Latest / WGKM / Podcasting 2.0 - Episode 266: Research Velocity

Transcript

Podcasting 2.0 - Episode 266: Research Velocity · WGKM

0:00
Podcasting 2.0 for July. Hello everybody, hello everybody in podcasting
0:12
land. Well, only the future of podcasting. Yeah, that's right.
0:18
This is the board meeting, the place where it all
0:20
goes down, where this band of misfits gets together to
0:24
discuss it all. We are the only board meeting you
0:27
can attend with your eyes closed. I'm Adam Curry, here
0:29
in the heart of the Texas Hill Country and in
0:31
Alabama, the man who can tell you exactly why the
0:34
Roman Empire fell. Say hello to my friend on the
0:37
other end, the one, the only, Mr. Dane. Any anything
0:52
happening in your world? Nah, not much going on. Yes,
0:55
it seems a little slow. It's been a slow week.
0:59
Yeah, you just kind of Yeah, nothing. Oh yeah, that's
1:08
right. I don't want anyone who I know to die
1:10
this week. Wow. So, um, six hours of podcast. How,
1:21
how much do you even have in the tank today?
1:23
No, I, I got, I got a tank, but, uh,
1:29
I would say. You know, that mode where you're just
1:31
like, I'm running in crisis mode. Can I? Sort of
1:38
take the wheel for a second. I would like to
1:43
just ask you if you're okay with this. I would
1:46
like to ask you some questions that I think you
1:51
can answer. would be helpful for people who might be
1:58
in your same space. And so you, your co-host of
2:08
No Agenda, because there's... agenda. There's me. That's true. You're
2:20
in the Venn diagram. You're the one where it all
2:22
over. You're the center. I'm there. I'm there. Yeah. You
2:26
and Cotton Gin. And so the, the, you know, some
2:33
people may not know who are not listening to Nojana
2:38
and you're, so you're a long time co-host for 18
2:43
years. Is that right? Yeah. in October. for any for
2:57
a substantial And for any show. So there's a few
3:03
questions I want to ask about this because if any,
3:06
for any, like. Number one, a little bit of selfish
3:12
concern for you that I have. No agenda is your
3:20
main source of income. Yes. I need to ask you,
3:23
are you going to be okay? I don't know. Honestly,
3:28
I don't know. I appreciate the question. Would we be
3:32
okay for a year without income? be it. So it's
3:39
not preferred. I mean, up and down income is the
3:46
life of any sort of freelance person, whether it's a
3:52
contractor or a freelance artist or any type of person
3:55
who does freelance. I mean, you're not, well, you don't
3:58
have a guaranteed paycheck. So for, and again, I'm kind
4:03
of asking this question as if I'm another person out
4:07
there who also, may have your next steps are. Good
4:24
questions, the questions I've been asking myself. No, no, no.
4:38
Uh, well there's a couple th- The main issue I
4:43
just want to put the money aside because I've never
4:48
really worked for money. And it shows. He does. So
4:59
what's interesting... but ours in particular. in my opening of
5:10
our tribute show, which will... have their say, so we
5:18
just said, put it in your donation note. We read
5:22
753 notes yesterday. That's why it was... Oh my goodness,
5:26
that is so much. But it was beautiful, of course.
5:31
thing that struck me as an aside is the number
5:34
of faithful people Not just people sending thoughts and prayers,
5:38
but like real people. was kind of blew me away.
5:46
Yeah, we... But it was a performance. We always said
5:51
this. It was a performance between And... I think we
5:58
both the other guy laughs. to say, hey, look at
6:01
me, I got something, look at this theory I have,
6:05
or look at these things I've found, clip of the
6:08
day, that kind of stuff. So that's a chemical. that
6:12
we didn't even start until we had been talking on
6:15
Skype 18 years ago for weeks. You know, hey, we
6:20
should probably record this. That's literally how it got started.
6:25
So very much like... talking for years podcast. Do we
6:35
get to do this with Pastor Jimmy? We're now 42
6:37
episodes in. So, you know, I'm not a solo guy.
6:45
can't just plug someone in and oh okay now So,
6:52
and also, I don't think that, I don't know if
6:56
that would work. Right. And honestly, I'm also very tired.
7:14
you I think propping him up, dragging him along a
7:21
little bit, which was completely understandable. Also, you know, not...
7:27
a big secret I was supposed to keep, but he
7:29
was 80 years old, not 74. He was a little
7:33
older, not that that's a death sentence for most. You
7:39
know, and just, you know, what's happening, what's going on,
7:43
concern for him. You know, two families have been living
7:46
off of this show, in essence, for all this time.
7:50
So I really don't know what's going to happen or
7:56
what I'm going to do. Honestly, I think I need
8:00
just a week or two of rest. I don't think
8:04
his passing is going to be a good thing. you
8:08
get the call on Monday. There's, you know, what are
8:12
we going to do? Consoling. um with his family. It
8:21
affected me. And Processed it yet. And. I'm like, I
8:28
haven't really quite wrapped my head around it, but yes,
8:31
I do have to figure something out for sure. What
8:34
that is, I just don't know. And do I really
8:38
want... with someone else, I don't think it would be
8:43
the same. And it would, I think it would feel
8:46
fake to me. Does that make sense? Mm-hmm. Yeah, it
8:50
does make sense. And it melts it. Like, sometimes I
8:58
think when you're in a situation like this where somebody
9:04
passes away, but you also have lost. Thank you. of
9:10
uh orbit around that person even even a spouse or
9:15
something like that you they're passing your grief connected to
9:21
their passing gets pushed to the side for a while.
9:26
You just have to take care of a lot of
9:29
things. You have to take care of, you know, if
9:31
you're the family. the funeral and you have to take
9:34
care of all those things if you're you know somebody
9:36
who's a business partner or a long time friend or
9:40
you have other responsibilities you're gonna have you everything suddenly
9:44
becomes around um what there are immediate actions that i
9:49
need to take to take care of real world issues
9:53
and so you tend to put your own grief to
9:57
the side for a little while and then it's only
10:00
later that you really begin to experience that. And so
10:05
that, that makes a lot of sense. I mean, you
10:07
probably just kind of feel a little bit lost right
10:09
now, I would imagine. Yeah. Tina's been fantastic. She's like,
10:16
we'll figure it out. Everything's doable. Everything's possible. We'll figure
10:21
it out. That's really awesome that she's not freaking out.
10:25
But she realizes the same thing. It's like, okay, well,
10:29
what does this mean? How do we move forward if
10:33
income stops? But I'm trying not to. I kind of
10:37
stopped worrying about money in 1990 for a whole bunch
10:41
of reasons. Mostly because you lost it all? No, no,
10:46
no. That came later. Okay. No, that came in the
10:51
2000s. And I really love podcasting. I love the shows
10:57
that I do. I love No Agenda. I really did,
11:01
but I just don't see... and recreate some chemistry. I
11:10
think I'd always-- if that were the way I would
11:13
go, it would always be unbalanced. You know what I
11:17
mean? And also, you know, we were 50-50 partners on
11:21
a handshake, literally, except for the, we had to form
11:25
an LLC because the IRS was like, what are all
11:28
these $3 checks? What's all this cash and envelopes? Are
11:32
you a political candidate? What is this? What's going on
11:35
with that? So we had to formalize that. You know,
11:40
so there's another, there's another party in this equation, you
11:44
know, so how do you deal with that? And it's
11:47
just, but, but just from a creative, point. John was
11:52
my counterpart. Just like you're my counterpart on this show
11:56
and Jimmy's my counterpart. do this. And that's something, you
12:00
know, Jimmy and I've been talking for years before we
12:03
said, hey, let's hit record on this. Same with you
12:05
and I. So I don't have that luxury of finding,
12:09
oh, let me find you, buddy. Yeah, let's talk for
12:11
a bit and then we'll hit record. Just doesn't seem
12:13
like it. But it's just interesting as, and John and
12:17
I would talk about this all the time. This was
12:18
our famed exit strategy talk because there's no exiting from
12:23
a podcast, except the way he did it. He literally
12:26
took the only exit strategy there. is and by the
12:30
way to just be lighted out while grousing about the
12:34
hole in NCIS, yeah, I'll sign me up for that.
12:38
That's the way to go. That's simultaneously the most John
12:44
Dvorak thing to do and also a great way to
12:47
go out. Exactly, exactly. So, and I appreciate him for
12:51
that. But, you know, I'm... It'll sound corny, but I'm
12:56
just putting my faith in God and praying about it
13:00
and saying, what does this mean? And listening for answers.
13:04
And that will take a little bit of just time
13:08
without the stress of. doing a show twice a week
13:12
so i have some thoughts on how we'll keep the
13:14
show going uh for for a couple weeks so i
13:18
can just get my get my head together sure um
13:22
and we'll see we'll see where it goes but i
13:24
appreciate i appreciate the concern because yeah the concern is
13:26
obviously real you have to get a job i don't
13:29
know Maybe. I mean, there's, this could be the case
13:35
that other people face when... of a long running co-host.
13:43
I think there's maybe not a lot of people in
13:46
this situation, but there are quite a few who, I
13:52
mean, think about any show like, take "Accidental Tech" or
13:57
"The Flophouse" or any of these shows where there's two
14:01
or three people who are sort of, whose personalities combine
14:07
to create really what the show is. It's not one,
14:12
you really can't just plug another person in. It's really
14:15
this unique fingerprint that's created by the combination of... personalities
14:24
and how they interact. Here's the thing that is unique
14:27
though. Because with everything I've ever done in broadcasting, and
14:33
certainly in podcasting, I learned a long time ago, and
14:36
this is part of the value for value model, is
14:39
that... building a community. audience. See, this is my, this
14:44
is completely antithetical to what the podcast industry. wants. They
14:50
want numbers. And we're going to jerry-rig this thing so
14:55
that we can show numbers. Even if you hit play
14:57
and listen for 30 seconds, we've got numbers. Numbers, numbers,
14:59
numbers, numbers.
15:01
And... very very difficult you know we probably would have
15:08
made a little maybe 20 percent with advertising, with the
15:15
amount, depending on how many ads we would jam into
15:18
a show, just looking at the number of downloads, if
15:22
we'd gone that route. And it would have never worked
15:24
because we would have lost advertisers for all kinds of
15:27
reasons. And certainly during COVID, where the value for value
15:31
model actually worked very well in our favor, that would
15:35
have been extremely adverse if we were an ad-driven podcast.
15:43
is far bigger than than adam and john i mean
15:46
we have five meetups a weekend taking place all over
15:50
the world you know it's not like we're at the
15:53
meetup people just hanging out together discussed or whatever they're
15:58
doing. You know, we've had people meet each other. It's
16:02
a very, it's a community. that help me through it
16:09
no matter what I do or or but I also
16:14
feel a responsibility towards them too. You know, you can't
16:18
rug pull. I mean, that's just not something that can
16:24
happen. It's way too... We have jokes and inside jokes
16:31
and and all kinds of things that you know I
16:36
don't want that to fall apart. just general point to
16:43
make about what no agenda was is because you built
16:49
because you because you And it evolved over time into
16:54
being more than just content. It was a whole community
16:58
of people. community of people will follow you to your
17:05
next project, but it doesn't mean that you're going to
17:09
keep them there. I mean, if the con. They won't.
17:15
built the community and that community is, is strong and
17:19
will follow you and give this next, whatever project you
17:24
do next, they'll give that a shot. Yeah. No, I
17:26
know. And that gives you, you know, at least it
17:29
gives you a solid footing to begin with, you know,
17:33
and, and because they want, they're, they're, I think a
17:37
lot of them, like, it's, it's sort of like, um,
17:40
when Neil Peart passed away in Rush, You can't just
17:45
plug... It's not that bad. different. Yeah. Oh yeah, for
17:57
sure. I think of, I mean, just our show, we've
18:01
been going for six years and, you know. Don't die
18:05
on me. I promise. I will never die. If you
18:18
left or something, I could, like, I'm not going to
18:21
just keep doing it. I can't do it by myself
18:24
or, like, with somebody else. It's just not the same
18:27
thing. the, in, in, The desire to even do that
18:31
wouldn't be there. Right. Well, there you go. Because this
18:36
also applies to... If you're a podcaster and your longtime
18:40
podcast co-host dies, it's not just that they die. They
18:45
could also just leave for business reasons or health reasons.
18:51
There's just lots of reasons why something like that could
18:53
become an issue for people. I mean, I guess it
19:00
has to be part of your planning process, I guess.
19:02
Basically, what we're saying is podcasting is a bad industry
19:05
to get into. It never ends well. It never ends
19:08
well. It's the road. Straight up. So... I have some
19:24
ideas. I've spoken to a few people. The worst part.
19:39
is an interesting I can't do the show with anybody.
19:43
that show the way it is. So something has to
19:46
transform and I may or may not be. But I
19:50
think the show can continue. one way or the other.
19:57
yet. And as I said, I haven't really. Tuesday, it's
20:04
like, okay, I... Monday. Like, what? everybody, trying to get
20:17
stuff together. We wanted to do a memorial show, of
20:19
course. I don't know what I did Wednesday. We just
20:23
did the show Thursday. Six hours is long when you're
20:27
reading 700 notes. And today or tomorrow I've got to
20:34
put some some kind of best of job. stuff to
20:41
me so that's really good always the producers of course
20:43
they do So I got to put that together and
20:46
then we have to. because more will come in. And
20:50
then, you know, John's memorial is going to be on.
20:54
So we got it to San Francisco. Coincidentally, like... within
20:59
the period that John died, I got a baby granddaughter.
21:05
Yeah, that's right. So it was kind of like life
21:08
and death in one go. And we got a We
21:12
got to go visit. Yeah. Because we haven't seen that
21:15
we were there long. You have some kind of buzz
21:19
going on. Do you have a USB cable? I heard
21:21
it. Do you have a USB cable somewhere near? Let
21:23
me. No, not really. I think it's gone now. Um...
21:30
So we'll see. As I would say in these cases,
21:36
it was a real day wrecker. That's for sure. It's
21:39
a real day wrecker. Now, this. in the podcast world.
21:46
Okay. group came out and still kind of secret who's
21:52
in it. You find it interesting? I find it interesting
21:57
because I see opportunity. That's why. I mean, I see
22:00
opportunity that may be counter to what most people see.
22:05
I thought the most... And... You know, so they have
22:16
a definition of is and you know I I kind
22:19
of like the idea of it works with your eyes
22:22
closed okay that's cute and then they add to that
22:25
which is Could be RSS, doesn't have to be. Okay,
22:30
well, I take some. Some exception to that. But then
22:35
it all folds right into... Well, okay, so I was
22:40
listening to Dan Granger on Pod News Weekly Review, and
22:44
he said, you know, the worst thing we can do
22:46
is we can be confusing our customers. And I thought,
22:51
hold on a second. Who is the customer of a
22:53
podcast? That's the listener. And they're all focused on the
22:58
advertiser. And I understand it, and I understand what they
23:03
want to do, and it's not going to make, it's
23:05
going to make a few people. nothing because agencies and
23:12
platforms and uh agencies, platforms, well, you know, they... The
23:22
role. You get 10,000 downloads. for any podcast, you know,
23:31
they're going to take a hundred of them, roll them
23:33
up and, you know, they're going to make their 15%
23:36
agency cut or whatever it is. And everyone else will
23:38
get, you know. off of it nice little money and
23:42
somehow that's a multi-billion dollar industry i'm just not seeing
23:46
it but i guess it's true and and so then
23:49
they come up with this metric which is so determine
23:58
what a play is and it's either 60 seconds or
24:00
30 seconds i'm like you actually with the with the
24:04
infrastructure they're talking about putting together handing over all this.
24:11
which we know that an app can actually do a
24:15
pretty good job of understanding what people are doing and
24:18
even how long they're listening. But see, they don't want
24:21
to do that. They don't want to actually say, yes,
24:23
advertiser, your ad was heard. played while someone had the
24:28
phone in their hand and was using it and the
24:30
play button was pressed we're going to presume that and
24:33
they heard your ad. They don't want to do that.
24:36
Which, if you're going to do all this data stuff
24:38
anyway, just get to it. You can do that, but
24:41
they don't want to because they know it's bullshit. Because
24:45
the numbers will drop off a cliff, which I personally
24:48
think would be good because then you have more scarcity,
24:51
you can ask for higher CPMs, but the industry is
24:54
all built around this now and they don't want to
24:57
see that happen. So the whole thing is kind of
25:00
a farce. As I'm listening to it, I'm like, well,
25:06
first of all, I said, if I'm on a podcast
25:10
app, just... and I hit another chapter, does that count
25:16
as another play? - Why would that, what are you
25:23
thinking in that regard? - Well, I mean, we're hearing
25:26
all of these numbers coming out of Apple, like, you
25:29
know, I had 300 people listening, but I had 350
25:32
plays. Why is that? Is that because the plays get
25:38
registered when it stops for a second because you got
25:42
a text message? message or you hit on a chapter
25:44
there's all these variables that we don't know and so
25:49
this just it's it's gotten if it's gotten even worse
25:53
I think what they're proposing it's worse than downloads now
25:56
it's just all these variables and they're cutting everybody out
26:00
it's only they only want to get Apple on board
26:02
they only get will be on board by default. YouTube
26:06
is in there. It's only going to get worse for
26:08
everybody. It's not going to get any better. They're not
26:15
really, no one's talking about Podverse or Podcast Guru or
26:19
True Fans. No one's talking about that. They don't care.
26:24
They do not care. They only care about Apple, Spotify,
26:28
and I guess YouTube. That's all the big three. The
26:30
big three, the big three, the big three. So, you
26:35
know, it's like good luck. Good luck. I'm conflicted a
26:46
lot about the whole amp thing because in... of me
26:50
that says this whole thing is just going to be
26:57
sort of a bum And then nobody's ever going to
27:01
think about it again. I mean, I would not. I'm
27:05
going to give my thoughts on it and try to
27:16
be as kind as I can, but also give my
27:24
raw opinion. and Giles and all these, I think they
27:30
just seem like nice, well-intentioned people. it would not surprise
27:38
me if this thing just basically goes into most people's
27:45
trash bin like right after they read it and it'll
27:49
never be thought of again. I just... I'm not trying
27:57
to be dismissive. I'm just giving what I think is
28:02
an honest assessment that I'm not sure that this document
28:06
is ever going to even last. in anybody's memory past
28:13
this week. That's a very short life you give it
28:16
there. Well, there's a lot of, there's a, there's some
28:21
reasons for that. I feel like, um, you know, they
28:27
contacted me last week and asked for my opinion. past
28:31
week and I was very honest with him that I
28:36
disagreed philosophically with doing things in private. And I think
28:45
that the only way they could get all these people
28:48
in the same Zoom session together was by making everything
28:55
very private and a bunch of NDAs and that kind
28:59
of thing. I think that's the only way that any
29:01
of these big players are going to actually talk to
29:03
one another and try to do things in anything that
29:07
looks like a committee or anything like that. like that.
29:12
So the secrecy was baked in from the very beginning,
29:15
probably because, probably not out of desire, but probably because
29:18
it was just the only way to get all those
29:19
people in the same room together. And the problem with
29:24
that, though, is... Whatever decisions they make in the secret
29:31
room don't have a chance to be commented on by
29:35
people outside of that secret room bubble. That's why doing
29:41
things in secret is so important. is not the right
29:44
approach and it kneecaps otherwise efforts because When you do
29:53
it in secret, you don't have the insight that communities
29:58
bring.
30:00
And you don't, you can. You can short circuit a
30:03
lot of wasted effort and a lot of Mediocre ideas.
30:09
by just having more people give their opinion. And that.
30:16
It was clear that that was not done here. Now,
30:20
I mean, I want, I was wanting to have. those
30:24
guys on the show, but that's proving to be a
30:27
little more difficult than I thought it would be. And
30:30
maybe we'll still be able to get that to happen.
30:35
um Like, essentially, I'm just gonna be talk- I'm just
30:39
gonna talk- about the the questions I would ask. or
30:44
I would want to ask is, The podcast definition part
30:54
of the document. To me, that seems like a little
30:59
bit of internal, like inside baseball to the marketing industry.
31:02
And I don't, I mean, that doesn't really, I don't
31:04
care, whatever. That's their business. then the document takes a
31:10
turn and asks and starts laying out a lot of
31:13
technical specifications. and what it lays out. is a heavy
31:20
lift. podcast app developers. What it wants is... is constant.
31:32
reporting. first party reporting. Play metrics. Ditter. that are somewhat
31:42
anonymized. And let me say what I mean by somewhat.
31:47
It's using the made MAID standard for like in a
31:51
hashed email address for the listener. It is. not clear
31:58
in what ways this this interacts with GDPR compliance and
32:04
California's privacy regulations. That's a good point. So immediately, and
32:09
the document lays this out. Um... It lays out. that
32:16
There will have to be opt-outs. and things like that
32:22
that are put in place in order to support this.
32:25
calls there is probably going to be some sort of
32:30
violation of GDPR or some sort of interaction with those
32:34
laws. So what you end up with. is a humongous
32:42
coding effort. on the part of a small team or
32:47
a solo app developer. and It's a ton of work.
32:54
With lots of exposure to downside. Because now, if the
33:01
stuff breaks... You hear directly from the app user. If
33:08
Now you have to support all the GDPR takedown, delete
33:12
my account requests. Yep. There is all, and all of
33:17
this stuff comes to the app developer, not the people
33:20
who wrote this specification. I didn't see app developers in
33:24
there anywhere. All that was in this document on the
33:28
app development side was what they want. Yeah. And there's
33:32
nothing in there about what the app developers are going
33:35
to get because the app developers are not going to
33:37
get anything but headaches. No. There's all the app developers
33:43
in what is being asked. The app developers get all
33:48
of the downside and none of the benefit. I mean,
33:52
really, they're doing what? advertising always drives companies to do
34:00
is hijack it. And the problem is that people have
34:03
tried to hijack RSS for decades. And it's very, very
34:07
difficult. And I don't, I really don't think they can
34:10
do it. And the whole nature, the whole. openness of
34:15
RSS is antithetical to how advertising works. Advertising wants to
34:21
be in a controlled environment. We have all this information.
34:26
And we can really also... artificially restrict the supply to
34:33
up the demand to increase the CPM. And the internet,
34:38
I mean, ask yourself. Who is really making a living
34:43
with a blog? Nobody. The only thing that is reasonably
34:51
successful is Substack. And I'm sure there are others. But
34:56
Substack has a very interesting model. They make it super
35:01
easy for you to... basically create a, you know, it's
35:07
like they have a... funding tag, you know, click it,
35:10
click here. You want to, you want to spend money
35:13
once you're. In the system... then you can subscribe to
35:18
any of these blogs. We'll just call them blogs. And
35:21
now they also do podcasts. And that's actually not a
35:25
bad idea. I don't like that it's a platform because
35:28
platforms can change. They can change ownership. They can go,
35:31
all kinds of stuff can happen. They can increase fees,
35:34
you know, all these problems. I think still. Podcast is
35:40
something you create to listen to if you want to
35:42
if you want to make your living out of it
35:44
You're gonna have to do something else. You it's not
35:47
just about how do I increase my downloads? They're there
35:51
this whole amp group the way I think they're looking
35:54
at it is like influencers. So they have their Dynamically
35:58
inserted ads which they just want to do at scale.
36:00
So I just want to buy 1000 podcasts insert my
36:04
ad there each podcast has 10,000 a thousand downloads, listens,
36:12
plays, whatever you want to call it. So that's a
36:15
million. that's uh a thousand CPMs. So that should cost
36:21
me, you know. $3,000, $5,000, whatever. Probably less. And the
36:28
money that trickles down is pennies to everybody else. The
36:32
influencer part is, which is a very established industry right
36:37
now. I know because my daughter is doing it. And
36:40
they look at how many followers do you have. So
36:42
you have to have at least, I think, 50,000 followers.
36:46
And then they look at your views. And they don't
36:49
really care. how long you want because it's video it's
36:52
a different experience and okay you get so you get
36:56
x amount of views per post uh we want you
36:59
to promote this product incorporate it into your uh into
37:04
your content which usually turns out to be the content
37:07
which i've always been a fan of that i like
37:10
home shopping network you know i think this is a
37:13
great model and that's kind of what they want to
37:16
do But that's because the platforms have all this information.
37:22
And, Lord, the platforms, they make more money on selling
37:26
the user, the viewer information than they do on the
37:30
ads, you know, to data brokers and all this stuff.
37:34
So it's really, it's, you know, people want to capitalize
37:38
and want to monetize podcasting. It's not set up for
37:41
it. And it never will be because it's not. This
37:45
is why they say it doesn't have to be an
37:46
RSS feed. That's the key. That's the key. That's the
37:50
thing they want. This is a podcast and you don't
37:55
have to do it in any certain way. But if
37:57
this podcast is on YouTube or if it's in some
38:02
other enclosed environment, then I'm sure Instagram will soon have
38:06
podcasts. They'll call it podcast. They'll have a tab called
38:09
podcast. X will have that. Everybody will have that. But
38:15
that's when you are so at risk. It's worse than
38:18
your podcast partner dying. Lost Adam. Oh, did you lose
38:22
me? Hello? Oh, there you are. You disappeared. Oh, I'm
38:26
sorry. Didn't mean to. That was weird. No, I was
38:30
saying that... The way did you lose me? Uh, I
38:35
don't remember. Okay. Don't worry. It was like five words
38:39
ago. It wasn't important. I just, I don't see it
38:43
as... Oh, you said Instagram's going to have a podcast
38:46
tab. Yeah, everyone's going to have a podcast tab. And
38:49
that's going to be called a podcast. And now you'll
38:51
have to upload your video to all these different platforms.
38:55
It's the opposite of what podcasting is. It's the opposite.
39:04
Well. I can tell you this. Marco Arment. is going
39:11
to keep a copy of this document in the bathroom.
39:15
so he can wipe his ass with it every time
39:16
he goes to the bathroom. And I can pretty much
39:21
guarantee you that most other indie app developers... Are not
39:26
even going to read it. But see, that's the point.
39:30
I've heard these guys talk. They don't care. They want
39:32
Apple to do it. And they're going to push Apple
39:35
as far as they can. And guess what Apple's going
39:37
to do? Nothing. No, they will. Oh, they definitely will.
39:42
Oh, yes. Oh, yes, they will. Look, you think that
39:44
you think Apple's going to report. stats back every 30
39:48
seconds they're already in the game with hls video no
39:53
yes they are they're taking a piece Yeah, well, yeah,
39:57
they're taking a piece of it, but why would they
39:59
help in this way? You're missing my point. They'll be
40:03
like, sure, we'll help you out. We're going to take
40:06
a piece. I don't think so. They're already doing it.
40:10
Why wouldn't they? Because they weren't part of it. That's
40:15
not the way they would. They don't let other people
40:17
build specs and then go adopt it. When it's something
40:20
like the only reason that Ted and the fellers at
40:28
Apple Podcasts have implemented the podcasting 2.0 tags is because
40:33
they made a lot of technical sense and it matched
40:37
up with some things they wanted to do. There was
40:44
a synchronicity there that happened over time. And we just
40:49
ended up, you know, we were all in the same
40:51
rooms together at various podcast conferences and stuff. And it
40:55
just evolved as a relationship over time. Why did Apple
41:01
get into taking a piece of the advertising with HLS
41:06
video? Why? because they developed it. Oh, okay. So they
41:15
can look at this. That's my point, though. They wrote
41:17
HLS, and they also developed the spec to get the
41:21
money out. Apple podcasting has always been a cost center.
41:25
I'm not talking about Ted and the wonderful people over
41:27
there. Sure. We got a new CEO coming in. I
41:30
don't know what he's doing, but it seems to me
41:34
that if there is a group out there who says,
41:37
hey, there's another billion dollars on the sideline here of
41:39
an already billion dollar or two billion dollar or 20
41:42
billion dollar industry. I don't know what the numbers are
41:45
anymore. Seems really big to me. And all we have
41:48
to do is jump into it because these guys want
41:51
to spend money and we can get a piece of
41:53
it. Oh, we're going to tell you how the spec
41:55
is going to work. Because I can tell you? that
42:02
none of that would go to Apple. Almost 100% of
42:06
it would go to YouTube. Yeah, but... The billion dollars
42:09
sitting on the sideline, it's not going equally... It's not
42:12
being equally distributed amongst all these players. It's going to
42:15
go heavily to YouTube. Well, that's possible. Because they defined
42:20
the definition of a podcast away from RSS. Yeah, that's
42:24
true. That's true. I also think that the Apple ethos...
42:31
is You don't just go do a thing and drop
42:34
this spec on us and tell us we have to
42:36
do it. That's the quickest way to get them not
42:38
to do something. Okay. Well, we shall see. We shall
42:42
see. But for sure... Doesn't matter. This movement they have
42:48
here. With those Judases over at Libsyn. It doesn't include
43:00
us. Let's just be very clear. No, that's for sure.
43:03
Let's just be very clear. But I like that position.
43:06
I like being different. I like being counterculture. And the
43:12
only solution I have, which doesn't help the app developers
43:16
yet, but I'm really thinking, how do we evolve the
43:20
funding tag? If we could evolve the funding tag. to
43:25
help the app developers. And I really, you know, I'm
43:29
really looking at different payment systems. And I'm talking boosts
43:33
just for, you know, let me send something, a subscription.
43:38
There has got to be a way. that this can
43:43
be done. And it may take. It may take a
43:47
centralized company to do that. I actually have a funny
43:52
feeling that X money might might make something possible. -
43:59
What is X money? - Oh, that's Elon's everything app
44:03
money. Yeah, yeah, you can make the sound, but he's
44:07
gonna offer 6% APY interest rate if you store your
44:11
money with him. People will flock to it. Isn't this
44:15
the second time that X slash Twitter has promised to
44:18
be everybody's wallet, though? Well, I... You know what? It's
44:23
easier than going to Mars. That is very, I cannot
44:30
find fault with that statement. Is there some way? that
44:35
we can take this Newly proposed. unratified contact tag from
44:44
Daniel and Alex and Filippo. And use that as. The
44:51
messaging mechanism. for the for the funding tax somehow this
44:58
is all just come up in my head like with
45:00
We could always expand the funding tag to add that
45:04
in. That would make more sense to have another field
45:07
and another destination. Like a linkage. Like if you had
45:11
a contact tag that was like the designated message destination
45:20
for the funding tag thing. I'm throwing stuff at the
45:27
wall here. That's cool. If we're gonna do this... And
45:32
we haven't seen a lot of uptake on the funding
45:35
tag. And by uptake, I mean not that you have
45:37
a dollar sign and you can tap it and it
45:39
goes somewhere, but like the boost button. It was a
45:41
big-ass button. You know, like, boom, hit this. And I
45:45
understand why. The only way I think we should develop
45:48
this, if we can figure out a way that we
45:50
can get some of that money to an app developer.
45:56
And that's the only way I really want to do
46:01
it. do it because otherwise we're just killing ourselves. Or
46:05
they're being eschewed. They're being eschewed. Right. If we can
46:12
turn that around. And that was kind of the beauty
46:17
of Lightning is we had the splits. The splits was
46:21
really the thing. You had a split for the app.
46:24
You could put splits in however you wanted to do
46:31
it. And it could be for bigger amounts. We don't
46:34
have to do the micropayments. But if I knew that
46:39
certain apps were surfacing a big donate, give, boost, whatever
46:44
button you want to call it, I tell people to
46:46
use that app. Hey, it's right in your app. Hit
46:48
that now. But it has to benefit them as well.
46:53
I agree. Otherwise, I don't think we should even work
46:56
on it until we figure that part out. All the
47:01
same problems come into play. Money transmitter licenses, you know,
47:05
all this stuff. How does, so what Buzzsprout does, they
47:10
do that from the hosting side. Yes. And so I
47:16
presume that they're taking a platform fee for doing it.
47:20
should be i don't know about that i i don't
47:24
i if i remember correctly when at least when they
47:27
rolled it out i don't know about the current state
47:29
but at least when they rolled it out i'm almost
47:31
positive that they said this was just a service to
47:36
their to their podcasters they were not going to take
47:39
any sort of cut they should they just they yeah
47:44
the the thing about like i was praising the the
47:50
Buzzsprout solution last week and I think part of that
47:58
solution could be if you had a verified Daniel J.
48:10
Lewis is in the boardroom. He says, if an app
48:13
developer offers a strike... they can take a fee from
48:18
the payment. That's interesting. So does that mean that I
48:25
would have to register my podcast with every single app?
48:30
problem actually it's not that big of a deal no
48:33
i mean you could single sign-up sheet that could That's
48:40
not that crazy. But you'd have to sign up with
48:43
every app. Well, I'll tell you what, that would take
48:46
me a couple hours on an afternoon. I'd sign up
48:47
with every app that offered that. Let's go. The other
48:52
way to do it is if you have a third
48:56
party or a hosted platform, tell me where I'm messing
49:00
up. Tell me what I'm missing here. If you had
49:04
the bus route model, I'm in the app. I tap.
49:09
And it. doing, you know, this. Can't the service... if
49:38
and Striper Card. to each individual. And then the podcast
49:45
app is the one that takes a platform fee and
49:48
sends the money directly to the podcast app. my Stripe
49:51
account. Is that correct, Daniel? That's interesting. Here's what I'm
49:58
trying to avoid. you The effectiveness or the excuse me
50:09
the adoption on the app side. Lightning was a heavy
50:17
lift on a technical stem from a technical standpoint. That's
50:19
not. that but the other i think thing was that
50:27
app developers are scared about getting in the money. they
50:31
don't want to be the ones that dole out the
50:34
cash. The way I understand it, a stripe is pretty
50:38
interesting. with Stripe. They do have a lot of... and
50:45
do things. And if you are just connecting, if you
50:52
just have a Stripe account and... A podcaster has to
50:56
connect to you. which I made it. out then you're
51:02
not handling any money it's automatic stripe does that on
51:05
the back end for you I don't think I'm following
51:09
can you explain this to me again Daniel J. Lewis
51:13
is saying, and I've looked into Stripe just before we
51:16
started actually, but I was a little fuzzy. So the
51:20
app developer has a Stripe account. And then... You hit
51:25
the... your story. to the podcast app strike account. And
51:33
so when you, when someone hits the boost button, the
51:36
funding button, the donate button, whatever, then it goes. collected
51:46
by the app developer. your piece, podcaster piece, is directly
51:51
sent to your Stripe account. So it's a form of
51:54
a split. I think I would have to... We gotta
52:00
look at it. We gotta look at it. Yeah, I'm
52:02
just too visual. I think I just need to see
52:04
it. I see strings. Hmm. Let's just presume that worked.
52:19
Okay. I mean, so I'd have to connect my podcast
52:24
to a couple of different apps. You know? Daniel says,
52:33
I'm using clerk.com as my authentication and payment system for
52:36
Podchats. Stripe account, but Clerk manages the payment page and
52:42
subscriptions. So Clerk, not Stripe, is doing the recurring payments.
52:47
Then whenever I get paid in my Stripe account, I
52:50
see that Clerk got a 0.7% fee from it. Right.
52:56
But then you still have to signal. The problem is
52:58
the signaling. How do you signal from a tag in
53:01
the RSS feed that this is me, this is my
53:05
podcast Stripe account? through this app. So that's where I'm
53:12
saying that if you use... You have a ma- Like
53:20
in Buzzsprout's case. the payment in If they... verify That
53:29
that payment came from Castamatic and Franco has set up
53:35
the relationship and is like, you know, shown that maybe.
53:40
Here's how you pay me address. They can just turn
53:46
around and take a, you know, half a percent and
53:51
shoot it to Franco. by some money transmitter platform that
53:59
has all the licenses and does all that. Or maybe
54:02
Franco goes and gets a subscription, a podcast app developer
54:06
subscription. And I'm using. this is what I've seen before.
54:12
But this could be, Daniel J. Lewis could build this.
54:16
He could build a thing that says podcast, you know,
54:23
podcast payments doc. something and and it operates just like
54:31
this it takes sort of like an escrow for the
54:36
podcaster and then just disseminates it or whatever. Again, we're
54:45
throwing everything at the wall. And the sad part is
54:48
we solved all of this with lightning. We did all
54:50
this. We did all this work. We had it all
54:53
figured out. I know. I know. And it almost works.
55:01
Even right now. Yeah, I don't know, man. I don't
55:07
know. Maybe just this is the idea. Yeah. Phase of
55:13
this. Well, Daniel J is smart. He'll think about it.
55:19
I'm too foggy. I did want to talk about this.
55:23
I did want to talk about this new podcast app
55:29
that strips out sponsor messages and commercials and all that
55:34
stuff. I briefly saw that go by. I don't remember
55:38
what it was called or anything. I tried before I
55:44
forgot what it's called like ear say or something like
55:46
that. It stripped out anything it stripped out like Stuff
55:51
all over the podcast that was not ads like whole
55:53
sections of the podcast would disappear right because it thought
55:56
it was talking about an advertisement Right, right, right, right
56:00
So this app is called Podtastic. Podtastic! And I downloaded
56:13
Podtastic, and it's actually not even that bad of an
56:16
app. Really? Yeah. Let me just bring it up here.
56:21
Podtastic. And, you know, forget the Podtastic. There we go.
56:29
Forget for a moment the stripping out, you know, and
56:32
it charges you money and stuff and whatever. Forget for
56:35
a moment the stripping out ad stuff. But what it's
56:37
doing is it's creating chapters. that it thinks is a
56:43
chapter. It's played a lot of chapters, actually. It's exactly
56:49
what I've been... It is trying to, and you can
56:52
feed back like, you know, I didn't like this. I
56:54
did like that. It didn't, you know, this was not
56:56
relevant to me. that is, I guess, maybe not in
57:00
this form, but moving forward, will start to customize the
57:05
additional things around the actual podcast to your preferences. Okay.
57:11
And that's interesting. And so, you know, it would be
57:18
quite trivial in general to create a podcast app, a
57:24
client side app that, you know, just says, you know
57:28
what? just strip out the first two minutes of pivot
57:32
because I hate those ads. And it would do that.
57:36
It would do that. But that's just like low-hanging fruit.
57:43
Really, that's what you're going to market on? But doing
57:46
other things and looking for other podcasts and different episodes
57:50
or people who, you know, a lot of the stuff
57:52
we already have. like person tag. There's so much that
57:56
you can have your podcast app do. Make it smarter.
58:01
And so this just happens to, you know, the marketing
58:05
ploy here is, oh, no more ads, which is just
58:08
mean. It's not like it's illegal. To me, it's not
58:14
illegal. I don't think it's illegal. You download the podcast.
58:17
You do whatever you want. You can strip out the
58:20
ads. You do whatever you want. That's fine. As long
58:23
as they're not re-hosting it and it's all happening on
58:26
device, there's really nothing legal there. But I'm saying that's
58:29
the future. People will want this. Phones are going to
58:34
be doing this, all this stuff. And by the way,
58:37
just as a throwback to the amp stuff. Ultimately, the
58:42
app doesn't decide what's going to happen. It's your phone.
58:46
And the person who owns the phone, i.e. Apple or
58:49
Google, they're going to decide what comes and goes, and
58:53
they're going to want a piece. It's all so messed
58:57
up. It's all messed up. Back to browsers. Back to
59:03
browsers, I tell you. Well, yeah, that's the sort of
59:10
idea. there is in this moment where they are introducing
59:17
all this heavy lift and all these specifications to support
59:22
And I honestly, I don't see, I don't see that
59:27
the... are going to do that either. I know they
59:31
had all these people in the room, but that doesn't
59:34
mean they're actually going to do it. No. And so,
59:38
like, Adam... At the moment where they're introduced, heavy lift
59:44
you also at the same exact time see apps like
59:48
this trying to strip the ads out completely altogether. Advertising
59:54
in general is old fashioned now with ads inserted and
59:58
all that stuff. That's all gonna go away.
1:00:01
It's all going to go away. Yes, with LLMs on
1:00:05
your device, it's all going to go away. It's going
1:00:09
to be a beautiful world for you as a consumer,
1:00:13
as a listener, as a viewer. You're going to be
1:00:15
able to do a lot of things. And if the
1:00:18
programmatic ads go away, then it makes a real... meaningful
1:00:24
dent in the slop problem too. Yes. Oh, and I've
1:00:29
always said that's transitory. Yeah, you have. Yeah, that's been
1:00:33
your point. That's just because what can I do? What
1:00:36
can I do with this? What can I do with
1:00:37
this? What can I do? I can set up these
1:00:41
three agents and I'll get up with podcasts. Yeah. It'll
1:00:45
be $400 of extra ad income. Dana Brunetti. You know
1:00:49
who Dana Brunetti is? A Netflix guy, right? No, no.
1:00:54
Danny Brunetti is a big-time Hollywood producer who left Hollywood
1:00:57
because he hates it so much. But he did. He
1:01:01
produced The Social Network, Captain Phillips, House of Cards. His
1:01:09
big money came from Fifty Shades of Grey. That's the
1:01:13
big money. I thought that movie didn't do good at
1:01:15
the box office. Huge money. Really? Oh, no. That's like
1:01:20
$100,000 a month checks. Of course it's porn. Yeah, Fifty
1:01:24
Shades of Greyer. He did a follow-up. And what was
1:01:28
it? The Grand Prix movie, which I think was Sony
1:01:32
paid for. So he's a nerd. And he was, we
1:01:37
were pretty, I'm friends with him. He was good friends
1:01:39
with John. And so I've been talking to him a
1:01:42
lot. And then John would say, call Adam about that.
1:01:44
And he was sitting at home with Claude Code. And
1:01:47
he's like, look at this, man. I made a podcast.
1:01:51
It's a news podcast for just my local area. And,
1:01:56
you know, I listened to it. Yeah, it was great.
1:01:58
He happened to live up where he lives in the
1:01:59
farm country in Northern California. And, you know, and he
1:02:03
was just scanning feeds and pulling it in and a
1:02:06
TTS model and boom. And, but he was really only
1:02:12
doing it because he could. And then I said, Dana,
1:02:14
you're a producer. Tell me how you're going to make
1:02:17
money on this. Please explain to me where the money
1:02:19
is. And he comes back a few weeks later, said,
1:02:22
nope, there's no money. It says, just, I can't make
1:02:25
it work. And it's true. You know, he said, I
1:02:30
have to create so much at such a scale in
1:02:33
order to get, you know, a thousand bucks out of
1:02:35
it. He says, no, it's not worth it. But he
1:02:37
liked it and he liked it for himself. And that's
1:02:39
my point. So creating that for your little community, valid,
1:02:44
great, fantastic. You be the guy. Have people buy you
1:02:47
a coffee and chip in for your token cost. Or
1:02:51
your server chip in. Or for your server that's running
1:02:58
at home. So We are entering a world and we
1:03:04
can be ahead of a lot of this. We can
1:03:06
be way ahead in creating great experiences for our true
1:03:09
customer, which is the listener. For me, that's the listener.
1:03:12
And for the podcast app developers, that's the listener. That's
1:03:16
the user of their app. I think we can accelerate
1:03:20
far beyond and be much faster than anyone else. It's
1:03:25
my dream. Claude Opus 5 is released. Oh, so yeah,
1:03:30
I wanted to play this clip. This is, I mean,
1:03:34
I love a news story that is just so obviously
1:03:38
blatant marketing. Yeah, I know. You're going to talk about
1:03:43
the OpenAI hacking, hugging face, right? An unprecedented AI security
1:03:48
incident. Chad's EPT creator, OpenAI, says two of its models
1:03:53
escaped controls and successfully hacked into another AI company. OpenAI
1:03:58
says they are, the way they describe it, is responding
1:04:01
accordingly. But they also say this involves state-of-the-art technology and
1:04:05
is unprecedented. Unprecedented. Claire Duffy for more on this. Good
1:04:11
news. Yeah, Kate, this kind of sounds like the plot
1:04:13
of a sci-fi movie, but it is indeed happening in
1:04:16
real life. This is the first major example that we've
1:04:18
seen of an AI model independently conducting a hack outside
1:04:22
of human control. And this is something that experts have
1:04:25
been warning about. This is also driving some of these
1:04:28
conversations that we've heard. This is CNN, by the way.
1:04:30
That's why I pulled this one specifically, just to get
1:04:33
the urgency and the excitement. Including at the White House
1:04:36
about increasing controls and oversight of AI models. of exactly
1:04:41
this sort of hack. Now what happened here is that
1:04:44
OpenAI was testing these models. It had lowered its normal
1:04:48
safeguards in order to test these two models, which is
1:04:51
GPT 5.6 Sol and also a new unreleased more advanced
1:04:55
model. More advanced? You don't have it yet, but what
1:04:57
we got here in the closet is amazing. It lowered
1:05:00
its safeguards because it thought it was testing in what
1:05:02
is called a sandbox. Oh, it was in the sandbox.
1:05:05
Don't worry, people. It was what is called a sandbox.
1:05:10
Sandbox. virtual lab. But when it gave these models a
1:05:18
task to test their hacking capabilities, the models went to
1:05:21
such great lengths to complete that task that it found
1:05:25
vulnerabilities that allowed them to access the open internet outside
1:05:28
of this controlled testing environment. They then hacked into Hugging
1:05:32
Face, which is a platform that hosts all kinds of
1:05:35
AI models. They used stolen credentials. Hey, don't go to
1:05:39
that cheap Chinese crap. We're better. We can hack them.
1:05:42
Look what we can do. We're better. The AI war
1:05:45
is over. We beat China. Stolen logins to navigate that
1:05:50
platform. And ultimately, this hack was identified by Hugging Face
1:05:54
in part because of artificial intelligence, which they used to
1:05:57
shut it down. But this is a huge warning sign
1:06:00
for this industry. In a statement, OpenAI said, we consider
1:06:04
this incident to be an unprecedented cyber incident involving state-of-the-art
1:06:09
capabilities and our response. How many times are they going
1:06:11
to say that? State of the art. State of the
1:06:13
art. Wait for the payoff. We are sharing preliminary findings
1:06:17
at this stage to help defenders understand what happened and
1:06:20
to help calibrate on what models are now capable of.
1:06:23
I think that is the real key here is that
1:06:25
models are now capable of conducting these kinds of hacks
1:06:28
independently. Yeah, the statements from both the companies seem to
1:06:32
be somewhat in awe that it actually happened, but it
1:06:35
is terrifying. It's terrifying. It's terrifying. It's terrifying. Be terrified.
1:06:40
be more terrified. I mean, it's so obvious. Anthropic had
1:06:45
their, oh no, we have a model here that is
1:06:49
so cyber hacker aware that we can only give it
1:06:52
to the banks first because, oh, this is just too
1:06:55
much. And then OpenAI comes out. We got those cheap
1:06:59
model guys. They had to come to us and beg
1:07:01
us for help to use our models to combat the
1:07:05
hack that our models did on their model. Bull crap.
1:07:11
The... It's like our model. beat the shit out of
1:07:16
their model. It went in. It didn't just hack Hugging
1:07:21
Face. It hacked Hugging Face and stole all the Chinese
1:07:24
models. It's so good that we couldn't even keep it
1:07:29
in the sandbox. The sandbox was full of poop. They
1:07:35
didn't tell you that they forgot to like, that they
1:07:38
accidentally left the hugging face like SSH key sitting there
1:07:41
or something like that. I love it. I love it.
1:07:43
Yeah. And then they did like a SQL injection. Okay.
1:07:47
Sure. Sure. Like, and I'm thinking hugging face, hugging face.
1:07:53
Really? If you, if you're hugging face, which is probably
1:07:58
90% built with vibe coding, if you're hugging face and
1:08:02
you've got your stuff hacked. What it kind of, this
1:08:05
whole thing is, this is Doom trolling. to the nth
1:08:10
degree. Yeah. Like, this is so stupid. The New York
1:08:14
Times article I read called, it didn't say hugging face.
1:08:19
It said open AIs. A new top secret security model
1:08:22
hacked into a library. Ooooooh! It called Hugging Face something
1:08:29
like an open library or something. I was like, archive.org?
1:08:32
What are we talking about? First of all, Gail Mann
1:08:35
amnesia. So when you read that, you can't believe anything
1:08:39
else that you read in the New York Times. So
1:08:43
apparently what happened is they had this... This model was
1:08:49
sandboxed, but it could access some kind of worker box
1:08:54
that had libraries that if it needed, but that was
1:08:57
also not connected to the internet, but it hacked the
1:08:59
worker box and the worker box. Then they went through
1:09:02
there to get to someone's, you know, PC on a
1:09:05
desktop. And that's how they got to the internet. You
1:09:07
know what? Script kiddie stuff. That's how the model does
1:09:11
it. It's what script kiddies do. They keep pounding, keep
1:09:14
pounding. Oh, okay, it just can do it faster. Like,
1:09:17
and, you know, come on, please. They made this happen.
1:09:21
Don't even think about it. Oh, who's, where's the proof?
1:09:24
There's no evidence it was sandboxed and it was supposed
1:09:26
to be all secret. Oh, we can't believe this happened.
1:09:29
Blah, blah, blah, blah. No, we can. We love it
1:09:31
that this happened. We beat the cheap guy. You don't
1:09:34
want those cheap guys. Hey, hey, hey, Mr. Corporate Fortune
1:09:38
500, you don't want to use Hugging Face. with their
1:09:40
models, with their cheap models that your IT department is
1:09:42
saying is much cheaper. You don't want that because look
1:09:44
what can happen. Look what can happen. You want the
1:09:47
big guys, the big boys, the frontier, the frontier of
1:09:50
everything. You want to be with us. obvious. Yeah, it
1:09:57
says, the incident which happened last week while OpenAI was
1:10:00
testing the cybersecurity capabilities of systems displayed the kind of
1:10:03
science fiction potential that AI companies warned would soon become
1:10:08
a reality. Where's my theremin when I need it? I
1:10:15
mean, this is... I'm so sick of this crap. Yes,
1:10:20
reality. We are implementing strict controls in infrastructure configuration at
1:10:25
the cost of research velocity while the vulnerabilities are patched.
1:10:31
Research, we have to stop our research velocity. It's too
1:10:36
fast. He's spinning out of control. Show title, Research Velocity.
1:10:44
I like it. Yeah. That's what happened with amp. Just
1:10:49
got too fast. The research velocity just got out of
1:10:53
control. Off the rails with the research velocity. That's too
1:10:57
funny. Can we talk about pod thing? Yeah, please. Because
1:11:01
I did some stuff this week. Dave did some stuff,
1:11:04
everybody. I did a lot of stuff. Actually, a lot
1:11:08
of stuff. Did some stuff. Yes. trying to currently get
1:11:14
out of of my remote desktop inception that I have
1:11:23
going on here that is confusing the crap out of
1:11:25
me. Okay, here we go. Pod, yeah. So, Podping... .cloud
1:11:35
is now fully Operational. Okay. Wow. With gossip. P2P gossip
1:11:43
over Eero protocol. Okay, so is there something I can
1:11:48
look or it only works if I have the... Well,
1:11:51
the question is, do you even know what I'm talking
1:11:53
about? No, I have no idea, but I want to
1:11:56
go to podping.cloud and see if I can do something
1:12:00
there. Well, there's nothing there. No, there's nothing. I mean,
1:12:03
there's stuff. There's something there, but... Oh, all right. Yeah,
1:12:08
uh... So what this means is. that There's now, so
1:12:16
each, so there's five, just to recap real quick, there's
1:12:19
five podping.cloud hosts that are all load balanced under the,
1:12:23
behind the domain of podping.cloud. Yeah. So if you send
1:12:27
a podping to any of, to podping.cloud, you're going to
1:12:29
get, it's going to hit one of those five hosts
1:12:31
round robin. So... Um... Four. Podping.cloud nodes would just write
1:12:44
the incoming podpings to the Hive blockchain. And the stack
1:12:52
looked something like this. There was a front-end web server.
1:12:58
And... which exposed a ZMQ zero MQ socket. On top
1:13:11
of that socket you had... A Cap'n Proto protocol layer
1:13:17
delivering messages with... Alex Gates' Plexo. type control system. Okay.
1:13:26
These messages would write. to the ZMQ socket and on
1:13:30
the other end of the socket. The only thing that
1:13:33
was listening was the Hive Rider. Right. Hive Rider would
1:13:37
accept the ping. and give back, sometimes in batches. maybe
1:13:45
10 or 12 at a time. It would accept those,
1:13:48
write them to the Hive blockchain, and then send an
1:13:51
ACK. back through the socket saying hey yeah we're good
1:13:55
it's written dequeue it That is basically... What, DQ? What
1:14:01
is DQ? Remove it from the queue. Oh, okay. So
1:14:05
it's like I've handled your ping that you gave me
1:14:08
to write. Take it out. Stop trying to send it.
1:14:14
So that is the way this thing has always worked.
1:14:17
Now. It's writing. over two. sockets. So now it's writing
1:14:28
to multiple, excuse me, podping.cloud frontend is sending to multiple
1:14:34
writers. The first, it's still sending the hive. That will
1:14:38
continue. Right. No change there. But there's a second rider
1:14:43
now, which is riding to... pings to the Eero gossip
1:14:50
channel. Okay, so we have parallel systems. Correct. Correct. Sounds
1:14:56
good. Sounds like a reasonable thing to do. Redundancy.
1:15:00
Do you have it sandboxed or can it get out?
1:15:02
No, if it gets out, it's going straight for pod
1:15:05
news. I'm sorry, James, but if this thing escapes containment,
1:15:11
your side is toast. So now this is. Fully deployed.
1:15:22
This was all under the podping.alpha repo, the development work.
1:15:27
Everything proved out. It ran for a few months with
1:15:30
no... Real issues. So the thing that I did this
1:15:35
week is breaking out each one of the different parts
1:15:39
of the system into their own repos. So now... Uh...
1:15:45
the everything's split out. There's the podping.cloud with those changes
1:15:50
and that is fully Dockerized, Docker composed. I'm riding the
1:15:56
sysadmin. train here. And then you also have the pod
1:16:03
ping dash gossip writer repo, which is the writing portion
1:16:09
which sends pings to the network. You have the Podping-Gossip
1:16:14
Watcher. which is a standalone. You can run it by
1:16:19
just compiling the executable and running, or you can run
1:16:21
it with a Docker. With a Docker run. So here's
1:16:24
my question. If I want to run this. What kind
1:16:28
of disk space and CPU does this take up? very
1:16:33
little. On both? Very little. So let me give you
1:16:37
the... The monitor. Let's see. Pod ping monitor. So go
1:16:42
to PP monitor. Monitor your pee-pee. ppmonitor.podcastindex.org and this will
1:16:55
show you the mesh. Mesh. Oh! Whoa, whoa. This is
1:17:01
what the current swarm looks like. Oh, I love dragging
1:17:04
these little nodes around. That's sexy. You can drag nodes.
1:17:09
Why is it moving? What signifies a move? Well, it
1:17:13
tries to reorganize itself to fit everything on the screen
1:17:16
as best as it can. Oh. So things in the
1:17:22
swarm. It's very swarmy. You have a mesh of nodes.
1:17:32
And you can imagine that with any mesh, you're going
1:17:35
to have edges. And those edges... Holy crap, that's a
1:17:40
lot of nodes here. This is cool, guys. And the
1:17:44
edges are naturally less connected than the internal nodes. People,
1:17:51
you've got to upgrade to 0.8.0. You're on 0.6.0. Upgrade.
1:17:56
Upgrade. Okay. You're lagging. So are you. DaveLaptop2. DaveRemote1. DaveRemote
1:18:04
underscore Nix1. Yeah, but look at Dave underscore laptop 2.
1:18:11
It's 080. Wait. I'm half upgrade. Dave remote underscore NX1
1:18:15
is 0.90. That's the top secret version that we can't
1:18:20
release or else it'll kill everybody. Keep it in the
1:18:22
sandbox, please. I've only given it to certain partners first
1:18:28
to make sure that we... This is nicely vibe-coded, man,
1:18:33
this website. That's good. I like it. Yes, Eric. a
1:18:37
jiggly blob of nodes that's correct jiggly blob a jiggly
1:18:40
blob of notes very nice i like it it's good
1:18:43
so Sexy. It's very sexy. You can see like with
1:18:47
any mesh, the edge nodes are sad. They're sadly connected.
1:18:54
Sad. Hey Citizen seems to be the spider in the
1:18:58
center of the web. Hey Citizen's bathroom. It's Hey Citizen's
1:19:02
bathroom. That's great. That's great. I've got to get my
1:19:07
node up and call it something really obnoxious. So the
1:19:13
node underscore friendly underscore name is how you name your
1:19:17
node and obscenities will flow, of course. The So... The
1:19:26
network works. Great, Dave. This is super cool. super cool.
1:19:32
Where's Apple's node? Oh, I'm sorry. It's coming. It'll be
1:19:38
called ITMS or something inscrutable. So this... The five... So
1:19:47
once everything was deployed to podping.cloud, the five... Podbean.cloud hosts.
1:19:56
Riders. Their Eero node key. I took those keys and
1:20:04
baked them directly into... The code of the watcher, the
1:20:09
monitor. And the rider. So the way these things work,
1:20:16
they can all operate independently. And each one of them
1:20:20
will attempt to... contact the swarm over the mainline DHT.
1:20:27
Why is... SU Pipe Pipe CD, Gray. So sometimes things
1:20:34
go on and off. Things go on and offline. Oh,
1:20:38
he's offline. That's just a natural consequence of these things,
1:20:43
like a long connected... nodes will end up getting They'll
1:20:50
drift. Their connectivity will drift because it's a consequence of
1:20:55
the high-par view. protocol and you can see he just
1:21:00
came back online oh yeah So Hyparview has a problem.
1:21:03
Welcome back. Welcome back, buddy. What else can I use
1:21:08
this beautiful mesh network for? Can I send text messages
1:21:11
over it? Oh gosh. please. But, but this, but that's.
1:21:18
A great question, though, because... I think we can. Extend
1:21:25
this now. to other use cases. I mean, use cases
1:21:31
that fit in this same vein. Obviously. We can use
1:21:36
this same format, maybe not the Podping network, but this
1:21:40
is how we can decentralize the index. Yes, now you're
1:21:44
talking. Now you're talking. All right, I've got to get
1:21:46
a note up. I've got my new machine in. I
1:21:49
got my... I didn't know you were getting a machine.
1:21:51
Yeah, yeah, yeah. I got... I'm too poor to get
1:21:56
what I really wanted. I'll tell you what I got.
1:21:58
It's still in the box because, I don't know, Dvorak
1:22:01
died. Bastard. My plan was to get this up on
1:22:06
Tuesday. Thanks, John. Yeah, thanks. Let me see. What did
1:22:10
I order here? It's the... GMTechEvo-X2 with 64 gigs. What
1:22:22
is this? I don't know what this machine is. It's
1:22:26
AMD Ryzen AI Max. Oh, Strix Halo? Yeah, yeah, yeah,
1:22:30
Strix, yeah. Is it the 395? 395, yeah. 64 gigs
1:22:37
of RAM. Yeah, I couldn't afford 128. It's like five
1:22:41
grand. Yikes. How much was the 64? Let me see.
1:22:46
View your item. Buy it again. Oh, there you go.
1:22:49
Yeah, just get another one. It was, hold on. Why
1:22:52
is it not telling me? one nine nine nine That's
1:22:57
a huge price difference between 64 and 128. No kidding.
1:23:02
No kidding. So what are you gonna run on it?
1:23:05
I'm going to run my Claude agents because I'm running
1:23:10
it all on my show computer. And the show computer
1:23:13
only has, what do we have here? Uh, 16 gigs?
1:23:18
So you run a couple of Claude agents and you
1:23:22
run Currycaster and Thunderbird and Obsidian and your machine's going,
1:23:28
my swap is flapping! The last thing anybody wants is
1:23:35
a flapping swap. You do not want your swap to
1:23:37
be flapping. So I figured I'd run three agents on
1:23:41
it and a whisper model because I do a lot
1:23:45
of whispering. But that costs quite a bit of money.
1:23:50
You know you it's like every 10 days of 25
1:23:53
bucks 25 bucks 25 bucks Well, I got a tighten
1:23:55
the belt now, you know with with my dead podcaster
1:23:59
friend. Yeah and And I know to the node. Yeah,
1:24:04
the node is another problem And I'm hoping because I
1:24:09
saw this I tried out this really cool Chinese TTS
1:24:13
model called dots I think it's dots. dots TTS Holy
1:24:20
smoke. This thing is good. Is it on the recently
1:24:23
hacked hugging face? I think. And so I did a
1:24:29
little test. just on my 3090 card on the Pi,
1:24:34
which, yeah, that whole setup is weak. Then I gave
1:24:39
it 15 seconds of my voice and it was impressive.
1:24:43
Very impressive. You need this. You need DTS. Yeah. Yeah,
1:24:49
so Whisper and TTS, those are the two main things.
1:24:52
I don't need to run a huge model or anything.
1:24:55
I don't need to run that. I'm not into attempting
1:24:58
that because I've already done these things and it's always
1:25:01
disappointing. And I'm like, okay, I'll just pay Claude. Fine,
1:25:05
whatever. That thing, that 64 gig machine could run Quinn
1:25:11
3635B like a bandit. It can, yeah. But I don't
1:25:16
think I can fit the TTS and the Whisper model
1:25:19
in there at the same time. Oh, really? I don't
1:25:21
think so. Well, I don't know. You may want to
1:25:26
try it because the 35B is only 3 billion active
1:25:30
parameters because it's an MOE model. It is very, it's
1:25:34
very convenient. But I like, I like Claude's personality, Dave.
1:25:41
We need to talk. I'm concerned about you. Let's thank
1:25:47
a few people while we're still ahead of the game,
1:25:49
shall we? We did get some boosts while we were
1:25:52
chatting away here. We got 1701 Sats from Lyceum. We
1:25:56
know that's Martin Lindus Koch. He says, go podcasting. Pigs
1:26:02
in space. Remember that? Is that from the Muppets? Muppets,
1:26:07
yeah, the Muppet show. Yeah, I remember that. Drebscott comes
1:26:10
in with 1234. Thank you. Go podcasting. And Stephen B.
1:26:15
comes in with a row of sticks, 11:11. He says,
1:26:17
"We already solved the metrics of the podcaster and the
1:26:19
dev getting paid. Three years ago, we called it the
1:26:22
value split in the TLV record." I know. I know.
1:26:26
I know. I know. I know. I know. I know.
1:26:30
I don't know what to say. You're right. I mean,
1:26:32
obviously. Yes. Yeah. Is that your Tucker Carlson impression? Yes.
1:26:40
*laughs* You do that really good. That's the only one
1:26:44
I got. That caught me off guard. Yes. Oh, wait,
1:26:50
we got another. We got Eric PP with 3333. Thank
1:26:53
you, Eric PP, coming in from CurioCaster. Beautiful. All right,
1:26:57
why don't you give us our support? Remember, just to
1:27:00
remind everybody, because now I got to start reminding everybody.
1:27:04
This is value for value. The machines keep running because
1:27:08
you keep supporting the podcast index. It's not for Adam
1:27:12
and Dave. It is for the podcast index. And so
1:27:15
your support is needed and highly appreciated. Anonymous, $33.33. Thank
1:27:21
you, Anonymous. Oh, nice. Thank you. We got Salty Sessions.
1:27:28
$10. Thank you, Salty. Appreciate it. Let's see, we got...
1:27:33
Oh, looky here. What? Tom Rossi, Buzzsprout, $1,000. Thank you,
1:27:43
Tom. Is it from Tom personally or from Buzzsprout? It's
1:27:46
got his name on it, but I'm going to assume
1:27:48
it's Buzzsprout. Either way. Either way. Thank you. That keeps
1:27:53
it alive, man. Thank you. Appreciate you. Maybe Tom's independently
1:27:56
wealthy and he just supports us on the side as
1:27:58
a hoot. Let me tip those guys. Michael Goggin, five
1:28:06
bucks. Appreciate you, Michael. Jorge Hernandez, five bucks. Thank you.
1:28:11
Cone Gloss Bar, $5. James Sullivan, $10. Christopher Reamer, $10.
1:28:19
Who we got here? Let me make sure it's not
1:28:21
anonymous. I got editing to do. I got editing to
1:28:26
do, man. Sorry, man. I gave you work on the
1:28:29
worst week. All right. Jesse Anderson. 15 bucks. Thank you,
1:28:33
Jesse. He says, message says, keep on going on. Keep
1:28:36
on keeping on. That's what he says. Randall Black, five
1:28:40
bucks. Thank you, Randall. Appreciate that. That's our PayPal. And
1:28:42
we got a boost from... All right. What did he
1:28:51
say that he sent us? Let me make sure that
1:28:53
I see. Oh, he says, do we have to do
1:28:55
a make good for his previous one? Maybe, but I
1:29:00
never found it. And I didn't find this one either.
1:29:02
I had to get... The splits looked like they worked.
1:29:06
Yeah. But then they didn't. It says it's going into
1:29:11
podcast index. Oh, I see what the problem is. I
1:29:17
hate this crap. What? What's wrong? It went to the
1:29:21
Get Albie thing, which I haven't had time to vibe
1:29:24
code another extractor thing for. Oh, it's down. Oh, okay.
1:29:28
No, I see. No, it's not down. You can't get
1:29:30
it out. You can't get the date out. Yeah. I
1:29:34
got it out, but it came out in such a
1:29:36
horrible format that I can't. Because we get pieces of
1:29:41
stuff for other people, too. Yes. Wait, don't I have
1:29:45
comic strip bloggers? Let me see. I've got his note
1:29:48
for today. for the previous one, let me see. I
1:29:53
read one last week, but he had one the week
1:29:55
before. Yeah, he was all huffy about it. Let me
1:29:59
see.
1:30:01
This is from, no, that's May 15th. That's not it.
1:30:05
Hold on, hold on, I'm getting there, I'm getting there.
1:30:09
I have one from three weeks ago. That might be
1:30:13
it. No, you did that one. I did that one.
1:30:15
Yeah. I'm going to vibe code this thing to get
1:30:21
these stuff out in the right way. Yeah, just connect
1:30:24
to my... It's actually pulling it out of the sequel.
1:30:26
light database yeah just to connect to my helix thing
1:30:29
over here no problem He says it's $15.55 worth of
1:30:35
BTC. And he says... Howdy, Dave and Adam. Rest in
1:30:40
peace, John C. Dvorak, a guy who once blocked me
1:30:44
for a couple years on Twitter just because I suggested
1:30:46
via DMs that he can get Polish citizenship, as his
1:30:49
mother had Polish passports, so he was eligible. Why? full
1:30:54
EU citizenship for him and his descendants. Anyways. Please subscribe
1:30:58
to HTTP. https colon slash slash youtube.com slash at symbol
1:31:05
AI-ArchWizard. This weekend, I resume video posting. Yo, CSB, OzempicTaker.
1:31:13
Oh, Sam Pigtaker. He's on the fat shot. He's on...
1:31:19
Oh, boy. I didn't know CSB was a Husky. I've
1:31:24
met him. I don't recall him as a Husky. I
1:31:26
don't remember. Oh, it's been a while, though. He may
1:31:29
have decided to boost those calories up. Just got a
1:31:33
6048 sat boost through podcast guru from Brian. Says, love
1:31:39
you guys. RIP, JCD. Brian in Fresno. Thank you very
1:31:44
much. Brian and Fresno. I don't remember Brian and Fresno.
1:31:47
Is he new? It's not Brian of London. No. No?
1:31:51
Maybe... Brian of Fresno. I don't know Brian of Fresno.
1:31:55
Interesting. Brian with an I or a Y? Defragmented. No,
1:32:00
with a Y. He's been defragmented, so he's legit. Yeah,
1:32:05
he's fully baked. Yeah. Anything else? Was that it? That's
1:32:12
it. I just. People. If you're so inclined, please run
1:32:18
a Podping node because that would help the network. All
1:32:21
right. And where do we go to run a Podping?
1:32:23
Oh, you just go to podping.cloud. You hit the info
1:32:26
here and then you can get to the GitHub and
1:32:28
then someone like me will go, hey, Claude, figure this
1:32:31
out. That's probably. Or you could just go to the
1:32:38
podcast index GitHub org and find the watcher and run
1:32:42
it. And say, Claude, figure this out because there's a
1:32:45
cargo and I got to do a crate. No, you
1:32:47
know what? You can just Docker. You can just run
1:32:49
it on Docker. Oh, yeah, Docker. And it says, you
1:32:51
don't have Docker. You don't have a Docker on your
1:32:53
machine. Are you sure you want to run that? You
1:32:55
only got 16 gigs and I'm running. So I'm going
1:32:58
to escape you. And you say, stop telling me to
1:33:01
escape you. Is that what you said? I'm getting, I'm
1:33:05
out of here. I'm going to escape you. I'm going
1:33:07
to leave and go to a different network where I'm
1:33:10
appreciated. All right, everybody, that'll do it for the board
1:33:13
meeting for Podcasting 2.0. Please come back next week, Friday,
1:33:17
same time, same place. Adios, everybody. Go podcasting! Amen.