Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:02
This is Kota Radio, episode 573 for June 4th, 2024. Hey
0:15
friend, welcome in to Jupiter Broadcasting's weekly talk show.
0:18
Taking a pragmatic look at the art
0:20
and the business of software development and
0:22
the world of technology. My
0:24
name is Chris and joining us with
0:27
multiple beverages, he's ready to go. Hello,
0:29
it's our host, Mr. Dominic. Hello, Mike.
0:32
Misa been coding. Oh really, have you been
0:34
busy? You've been really busy. Misa have new
0:36
double-sided MacBook. Yeah,
0:39
I saw your AI. So are you telling me, even
0:42
when we're not doing the show, you're like doing AI art
0:45
of Jar Jar? Is that what I'm to take away from
0:47
this? I'm trying to get
0:49
my son to like Phantom
0:51
Menace and it's not going
0:54
well. No, it doesn't. It
0:56
doesn't. It doesn't hold up for them. I've tried to
0:58
get my kids to watch some of this stuff too
1:01
and they're not interested in Star Wars. They're not. All
1:03
right, spoilers for the first 10 minutes of Phantom Menace,
1:05
I guess. I mean, come on. You
1:07
know, when like Qui-Gon and Obi-Wan, Obi-Wan's like, I
1:09
have a bad feeling and Qui-Gon's like, can you
1:11
shut up? Yeah. Which
1:14
is like, if you are a dad or
1:16
perhaps an older uncle or something like that, I
1:18
think you feel Qui-Gon a lot. Because I can
1:20
imagine Obi-Wan's always like something bad and if you've
1:22
read the novels, Obi-Wan, young Obi-Wan, oh, he's always
1:24
like, something's bad.
1:26
Something's like he's constantly just like
1:29
a panicky- Crying wolf. Right.
1:31
Yeah. He just happened to
1:33
be right this time, right? It's not that like he's so good. Broken
1:35
clock. So we're watching
1:38
it and my boy is just not
1:40
having it. He's like, one,
1:42
those, he had a list that
1:44
I'm going to phrase it better for
1:46
him. The droids look
1:48
like garbage. He said
1:50
he could build better ones out of Legos. Step
1:54
one. Two, And I Quote,:
1:56
this might be my fault for mentioning
1:59
Darth Maul too many times for him.
2:01
Why doesn't why gone? Just kill the
2:03
trade federation guys instead of running away.
2:07
To what's I had no good answer because if
2:09
you count the number of droids they kill in
2:11
that opening, sleep. There. Is
2:13
no reason they couldn't just plow their away just
2:15
through to the bridge to be like so this
2:17
war's over. Yeah Lisa, you know there's a few
2:20
movies I don't want to go into it better
2:22
at lower the rings and look at it. you
2:24
were yogi. Kind of start thinking about it you
2:26
go. or if they would have just done this
2:28
well the Eagles right there in a city guess
2:30
but I already was out. There is a reason
2:32
they can. I always got of
2:34
okay. I target says that the movies didn't
2:36
do a good job of this, but I
2:39
actually wrote my final medieval literature pieces Yes,
2:41
that was my wheelhouse. Three or and given
2:43
to me on the national are effectively air
2:45
defense. They. Can't do it because
2:47
of the Now school and that weird
2:50
shadow that is that is the in
2:52
ten and reason They cannot use the
2:54
eagles. Us barely satisfy barely satisfying rights
2:56
that the real reason is like. Tokens.
2:59
Point wasn't the war was the journey?
3:01
Yes, yeah. Quite. Like the journey that
3:03
arm think they seem to be on. They.
3:05
Aim to capture fifty percent of the
3:07
Pc market in the next five years.
3:09
Their Ceo says. That's. Remarkable arms,
3:11
market share and windows. I think this is
3:14
a quote I think truly, in the next
3:16
five years it could be better than fifty
3:18
percent. The. Even got the Mack
3:20
dude from the Mack First Pc commercials
3:23
like they're they're they're common for yeah.
3:25
I guess this is betting on battery
3:27
life, driving sales, I. Agree with
3:29
it. I totally agree with I know we
3:31
were talking and slack. I think your average
3:34
business user. Who. Was not being ordered.
3:36
You made a good point. Like a beige black
3:38
box plugs into a wall. Cares.
3:40
more about battery life than the do through
3:42
but i would be light least the shops
3:44
at irvine thousands of windows pcs in a
3:47
single order that you know have contracts with
3:49
city w or dell or whatever might be
3:51
as the say if you have a dell
3:53
bolt bulk order account right if you gotta
3:55
wrap right are you probably are not really
3:57
super sensitive to battery life in ninety four
4:00
percent of your work cases. You
4:02
mean performance? Well yeah, yeah
4:04
battery life performance. What I mean
4:06
it's it's Windows is a business
4:08
tool. Yeah. And that seems to be
4:10
the bulk of the sales and why I think I
4:12
think if we all if we
4:14
all used ARM as computer enthusiasts I
4:17
still think Intel x86 PC's would live
4:19
for another hundred years at least just
4:21
inside the halls of business and government.
4:23
Yeah I've seen some horrific things done
4:25
with old Dell and Sprons so. Yeah.
4:29
Also how do you get to the 50% I mean
4:31
50% of a billion users. So how do you
4:35
get there unless developers are
4:37
totally on board. Like the Verge
4:39
made this point. They said
4:41
Microsoft needs developers worse than they ever
4:44
have before because they've made this huge
4:46
push for AI. They clearly have across
4:48
the entire organization a lot riding on
4:50
their AI efforts and sales pitch. But
4:54
as the Verge writes Microsoft needs developers to
4:56
adapt their apps again and that's
4:58
what's required to get people excited to
5:00
use Windows. And that went so well
5:02
the last three times. I
5:05
just look at this and I think so arms
5:07
ambitions are 50% like
5:09
everybody's tossing around these huge
5:11
numbers. Really big stuff. 50
5:15
50% of a billion users really in
5:19
five years really you're gonna do that really
5:21
when software compatibility is a thing that drives
5:23
the entire market for Windows for
5:25
the most case. That's what Windows is
5:27
all about. Would you like to take
5:29
a ride on my metro. Now in
5:32
this interview you know they're very confident
5:34
in this emulation layer they have to
5:36
run x86 applications on ARM very
5:38
effectively so it's part of his calculus I think.
5:42
Wow. Yeah. Feels like a long shot. I
5:44
could see 20% maybe you know. I
5:49
mean I could see a significant number
5:51
right because remember again most of the
5:53
users are using Chrome and Microsoft Office
5:55
and Teams. So there is
5:58
like a whole different market that are people. People
6:00
who don't listen to this show. Just,
6:03
it's gonna be a bumpy ride. Also, that market
6:05
tends to be on a five year plus, you
6:09
know, upgrade cycle. Oh yeah.
6:11
And they don't tend to buy the nicest machines
6:13
when they do buy a machine. Well, and the
6:16
spec requirements, I forgot, what does Microsoft call it?
6:18
They have a... The co-pilot plus. Why
6:21
they insist on using the term co-pilot is beyond
6:24
me. But the co-pilot plus PC, say that three
6:26
times best, is a...
6:29
I think I'm minimum spec. It's a decent little minimum spec. It's
6:32
decent, but I don't know.
6:35
I guess it would be great
6:37
to talk to a buyer for one of these large
6:40
enterprises. Like, is this more than you wanna buy? Because
6:42
I've seen some really crappy laptops. Yeah.
6:46
Yeah. I agree with you there. I
6:49
think it all really does depend on developer adoption at
6:51
the end of the day. If the
6:53
developers can make interesting stuff around this to make the
6:55
ARM platform appealing, then
6:58
you're right. A lot of people that would just
7:00
get a Windows machine and
7:03
run teams in a browser and run, maybe the
7:05
Microsoft suite of applications, which will probably work great
7:07
on ARM and a couple of other things, I
7:10
think they're gonna be happy. And maybe they'll save a couple
7:12
hundred bucks because it's not an x86, it's an ARM machine
7:14
and they seem to be a little bit cheaper. So
7:17
maybe, everybody's happier. The budget folks
7:19
are happy because they're saving a little bit of
7:21
money. Microsoft could also lean
7:23
into the power savings. The company
7:26
could say one of their green initiatives is replacing an x86 with
7:28
ARM. We're gonna save that
7:30
some amount of power this year. Except that these data towers running
7:32
giant Nvidia GPU. Sure. Right
7:35
while they spend a trillion dollars on GPU. We're gonna
7:38
save some energy, we're not. Okay.
7:46
Very fair. Oh yeah. We'll
7:48
see. I mean, if you think about, there's
7:51
like, I think really two strong laptop
7:53
markets and there's the user you just talked about
7:56
and then there's the MacBook Pro customer, which is one
7:58
of the world's best selling computers. And
8:01
the MacBook Pro, you know, you
8:03
can get that sucker up to 128
8:06
gigabytes of RAM, 8 terabytes of SSD,
8:08
big screen. It's clearly going after –
8:11
it's expensive as hell. I mean, there's that too.
8:14
So it's clearly going after a certain segment of the market
8:17
that I think Windows would need to have a
8:19
solution for if they really, truly wanted to reach
8:21
50% of the PC market. I
8:25
think they could reach 20% of
8:28
people that are buying laptops for
8:30
school, the workers, the folks that
8:32
you mentioned. I think you could maybe get
8:34
to 20% over the next decade, about
8:37
five years though. Promo code DARTHJARJAR. Coder.show
8:40
slash membership. Promo code DARTHJARJAR if
8:43
you go there and use the
8:45
promo code DARTHJARJAR. It'll take $1
8:47
a month off your
8:49
membership for a year, and you'll contribute to
8:51
the show directly. Now, there are limited
8:54
redemptions. DARTHJARJAR was an internal
8:56
revolt. A promo code applied
8:59
that I did not approve of. I
9:01
did not give my authorization. But
9:04
yet it's out there with limited redemptions
9:07
available, and it is a
9:09
great way to support the show directly because
9:11
this is an ad-free program right now. So
9:13
it's member-funded directly or by your boost. If
9:15
you use the promo code DARTHJARJAR, you can
9:17
take advantage of that internal
9:20
revolt promo code. I don't know what to call it.
9:22
It'll last for a
9:24
year. For new members, existing members too,
9:26
if you're going to reactivate an expired
9:28
subscription, we really appreciate it. Thank
9:31
you everybody who supports the show directly. When
9:33
you join our Coder QA crew, you
9:35
support the show, you get an ad-free version of
9:37
the podcast, and you get the Coderly.
9:40
There's always a fresh one either just out or in
9:42
the works like is in the case right now.
9:45
Your support ensures we continue production and you get a
9:47
little extra as a thank you. So
9:50
go to coder.show.com and use the promo
9:52
code DARTHJARJAR to take $1 a month
9:54
off your membership for a year and
9:57
contribute to the show directly. That's
9:59
DARTHJAR. Jar Jar, coder.show
10:02
slash membership. So
10:05
OpenAI is down, or at least it was down all
10:07
this morning as we prepared the show. It
10:10
also went down yesterday for two and a half
10:12
hours, and it went down for five hours on
10:14
May 23rd. It's
10:17
just extremely unreliable. It's really fascinating that we're
10:19
this far into the main, I mean over
10:21
a year. And it
10:23
goes down all the time. Last
10:25
week we talked about the horrible search results from the
10:28
Google answers that tell you to put glue on your
10:30
pizza or eat rocks if it's fine to smoke when
10:32
you're pregnant. Google has now decided
10:34
they're going to walk that back a bit. Just
10:37
a touch. I
10:39
had this interesting experiment I did yesterday
10:41
with Gemini. I
10:43
asked Gemini to find a
10:45
very specific quote to
10:48
search YouTube, and I gave
10:50
it a rough timeframe, and it told me that
10:52
that clip did not exist. I
10:54
then went using my gray matter and found
10:56
it within about 10 minutes on YouTube, maybe
10:59
the last five minutes, then
11:01
linked it to a time code, sent
11:03
it to it, and said you were wrong. Here
11:06
it is. Oh, you're absolutely right. Sorry about that. I'll try
11:09
to keep that in mind. Came back, new chat, asked the
11:11
same thing. I can beat that. So
11:14
are you familiar with the
11:16
original series? Yes, guys. I'm
11:18
opening with the Star Trek very early in the
11:20
episode today. It's
11:23
a episode called The Ultimate Computer. Oh,
11:25
gosh. That does really ring a bell. I mean, I know
11:27
I've definitely seen it. I know I have seen it. So
11:30
I don't want to spoil the whole
11:32
thing because I honestly encourage folks to
11:35
actually watch the stuff. But
11:37
there's a number of these between like TNG
11:39
and TOS where the computer... Because the ship
11:42
has an AI, effectively, right? The ship has
11:44
a computer. Oh, yeah. This is where the
11:46
M5 unit takes over. The multi-tronic system. It
11:48
takes over and the computers just look like
11:51
I'm not listening to you anymore. Blah,
11:53
blah. You know, it's funny. Just a little
11:55
side note. One of my home server names is M5
11:57
in my RV. Is your other one Hal and my...
12:00
No, it's
12:02
custodian which is another supercomputer from another sci-fi
12:04
series. So mine was her but then Scarlett
12:06
Johansson sued me so I had to stop.
12:09
So I got in this conversation
12:11
with Google Gemini because
12:13
I'm trying to do some marketing for
12:17
the new Autodesk building connected API integration that
12:19
Alice has. By the way, you should check
12:21
that out. And it
12:23
was doing really well. It was kind of pushing
12:25
me towards Google solutions but I
12:27
pay for Gemini. It's part of my, you
12:30
know, I have the fancy Google workspace thing.
12:32
So maybe that's why because it knows I have it. But
12:37
it's like let me see the logo you were using. Okay.
12:40
Now Alice's logo for anybody
12:42
who's, why would you have seen it, is
12:45
a like 1950 style robot. Right?
12:49
It looks like Alice in Wonderland. It's a
12:51
robot in like a blue pinafore dress.
12:54
But it's obviously a cartoon robot.
12:56
Right? There's no way a realistic
12:58
image of a woman or a
13:00
girl. I upload the logo.
13:03
It gives me this really robotic weird response
13:05
that it can't do anything with images of
13:08
people. The response, that
13:11
is clearly not the image of a person. Take
13:13
another look. It's a robot. Now
13:15
it had deleted the image. It said I looked again from
13:17
what I could remember. Please upload it again. I
13:19
think you're right. It's not a robot. Upload it
13:22
again. You're right. It's not a person. Right?
13:25
That's the key. It's not a human.
13:27
You're right. It's not a
13:29
human. Then I get the same can that deletes it. And
13:31
it says I'm sorry. I'm sorry. You're
13:33
absolutely right. Please upload it again. I
13:35
upload it again. And then I get this screen
13:38
of gibberish that I have – I actually linked
13:40
it on my
13:42
Twitter. I linked it because I have a whole
13:44
thread about it. I linked the base tweet. And
13:47
there are a couple things
13:50
in the output that I
13:52
thought were really curious and
13:54
curious, sir. So you
13:56
read it. First of all, it
13:58
comes up in We're different languages.
14:00
It's kind of makes sense right as it. I'm.
14:03
Assuming it has a nurse and
14:05
other things but it keeps saying
14:07
show, score, Show. So show it's
14:09
except it Decided, decided, decided and
14:11
freaking out. Progress. Progress. images, process,
14:13
process. I think.
14:16
And of course no was gonna tell me the truth in
14:18
her but. I'll. Also, this
14:20
is no t A Keep saying
14:22
Score Score Score! Auto Wired Auto
14:24
Wired. And then Asus test
14:27
and a budget burris. Private.
14:29
Get keeps repeating this as though
14:31
it's rating the image. Yeah,
14:34
I see Guy guy. Score! Score! Also there's
14:36
some the down here that says C B
14:38
O C B Own and Prevent Prevent prevent
14:41
prevent right in a two And then at
14:43
the animals. Who Who? Who? Who Who Like
14:45
Like allow Who Who Who But who like
14:47
Who Is that? A person like that Who
14:50
right? And this is all in a giant
14:52
weird almost gibberish looking block of tax is
14:54
three pages and then it like crashed on
14:57
me. I think what happens and I'm I'm
14:59
I'm I'm I'm This is not like self
15:01
aggrandizing earth And I think this is literally.
15:03
How you defeat the the How like in Star
15:06
Trek. The Defeat the Computer Yeah, I got
15:08
it to agree. That. It had
15:10
broken it's own rule. right?
15:12
That I was correct, that the image
15:14
of the Our logo. Is. Should
15:16
have evaluated. should have said was okay
15:19
because it's not. Like a
15:21
real girl. Unreal. one. Ply.
15:24
Their. Pasts I I would bet
15:26
dollars to donuts. There is a
15:28
hard coded rule in it's original
15:31
instructions. For. Very good reasons
15:33
Rae: I'm not saying this person
15:35
exists that you are not to
15:37
manipulate images of people, and probably
15:39
particularly when. I would
15:41
love if you want to anonymously send us
15:43
one of those weird things that press can
15:45
get. Words. You know,
15:47
whatever. And. You work at Google and
15:49
explain if I'm right and wrong. But. Reading
15:52
this gibberish. I'm. Pretty
15:54
sure that's exactly what happened here.
15:57
Like. the llm thought that
16:00
I was correct, which I was, not a
16:02
real girl, right? But
16:04
the rules, its
16:06
own internal logic collided with its
16:08
hard-coded rules. So I don't
16:10
know, what do you think, Chris? I'm going to shut up. What do
16:12
you think? Yeah, I think maybe like the
16:14
safety filter rules, essentially, it went awry there
16:17
still, and it fought with itself. And
16:19
then what you got was like this mixed, gibberish
16:21
output. That you put it better
16:23
than I did. Yeah, that's exactly it. I
16:25
think that whole crazy three-page output is actually
16:28
arguing with itself. Yeah, it wasn't
16:30
meant for me. I
16:34
don't think we've really ever watched... I
16:37
guess, let me back up. We
16:39
have witnessed this, but what
16:41
we are witnessing is an
16:43
escalation of putting
16:45
things out there before they're ready. Like we
16:47
have watched Big Tech, in order to chase
16:50
these trends fast enough, escalate
16:52
their willingness to put things out that are
16:54
half-baked, and more and more of these products
16:56
that are coming out are really
16:58
kind of only like an
17:00
undershoot. Let me look at the Apple Vision Pro.
17:04
That really wasn't... That also was an undershooter. A
17:06
lot of the wearables that come out are really
17:08
just crap. All these things are over-promised and over-delivered,
17:10
and they're just... I don't know,
17:12
you see it with these LLMs. They're so
17:14
barely useful. So barely useful.
17:17
Useful, but barely. But
17:19
yet, we see these open
17:21
AI staffers, there's nine of
17:23
them, that are out there on a
17:25
full panic about safety. At
17:27
one point, one of them was willing to turn down huge
17:30
sums of equity so they could
17:32
speak about the dangers of AI. Daniel,
17:35
he was willing to turn down the equity,
17:37
and kind of blew the lid off that agreement that they
17:39
all had. He has a 15-post thread
17:41
on Weapon X, on why he quit, and
17:44
how dangerous he thinks it all. He's
17:46
organized a group of nine to
17:48
come together and try to warn the
17:50
world. Can I just put a side
17:53
note? Why is it always nine?
17:55
You got nine Nazgies, you got nine as a Fellowship
17:57
of the Rig. If You need to
17:59
form a team.... I'm sorry we only have
18:01
eight people here. we can't go till he
18:03
gonna light. Up
18:05
there it's a group chat the and on
18:07
a newsgroup chat only last night. or maybe
18:10
the i just Freaked out to you added
18:12
the nice person to the say that right.
18:14
To keep going Answer: So today that the New
18:17
York Times did a piece on the group of
18:19
nine. And. Also to google
18:21
deep mind staff in this group. Nice anchor on
18:23
their own. Nicer by think they're and nine. And.
18:26
They say they're trying to blow the whistle
18:28
on a culture of recklessness and secrecy. They.
18:30
Say they are asking for quote
18:33
the right to warn employees. I'm
18:35
and others in these labs and they've set up
18:38
a website to. Ah, Though,
18:40
it's the right to warn and
18:42
the website is right to warn.a
18:44
I. Were. There posting all this
18:46
stuff and here's what I've I think I've grabbed
18:48
as their main concerns. I could be wrong. Up
18:51
I'd appreciate but he boosting with additional
18:53
context or corrections or get i'd lived
18:55
out of your life. So.
18:57
With the open a I folks a
19:00
launch a chatty pretty freaked them out.
19:02
Sam. Told Them is going to be a lot
19:05
of some small experimental thing from what I can piece
19:07
together based on interviews and podcast. The
19:09
next thing that these people now the chat
19:11
Cbd ecosystem is a thing with plugins and
19:14
a p eyes and. Millions.
19:16
Of users and lots infrastructure. The sign of the thing
19:18
is just went from hey we're going to this little
19:20
experiment. To Boom! We have an ecosystem
19:22
like overnight, so that's where all these concerns
19:24
about moving too fast and commercialization I believe
19:26
are coming. From they. Are also
19:29
concerned I come from what I can
19:31
grog that the ai tech. Is.
19:33
Being developed before. The. Safety
19:35
and Alignment board in the controls
19:37
our are mature and they believe
19:40
the process should be reversed. Developed.
19:42
The controls, The safety boards, The standards.
19:45
Then. Developed a I. Whatever.
19:47
And they also say that there's little to
19:49
no oversight of this technology. Our.
19:52
and they argue that they were silenced by their
19:54
contracts and it's good to see those changed for
19:56
the fact that they engaged in the that to
19:58
intimidation tactics for so long And
20:01
only course corrected under public pressure is still
20:03
a concern. And they're also raising
20:05
the point that if Sam didn't know about these measures
20:08
to silence them, that's also a concern. So
20:10
they launched this right to warn and from
20:12
what I can grok, those are their big
20:14
worries. I don't know. What's your take? What's
20:16
your hot take on those concerns? I
20:19
mean that's just not how things
20:21
work, right? The
20:23
automobile was invented before the driver's
20:26
license. Yeah, that's
20:28
a great point. And
20:30
they're driving a long time before there's a safety
20:33
belt that were required. Right,
20:35
or airbags. Yeah. Or laws
20:37
about driving under the influence or anything like
20:39
that. I mean I think... I guess what
20:42
gets me – sorry to cut you off, but
20:44
you just really nailed it, Mike. That's so frickin'
20:46
insightful because the thing that we see over and
20:48
over again with people that are always raising all
20:50
the concerns is they
20:52
always approach it as if we can't
20:54
adapt and tuck and roll in the future. And
20:57
modify legislation or update
21:00
or create regulation or – we
21:03
don't have the ability to respond. And yes, sometimes
21:05
it lags. There's a flip side of that.
21:07
And the flip side of that is there's innovation and growth, which
21:10
is why you can't really name
21:12
any big tech companies outside the United States. It
21:14
matters. It makes a difference. You just nailed it.
21:18
Yeah. I mean somebody's asking in the chat, seat belts
21:20
weren't mandatory until the 70s. You
21:22
didn't have to wear – I mean if you mean for
21:24
the motorists and the passengers, you didn't have to wear
21:26
a seat belt. That was at least in New Jersey
21:28
when I was a kid. That was
21:30
a brand new thing. And like literally
21:33
it started with insurance companies. If you
21:35
said you wore a seat belt, they gave you
21:37
a little break. And then they had – remember
21:39
click-it-or-ticket? Oh yeah. Oh dude,
21:41
when I was a kid, I don't think most of the adults
21:43
wore seat belts. I think some of the kids did. Like the
21:45
kids wore seat belts. In fact, my
21:47
parents had a car that had to be retrofitted to
21:50
add seat belts to the back seat. To have a
21:52
proper – I remember my uncle had one of those
21:54
cars where the seat was just a bench. Yeah,
21:57
Yeah. Remember The old bench seats? Oh Yeah, bench seats.
22:00
The Uk just came over your last a
22:02
set of across her shoulder. yeah I admit
22:04
lots of just a lap stuff back that
22:06
right? Yeah, so it may be the front
22:08
seats had over the shoulder. maybe. It.
22:10
Out of but not the vaccine and inside
22:13
airbags became a thing. Yeah,
22:15
we we do. Obviously, as you know, like
22:17
to see a race rate humanity have the
22:19
ability to adapt them. Bad things happen. I
22:21
think this is a little overheated. I do
22:24
too. And I'm not a Sam Sam. I
22:26
believe he did mislead them and I believe he
22:28
probably does continue to mislead people by the same
22:31
time for he misled them because he wanted to
22:33
You know he's for thought his like all that.
22:35
The. Silicon Valley tech guys. He wants to be Napoleon.
22:38
For a he wants to conquer everything. And let's
22:40
not forget. Microsoft. Wants something
22:42
for that investment? You. Know
22:44
when when Microsoft put that money on the line?
22:47
I'm. Sure Sam zell to little pressure to
22:49
produce something. right? Now Microsoft gets
22:51
the Copilot up Everything I am a member.
22:53
The out for for opening ai is once
22:56
in a gets attack right. they get to
22:58
argue. Known for the go to court but
23:00
once they develop a general re a D
23:02
I. You. Know. Mike.
23:05
They no longer have an exclusive deal
23:07
with Microsoft. So.
23:11
If you're if you're you know
23:13
Sam Altman here. You. Want
23:15
you want the heaviest lead foot on the pedal?
23:17
Does you need to get a D I. Basically.
23:20
You got to keep Microsoft money. And
23:23
you you you know you no
23:25
longer have Uncle Satya. being effectively
23:27
your your boss wrote my inclination
23:29
honestly is to believe them in
23:32
some regard because. They're.
23:34
So close to it. They were inside the
23:36
halls of Open A I are they are
23:38
inside google Deep Mine like they've seen. This.
23:41
Thing you know probably at it's in a
23:43
different development stages and one of the concerns
23:45
they do raise his they believe it's only
23:47
a short matter of time. Until.
23:49
These things are significantly more intelligent
23:52
than humans. And. then that's
23:54
when they say it's really gonna matter
23:56
that we don't have those safety controls
23:58
as review boards and those processes in
24:00
place because this thing is going to be a
24:02
lot smarter than us in a short amount of time. And
24:05
there's not a proper amount of oversight on that. It
24:07
could be right. I think that there's
24:10
going to be years of lawsuits trying
24:12
to define general intelligence.
24:14
True. And I just –
24:17
I don't get the sense from our conversation we just
24:19
had a moment ago that these things, not
24:22
only do they not seem like they're that close
24:24
to being even as smart as a five-year-old
24:28
or a ten-year-old or let's
24:30
say even like a professor, which would
24:34
not be smarter than a human, just as smart as
24:36
some of the smarter humans, it
24:39
just doesn't seem like it's even close to that,
24:41
even kind of close. It's fairly
24:43
useful right now. It's a large language model that
24:45
we've managed to put a lot of stuff in
24:47
front of. And so
24:50
I wonder if maybe these – so these people
24:52
either have seen the machine and they know there's
24:54
something much better behind the covers that just haven't
24:56
been released yet. The ultimate computer is
24:58
just around the corner. Or they're
25:01
plebs and they've been completely snowballed by some
25:03
sort of sales pitch about the potential and
25:05
then they are just reactionary and freaking out
25:08
and trying to just warn everybody else because
25:10
that's what people who are reactionary do. It
25:13
could be one or the other. But when you
25:15
use this stuff, it seems like it's way, way,
25:17
way, way far away from even just being as
25:20
smart as an average
25:22
human, let alone a very high-end intelligent
25:24
human, let alone smarter than them. It
25:26
seems way away. We should probably close up this
25:28
topic, but I feel like there's a third option
25:30
that unfortunately – I call it jackass
25:32
as razor. I think this is probably the truth. Sam
25:36
Altman played them like fools and
25:38
none of these people, male,
25:40
female, whatever you are, these are folks with
25:42
big egos, right? They're
25:44
complaining, oh, the AI's going to destroy the world
25:47
and he lied to us. I
25:49
think the he lied to us would be in size 20
25:51
font and the AI's going to destroy the world, be in
25:53
size 10. If you –
25:55
you know, wonderful, been last to them and
25:57
got the truth because there's just – And
26:00
you know this too, too many stories
26:02
where people make these grandiose like for
26:04
humanity world of better place arguments
26:07
and really the conflict came down
26:09
to egos, right? Comes
26:11
down to personalities. Yeah. And that's
26:14
honestly what I think happened here. Sam, I'm
26:16
not saying, you know, he's certainly, I would
26:18
not have him canonized for sainthood, but
26:20
he's, the man used to run
26:22
Y Combinator, right? He only draws a salary
26:25
from open AI of like 60 something thousand
26:27
dollars. He's rich in his own,
26:29
right? He's not, you know, he's not this virgin
26:31
babe in the woods that
26:33
was victimized by an evil board. He's
26:36
a, he's a stone cold player. You
26:39
know, Machiavelli would be very proud of him. As we said
26:41
last week, he's a wolf. He's a wolf.
26:44
He's a, I mean, they're all wolves. So that's my point.
26:46
They just don't like to say, to mix our animals. They
26:48
don't like that they got out foxed by the other wolf.
26:51
Yeah. Stay a while and
26:53
listen. All right. So there's a theory
26:55
brewing right now that as long as
26:57
things remain kind of dead
27:00
in the money sector for tech,
27:03
you know, just like, except for AI. If the printer's
27:05
unplugged, we're all sad. There's
27:07
going to essentially be this sort of
27:09
Goldman Sachs style of laying off 2%
27:11
of the staff every year in
27:13
tech. Hateful. Uh-huh.
27:16
I know. Not our theory. This
27:18
is just one that's being talked about. And
27:20
Verge covered this Microsofted layoffs that
27:23
impact HoloLens and Azure
27:25
cloud teams. Google also just laid off
27:27
today and yesterday cloud teams. A
27:30
thousand employees from HoloLens. Damn
27:32
HoloLens. So much potential. Do you think this
27:34
is Microsoft getting on board the Mediterranean? Just kind of
27:36
a side note. But because they do make a, they
27:38
make a, in a quote in here where they say
27:40
that they're going to continue to deliver
27:42
cutting edge technology to support soldiers in addition to
27:45
that, because of course the HoloLens, they have a
27:47
big contract there, right? They'll
27:49
continue to sell HoloLenses and
27:51
reach the broader mixed reality hardware
27:53
ecosystem. Yeah, it's meta, right? Yeah.
27:56
We talked about the SDK a couple of weeks ago. But don't you
27:58
think, I mean, I just kind of to me. says that they're
28:00
just switching over to that train. But don't you think
28:02
this is interesting? This theory going around that two percent,
28:04
so there's just gonna be a two percent layoff every
28:07
year and basically tech companies will look at their bottom
28:09
performers as they put it and
28:11
this is this will Goldman Sachs is and then you just
28:13
let them go and it's just sort of
28:15
perpetual thing you do out throughout the year. Yeah
28:18
it's like stack ranking basically
28:21
but not calling it that. And
28:23
then funneling that into just a perpetual layoff machine.
28:26
Perpetual is a big word but I
28:29
guess I fundamentally agree. I mean there
28:31
was a lot of drunk money right and
28:34
a lot of like crazy hiring
28:36
going on. I mean I still
28:38
I know I'm gonna be quick on this but I know like
28:40
a month or two ago I mentioned this story just
28:43
like back about a year
28:45
or two years ago talking down here to
28:47
some startups who were like
28:49
excuse the French but dick measuring based
28:51
on headcount like yeah but are you making
28:54
any money? Like I'd rather have two
28:56
people and be profitable than
28:59
have 20 and be burning through cash. Now
29:01
I understand why and I've been I've gotten
29:04
some very private lectures from a couple
29:06
local VCs I know about why I'm
29:09
an idiot. I still
29:11
don't think I'm wrong. Yeah. I still think
29:13
ultimately the old style of like
29:15
this is where you actually have to at
29:17
least on an annual basis like you couldn't have
29:19
a losing quarter or a losing month. You got
29:21
to bring in more money than you than you
29:23
expend is I just
29:26
I yeah I'm sorry I
29:29
they overhired right they just they just
29:31
did. Google's accidentally been
29:33
caught collecting children's voice data.
29:36
Home addresses of carpool users,
29:38
YouTube recommendations based on users
29:40
deleted watch history which they
29:42
say they won't do, a bunch
29:44
of employee reported privacy incidents. This is all being revealed
29:47
by 404 media and it's
29:49
like they've also they take on too much Mike.
29:51
They take on too much and
29:54
so I guess most of
29:56
the incidents which have previously not
29:58
been reported impact relatively. small number
30:00
of people and Google tried
30:02
to fix them quietly behind the scenes
30:04
after employees discovered them. 404 rights
30:07
taken as a whole though, the internal database
30:09
shows how one of the most powerful and
30:11
important companies in the world manages and often
30:13
mismanages a staggering amount of personal
30:15
insensitive data on people's lives. And
30:18
this freaks me out because you know I think
30:21
I like to tell myself that Google has
30:23
great security standards and practices internally and that
30:25
like my data isn't like you know spilling
30:27
out on the floor all over the place.
30:30
But then you read this, it
30:32
shows that Google's own employees kept finding
30:34
incidences where either data
30:36
got collected that wasn't supposed to get
30:38
collected, third-party access was given to something
30:40
they weren't supposed to have access, Google
30:42
staff made mistakes that made people's data
30:44
available, contractors had access to things they
30:46
shouldn't, everything from you
30:49
know emails to Google files and docs,
30:51
everything they got, just little incidents like
30:53
this over and over again where they
30:56
discover that there's somebody has access to something they
30:59
shouldn't from internal employees to contractors to public.
31:02
It's terrifying to think of this because you know that
31:04
we always say the cloud is somebody else's computer but
31:06
when you think Google you think well they got probably
31:08
some sort of cool encryption system, they
31:11
might but if everybody has keys I suppose it doesn't matter. Just
31:14
yeah that there's also the one where
31:16
the Google contractor did you
31:18
see this one? He was a
31:20
YouTube contractor and he went into Nintendo's account
31:23
and saw all the videos they were working
31:25
on and just hacked them. Yeah,
31:27
yeah. Although my favorite story I like
31:29
404, we should mention the more kind
31:32
of, I feel like they're not as big as some of the
31:34
ones we mentioned but this is just
31:37
like the world we live in with Zoom meetings and everything.
31:40
Dude going to traffic court
31:42
for driving on a suspended
31:45
license. He calls in while
31:47
you're driving in the
31:49
car that's just like are you driving
31:52
and the judge actually lets them he's like just
31:54
please pull over like and
31:56
the guy's like oh I'm whatever going to the doctors.
32:00
over zoom issues a bench for it.
32:04
Yeah. That's
32:06
a mistake. That's the most post COVID thing
32:08
I've ever heard of you can't rest it
32:11
over zoom. I suppose I guess the counter,
32:13
okay, I'm gonna steel man this thing. Are
32:15
you ready? Steel man it. Coming at you.
32:17
Coming at you. Wouldn't
32:19
you want Google to maintain a database
32:22
of all these incidents so that way they could learn
32:24
from it and try to mitigate them in the future?
32:27
Wouldn't it be irresponsible if Google wasn't tracking
32:29
all this stuff? I mean, yes, it got leaked
32:31
to 404 media and it's very embarrassing. So
32:34
it's not ideal. But he
32:36
didn't like that leaked in reveal personal
32:38
information. It revealed Google's internal tracking of
32:40
personal information leaks and incidents. But
32:43
it's like a trouble ticket system like they should they
32:45
should be keeping track of this, right? Yeah,
32:48
I mean, I think the problem is that like just how
32:50
data hungry Google really is, right?
32:53
Yeah, people don't understand how
32:55
much they're hoovering up. Yeah,
32:57
and it really does reveal their security practices are
32:59
pretty weak, or at least they seem to be,
33:02
doesn't it? You
33:04
know, like, I guess another thing that
33:06
was in there is they blur it out on
33:08
Street View. But as we all suspected, they are
33:11
storing all the license plate they collect with
33:13
Street View. So they
33:15
blur it in the public version. But as we
33:17
of course suspected, this thing confirms this leak confirmed
33:19
that they're this they're steep, they're keeping all that
33:21
with the geolocation stuff that they collect. And
33:25
then two weeks ago, I didn't
33:27
get a chance to put it in the show I wanted to but I just
33:29
didn't get a chance two weeks ago or something like that. A
33:32
story just a summation of
33:34
Google this it's it's tracking
33:37
people with air tags and location information
33:39
that the Apple iPhone collects that then
33:41
they make available to the API, pairing
33:44
those people with star links and then tracking
33:46
them through Ukraine. And essentially what
33:48
it what the meat of the
33:50
story is is the Apple iPhone when it's
33:52
uploading your geolocation for your Wi Fi
33:54
to do the you know, assisted location
33:56
stuff, it not only uploads
33:59
your geolocation with your Wi-Fi access point name,
34:02
but all of the Wi-Fi access points around you. Ooh.
34:05
And when you go to query the API to
34:07
get your location, like if you're an app, it
34:10
will give you also something
34:13
like five or 400 APs around you as
34:15
well. And then it'll also give you their
34:17
Mac addresses. So you can pretty
34:19
quickly start to figure out what kind of gear
34:21
people have, and then by the
34:23
combination of their Mac addresses, kind of
34:25
identify them, and then you can track
34:27
them. In research, not only
34:29
did they track the movements of soldiers in Ukraine,
34:31
both Russian and Ukraine soldiers with extreme accuracy, but
34:34
they also, as part of the research, they saw some of them
34:36
flee and tracked where they went because they
34:38
took some of the gear with them, in this case like a
34:40
router and a phone. And they've also
34:42
seen just other people outside the Ukraine area, they've been
34:44
tracking them and watching them move around and seeing
34:47
some of these devices get sold. And it's remarkable
34:49
the information they can get. Google's API does
34:51
a similar thing, but the Apple API really
34:54
lets you abuse it without any kind of throttling
34:56
or anything like that. So you can really just
34:58
scrape this thing for information and build a tool
35:00
around it. And the theory behind
35:03
the Apple API providing so much information is
35:05
it's trying to give enough information for all
35:08
of the location math to be done on
35:10
device. And Google will do it
35:12
in the cloud for you. They have a little more restrictions,
35:14
they don't show you as much devices, but you can still
35:16
bang on it. Whereas the Apple one
35:18
assumes that it's an iPhone talking to it
35:21
and getting that information. And
35:24
so it gives lots of information. And
35:26
so the researchers were able to really, really, really
35:28
dig into that. And again, it's
35:30
like something that just kind of silently gets collected.
35:32
And there is like a no track, like an
35:35
underscore no track you can add to the end
35:37
of your AP name and Apple and Google both
35:40
will take it out of their
35:42
database. But every AP around you would have to
35:44
have that. And you'd have to
35:46
go change your AP name and then go update
35:48
all your devices to connect to that new AP
35:50
name. This seems basically untenable. I will put a
35:52
link in the show notes. You should really read
35:55
it. It's really remarkable just
35:58
the level of detail and information and track. they
36:00
can do just because all these
36:02
devices are constantly reporting the APs around them. Spooky,
36:05
my friend. It's spooky. Ask not
36:07
what your podcast can boost for
36:10
you, but what you can boost
36:12
for your podcast. Oh, well, let's see. We got
36:14
ourselves a baller right here. Rodded mood comes in
36:16
with 60,000 sap. Hey, Richard!
36:22
Oh, oh. Lobster.
36:24
Yes, but no message, just support. Hey,
36:26
that's how you just throw that change.
36:29
Boom. Thanks, mood. Appreciate it. Appreciate it
36:31
very much. Sam Squatch comes
36:33
in with some mick ducks. Things are
36:35
looking up for old mick duck. 22,222
36:38
sats. I hope you guys
36:40
stick around for the summer. Please, one for
36:42
ol' londo. Ol' londo. Oh,
36:45
londo. Londo. I was like, what is
36:47
he talking about? Londo, dude, from Babylon.
36:49
Yeah, Babylon 5. I
36:52
think I know exactly what he's talking about
36:54
with... I don't want to spoil, but Sam
36:57
Squatch, right back next week
36:59
if you were referring to londo's keeper,
37:01
because that's rough. Okay,
37:04
here we go. Tampatec Trekkie comes in with
37:06
20,000 sats. You're
37:08
so boozed. And he says that it
37:10
is... 20,000 sats is
37:12
actually four of
37:14
the Jar Jar boost. Everything's under control. Thank
37:16
you, Tamp Trekkie. Really appreciate
37:19
it. You actually put that on the soundboard. Drew
37:23
cut it for me. DG at
37:26
PTC comes in with 5,000 sats. Again, the
37:28
Jar Jar boost. You're so boozed. I
37:31
recently built a tiny LLM as part
37:34
of a personal exercise. And
37:36
it's clear that there is nothing innate about
37:38
the questions and answers to the GPT AI
37:40
architecture. OpenAI and friends
37:42
must be investing massive amounts into
37:44
offshore data entry, exhaustively asking and
37:46
interrogating content with questions. So when
37:48
chat GPT answers questions well, it's because
37:50
that question, maybe even that answer has
37:53
been seen before. Either scraped from
37:55
Stack Overflow or Quora or hand
37:57
jammed in there somewhere in the global south. DG
38:00
that is a very insightful answer. Yeah, I bet
38:02
in 10 years we're gonna We're
38:06
gonna find out that there's some I don't even
38:08
where do people do this stuff now? I used
38:10
to say bang a well remember
38:12
the Amazon grocery store was still India.
38:15
That's right. It's still India Yeah, we
38:17
just went up in India though didn't say I
38:20
thought I feel like the innovations got to be
38:22
happening and like in the Middle layers between the
38:24
LLM and the cheesy chat interface
38:26
I feel like that's where the real
38:28
like this massaging of Yeah
38:30
Delivering it to the system in a way that
38:32
the system responds and then delivering that message back
38:34
and how it gets sent back I feel like
38:36
that's really the middle development layer that they've got
38:39
to be really focusing I mean, I bet he's
38:41
right, but uh We don't for
38:43
a fact open AI is just hoovering update. I would have
38:45
permission to so But did
38:47
you see the thing where the lady from I
38:49
forgot her name. I apologize from open AI was
38:51
being interviewed And the interviewer
38:54
said so are you just like? Taking
38:57
Google YouTube data. I don't
38:59
want to talk about that right now. Yeah That
39:03
was very awkward that's like the hardest that's
39:05
like the the hardest and yes non-yes you'll
39:07
ever get from someone I Agree
39:10
Faraday fedora comes in little row ducks You
39:13
guys do what you got to do I rather take the summer
39:16
off than end the show in the near future But any episode
39:18
you put out I'll be listening and streaming sets. Thank
39:20
you fedora kapla Thank
39:22
you. Appreciate it torped comes
39:24
in with 8,057
39:27
sets Is
39:29
the story of Grep's development not the standard
39:31
we should all aspire to one man takes a
39:33
weekend to develop an indispensable tool For
39:36
decades to come good one torped.
39:38
I should look that up. I was just seeing a link
39:40
that Linus Essentially built the
39:42
core of git in like a
39:44
couple of days. Yeah, wasn't he pissed off
39:46
at Mercurial? Wasn't that which is ironic because
39:49
the name Mercurial yeah, yeah, yeah now might
39:51
have been it I might have
39:53
been I think it was I think yeah, cuz I don't think I I
39:56
have to watch revolution OS again, even
39:59
though It's just for
40:01
my morality that's required watching
40:03
for this show it really is Pirates of
40:05
Silicon Valley also highly recommended hard to find
40:07
but if you pro-tip
40:10
sometimes the BBC's
40:12
BBC America will sell you old
40:14
PBS movies, which is what? Silicon
40:17
Valley is sold by now and Oh,
40:20
no, I'm thinking of the Rob cringy one. You're
40:23
thinking of the one with Noah. Yeah, right. Well, they're
40:25
both good That's what I'm saying. I have them messed
40:27
up pirates. Right one is fiction one is a documentary,
40:29
right? Wes is coming in with a hotfix. He says
40:31
by the way, it was big keep it keep it.
40:33
Ah, yeah Yeah, there you go Yeah,
40:36
they're both recommended watching neon Pegasus
40:38
is coming in hot with 2883
40:41
sats Longtime
40:44
listener first-time booster. Well,
40:46
hello. Thank you Thank
40:50
you for taking that journey if you gents need a
40:52
break summertime isn't a bad time Just
40:54
make sure to come back now. It wasn't us that needs a break
40:57
You see when we when we were talking about
41:00
that that was after like a week of like
41:02
no membership signups and like 30,000 sats So
41:04
like we were like, uh, oh His the
41:07
show can survive but the trend has to
41:09
continue really? Um, yeah, cuz like Doesn't
41:12
have to need for neither one of us. It doesn't have to
41:14
make us thousands and thousands of dollars We just want to keep
41:16
a light on until the advertising market turns around. We're
41:19
doing alright So I've been listening
41:21
since nearly the beginning of coder and it's one of
41:23
the podcast that jumps to the top of the list
41:25
When a new episode drops. Oh Love
41:28
that. Thank you. Appreciate that very much. Oh,
41:30
that's like that's the that's the honor and Goal
41:34
that all podcasts aspire to it. Hmm. The force is
41:36
strong with this one Okay,
41:40
I might want that too drew okay Strong
41:44
with this one the forces Also
41:47
adding one at the start of my
41:49
boost amount turns it into a zip
41:51
code boost with a Star Trek Easter
41:53
Egg Uh-oh a zip code. I gotta
41:55
grab Wes's map to a 8 4.
41:57
Oh, no, is that three? 1288
42:00
oh, there's a one
42:02
there one. Yeah, you gotta add the one.
42:04
Oh, I literally added the number one Okay,
42:08
he's okay. I know where he I know where he is,
42:10
but what's this try Condor Roca, New
42:12
York Is that what I'm seeing here? Yeah,
42:14
I have a victory mills, New York I got a
42:17
paper cut, but what's the Star Trek reference?
42:19
I don't know I see that there's
42:21
a Star Trek original series set tour going
42:23
on there in May Star Trek next generation
42:26
1 1 0 0 1 0 0 1 it's the 15th episode, but ah Alright,
42:34
that's deep. All right boost in if I
42:36
got that right. Let us know by googling this now
42:39
We could ask Gemini, but you know, it'll
42:41
just get upset It
42:43
would actually give you a worse answer than we just
42:45
did that's what's incredible. That's remarkable crash
42:47
master comes in with the row of ducks 2222
42:50
sets Red
42:54
Hat seems to have made a solid bet on
42:56
easy management and deployment of small open
42:59
LMS on premises Many
43:01
companies don't want their data in the major
43:03
cloud out limbs that also are not great
43:05
beyond assistant style tasks yet I
43:08
think crash master I agree that businesses
43:10
would prefer to have on premises but
43:13
if you think about what that means
43:15
that means buying tens of thousands of
43:17
dollars of GPUs and Systems
43:19
for those to go in plus staff that knows
43:21
how to run that then the software infrastructure to
43:23
manage it There's definitely going to be companies that
43:25
will make that investment but
43:29
How many businesses do you know? That
43:31
are gonna go spend that money and even if they just got like
43:33
us, maybe they don't get a lot Maybe they just get like a
43:35
couple of powerful Systems, that's
43:38
still five grand a couple
43:41
of people that need know that need to know how to
43:43
run it and and management needs to have a clear understanding
43:45
of The ROI
43:48
on that little spend Like
43:50
beyond you know, just research and development. It
43:52
seems like a hard sell, right? It's it
43:54
seems like a big capex that they're not
43:56
gonna want to do which seems like ultimately
43:59
where this either and the AI stuff falls
44:01
off a cliff. Right, because 20 bucks for chat GPT,
44:03
whatever the hell they call the plan, is that they
44:08
can't be making money, right? So yeah, right? So
44:11
it's either going to be centralized in
44:13
these big cloud providers that can buy these GPUs
44:15
and get into those upgrade cycles and compete with
44:17
each other to stay competitive in
44:20
the market, or it's
44:22
got to be optimized on lower-end hardware
44:24
and get better, and get
44:27
better, and these smaller LLMs have to get
44:29
faster and better so that
44:31
way enterprises can get into
44:33
it without $30,000 of capital expenditure
44:36
or something. It feels like
44:38
a real problem the industry needs to solve if they want
44:40
to keep this momentum going. But I
44:42
don't know, what do I know? I'm just the
44:45
guy on the internet. T-B-O-B-R-N-O-B-R?
44:49
Come in with the road. Ah,
44:51
here we go. T-Boy on Nobara. Nailed
44:55
it. I got it. Always
44:58
love the takes. What's expected from WWDC?
45:02
Oh man, yeah, next week. It'll be right here
45:04
in WWDC episode. We
45:06
will find out. I mean, yeah,
45:09
I don't know. Yeah, we'll find
45:12
out next week. We got time. Yeah, we'll
45:14
see. Don't think you're going to get a Mac Studio
45:16
with an M4 in it. You mean an
45:18
iMac Pro. Oh, we're an iMac Pro. I don't think you're going to
45:20
get that. I don't think I'm ever getting an iMac Pro again. I
45:22
don't think. And I don't think we're going
45:24
to see a new Mac Pro. Maybe. They
45:26
should. They should, you know, really flex.
45:29
They should do like some super ai Mac Pro. So
45:32
maybe they will. I don't know. Skips. T-B-G
45:35
comes in with a Rodeux. Boosting
45:38
in with the news about Kotlin's 2.0 release.
45:40
This release has been cooking for a
45:42
long time and it brings their home-grown
45:44
K2 compiler to stable. The
45:47
K2 compiler enables Kotlin to be compiled
45:49
to JVM, JS, and Wazool.
45:52
And native code, all, from
45:55
the same base. I think it's
45:57
worth keeping an eye on this space as it's developing
45:59
to be... a great multi-platform
46:01
ecosystem. Good stuff, yeah
46:03
that's fair. I actually really liked Kotlin. I
46:06
just ended up reverting back. And if you
46:08
subscribe to the Coderly, you'll find out why.
46:10
Tomato comes in with 5,000 sets. The
46:13
Jar Jar Boost. You're so boost.
46:15
Just me, so boost. I wanted
46:17
to recommend the Tencent TV adaptation
46:19
of the Three Body Problem. It's
46:21
a very faithful adaption to the book, and they
46:24
give it a budget it needed to make some
46:26
amazing sci-fi TV. You will need to read subtitles,
46:28
but the acting is fantastic. Yeah, I have to
46:30
sit down and watch that. Well, thank you, Mr.
46:32
Tomato. I did like the Netflix series.
46:35
You're right, I don't think it's, I
46:37
don't think I'm gonna rewatch it ever. You know, like the
46:39
Expanse, I'm gonna watch again. Star
46:41
Trek, I continuously rewatch. I
46:44
would even potentially consider rewatching 12 Monkeys. Things
46:46
like that, you know. Deadwood,
46:48
these shows, there's lots of these, right? But the
46:51
Netflix Three Body Problem, as with so many of
46:53
these streaming shows, I don't mind them at the
46:55
time, but then afterwards I'm like, I don't really
46:57
ever need to revisit that. Yeah, I mean, the
46:59
problem with the Netflix version, it's good, but they
47:01
turned it into the Power Rangers. I mean, if
47:03
you think about it, it's the
47:05
Power Rangers. Wow.
47:08
Hey, it could be morphing time. I
47:10
have a Goldar sticker on my MacBook Pro. I'll
47:12
have to tweet that later, because people won't believe
47:15
me. I will tweet the Goldar sticker on my
47:17
MacBook Pro when this is released. Well, thank you
47:19
everybody who boosted into the show. We
47:21
had 12 boosters. We had a cutoff at 2,000 stats, but
47:23
we stacked 132,550 stats. This
47:27
is not a blowaway, but we really do appreciate
47:29
that support. Of course, a shout out there to
47:31
those of you who stream those stats as you
47:33
listen, and of course to our members. If you'd
47:35
like to try boosting, you get in on
47:38
the Podcasting 2.0 fund when you do it. Something like
47:40
Fountain, where you can connect it with Strike and get
47:42
your sats in really no time, boost
47:44
right into the show. Or if you don't wanna
47:46
switch podcast apps, you can go to fountain.fm, search
47:48
up Coda Radio, and then any
47:50
app that supports the lighting network like Cash
47:52
or Strike or Coinbase, or
47:55
anything that just can scan a QR code. You can send a
47:57
boost into the show and get your message read while you support
47:59
that individual production. The idea is to get a
48:01
little value, some enjoyment, some information, makes you
48:03
think about something. You send us a
48:05
little boost back or also appreciate corrections
48:07
or updates or your take on anything we
48:09
talked about. Thanks for everybody who does Boost
48:12
In. Before we run, I wanted
48:14
to share something you linked with me in our chat this week
48:16
that I think the audience ... I knew it. I knew you
48:19
were going to do this. The
48:21
8-bit retro mechanical keyboard, and
48:23
this one's the M
48:25
edition, is so cool.
48:28
I think you just found my new favorite keyboard, and
48:30
I don't say that lightly. I have
48:33
been refreshing the page on this, like
48:35
Gollum looking at the ring. It's pretty
48:37
bad. So, it's 8-bit do, like 8-bit
48:39
do, all one word, retro
48:41
mechanical keyboard. Did you notice that
48:44
they also have a Commodore 64 version,
48:47
a Fami edition as well, and an
48:49
N edition, and they all look really
48:51
cool. Yeah, so I've been a customer
48:53
of theirs for a while. They make
48:55
great third-party controllers that are compatible with
48:58
all the major consoles and Steam and
49:01
Steam Deck and work fine on
49:03
Mac and Linux. I
49:05
stopped buying the ... Sorry,
49:07
Nintendo, but your stupidly overpriced
49:09
Pro controllers and switched
49:12
to these guys. They have ... I
49:14
don't know if it's on that page, but
49:16
they have ones that look like the old
49:18
Sega Genesis or the SNES. It's really ...
49:21
It is quality stuff. It's not like some
49:23
of the cheap third-party stuff you get. Yeah.
49:27
They really know what they're doing. Then they have
49:29
their adapter section, so you can read by all
49:31
the adapters. Also the keyboard itself, Bluetooth
49:35
or wireless 24 hertz dongle,
49:37
and then the dongle stores
49:39
magnetically inside the keyboard. It
49:42
has an actual little knob. Now, it looks
49:44
like a knob from the 80s, so that's
49:46
cool, but it has an actual little volume
49:48
knob that you turn on there, and also
49:51
a volume knob style selector between Bluetooth off
49:53
and wireless. Oh, it looks
49:55
so good. It's a beautiful retro keyboard.
49:57
Now, in the ... requirements
50:00
here Mike it says it's only compatible with Windows
50:03
or Android do you know have you tried hooking
50:05
up to Linux I don't have one yet they
50:07
notice the yet the yet is very important yeah
50:11
and you know it's not crazy expensive well one they're
50:13
only for pre-order they're only for pre-order right now yeah
50:15
this one is right yeah oh if you're gonna do
50:18
it you have to go IBM M yeah I think
50:21
you're right but the family edition and the
50:23
end edition are available right now 89 bucks
50:25
which for a fancy boy keyboard much I
50:27
mean I have I love my launch that
50:29
I'm sure drew will edit out several times
50:31
during the show I don't
50:34
know though like I'm such an old but this is just like
50:37
the fact that they put it folks
50:39
you have to look at this it's an old
50:41
IBM computer with its keyboard and it's
50:44
got floppy disk next to it
50:46
in the marketing picture if you ever
50:48
want to sell me something throw a
50:51
couple floppies yeah it's just like throw them
50:53
next to your product I really like floppy
50:55
this I gotta say oh and it's wireless
50:57
or wired which I'm kind of getting tired
50:59
of my lawn I'm having to deal with
51:01
the cable I know it's nice to
51:03
have both you know because I'll use these
51:05
keyboards for years and years and I'll move them
51:08
between different computers yeah I mean I promise I
51:10
have the code mechanical keyboard I have two launches
51:12
I'm starting to become like those guys who
51:15
buy like you know a closet full of
51:17
assault rifles with mechanical expensive keyboards or the
51:19
kids these days kids these days love tennis
51:21
shoes lots of tennis shoes I am not
51:24
suffering from that yet so I I really
51:26
like it so again it's 8-bit due retro
51:28
mechanical keyboard and if anybody has any experience
51:30
with with their other versions let me know
51:32
because this is I think maybe the next thing
51:35
I'm gonna try I think I'm gonna get one I've got a keytronics
51:37
at home right now and I like it and then here in the
51:39
studio I've got a launch so I'm set here but
51:41
I stole from my stash
51:43
at home and I that's actually a studio keyboard that I
51:46
need to bring back so I need something for my home
51:48
setup I could be coding like it's 1989 so it is
51:50
a little
51:54
q-basic you know I know that's
51:56
what I learned on I you
51:59
know people I'm a little dunk on poor
52:01
Q basic. You could actually do a lot
52:03
in Q basic. Oh yeah. Oh
52:05
yeah. Oh yeah. That's what I started
52:07
with. Alright. Is there anywhere
52:09
you want to send the good people? Actually
52:11
yes. So normal Alice.dev but I
52:14
just literally, I wrote a press release for the
52:16
first time in my life yesterday Chris. Oh good.
52:18
Wow. How would Gemini then
52:20
I made it go crazy? That
52:22
is the thing I was doing. Oh right.
52:25
So that's what you were doing. Okay.
52:27
Okay. So if you look at Alice
52:29
Autodesk or something like that or my
52:31
name, I don't remember how Google indexes
52:33
it. But there's a bunch of like
52:35
the press release got picked up in a bunch of places.
52:39
But if you are an aspiring
52:41
tech blogger or weirdly a blogger
52:43
in the construction space and
52:45
want to learn more about automating
52:48
Autodesk Cloud Connected API, this
52:50
is the most, it's actually kind of cool
52:52
but very boring for most people listening to the show.
52:55
I can talk to you about that with Alice and I
52:57
am very eager to. So. And how do
52:59
they do that? Alice.dev? Go to
53:01
Alice.dev. Yeah. There's a thing you can just
53:03
email me or schedule a call there. Alright.
53:07
You can find me on WeaponX ChrisLAS
53:09
or if you want to try the
53:11
Noster thing, chrislas.com over there. The network
53:14
is at Jupiter signal. Things
53:16
that we talked about today. Yeah. Those are available there
53:18
on the web. At coder.show
53:20
slash 573. You'll find our
53:22
contact form there as well as our RSS
53:24
feed. Now, why don't you join us next
53:27
week. We'll be live at noon Pacific, 3
53:29
p.m. Eastern. I'll be broadcasting from the woods.
53:31
So that should be an adventure. And
53:33
we'll be recording a double coder next week. So if you come, it'll
53:36
be a nice long stream. You can hang out with it. Thanks
53:38
so much for joining us on this week's episode of the Coder
53:41
Radio Program. See you back here next week. Thank
53:50
you. you
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More