going to
have a trillion parameters orsomething. And it's like, how
large is large too large? Or isthere no such thing?

Chris (13:12):
So, yeah. I mean, one of the things that we've we've been
talking about for a while now isthe fact that like, it used to
be in the early days of the GBTseries from OpenAI that you saw
distinct capability differencesas you went from three to 3.5
and to four and that kind ofstuff. But there's also been
We've seen kind of plateau. It'salmost like you're seeing that a

(13:34):
lot of the It's not just a modelthing, but also some of the
infrastructure that's beingbuilt around it has made it much
more accessible in terms of itsproductivity and its usefulness,
and there's less of a frictionwhen we're trying to use models
at this point. So I do thinkthat there is no infinite rides
on terms of the number ofparameters you have to do.

(13:56):
I think that that does levelout. And also, if you're going
to have that many parameters,being able to use that
productively from an inferencestandpoint, the world is turning
out to be a mini model worldinstead of a giant model world.
And I'm not sure that a lot ofpeople in the general public

(14:16):
that aren't people like us thatfollow this closely really
realize that. I think when theythink AI, they're thinking chat
GPT, because it's what theyknow. And one model to rule them
all, one model to bind us.
And I'm not at all That's notwhat I think is the world. I
think the world is many, manymodels contribute to solving a
problem in various ways. Here weare in 2025 deeply into the age

(14:45):
of agents. So it's no longerjust models, but now agents with
models that are acting on yourbehalf. And I think the reality
is it's a mini it's a mini agentfuture that we're talking about
here.

Adam (14:58):
Before we go there, I gotta ask you, because we're
talking about companies andpredictions and potential here.
Have you tapped into or heard ofthe next Jeff Bezos thing
Prometheus and the startup he'scheering, co founding, etcetera?
Are you are you tapped intothat?

Chris (15:14):
I'm not I'm not up to date on the details.

Adam (15:15):
That's like

Jerod (15:16):
hot off the press, isn't it? They announced

Adam (15:17):
Yesterday's news basically. Today, today's news.

Jerod (15:21):
I think there's like a perpetual Bezos Musk pissing
contest that goes on. And thisseems like the next one. He's
like, you have x AI? I've gotthis thing.

Adam (15:29):
According to TechCrunch, Jeff Bezos reportedly returns to
the trenches as co CEO of new AIstarter Prometheus. Project
Prometheus. So he hasn't doneanything from a CEO aside from
shareholder, you know, etceterabehind Amazon. He's been just,

(15:50):
you know, getting swellessentially. Just getting swell
and going to space.

Jerod (15:54):
On his yacht.

Adam (15:55):
Yeah. Yeah. As you would if you were

Jerod (15:57):
But he's been doing the space stuff. He's been doing
Blue Origin.

Adam (15:59):
That's what I said, like, getting swell and going to
space. Oh, that's what he's beendoing. Yeah. Yeah. So this is
kind of cool that I suppose thenext big thing could be from
him.
So maybe the next time we talkChris, you can give us your non
prediction prediction.

Chris (16:15):
I can slide out of that one too.

Jerod (16:16):
Yeah. Do we go by Amazon right now? That's what I wanna
know, Chris.

Chris (16:20):
So I'm probably the wrong person to talk to about this,
not only because of theprediction that we just talked
about, but also I wanna pointout, am honestly This may sound
really counterproductive asPractical AI co host on this,
but I think Daniel's the sameway. We're less interested in

(16:41):
the big, big names coming outwith their latest big things,
because there's so much amazingwork being done by real people
out there. Take that, Bezos.There's the chat Yeah, plastic
Jeff Bezos, like, Hi, I'm JeffBezos, and Elon Musk and all
these guys. I'm just like,they're always one upping each

(17:03):
other and they do some bigthings.
But I think 99% of the press isgoing to these people.

Jerod (17:12):
Six people.

Chris (17:12):
But I think 99% of the real productive work in AI is
going to all these invisiblemasses of amazing people that
are doing this stuff every day.And if I could get the
mainstream press to refocus, I'dbe like, look around. There's

(17:34):
just astounding, amazing thingsthat are happening, but they're
not happening by these famousfigures. And these guys, yes,
they have tons of money, they'resuper, super ultra wealthy
beyond imagination. And they canthrow their money around and
stuff.
But you kind of mentioned, it'skind of the pissing contest, for
instance, between some of them.

Adam (17:54):
And

Chris (17:54):
I just like, there's so much cool stuff out there that's
the latest, you know, the latestBezos, know, Yeah, Elon massive
thing.

Adam (18:05):
I mean, 6,200,000,000 behind this thing is

Jerod (18:09):
quite It's like crazy.

Adam (18:09):
Quite an investment in there that he's raised for. It's
$6,200,000,000.

Jerod (18:14):
What are they doing?

Adam (18:15):
What's their deal? It's only speculative at this point.
It's only got a name. ProjectPrometheus, Jeff Bezos, co
founder I believe is Vic. Wouldonly mess up the last name.
B a j a j is the last name ofVic.

Chris (18:31):
Can you imagine being able to throw $6,200,000,000 at
something that you don't reallyknow what it is yet?

Jerod (18:37):
Right. Well, I think he knows.

Chris (18:38):
I'm just saying.

Adam (18:39):
I don't know if we know. I think you've already checked for
6.2 bill or you even raisedthose funds.

Jerod (18:46):
The reason he announced it is to get better raises.

Adam (18:48):
That's right. Some version of more money. Get people
interested.

Jerod (18:52):
So Chris, you probably can't convince the mainstream
media to ignore the 800 poundgorillas, but you can convince
us. So here we are, we're ready.What's cool? What's underneath
the covers? Or what's theinvisible stuff that people are
doing that you and Dan and weshould be interested in?

Chris (19:09):
So I think it's funny. I'm going to say something that
I said the other day, and I'mstarting to say it more and
more. But I think people easilylook around wherever they are in
the world politics are, and itfeels like a difficult moment.
And it feels like there's allthese things you can point at

(19:32):
and say, we're going through areally tough time, and it's
tough, and everyone's trying tofigure out. But I want to offer
a counter narrative to that.
We're also at this moment wherethis stuff, the AI, and there's
a hardware revolution going on,and there's a robotics
revolution going on alltogether, and they're all

(19:53):
connected, they're powering eachother. And I think we live in
the coolest moment in humanhistory right now, like we are
sitting in it as we speak today.

Adam (20:04):
So I agree.

Chris (20:05):
What's happening right now is with all of these
different relevant capabilities,the robot people and the AI
people and the software peopleand the hardware people, it's
all coming together, and you cando amazing stuff today that even
a year ago we couldn't do. Imean, it's like if you think
about before now, we'd kind ofhave several years of little

(20:30):
software eras when we weregetting into certain ecosystems
with a language or whatever, andthey'd kind of run for a few
years. But right now, it'schanging so fast, and the
capability is coming so fastthat aside from the big 800
pound gorilla types and stuff,everybody can get into this
stuff. And so I think we're at amoment right now where it's

(20:54):
really going to start beingpervasive in everyone's life in
a bigger way than it has been.Not just like, I'm gonna open my
phone up and talk to chat GPTkind of way.
Because that was unimaginable ifyou think about it just a few
years ago. It hasn't been longsince. That was an unimaginably
amazing thing to do. But wedon't even think about that now.

(21:16):
We do it all the time.
Don't even think about it now.But like physical AI and the
fact that robotics have come sofar in the last few years now
and there are In addition toNVIDIA, there are many other
chip makers that are coming onscene to support AI, and some of
them are doing more of thededicated AI chips, and others

(21:37):
are doing more like combiningdifferent types of chips so that
you have that. And some aregreat for data centers, big
cloud data centers, and othersare great for edge devices and
tiny little constructs. And Ithink you're gonna see so much
happening in the marketplaceright now that are coming from
startups. They're not comingfrom the 800 pound gorillas.

(22:00):
They'll have their fair share.At 6,200,000,000.0, they better.

Jerod (22:03):
Yeah, better do something with that.

Chris (22:05):
Yeah, you're going to see amazing capabilities coming out
of fairly small companies.Speaking back again to Daniel
Witenack, my co host and part ofour family in this, he started
his own company, which is kindof supporting that. And that's
what I like seeing. He hasPrediction Guard, which is kind

(22:25):
of supporting open modelapproach. And I think that in
general, that approach ofanybody can go, whether you're
using a cloud environment orstartup like Daniel's or
something like that, you can goproductively pull down models
from Hugging Face, which I likento GitHub for the way GitHub has

(22:46):
always been for software,combine a bunch of different,
fairly sophisticated open sourcesoftware packages and do some
amazing things without6,200,000.
You can do it as a collegestudent in the dorm,
figuratively speaking. Andthat's the thing that really
excites me is that, is theability to everyone becomes a

(23:07):
maker, if you will. Everyone outthere can become once upon a
time, were kind of like, hey, wehave the internet. Everyone can
be a software developer. All thestuff you need to learn is
online.
There's all these resources. Alot of it can be done for free.
It doesn't matter where in theworld you are. Well, now
everybody can become a maker.Everybody can access these

(23:28):
different things and go dosomething great.
And I think that's the fact thatwe all have these Roomba type
things, these vacuums in ourhouses, everybody is now
completely used to that. But Ithink we're right on the cusp of
having lots of little deviceslike that in our houses and our
businesses that are doing allthese things, which eventually

(23:50):
will get us into this notion ofthat we're going to talk about.

Jerod (23:55):
Ready for the little robots. I don't want the big
scary robots, but I like thelittle robots that help you

Adam (23:59):
do

Jerod (23:59):
things. The neo thing is weird. We don't have to talk
about that. But that was kind ofstrange.

Adam (24:04):
Was it neo? It wasn't neo. John Mnemonics. You think Johnny
Mnemonics?

Jerod (24:08):
Yeah. What's Johnny Mnemonics?

Adam (24:09):
Well, Johnny Mnemonics was like he had man, I can't
remember this one, but it wasthe same actor, Keanu Reeves.
And I believe he had, like oh,he had something in him and he
was carrying data. And it wasVaguely, I recall this.

Chris (24:22):
Yeah. It's been a while.

Adam (24:23):
The idea of a mule, but not drugs.

Chris (24:26):
Yeah, it was That was back when he was young.

Adam (24:28):
Yes. Yes. I thought you were talking about Johnny
Mnemonics.

Jerod (24:31):
You jumped right to the matrix, which makes sense, Adam,
because most of my referencesare the matrix. But I was
actually talking about this newrobot in your house that cost
$20 and it's controlled by ahuman currently.

Chris (24:42):
I saw that but I still don't think that's going to be
the thing.

Jerod (24:45):
No, don't think I would say that's kind of weird at this
phase like that's it's a generalpurpose like it does laundry, it
does your dishes, and it's likea humanoid full size similar to
what the optimists think they'rebuilding. And yet it's at this
point because they need data totrain these models better. It's
not at all autonomous. It'scontrolled by a human with what
I imagine is like asophisticated joystick, probably

(25:07):
overseas.

Chris (25:08):
It's kind

Jerod (25:08):
of creepy when you think about it. Super creepy.

Chris (25:11):
Your grandma's in there with a stranger in the form of a
robot. The Wall Street

Jerod (25:15):
Journal did a great video about it. Joanna Like, Stern
told it to do the dishes orsomething. It took like three
minutes to load a cup into thedishwasher, which is a fifteen
second task. Anyway, let's notdo that. Yeah, it's not there
yet.
I feel like that's being too bigin general purpose. I feel like
more specific, small, theRoomba, it's gonna vacuum.

Chris (25:36):
The Roomba is the future. That was an early thing, but
it's purpose built for a veryspecific thing. And there's a
whole bunch of them on themarket, a bunch of different
makes and manufacturers andstuff on the market. And we can
go through and debate what'sbetter and all that kind of
stuff. But I think you're seeingthat times many, many, many

(25:59):
things across all sorts oftasks.
And they're cheap. And evenRoomba type vacuums are too
expensive right now. Think withthe cost of robotics coming down
and accessibility, then it'slike if you think outside this
and just walking into a retailstore or getting online to

(26:20):
Amazon or whatever and justbuying something that once upon
a time might have beenexpensive, and now it's $30. And
I think that in this day andage, that $30 purchase, I think
that getting a robot that'll dothis and that and the other, and
the fact that they haveeventually, you have families of
robots that can do differentthings, And you can put it in

(26:42):
swarming mode and just say, Automy house in swarming mode, as
we'll get into. And they justcoordinate and do all the stuff.
They're sensing you. They'removing around you. You're doing
the thing. And that's real life.Aside from just the vacuum, lawn
and garden care is getting takencare of, your security around

(27:03):
your house, your roof and gutterinspections.
Nice. It's integrated into yoursmart home stuff. You don't have
to worry anymore about whereyour packages were left by the
delivery driver because thoserobots or the swarms that are
managing your house are justdoing that. And and it's not

(27:23):
insanely expensive. People arelike, yeah.
Yeah. Where am I gonna get the6,200,000,000.0 from bases to
buy my swarm for my house?

Adam (27:29):
And I'm

Chris (27:30):
like, no. No. It's not. You're gonna have the Christmas
deal. You know, we're coming upon the holiday time.
That's right. You're going toget online, and you'll have all
the different packages aboutwhat level of swarming do you
want. This one is an 18accessory swarm package that you
can come. It's gonna handle yourIt's gonna do this. And you're
trying to choose.
You're like, Well, I don't know.I'm gonna spend more for my kids

(27:55):
on that. But there's great auntLouise, and we only talk to her
once every five years. And Isend her kind of a token thing.
So I'll send her the four itemswarm package that she can add
into whatever she's alreadyusing because it's all open
stuff.
And that's that's gonna benormal, and we're not that far

(28:17):
from the opportunity, and it'snot the 800 pound gorillas that
are gonna bring It's gonna bethe billions of startups out
there. They're each doing alittle piece of it, and their
swarm components and stuff we'reable to communicate. That's the
future that we're gonna build.

Adam (28:32):
Well, I'll tell you one thing. You've definitely put a
lot more pressure on the idea ofHomeLab. That's for sure.
Because that's all HomeLab.Those are ton of DNS queries out
there.
Probably a ton of telemetrybeing tracked. A lot of

Chris (28:46):
Yeah.

Adam (28:46):
A lot of things you may or may not be concerned. Those are
things I think about when Ithink about adding more and more
devices to my home. Gosh, man.

Chris (28:55):
So separate, I have a slight side story, but it
contributes to that. So about ayear ago now, almost exactly a
year ago, we bought and movedinto the house that I'm in now.
And the guy that we bought itfrom, he and his wife, he was a
fanatical home automationperson. And so we moved in, not

(29:17):
because of the automation, thatwas incidental, but it's helped
me move from just more of aprofessional kind of thing.
We're talking AI in aprofessional kind of, to
thinking about stuff around thehouse with all the sensors and
the cameras and stuff.
And we have all the varioustypes of home automation stuff

(29:39):
that you see out there combined,CASAS here. We have many, many,
many dozens of Casa's devicesall over the place.

Adam (29:47):
And Casa's the brand from Lutron. Is that right? Am I
picking that up?

Chris (29:50):
Right. It's from TP Link, actually. TP Link. But that's
just one. There's a whole bunchof them in Apple Home and Google
Home.

Adam (29:57):
Was thinking Casita. Casita's from Lutron. Those are
the light switches.

Chris (30:00):
But yeah, the Lutron does light switches, but there's some
common protocols that they allwork on. And I'm starting to
see, because I didn't have to gostart it from scratch and
because I inherited what thisguy had already kind of put
together and then had to figureit out and make it work, and
suddenly I'm like, Oh gosh, itwould be really easy to add this

(30:22):
And when we're talking aboutthis robotic future, even in our
homes, not just a commercial orindustrial or whatever thing,
but in our homes, it's so easyfor me to see that now. Because
I realize I already have a goodbit of infrastructure here, and
it's not expensive and it's not.It just takes a little bit of
effort. And if they can makethat easier for people to get

(30:43):
into, it's a done deal.
We already have Wi Fi and allthe other things. And then you
start adding things to plug in.It's like Legos. It's like home
automation Legos in your home.

Sponsors (31:11):
Well, friends, you know that feeling your team has
solid ideas. You got some goodstuff in there, but there's this
gap between these brainstormsessions and actually shipping
the stuff. Weeks of back andforth, scattered feedback.
You're in chat GPT friends andcolleagues are in Claude and
somebody else is in Claude codeand it's just everywhere. Right?

(31:33):
You're throwing AI at it andit's not solving the problem. It
just gets messier. Well, Miro,they flip the script. Miro's
innovation workspace, it ispowered by AI and it turns work
that normally takes weeks intodays. Here's the magic.
Miro AI sidekicks aren't thereto replace your thinking,
they're there to extend it. Andcustom sidekicks that think like

(31:54):
product leaders, agile coaches,or product marketers, they
review your materials, theyhighlight what to double down
on, and they catch gaps for youso you're not spending weeks,
hours, days, months, whatevergoing down the wrong road
without details. It integratesdirectly into your workflow as
an extension of your team'scapabilities, not a replacement.
You can generate insights fasterwithout tool switching. Miro AI

(32:17):
sorts through everything, stickynotes, research, ideas,
different formats and combinesthem into structured research
summaries, product briefs,sentiment analysis, you name it.
Test 20 concepts rapidly withMiro prototypes, generating
instant prototypes right fromyour board so you can iterate on
variations and feel confidentabout feasibility before you get

(32:40):
into the heavy builds. And thisopens the door for teams to move
at a brand new speed, the speedof their ideas, the speed of
enhanced thinking. Spend timebuilding the right things, not
digging for information. Getgreat done. If you're ready to
move faster, check out ourfriends over at Miro.
Rhymes with hero. Miro.com. Onceagain, miro.com.

Jerod (33:09):
Speaking of Legos and home automation, IKEA just
announced a whole new set oflike 27 smart home things coming
from IKEA.

Chris (33:17):
I saw that. Talk about

Jerod (33:18):
bringing it to masses like that's the kind of thing
that IKEA brings to the massesnow is they make it very simple
and straightforward and Legoesque in order to and it all
runs on Matter, which is I thinkthe open standard for
communication between thesethings.

Chris (33:31):
And so like Matter's in an interesting place and that
like, I only buy things thatthat are mat that have Matter
integrated in. And for forlisteners and viewers, the
Matter is a protocol that allowsdifferent makes and models of
automation to work together overa common protocol, and it's
local based instead of cloudbased. But not everything does

(33:51):
it yet, so it's still kind ofworking. It's been very slow. It
took a long time to kind of comeinto play, but it seems to be
having a second wind right nowbecause of all this new
capability that's coming about.
And so every new thing I buy,whether I'm using Matter yet on
that or not, I have to haveMatter so that as I go forward,
I can integrate into that. Butyeah, everything is local, it's

(34:15):
Matter. And I'm finding withtoday's craziness out there that
I'm moving more local and alittle bit more out of the
cloud. And so Matter is becomingincreasingly important from my
standpoint.

Jerod (34:26):
Well from the startup perspective and the swarming
perhaps at least the droningperspective, you'll be happy to
hear Chris that we do have astartup coming on soon zipline
who are now moving deliverydrones into production. They
actually have a delivery dronesystem that is started off
delivering medical needs inAfrica, vaccines and stuff like

(34:48):
that. And now they're movinginto The States and they're
doing food delivery, small itemdelivery, small package. So you
think your Chipotle burrito,that kind of thing. Yeah, your
you know, pounds eight

Adam (34:58):
or less.

Jerod (34:59):
Yeah, eight or less. It's super cool stuff. And they've
got it to where they're actuallyrolling out into into
commercialization now. Sostartups are making moves in
this direction. And now there'sour I assume in each city they
have a fleet of these deliverydrones.
Obviously each drone is operatedon its own. I assume eventually

(35:19):
autonomously. It actually seemslike a simpler problem than
autonomous cars because theairspace is just pretty open,
right? Like you got problemslike wind and snow stuff but
like that, it's gotta be easierthan cars.

Chris (35:34):
Yeah, generally. And so it's a different problem. So
it's easier. It's a little bitof both. It kind of depends on
how you're looking at it.
With cars, and we were justtalking to Waymo again a few
weeks ago in Practical AI aboutthis, so this is very top of
mind for me. With cars, yeah,there are a lot of challenges,

(35:54):
and you have the notion of thechild running out or the ball
bouncing out. There's a lot ofstuff that's right there, but
also how you're navigating isvery well defined in terms of
the streets and stuff like that.Air becomes more
three-dimensional, and so thechallenges are different. But so
long as it's not highlycongested, I would agree with

(36:16):
you that it is generally easierthat you can kind of move from
here to there, and so long asyou have good collision
avoidance and some othercapabilities for navigation
there, then you're probablydoing Okay.
Though that changes withswarming, because swarming
brings in close collaboration.

Jerod (36:33):
Yeah, so define swarming then, because I think of killer
bees when I hear swarming. And Iassume with drones, you're
talking about a bunch of dronesnearby each other then.

Chris (36:41):
You are. It's not just a physical distance thing, because
is physical distance is arelative thing, depending on
what it is you're trying to do.But it also, it's really more
about behavior. And so we candive into that. But before you
say that, I think that's a lineof thought we should go down is
that, as you guys know, I'mreally into animals.

(37:04):
We were making jokes earlierabout bazillion dogs and stuff
like that. I'm a licensedwildlife rehabber, and I study
animals. And in the context ofswarming, mother nature has
perfected not just swarming, butthere are many different types
of swarming from differentspecies. And so I have a set of
species that I tend to look tofor swarming purposes and say,

(37:27):
if I want to swarm with thistype of technology or this type
of platform, how do we getstarted on that? How do we get
inspiration or look for someinsights on the technology?
Well, you can look to certainspecies that are similar to the
technology platforms you'reinterested in terms of how they
move around and do stuff andsay, Well, how has nature solved

(37:49):
it there? And I definitely dothat a lot. It's not uncommon
for me to go into tech meetingsand start off with lots of
pictures of animals and stuffand people like, What's going on
with this?

Jerod (38:00):
Who is this guy?

Adam (38:00):
Are you thinking like fungus, bees, and Some bees,
bats,

Chris (38:07):
birds, starlings, those huge, what are called
murmurations of starlings.Answer awesome. Ants are
awesome. When I'm thinking aboutrobotics on the ground, meaning
what we would call a UGV, whichis an unmanned or uncrewed
ground vehicle, Ants are amazingin what they can do. And so

(38:30):
they're an awesome thing to lookat.
But I'll start with thedefinition that I use. Given the
fact that I work in the militaryintelligence space, My
definition sounds It uses thatjargon, but it really Don't get
caught up in that. Can beapplied to residential. It can

(38:52):
be applied to commercial. It canbe applied to industrial.
So don't get caught up in thisspecific wording. So I'm going
to read it in front of me. It'sone really long run on sentence
that's very specific in whatit's trying to imply. It is,
Swarming occurs when numerousindependent, fully autonomous,
multi agentic platforms exhibithighly coordinated locomotive

(39:14):
and emergent behaviors withagency and self governance in
any domain, which could be air,ground, sea, undersea, or space,
functioning as a single,independent, logical,
distributed, decentralizeddecisioning entity for purposes
of C3, which is command,control, and communications,

(39:34):
with human operators on the loopto implement actions that
achieve strategic, tactical, oroperational effects in the
furtherance of a mission. Solong, long, long sentence, but
it hits a bunch of very preciseconcepts and integrates them in
together.

Jerod (39:52):
I can tell each word was selected there.

Chris (39:54):
Yeah, a mission might be, instead of thinking military, a
mission might be getting apackage to your house. That
might be the mission. And thatdoes have command, control, and
communications involved. So itdoesn't have to be the military
esque jargon that

Jerod (40:12):
we're talking Yeah.

Chris (40:13):
Applies to any these commercial, industrial,
residential, military, whatever.So

Jerod (40:23):
that's a lot.

Chris (40:24):
It's a lot. It's a lot. And if you want, I can kind of
break down a high level whatsome of those mean.

Jerod (40:29):
Yeah, I think my broad takeaway, we can talk about the
individual words, because I knowthey're very specifically
chosen, like independent,logical, distributed,
decentralized, decisioningentity, like stuff like that.
Can tell each word was selectedfor a reason.

Adam (40:42):
Yes.

Jerod (40:42):
But I think my grand takeaway of a swarm is kind of
the E Pluribus Unum. It's like,okay, all these things are
individual and autonomous, butthey're all acting as one.
They're acting with one purpose.

Chris (40:56):
That's a fantastic insight that you have. And that
is the key to it is Swarm issuch a buzzword. We always have
buzzwords in this AI andsoftware space. There's always
the buzzwords of the year. AndSwarm is certainly a huge
buzzword right now.
And almost without exception, Iwill turn around and tell I can

(41:18):
go back to my definition,assuming that you want to accept
that as the definition ofswarming, and I can defend that
fiercely.

Jerod (41:24):
I cannot attack it. Can you attack it, Adam?

Adam (41:28):
I wouldn't

Chris (41:28):
try. You're the expert But here, I would say, all you
people who are talking swarming,no, you're not. It's not
swarming. What you're describingis all sorts of things that lead
to swarming. There's a wholebunch of incremental
capabilities that wouldeventually, as you add all those
capabilities together, theyculminate in swarming.
But the chances of somebodysaying that what they're doing

(41:53):
out there is consistent withChris Benson's definition of
swarming So is pretty what yousaid was right on. And that is
that just as you see in naturewith those ants, every little
ant has its neural capability,shall we call it, what it's

(42:14):
doing. But at the end of theday, they're functioning to get
a mission done, a job done,something productive for the
colony. And they are all lendingthemselves to that greater good,
even if some of them may notsurvive that kind of thing. They
are functioning as a singleentity, and it is the entity

(42:37):
that's trying to get the thingdone, not the individual ants.
The individual ant may be likewe have a crack in the ground,
and we have to get from thisside to that side, and they
build an ant bridge. We've seenpictures of that. And that one
little ant may have the job of,I'm holding onto the ant on this
side, and the ant below me isholding on there. And then they

(42:59):
have that going on as well. Andwe're all creating this ant
bridge over a chasm that none ofus individually could span.
But by working together for thatswarm approach, which is make
that accessible, they are doingsomething well beyond what any
of the individuals can do. Theyare super ants in that way. And

(43:21):
that's what I'm getting at isthat ability to give up your
individual identity as a memberof a swarm for the purpose of
the overall swarm's intent. Andthat swarm itself has an intent
that is a swarm level thing,kind of to your point, Jared.
It's like, that's not the thingon any one brain.

(43:43):
But when you put all thosebrains together, or technology
that represents that, there's athing that the overall thing is
trying to do as a single entity.

Adam (43:53):
Its power is a number. It's like, I saw my kids love
ants, animals, you know, all thethings essentially. Venomous
plants that kill things, youknow, that stuff entertains them
dramatically. Venus flytraps,things like that. And we watched
this show.
It's kind of a documentary, butit's also kind of dramatic. And

(44:16):
John Cusack was the narrator,and it's a movie called The
Besieged Fortress from 2006. Andthere's an ant type that I wanna
mention to you. It was actuallyI ruined the plot a little bit,
but it was ants versus termitesessentially. And it was very,
very well done.
If you've seen it, Chris,obviously say something.

Chris (44:35):
I have not, but I'm gonna check it out now.

Adam (44:37):
It's it's a 100% worth it. It is phenomenal. It's probably
gonna visualize for our entireentire audience the things
you're talking about because theparticular ant, I guess you call
it the name of the ant, Isuppose, is is how you describe
it, were driver ants. And thesedriver ants are are so swarm

(44:57):
like. You know, they don't thinklittle.
They they they create rafts forthemselves. The entire colony
can can float. I mean, you canput them underwater, they won't
die. They will like create thisbubble. They are just basically
resilient to to the nth degree.
And if you're in their path,you're dead. Like, no matter
what you are, snake, a rat, abug. They're gonna overwhelm

(45:19):
you. Oh, yeah. They they drivein numbers.
They're called driver ants, andthey are truly truly incredible.
And this whole entire dramaticdocumentary narrated by John
Cusack is phenomenal. TheBesiege Fortress, I would highly
recommend it. 2,006, amazing.But these drive rants probably
elicit a lot of the a lot of thequalities and and

(45:40):
characteristics that you'rementioning because they act like
if you're in their path, it'snot as if they're one, it's
they're many, and they acttogether, and Yeah.
It's

Chris (45:49):
I mean, it brings a whole capability. Whether you're
talking to ant or whether we'rehumans with our technology doing
this, you're basically inventinga whole new category of what's
possible by introducing this.And because of conflict of

(46:09):
interest and I stay away from myemployer, Lockheed Martin,
generally I'm delicate ondefense and intelligence stuff
anyway when we're talking inpublic. But the notion of, if
you were to look on the militaryside for just a second, at a
high level, there's the notionof mass. And if you go back and
warfare, many people would say,Okay, let's build up mass to win

(46:33):
against an enemy.
And then as things progressforward, we learn that maneuver
could kind of out You could goaround mass and you could hit it
from different ways. And somaneuver as a capability started
trumping what was possible withmass. But swarming becomes a
whole new thing, is that you'rekind of getting the best of mass

(46:55):
at individual small scale, butyou're getting mass and you're
getting hyper maneuverability,and so it's able to trump that.
So in that domain, in that kindof military world, it brings
about a whole new capabilitythat never existed before. And
similarly, when you move intocommercial and industrial, and

(47:16):
we talked about this superautomated house a few minutes
ago, you're bringing aboutthings that just were not
possible before.
You could have little pieces ofit that were possible discreetly
from a source, but the notion ofthis integrated solution that
would just kind of go attack areal world problem and overcome

(47:38):
it, kind of going back to yourdriver ants, is a new capability
that the world will enjoy goingforward across all different
types of domains. And so I thinkthat's the magic of swarming
right there, is it's differentfrom a fleet. I think a lot of
the times where people throw upa whole bunch of things like
drones, so that's the thingeveryone knows, will throw up a

(48:00):
whole bunch of drones in theair, but it's not really a
swarm. It's a fleet of drones.That's what it is.
And each one requires individualprogramming to go do this or do
that. There may be somecommunication between them
potentially, depending on whatthey're doing, but they're not
thinking almost like a brain,like an abstract brain
themselves. They're not lookingand dynamically handling what's

(48:22):
happening in the real world inreal time and saying, this is
changing right here, right now.As a swarm, I'm going to go do
that. They can't do that.
They're fleets. They canrespond, but it's going to take
inputs. It's going to take somecollaboration between them, but
it's going to take a lot ofguidance from afar to make that
happen. And that's thedifference in mass numbers in a

(48:45):
fleet versus what a true swarmwould be, is that that
capability and that intent andthat emergent behavior is really
key to identifying a swarm. Andyou do see that in mother
nature.

Jerod (49:00):
So let's take a recent phenomenon, which is the drone
light shows, you know, wherethey go out and let's say
they're making a dragon. Is not

Chris (49:09):
a swarm, but yes.

Jerod (49:10):
Well, I was gonna ask depends on the intensity.

Chris (49:12):
Not a swarm. See how I did that? Just say I did that.
Not a swarm.

Jerod (49:16):
I Well, was going to ask, it depends on how it's
implemented, isn't it? Couldn'tyou swarm to accomplish a
dragon?

Chris (49:20):
Absolutely could, but nobody has. So these days, what
they're doing is you may seethese light shows where they
have thousands of drones in theair, but each one of those is
following a preprogrammed path.There might be some limited
communication when they're veryclose in case of their winds and
things like that in terms ofanti collision. But what I would

(49:45):
say is if you were to do the bigdragon that you talked about as
a swarm, the swarm would figureout how to do it in real time.
It's actually using thatdecisioning entity that we
talked about in the definitionand saying, my mission is to
produce a dragon over this areafor people to watch.

(50:05):
And it would go do that. Itwould go figure out where all
the pieces need to be for thatdragon to come about. That's
true swarm behavior. If youthink about animals that are
getting out and doing something,they're not producing dragons,
but they're going out and doingsomething in a swarm, There's no
external thing saying, Swarm ofbees, I'm telling you to go do

(50:29):
this. And you need to make anadjustment there and all.
They figure it out in real timein the swarm and make whatever
it is that those species aretrying to achieve, it happens.
Emergent behavior that's realand in real time that supersedes
the individuals. And that's whatI'm saying. The light shows
fleets of drones that are beingprovided instructions, often

(50:52):
essentially a three-dimensionalvectoring trajectory on what me
as an individual drone would do,regardless of what all these
others are doing.

Jerod (51:01):
Okay. So even inside of emergent behavior, let's say in
an ant colony, you have roles,you have leadership, there's
some sort of like, there's somesort of mission that comes from
somewhere.

Chris (51:16):
There is.

Jerod (51:16):
And I assume now we're getting to the part where it's
like, okay, how do you makethese things? Because as a guy
who's makes fancy websites hisentire life for a living, this
sounds really hard. I just feellike, if I had a new job, hey
Jared, your new job is to builda swarming technology of
autonomous whatever, I'd belike, nope, not going to even
try that. Because that justsounds very, very, very

(51:38):
difficult. Where do you start?

Adam (51:41):
Like, how do you do it?

Chris (51:43):
That's a great question. And not only that, but you've
identified the thing that youjust said in your vulnerable
moment there in terms of like, Idon't even know where to go.
That's what almost everybodyThat is why it is a problem yet
to be solved. And there manygroups, companies, individuals
out there working on it,including me. This is my

(52:08):
passion.
And all of us, at some point,some of us might have had the
benefit of coming from robotics.But just like many other skills,
that also carries some baggagewith it that you have to unlearn
to do it. And that's one of thepros. So when I talk to people,
I've been doing drones fortwenty years. I know everything.
And I'm like, well, I'm like,that's good in some ways, but-

Jerod (52:29):
Not a swarm.

Chris (52:31):
Not a swarm. And not only that, but sometimes it's

Adam (52:35):
the- Let

Jerod (52:35):
me get

Adam (52:35):
you a T shirt says not a swarm. Ask me anything, not a
swarm.

Chris (52:39):
That fresh learner's mind though often does it. And so
it's a complex problem and youhave to break it down into its
constituent parts. And there's awhole bunch of layers, because
there's things that have tohappen at the member. Like if
you talk about the individualant, at the member level,
there's a whole bunch of things.It's got to navigate.

(53:00):
And that's kind of like where weare on drones today, in the
sense of if you go buy one,we're going go out, you go out
to the toy store and you buy adrone today, or order one online
these days because toy storesare not so common anymore. So we
ordered the drone online, andthat has basic navigation. And
there's a whole bunch of tasksassociated with that, and that's

(53:22):
where most of the robotics worldhas been, obviously, over the
years. But as you move intocommunication between them and
what kind of tasks happen, youkind of move up to a level.
There's a local drone level in alarger swarm.
And then there is the how do allthose locals operate together.
So you kind of steadily move upin abstraction till you have

(53:46):
that notion of this emergentthing, which is really It's
really quite a challenge ofbecause there's not a master
member. There's not the boss.Some You may have a queen bee.

Jerod (54:01):
Would it be easier if there was a boss, though?

Adam (54:03):
It would.

Chris (54:04):
So it depends on what you're trying to do. I would say
that's the step below. If you'redoing almost everything a swarm
can do, but you still have somecentralized control, there's a
couple of levels below that.While I can't

Jerod (54:15):
Not SWORM.

Chris (54:16):
Yeah, I can't share it today. I invented I created a
document that helps people at mycompany evaluate these
technologies at differentlevels. It's called a maturity
model towards swarming. And theycan look at anyone else.
Somebody has put something outthere, and we can evaluate it

(54:38):
based on that criteria aboutwhat exactly it does.
And I need to see if they'll letme release it publicly because I
think it would be useful.

Adam (54:44):
Let me see if I can maybe break down an idea. And I I
don't have your depth, but if Iwere thinking about this problem
and and obviously when wecompare ants, so in the case of
the driver ants, just becausethat's my example that I have
some clarity on at least, theydo have a queen and the job is
to protect the queen. It's likeif the queen disappears, they

(55:05):
will elect or attempt to elect anew queen, but there's always
somebody in charge essentially.But if that's not if that's not
a swarm, then the way I mighttry to create a boss would be
through consensus. Because ifyou're a controlling entity
that's connected and so you knowall your parties in this

(55:27):
connected mesh network orwhatever you want call this one,
you know player b versus z overhere has new information the
swarm needs to know toconsensusly, if that's even a
word, to have consensus on thenext decision.
And so we may, as a swarm, electa new, not so much boss, but a

(55:48):
primary information source thatchanges the way the swarm acts
as an entity. And so it's sortof self evolutionary.

Chris (55:55):
You're hired because you're on the right track.
That's it.

Jerod (55:58):
So aren't they just making their own boss then,
basically?

Chris (56:01):
So that's the thing. So the queen, like in the case of
the queen, yes, there's a queenwho is the general, the one in
charge, but at the same time,she's actually not making all
the decisions. A lot of it isinstinct that is being played
out.

Adam (56:19):
It's preservation in that case, right? The queen is not
the boss in terms of leadershipand knowledge because the drones
have the knowledge, The droneants out there doing the work.

Chris (56:27):
That's right.

Adam (56:28):
She is the preservation system for the It's a necessary
component of many.

Chris (56:33):
So she's not a master, like a master direction giver.
That's not her. Her role is, asyou said, perpetuation of the
colony versus she's not drivingthe specific actions of drones.
Those are built in. MotherNature has imbued the members
with that, and they understandhow to do that.

(56:54):
But to your point, Adam, thatnotion of kind of consensus, are
different approaches to it. Wecan use some different words
because there are differentalgorithmic approaches of
consensus, election, things likethat in terms of saying, well,
we have a distributed computegrid that is our swarm, that is

(57:17):
imbued in our swarm, And how dowe arrive at single overarching
directives that perpetuatethemselves downward through the
swarm and which change as theygo down? Because this is the
overall, this is what we need todo. There's a mission. There is
a high level sense ofabstraction about, well, to

(57:38):
accomplish the mission, you mustdo A, B and C.
But A has 10 steps to it, youknow? And some of the swarm
members are going to take theassignment of doing those, and
others are going to say, well,I'm going to go off and do these
other things that are part ofthat, that might have been part
of the B category. And so theyhave to self organize in the way

(58:00):
to do that in real time, becausethis is a physical technology.
So it's one of those. And thereare sensors coming in.
Things are changing constantlywithout. And so you are with
your sensors, whether you're abiological being or whether
you're technology, you're havingto take all that new information

(58:20):
in. You're having to dodistributed computing and
decisioning through algorithmicapproaches and select members to
accomplish all the things aspart of that overall mission
that you're doing. And it'squite complex. I mean, it's a
very complicated thing as we sithere in 2025.

(58:41):
I think we'll nail it gradually.I think we'll nail it in
iterations. And I think somebodya century from now will be like,
yeah, well, of course we didthat. But today, it's a tough
problem to solve.

Jerod (58:54):
So at what level do the humans interact? So let's
imagine that you've created aswarm of vehicles. And it's a
legit swarm. It's not swarmswarm.

Chris (59:05):
A legit swarm.

Jerod (59:06):
Yeah. I was thinking about that as you said that, you
remember the old Jeff Fox whereyou're thinking like, you might
be a redneck, we could do awhole line of like, might not be
a swarm. Like if you've got aboss, you might not be a swarm.
If you've got a path they gaveyou to fly, you might not be a
swarm. But let's say you haveone and this is like, Chris
approves.
And it's a bunch of drones,let's just do that. At what

(59:28):
level does the drones receivetheir mission from the humans?
Like, is it very generic? Or isit very specific?

Chris (59:35):
It can be either. It depends on what you're building
toward. And swarms havedifferent purposes. So remember,
a swarm is not a generic thing.They're purpose built for
certain capabilities.
And you do have that C3, whichis command, control, and
communications, that's inherentto that. And one of the other

(59:55):
phrases I use, which peopleoutside of the military context
may not be as familiar with, ishuman on the loop.

Jerod (01:00:00):
Not in the loop, but on the loop.

Chris (01:00:02):
Not in the loop, but on the

Jerod (01:00:03):
Does that mean?

Chris (01:00:03):
Are two different things. An in the loop is where a human
is controlling a technologydirectly, and they're making it.
So a human in the loop may say,make a choice for a task. So
they may say, yes, I'm going tonow have you drop that package

(01:00:24):
on that person's front door. Andyeah, it's clear.
We've looked at it. It's safe.There's nobody in the way. And
we're going to have you put thepackage on the front door
because it's safe. And we didnot want the drone to do that
until me as a human verifiedthat that was Okay for us to do
so that we didn't hit people orhit things.
On the loop, you are essentiallytasking that. It's kind of a the

(01:00:48):
human has a supervisory role andmaybe a mission giving role.
Like your mission swarm is todeliver the package to that. Or
maybe more, it might be, here'sa bunch of packages to the
swarm, and I want you to go tothis neighborhood and deliver
all these packages to the righthouses. And that is the mission.

(01:01:10):
And then the swarm understandsthat geographic layout. It
understands the real worldenvironment it's in, and it
figures out which member theyeach pick up a package, and it
figures out how are they goingto do that. Some of the packages
are more than the eight poundsthat Adam talked about. Some of
them are 60 pounds, and it takesmultiple swarm members to get

(01:01:30):
that package airborne andcollaborate. And so as they go
into that environment andthey're looking, I've to get
this package, to that address,and oh, by the way, that address
might have been reachable by afour pound package on one sworn
member acting alone, but it'snow 60 pounds.
We have multiple sworn members.And even with all those sworn
members, it's outside of ourrange. So how do we address

(01:01:53):
getting it outside the range,given the fact that we have
other concerns that may belimiting that? The Swarm would
work that out through itsdistributed computing and
collaboration that we justtalked about, that, you know,
where kind of comes to thatconsensus on how it's going to
collectively solve the problem.Does that make sense?

Jerod (01:02:12):
Yes, I think that it does. I'm wondering if maybe I'm
sniffing danger eventually,because

Adam (01:02:19):
Oh, go for it.

Jerod (01:02:21):
Well, because at a certain point. You give a
directive and maybe thatdirective is completely benign,
like you have a swarm ofcleaning bots, you know, your
house and you say, Okay, bots,you know, clean the bathroom.
And that's as far as you getinto it, you're on the loop, but
you're not in the loop. And sothey go about doing that. And

(01:02:41):
we've accomplished Chris Bensonlevel swarming.
So I now have numerousindependent, fully autonomous,
multi agentic platforms in mybathroom exhibiting highly
coordinated locomotive andemergent behaviors with agency
and self governance. Right? Soat a certain point, couldn't

(01:03:03):
they just say, this toilet'sreally dirty.

Adam (01:03:07):
What if we just removed it?

Jerod (01:03:08):
Wouldn't that be the bathroom would be even cleaner?
And then they all decide thatyes, that's a great idea. I come
back, I don't have a toilet.

Adam (01:03:15):
That

Chris (01:03:15):
is just Adam had a great moment a moment ago, and you
just had a great moment there,Jared. Oh, thank was Most

Jerod (01:03:22):
of my great moments in those toilets.

Adam (01:03:24):
One is all I get, just one for sure.

Chris (01:03:26):
You're cleaning up your act, man. But yeah, so that's a
great thing, and that comes downto you're not giving What you're
really telling about in swarmsis when you get down to the task
level, then you're talking maybenot about the whole swarm making
a decision. It might be a few ofthem that are addressing a task

(01:03:47):
and figuring out at a morelogistical level, operational
level, And how am I going to dothat is one of the things, is
that when you're doing We'reback to AI safety and AI
training

Jerod (01:04:01):
on this,

Chris (01:04:01):
is that maybe removing the toilet in most cases is not
an acceptable thing. So we needsome technology based guardrails
there, but that's also where,depending on the circumstance
you're looking at, that human onthe loop needs to be able to go,
no. There's kind of a killbutton, if you will,

(01:04:22):
figuratively speaking, meaningkilling of the swarm, not
killing a person, just to bevery clear so that nobody
misunderstands me.

Adam (01:04:29):
Super clear. No killing here.

Chris (01:04:31):
No, no. We're not talking about killing people here.

Jerod (01:04:33):
That's why I use a bathroom and a toilet as my
example.

Chris (01:04:36):
In this context, it might be don't kill the toilet kind of
thing. That's where the human onthe loop, an oversight, where we
still have these amazing,capable human brains that have
they can't do everything thatdigital technology can do, but
digital still hasn't yetarrived. It will, but it hasn't

(01:04:57):
yet arrived at what ourcapabilities are. And we can
look at it and go, taking thetoilet out is not acceptable to
the homeowner. We're not doingthat.
Maybe if you're Jeff Bezos,maybe instead of cleaning the
toilet, you just remove it everytime. Maybe Jeff Bezos will have
the toilet removal every day.The drone swarm goes into Jeff
Bezos' bathroom and it justtakes the toilet out and it puts

(01:05:19):
the new

Adam (01:05:19):
one It's a sea of toilets back there. It's like a pile of
them.

Chris (01:05:24):
There you go. Well, it cleans it all up because you
know what? When you can throw$8,000,000,000 at something you
haven't really identified yet,you can probably afford to have
your toilet

Adam (01:05:34):
6.2.

Chris (01:05:35):
Oh, 6.2. Made such a big difference in my head.

Adam (01:05:40):
Yeah, sorry about that. 8,000,000,000 is close, but
let's just round up to8,000,000,000.

Chris (01:05:43):
Let's just round down to see Okay, we're there. But yes,
so other than Jeff, I don't wantthe drone swarm taking my toilet
away. That would get rather

Jerod (01:05:51):
I'm with you.

Chris (01:05:52):
Yeah. It's a it's a little bit too much.

Sponsors (01:06:11):
Well, friends, when you're building and shipping AI
products at scale, there's oneconstant, complexity. Yes.
You're bringing the models, datapipelines, deployment
infrastructure, and then someonesays, let's turn this into a
business. Cue the chaos. That'swhere Shopify steps in, whether
you're spinning up a store frontfor your AI powered app or

(01:06:32):
launching a brand around thetools you built.
Shopify is the commerce platformtrusted by millions of
businesses and 10% of all USecommerce from names like
Mattel, Gymshark to foundersjust like you. With literally
hundreds of ready to usetemplates, powerful built in
marketing tools, and AI thatwrites product descriptions for

(01:06:53):
you, headlines, even polishesyour product photography.
Shopify doesn't just get youselling, it makes you look good
doing it. And we love it. We useit here at Changelog.
Check us outmerch.changelog.com. That's our
storefront, and it handles theheavy lifting too. Payments,
inventory, returns, shipping,even global logistics. It's like

(01:07:14):
having an ops team built intoyour stack to help you sell. So
if you're ready to sell, you areready for Shopify.
Sign up now for your $1 permonth trial and start selling
today atshopify.com/practicalai. Again,
that is shopify.com/practicalai.Could we get into this scenario

(01:07:40):
where it's like Aladdin and thegenie? And he's like, hey, you
know, I want make me a prince.Right?

Adam (01:07:48):
I think it was the first no. First one was to he tricked
the genie to get him out of thecave. We'll we'll skip that one.
And then the second one,technically, the second wish was
make me a prince. And he didn'treally make him a prince.

Chris (01:08:01):
That's great memory. Gosh. Like, I've seen the movie,
but like, you're really bringingit back to me.

Adam (01:08:05):
I'm Well, got a good brain, you know, over here.

Jerod (01:08:09):
My brain is That's a second great moment

Adam (01:08:11):
right here. That's right.

Chris (01:08:12):
Two in one moments here with me. So okay. Keep going.

Adam (01:08:15):
He doesn't really make him a prince. He just clothes him as
a prince. He mimics a prince. Hedoesn't really give him royalty.
He doesn't really give himlineage.
Mhmm. And I guess I'm sort ofsidetracking to some degree just
to be accurate about my Aladdinreference. But point is is
there's times when he adds Withit. Or I guess in in all of the
lore around the Aladdin figureand a genie figure where you ask

(01:08:36):
the genie for something, but youhave to be careful. That's where
this term comes from.
Be careful what you wish forbecause you wish for something
without the full awareness ofthe agency behind the genie or
behind the swarm, and so youmight get your toilet removed.
Is that a concern? Like how

Chris (01:08:53):
Totally.

Adam (01:08:53):
How are you guarding against that? How do you guard
against that without the QMLloop or the kill switch? Is
there an OS? Like, I don't know.How do you how do you guard
against this genie issue?

Chris (01:09:03):
I know. I think there's a lot, and I think it's at many
different levels. It's a realthing that we talk about in real
life today without havingachieved full Chris Benson level
drone swarming. And that we talkabout that in terms of AI safety
all the time now. That's a hugepart of the AI world, is what is
AI safety?

(01:09:23):
How do you keep unintendedconsequences from coming to
pass? I think anyone who'sreasonable will recognize that
some of those will still come topass out there. You can put
guardrails around things, andyou can even ask AI to put
guardrails around other AIs aswe're doing because we're using
the tool to build the tool. Butwe will have bad outcomes across

(01:09:45):
the board, just as we alwayshave with software and always
will. And so I don't have themagic bullet on that, but there
is training the distributedswarm brain, this abstraction of
computing, of grid computing,where they're all doing this and

(01:10:06):
using their algorithms.
That will Where it goes wrongmay happen different places. We
often talk about today's LLMscoming out with inferences that
are suboptimal. Sometimes it'squite funny, sometimes quite
tragic, actually. But that willcontinue to happen. We have
software issues.
We're also moving into thephysical world where if you have

(01:10:31):
these physical agents that areimbued with a whole bunch of AI
agents that are doing stuff andthey're acting as a member of a
larger swarm, there's a lot ofplaces where things can go
wrong. So there's going to be alearning curve on that, and
we're going to have problemsalong the way. So I certainly
wouldn't want I know for a lotof listeners and viewers, they

(01:10:54):
probably think a little bit piein the sky. Not everyone's going
to believe that this is probablysooner than they would otherwise
expect. But we'll get through itand stuff.
And we'll try it. We'll do thebest we can. And the responsible
people will put a lot of safetyaround it the best they can. But
we will make mistakes.

Jerod (01:11:13):
Where do we stand? Where are we in this initiative to
create this thing or thesethings?

Chris (01:11:20):
So I think like many things, you'll see it coming
from specialists. There's awhole area of expertise that you
develop around trying to solvethese problems. And some
companies are specializing that.And just like other things,
you'll see that. But I thinkover time, especially given the
fact that it's not one industry,it's many industries, there'll

(01:11:43):
be many players.
I think one of the things tomake this happen isn't just can
we get there? Because if youthink about it, once you can get
there, almost everybody kind ofdoes a close copies of that.
Once we had our first chat GPT,it wasn't long before that we
had competitors and other modelsthat were nipping at its heels.

(01:12:03):
And I think you'll see that hereas well. But I think it really
comes down to gettingorganizations and motivated
individuals into it so that theyare producing some level of
whatever's productive in whatthey're doing.
In their industry, in theirworld, what's productive and
costs will drive down. And Ithink as those costs drive down,

(01:12:23):
that's where you see it reallypushing out into lots of
different places in life. So alot of it isn't just a
technology question. It's aneconomics question. But I think
the pervasiveness of it willdrive that.

Adam (01:12:37):
Let's get since we helped create a show called Practical
AI, let's get practical. Yes.You'd mentioned this is
obviously burgeoning. You'recoining this. I I kinda feel
like swarming is the protocol.
Maybe this is a specification orsomewhere, and the
implementation is more of aproduct potentially. But take us
into the practical nature of,let's just say, over the next
three years. Will we seeswarming of any sorts in a

(01:13:00):
consumer level, home lab, put itin my home level? And if we do,
like, be realistic, practical,if you can. Like, what will it
be?

Chris (01:13:10):
At the level of the definition that I provided,
which is a very high bar, Ithink you'll see lots of things
that are calling themselvesswarming things developing
within that two to three yearhorizon. I don't think many of
them will rise to that level.There'll be quasi swarming
capabilities that you'restarting to see in consumer and

(01:13:32):
commercial products and stuff. Ido think, however, there are so
many really smart minds aroundthe world working on swarming
because by opening up an entirenew category of capabilities
that don't exist today, thatpeople already have productive
use cases in mind for, there's alot of money to be made there.

(01:13:55):
You have not only commercialentities and motivated makers,
but you have nation states thatare highly motivated to do that.
And it's a it's a big scientifictopic of research. I think
you'll see it probably first inareas where people can throw
lots of money at it. And if wedo talk about in the commercial
space, our 800 pound gorillas,you're more likely to see it in

(01:14:19):
a narrower case of use casesthere. I think in the military
space and intelligence space,you're likely to see it there
because you have the economiesof nation states that don't want
to be left behind. If we're notable to produce a swarm first
are very closely followingwhoever is first, then we have a

(01:14:42):
national security issue here interms of what's possible.
And so I think you'll see nationstates prioritizing that
probably in very closecollaboration with commercial
entities, which is really commontoday. I mean, if you look at
certainly how both the USgovernment and most of our
allies, as well as the Chinesegovernment, there's a lot of

(01:15:05):
overlap between nation stateresources and commercial
entities that have specialknowledge and skills working
together to produce that stuff.So I think those types of
collaborations are likely to bethe first ones, largely because
they can throw resources at theproblem until you get there. I
think the key is it's thinkingabout the problem the right way.

(01:15:26):
And I think that's where peoplestruggle is breaking down that
complexity that we were talkingabout earlier, you know, that
Jared pointed out, and saying,how can we discreetly address
those points of complexity in away that you can then pull those
many solutions together toachieve the grandiosity of the

(01:15:47):
definition that I provided?

Adam (01:15:48):
Let me see if I can not predict, but this is where I
would

Chris (01:15:53):
because I'm not gonna predict anything. Yeah. Yeah, we
both We know that

Adam (01:15:58):
think the two of you, I think will agree with what I'm
gonna say here. I think the areawhere I like to see this type of
swarming is in energyconservation. And so I think
there's multiple devices in myhouse that consumes energy from
a HVAC system above me that bothheats and cools my home to the
lights that power my house to,let's say, a kettle that is

(01:16:21):
electrified, all the things. Iwanna give my home the task of
being energy conservative.

Chris (01:16:28):
Right?

Adam (01:16:29):
This this swarm. I wanna have a swarm of devices that
help me be that. And it can ithey, Adam and the family are not
here. It makes sense as anagency to be to be conservative
with our energy use becausethere's no one here to do it.
And that's where the you can dolike individual device level
smart home automation, which ishere today.

(01:16:52):
Matter supports that. It's not aswarm though, right? It's not a
swarm.

Chris (01:16:56):
That's right.

Adam (01:16:56):
So I would like energy conservation to be my first
swarm tactic. The next would beI live in Dripping Springs,
Texas just outside of Austin,Texas, and we always have water
challenges. Right now, we'realways in some version of a
drought. There's actually a bigbet on the wall on Wall Street
against Texas running out ofwater. Like, there's a bet

(01:17:18):
essentially shorting Texasrunning out of water at some
point.
I just heard this headline.That's a headline only. I don't
know what the truth is behindthat, but I heard it so it must
be true. Okay. So the next thingis water conservation.
Help me as a household, maybeeven help me as a neighborhood,
a sworn neighborhood, beconservative when it comes to
water conservation. So my childgoes to flush the toilet or

(01:17:42):
yeah, don't know, some sort ofaction tries to take place, but
the swarm is like, hang on asecond, we're in a conservative
nature. We're gonna use the oneor the point five gallon version
flush versus the 1.2 becauseit's a, you know, it's number
two. There are some reason,right? But for whatever reason,
like we now have new tech in myhousehold that gives me things
that really matter energy,water, and then the last one for

(01:18:05):
me is food.
There is so much food waste inAmerica, tremendous amount. I
know I for sure buy some chickenonce, twice a month and I'm
killing chickens constantlybecause I'm wasting my chicken
not making it. So I don't knowif that's a problem that's me,
but at some point my tech, myswarm tech can help me solve
those three key things, energy,water, and food. And I think you

(01:18:27):
start there because that's whatmatters. My laundry, kind of a
me problem.
Maybe my washer can say, hey,you put a white in with a darks,
probably not smart, eject it oralert me. You know what I'm
saying? But like, I don't needhelp with laundry. I don't need
I mean, I like my iRobot andvacuum. That's cool.
But I think that the thing Iwould wanna conserve on is those
three things. That'd helpful.

Chris (01:18:48):
I I and I think I think you'll see that. I don't think
it'll just be the swarm doingthat though because, like, even
today, you know, if you startwith where we're at right now
and talk about the fact thatenergy monitoring is really
common within a lot of theseexisting devices.

Adam (01:19:02):
Not a swarm.

Chris (01:19:03):
Not a swarm. Not swarm yet, but we're getting there.

Adam (01:19:05):
We're getting

Chris (01:19:06):
keep saying

Adam (01:19:06):
it. Sorry.

Chris (01:19:06):
So bear yeah. Bear with me for a second. So we we have
the we have what we can alreadydo at the individual device
level. And then as we reallystart viewing our homes with AI
agents, which is gonna happeneven before the swarms are
hitting.

Adam (01:19:21):
So soon. Yeah. That's next.

Chris (01:19:22):
You're gonna have AI agents doing lots of different
things, including themonitoring, and those AI agents
will be monitoring your matterdriven devices thinking, Oh, we
need to make some adjustments.They'll be communicating with
the devices that are beinggoverned by that. And so they're
able to get you a great deal ofthe way down that use case that
you just talked about. But thereare also going to be things in

(01:19:46):
your home where things like forenergy, the energy conservation
thing, you mentioned things likeairflow and temperature, where
it's not an explicit devicethat's matter enabled and has
the energy monitoring built in,but it may be like that corner
of the room is cold. And in thatcase, that swarm that's

(01:20:07):
monitoring the house and maybeit has other functions that
aren't just monitoring.
Maybe it's doing a cleanup. It'sdoing the cleaning job, but it
also notes that, hey, thiscorner is not getting good
airflow. The temperature'schanging. To your vision, Adam,
that you just talked about,that's where the swarming
capabilities of having differentdevices work together will do

(01:20:29):
it. Now, an individual robotcould also detect that device.
It doesn't have to be a swarm.So you really get it for a swarm
to be effective there. You'rereally going to be looking for
how does a cluster of membersworking dynamically together get
me something I don't alreadyhave? And I think that's the
question to answer in that usecase, if you're actually wanting

(01:20:49):
to introduce the swarm to it.

Jerod (01:20:52):
Well, we humans have our own form of swarming. It's
called open source software. AndI'm curious if there's a place
where people who are aspassionate or maybe even just
potentially interested in thisinitiative, this movement, this
I don't know, this next bigthing of swarming tech. Is there
a place they can gather? Isthere a framework?
Is there a conversation? Isthere anything in the world of

(01:21:15):
open that people could gatheraround?

Chris (01:21:17):
There are. And I probably should have brought a list maybe
in the show notes, we can addsome stuff in. Some of the
things that I often tell peopleto start off on is robotics has
been a big part of this, kind ofrobotics role being part of
developing to the swarm is ROS2exists. ROS stands for ROS2. ROS

(01:21:40):
is the Robotic Operating System,which is open source, and it is
the most widely used roboticsoftware technology out there.
It's not the only one. There aremany, and some of them are
closed and some of them areopen. But there's tons of books
now on Ross. And so I often,when people are interested in
this and they're like, but howdo you do Aside from the swarm,

(01:22:02):
I can't even make a single robotwork, what do I do? Well,
there's tons of informationabout that.
Start off maybe not solving theoverall swarming problem that we
were describing as remaining ahard challenge, but start with
something more accessible. Youget on we've mentioned Bezos so

(01:22:23):
much, Amazon and others, andthere are a lot of maker kits
that you can get that are openmaker kits. You have Ross.
They're very similar in terms ofBut if you want to not do robots
and you want do drones, There'sa whole bunch of open source
drone stuff. And then the thingthat I love doing, I do this all
the time, is diving in on GitHubat different software

(01:22:46):
communities that support openspecs and stuff.
There's tons of repositories onGitHub that are designed to do
this that just interested peoplesaid, want to go scratch an
itch. I want to solve a problem.And I go there, and I'll then
also go to Hugging Face and lookfor small models that may, if I
need AI in the mix, cancontribute because really, you

(01:23:08):
know, small models are where thefuture is, you know, it's not we
talked to the very beginning ofthe conversation about the giant
versus the small, go for smallstuff you have you there's very
likely that you have a GPU in athome, it may be in your laptop
or something that you can buyfor a couple of $100 that they
can do all sorts of coolinferencing with an existing

(01:23:29):
model that you can then go dosome of the stuff with. So with
open source, that's the place togo. That's where I think that's
where I think the majority ofinnovation is really driving
from.
And it's a good place to startand figure out what is
interesting to you. And eventhat area, I'm really into, I'm

(01:23:49):
gonna also pitch a language thatI'm into, which is Rust. I
mentioned Go beginning of theshow. Love Go. I and I use that
in a in a lot of environments,but I've been using Rust as a
replacement for c c plus plusbecause and it's great for
embedded, you can use it with nooperating system at all.

(01:24:09):
And it's fast as can be. And soI've been that's been like when
I go play on my own, aside fromlike work work stuff in this
area, I'm always I mean, I'mevery day I'm looking at all the
innovation in the Rustcommunity, to do small little
projects that I can do for fun,that drives my own passion
forward. It doesn't have to be agiant 800 pound gorilla, or

(01:24:33):
defense industry or whateverkind of thing. Can be something
that the kid in you or maybe thekid in your house can go do on
their own.

Adam (01:24:42):
Give some shout outs to, I guess, some crates or some
projects out there in the restof world. Think probably Tokyo
or Tokyo probably is one ofthem. Saturday is probably one
of them. What else should youplay with?

Chris (01:24:52):
Tokyo is really good because it allows you to kind of
multi threaded things, manythings happening at once, which
is really important in robotics.And so that's really taken off.
Is I'm trying to remember thename Embassy is the name of it.
Was trying to remember. Forembedded, it is a runtime in

(01:25:12):
Rust that allows you to do awhole bunch of embedded
capabilities without writingeverything from scratch.
It kind of gives you thisframework. And so you can go get
a Raspberry Pi, even one of thesmall ones, I think, the Nano
and stuff that isn't theresupporting the OS and use
Embassy to create an executablethat runs on something that's

(01:25:35):
too small for an OS. And so Ilike exploring all these
different possibilities in termsof how you're gonna And when I
said Tokyo being multithreaded,it wasn't multithreaded. It was
concurrency. I said the wrongthing, so I just wanted to
correct that before we get toofar.
But being able to do highlyperformant, concurrent things on

(01:25:56):
very small pieces of hardwareout on the edge is a real thing.
Like five years ago, it justwasn't possible to do anything
like what we're doing now. Butthe beginning of the show, we
talked about the revolution ofall these different areas coming
coming together. Well, nowanybody can go use different
languages, but in my case, Rust,and find small bits that cost me

(01:26:21):
$10 out there and put someunique software and do something
that scratches my itch that noone in the world has done. And
it's no longer out in the cloudor out on some computer.
It can be something that I'mcarrying around on my body or is
literally a robot. This is allreachable now. And so that's

(01:26:42):
really what I would encouragepeople to do is the future I get
asked all the time about thefuture of AI, and I really think
the next big revolution in AI isgoing to be physical AI, is AI
imbued in all these things inour life that we've been talking
about, that we refer to as onthe edge in the software world.
But that's going to be the newnormal. And now you can do that

(01:27:05):
without any real budget on yourown, anytime, from any place in
the world.
So this, if you want to gocreate the future, and I said
this is the coolest time we'veever lived in, boy, you can go
create that right now no matterwhere you're living and no
matter what your budget is. Sothat's what go do it. If you're

Adam (01:27:23):
tinkering with Rust right now so let's say you're done
with this podcast, you're offfor the day. Let's just say
magically, have nothing to do.You're gonna go pick up your
next or your current Rustproject. Maybe you've got a new
model you wanna play with. Whereare the places you're going?
You mentioned Hungry Face. Whatare some of the stack that
you're tapping into?

Chris (01:27:43):
So there's the swarming stuff that we've talked about
and trying to figure outrobotics and all that. We've
talked about home automation.Think that feels, for answering
this question, that's anaccessible thing that I like to
do now. So as I've picked upthis kind of home automation
stuff, I'm trying to figure outwhat can I do? I go get some
Raspberry Pis or I can use aslightly larger, like a mini PC

(01:28:07):
to do something in the house.
None of this costs much. And I'mnow, on my day to day, when I'm
just at home and I'm notthinking about the day job, if
you will, I'm looking at all thethings that I do with my family
and thinking, wow, I can go picksomething to handle that. So
like almost all the lights inour house are automated. A lot

(01:28:29):
of the appliances are automated.We have voice command from
anywhere in the house where wecan tell a particular assistant,
Go do this, and it happens.
I've been starting to integrateAI agents into that workflow.
Now that that is becoming superaccessible with all the There's
so much open source that havemade agents very easy to do, and

(01:28:53):
you can get small models offHugging Face and run it off
compute that you have in yourhouse already. And so that's the
kind of thing that I like to do.And I think it's amazing because
it's gotten people in my familywho are like, Oh my god, Chris
is doing technology again. Like,have the family members.
They're like, Yeah, yeah, Idon't want hear it because
you're talking about that witheveryone else all the time. But

(01:29:14):
now they're like, they're usingthat and they're getting
interested in like Yeah, they'relike,

Adam (01:29:18):
Tell me more, Chris.

Chris (01:29:19):
Yeah, they'll start

Adam (01:29:21):
Not a swarm, though. Tell me more.

Chris (01:29:22):
Not a swarm. But my wife will say, How could we automate
this to make it better? And Icouldn't get her to think Yeah,
of she didn't want to get was mything, and just stop talking
about it, Chris. My daughter isstarting to get really She's 13,
and she's really starting tothink about what can we do. And

(01:29:44):
it just sparks the imaginationbecause it's real and it's
tangible.
So that's why I get to go dosomething. Just decide today
you're a maker. Go get somecheap stuff. Have a vision.
Recognize that every part of itis either free or only a few
bucks, and just go do somethingin your imagination.
If you can't think of anything,there's tons of websites with

(01:30:06):
maker projects out there, andfind something that you go, Oh,
god, that's cool, and just go

Adam (01:30:11):
do it.

Chris (01:30:12):
Even if it It doesn't have to be the greatest thing in
the world. Just go do it. Andthen you're helping push all
this stuff forward. You arediving into the future and
making this stuff happen. Andthat's why this is the greatest
moment in the history of the Itreally

Adam (01:30:24):
is. I mean, we went from photography or from painting
photos to photography in a blinkof an eye. And now we're
thinking, gosh, I just wouldn'tlike paint the picture that way
ever again. I would just takethe photo because that's the
way. Yeah.
It's a cool moment in life. I'msuper curious about one one
particular area that youmentioned. You mentioned voice.

(01:30:45):
Are you leveraging Alexa orleveraging the behemoths or are
you home assisting in it andyou're doing something with home
assistance?

Chris (01:30:52):
So I am moving. We have been for a while, Alexa all over
the place. And given the factthat I am increasingly concerned
about privacy, just in terms ofsurveillance is so built into
everything now that I amgenerally moving from cloud

(01:31:16):
based systems into more privatesystems that are completely
under my control and local andstuff. And I realize that may
not be for everybody. I thinkpart of that is because I work
in a world that is obviouslytouching on intelligence, and
I'm more aware of what'spossible from a surveillance
standpoint than probably mostpeople are, and how pervasive it

(01:31:38):
is.
And that makes me obviously wantto kind of protect our own
privacy a little bit. So I'mkeenly interested in automation
that's not specificallycommercial cloud dependent.

Adam (01:31:50):
We should circle back in the new year for a deeper
conversation. I'm sure you'llhave some time away, maybe new
progress, new projects, and newinsights. Because these are
things I'm about to go into inmy curiosity is I haven't
automated anything in my house.They're like, Adam, you're such
a nerd, you care about Home Lab.I'm like, yeah, I don't care
about that part of the Home Lab.

(01:32:11):
It's a different area of theHome Lab that I'm trying to
conquer.

Chris (01:32:14):
I didn't either. For me, the kick in the butt was buying
a house that already came with alot of automation in it. And
it's not just catching up onthat and learning. There was a
certain ramp, I had to level up.But then there was also the It
starts getting your imaginationgoing.
You knew in the back of yourmind you could do this, but now

(01:32:35):
you're living it. And thenyou're thinking about the next
five things after that. And Ithink that's it. Once you do a
little bit, it wets yourappetite, and you start seeing
all the possibilities. Andthat's what it took for me,
professional technologist, but Iwasn't really doing it until a
year ago.
And now this last year has justtake off.

Adam (01:32:54):
Being able to host models locally, have that privacy, the
fact that Home Assistant is sopervasive and so massive as an
open source project that theyhave a you could tap into via
the API, you know, whateverlocal, you know, models you have
running for inference, They havevoice capabilities. There's just
so much happening there. Whygive that data to, you know, to

(01:33:18):
Amazon? It's not that they'rebad. It's just that I have
preferences, and the preferencesdon't involve me telling what I
want, and then now I get hitwith ads for x, y, and z as I
scroll the Internet.

Chris (01:33:29):
People often complain about how creepy it is that
you're almost just thinkingabout something and then it
shows up in your Amazon cartkind of thing, you know, or
Google or whatever. And and likebut you're do you're doing that.
You're giving them that powerover you. And so to some degree
and it's not happened all atonce, but I'm taking
responsibility for the fact thatthat's been my choice because it

(01:33:51):
was the easy way to go, becausethey were providing this
ecosystem. I didn't have to domuch.
It just happened. All I had todo was let them was say yes
every time they send the updatedterms and conditions, and they
would take my data and dowhatever they wanted, and there
they are. And I've kind ofgotten to that point where I'm
done with that, and to somedegree and and turning around.

Adam (01:34:14):
Just gave me an idea, Chris. They you know, somebody
should I don't if this isactually a good thing or not,
like, AI is great at scanning anentire document, like a terms
and conditions. There was adocumentary, I think, on Netflix
about this that if you try toread all the terms and
conditions you would agree to inmodern society, you would spend

(01:34:34):
more than your entire life justreading terms and conditions.

Chris (01:34:37):
It's a lot

Adam (01:34:38):
of terms updates and or literally scrolling them to say,
yes, I accept, is not is notpossible. It's not realistic

Chris (01:34:46):
Yeah.

Adam (01:34:46):
Of a request from the people. So we're agreeing to a
lot of things just out of thenature that we don't have the
time to do it.

Chris (01:34:53):
And you're not going to if you're trying to get
something done. And now you haveto do through terms and
conditions to get somethingdone. They do it at that moment,
because they have you, they knowyou have to get something done.
And what are you gonna do? Go,well, I had to do the thing.
It was really important, but nowI can't do it because I'm not
gonna do terms and conditions.

Adam (01:35:10):
Correct. You're correct now.

Chris (01:35:11):
Yeah. So I'm I'm starting to invent my own world.

Adam (01:35:14):
Whereas the kids say cooked, you're cooked.

Chris (01:35:15):
Yeah. I'm starting to invent my own world where I I'm
not bound in that little prison,if you Well,

Adam (01:35:23):
that was cool. Thanks for deep diving on the swarm, not a
swarm, Rust, all the things.Make sure, if you don't mind,
some of the things that you canlink us to in the show notes,
I'm sure you got lots of links.Just spam us with all your
links. We'll put them

Chris (01:35:38):
in the

Adam (01:35:39):
show notes for everybody.

Chris (01:35:40):
Fantastic. Thanks for having me in, guys. It's been
great catching up with you and afun conversation.

Jerod (01:35:46):
Tons of fun. Go listen to practicalai. Practicalai.fm. If
you want more, Chris, that'swhere you find it.

Chris (01:35:51):
Thank you, Jared. I'm so glad you did that because we
would love for people to jointhe conversation. And we all
It's one big happy family, aspeople can see here. I love
Changelog, and I hope some ofthe Changelog people who haven't
given us a shot will give us ashot and join our conversation.
So

Adam (01:36:08):
you go. Practicalai.

Chris (01:36:10):
That's s it.

Adam (01:36:12):
You go. And be square as they would say in the eighties
or nineties, which is cool now.It's cool. It is cool now.

Jerod (01:36:19):
Yeah. The eighties

Adam (01:36:20):
and nineties are cool again. Good stuff, Chris. Bye,
friends. Bye, Chris. Bye,friends.

Chris (01:36:23):
Thanks, guys.

Jerod (01:36:31):
Alright. That's our show for this week. If you haven't
checked out our website, head topracticalai.fm, and be sure to
connect with us on LinkedIn, X,or Blue Sky. You'll see us
posting insights related to thelatest AI developments, and we
would love for you to join theconversation. Thanks to our
partner Prediction Guard forproviding operational support
for the show.
Check them out atpredictionguard.com. Also,

(01:36:54):
thanks to Breakmaster Cylinderfor the beats and to you for
listening. That's all for now,but you'll hear from us again
next week.

All Episodes

Episode Transcript

Popular Podcasts

Dateline NBC

Are You A Charlotte?

Stuff You Should Know

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}Chris on AI, autonomous swarming, home automation and Rust!

Episode Transcript

Popular Podcasts

.css-r6mb8g{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:1;overflow:hidden;}Dateline NBC

Are You A Charlotte?

Stuff You Should Know

All Episodes

Chris on AI, autonomous swarming, home automation and Rust!

Dateline NBC