Latest / KDAV Radio / Podcasting 2.0 - Episode 274: "Vibe Cloning"

Transcript

Podcasting 2.0 - Episode 274: "Vibe Cloning" · KDAV Radio

• 0:01
  Podcasting 2.0 for October 9th, 2026, episode 274, Vibe Cloning.
• 0:06
  Vibe Cloning. Hello everybody, welcome to Podcasting 2.0. It is
• 0:13
  time once again, you know what it's time for, we're
• 0:16
  going to talk about all things podcasting. We are the
• 0:18
  place where the board meets, the board of directors to
• 0:22
  determine what's going to happen to podcasting or something like
• 0:25
  that. However, we are always the board meeting where our
• 0:29
  total liability to you is zero. I'm Adam Curry here
• 0:32
  in the heart of the Texas Hill Country and in
• 0:34
  Alabama. The man who wrote a Terms of Service you
• 0:37
  can read in three minutes. Say hello to my friend
• 0:40
  on the other end, the one, the only, Mr. James.
• 0:45
  Dave Jones! correction i did not write anything well hey
• 0:53
  i didn't write anything either a robot wrote this and
• 0:57
  then i've added a few lines that was fantastic i'm
• 1:01
  like hey dave we should rewrite our terms of service
• 1:04
  okay let me let me give it a shot uh
• 1:06
  Three minutes later. Yeah, yeah. What do you think? Yeah,
• 1:11
  this looks pretty good. Yeah. I'm like, I was like,
• 1:14
  rewrite our terms of service. Uh. Here's the main five
• 1:18
  points I want to hit. Okay, that looks good. You
• 1:21
  good with that, Adam? Yeah. Yeah, I'm good. Looks good
• 1:23
  to me. But that was the best part. Our total
• 1:26
  liability to you is zero dollars. That was hilarious. That
• 1:30
  is completely the robot. I knew where it came from.
• 1:35
  I'm like, oh, man, this is just too funny. That's
• 1:37
  just too good. We owe you nothing. We owe you
• 1:39
  nothing. And you will be happy. You will be very
• 1:41
  happy. Yes. You will be owed nothing and be happy.
• 1:45
  So you are on my own personal guest box system.
• 1:51
  Yeah, what have we done here? You hit me with
• 1:54
  a new link this morning. You're like, we're not clean
• 1:56
  feed today. We're on this. Yes. You vibe-coded something? Yeah,
• 2:00
  well, in fact, I should probably do a little screenshot
• 2:04
  for the boardroom. They always love stuff like that. It's
• 2:06
  not quite as impressive as I would like it because
• 2:09
  there's no clips lined up or anything, but at least
• 2:11
  you'll get an idea. Hold on. on print screen yes
• 2:14
  okay print screen here we go is print screen the
• 2:18
  omarchie way to do a thing yes yes that is
• 2:21
  that's a full desktop screenshot yes um let me see
• 2:25
  i'm gonna do two screenshots this is one with your
• 2:31
  processing on upload And then I'll do one with... The
• 2:37
  cart. There we go. Print screen. Oops. Yeah, I did
• 2:41
  it. Yeah, this is the Currycaster 2. We're on a
• 2:46
  2.0 revision. Yeah, we are. Actually, 2.3.7. Oh, you're doing
• 2:54
  minor revisions. Oh, yeah. Yeah. So it now has built
• 2:59
  in live streamer, built in recorder with editing, the editing
• 3:05
  the way I like it with keyboard shortcuts and all
• 3:08
  that stuff. And has a clean feed replacement, which is
• 3:13
  built entirely on the same codec that clean feed uses.
• 3:19
  Which is what? It's some open source. I forget what
• 3:23
  it is. Actually, I can ask the robot. Ogvorbis? No,
• 3:28
  it might be. What's the codec we are using? I
• 3:35
  forget what it is. Hold on. It'll tell me. Um...
• 3:41
  Here it comes. No, no, no. The... For the guest
• 3:46
  room. Hey, I never saw a live come through on
• 3:51
  the Mastodon. Did you ping it? I'm pretty sure I
• 3:54
  did. Let me double check. Uh... Let me see. Episode.
• 4:00
  Yeah. Hmm, that's interesting. The live bot didn't retweet it.
• 4:06
  I can hit it again just to make sure. I
• 4:09
  wonder if my – I didn't get a notice from
• 4:11
  Cast-O-Matic either. Let me hit it again. You never know.
• 4:15
  It's amazing. Hit it twice. Just go ahead and hit
• 4:16
  it twice. Yeah, I hit it. Okay. So the codec
• 4:21
  is the – Yeah, Opus. Opus. Opus. Is that a
• 4:27
  Google thing? I don't know where it came from. Why
• 4:30
  is everything called Opus? Including some classical music piece. It's
• 4:36
  all Opus. Who needs that nonsense? Yeah. Come on. Opera
• 4:39
  Opus. And a bad movie with Richard Dreyfuss from the
• 4:43
  90s. So really all that CleanFeed does for you is,
• 4:47
  you know, they have a, you got to have a
• 4:49
  switching server, like a relay type thing. That's at currycaster.com.
• 4:53
  So I'll give you that. Like, what do you think
• 4:54
  of the web interface? Isn't that cool? Oh, it's nice.
• 4:57
  I've got, I've got some, I've got input. Yes. And
• 5:02
  a big mute button. You got that. Yes. Yeah, that's
• 5:06
  a bonus. I like the big mute button. I would
• 5:10
  say that... The level meter... bar that gives you instructions?
• 5:21
  Yeah, it's a little hot on the draw. I mean,
• 5:25
  like it's a little quick. So, I mean, as soon
• 5:28
  as my level drops, it's like, I mean, this thing
• 5:30
  is losing its mind right now. It's like level is
• 5:32
  good, level is good, a touch high, back off slightly,
• 5:34
  level is good, level is good. Hey, you're not talking.
• 5:36
  Too loud. That'll be short. We're not hearing you. You're
• 5:39
  not talking. I mean, it is within a millisecond that
• 5:42
  thing is all over the map. So it's a little
• 5:44
  too twitchy? Would you rather not? see that at all?
• 5:47
  Maybe just make a hide button for that. I think
• 5:49
  I would rather I like it, but I think it
• 5:51
  needs to be I think it needs to have like
• 5:54
  a um Just slower, just like a little relax. Don't
• 5:59
  react so fast. Yeah, a little more laid back. Laid
• 6:02
  back, laid back. I got you. It needs to take
• 6:04
  the Donna Summer a break. Yeah. So I got this
• 6:07
  thing. You can do just one mic. You don't need
• 6:10
  a mixer. Or you connect it to your Rodecaster. It
• 6:14
  saves the state of your... Clip directory and your show
• 6:18
  note, your prep and your your cart deck and the
• 6:22
  room and everything with a save. And, you know, so
• 6:26
  you can switch shows really quickly. I just went to
• 6:30
  town on this, man. You have like a command center
• 6:36
  is what this is. Yeah. Do you think it's too
• 6:38
  complicated? This is a command center. Well, yeah, this is
• 6:41
  what I need. You know, I got all the players.
• 6:44
  And all the mini map as tool tips for everything.
• 6:47
  I can mark things, you know, like, oh, I want
• 6:49
  to mark that in the show. When you play a
• 6:51
  clip, it automatically marks that as that clip was played.
• 6:55
  So it's easy to go back. You can export that
• 6:57
  all as chapters. I mean, the whole thing is like
• 6:59
  a podcaster's wet dream. That's it. Or maybe it's just
• 7:07
  mine. I don't know. It's possible people will go, huh?
• 7:12
  Yeah, I think most podcasters would look at this and
• 7:14
  be like, oh, what is that? Well, but it has
• 7:17
  all the things you need. You don't need to have
• 7:19
  a separate recording software and editing software. You don't need
• 7:22
  to have a separate streamer. You know, you got your
• 7:25
  cart deck. You don't need to have clean feed, no
• 7:28
  subscription. Although, of course, you'd have to have the switching
• 7:32
  station, which I guess only I have right now. Can
• 7:36
  we talk about this? I want to go straight into
• 7:39
  this exact topic. Okay. Okay. Um... So as... How do
• 7:52
  I start this game? There is a big... Well, you
• 8:01
  called me off guard last week, and you sent me
• 8:04
  into a mental tailspin over the course of this week.
• 8:07
  I felt kind of bad about it, but also I
• 8:10
  was a bit mad at myself, if I can just
• 8:12
  step back for one second, because Sam Sethi interviewed me
• 8:16
  this morning for Pod News Weekly Review. Power, power. For
• 8:22
  next week. I guess. And now I'm like, oh, man,
• 8:27
  I'm kicking myself because instead of the conversation starter I
• 8:32
  wanted, which was discovery, personal discovery of podcasts, it devolved
• 8:41
  into... advertising is over. And I'm like, ah, that's exactly
• 8:44
  not what I wanted. It's my own fault because I,
• 8:47
  I suggested it. Um, uh, and you know, and luckily,
• 8:52
  cause I've post, I posted on X and a couple
• 8:54
  other people are reposting. There's a couple thousand people using
• 8:58
  the podcast index robot.com. And they're not, I mean, I
• 9:01
  don't know what they're doing on their end, obviously, but
• 9:04
  I can see kind of what they're searching for, I
• 9:08
  guess, and what tool calls are being made. It's just
• 9:13
  I was personally upset that I even went there because
• 9:18
  it was like unnecessary, and it always – put something
• 9:22
  new and cool into a bad light. And that, of
• 9:24
  course, is immediately where Sam Sethi started off. Like, well,
• 9:27
  advertised killer. And that's what they talked about on Power.
• 9:33
  And then completely disconnected from that was, and then Adam
• 9:38
  made an MCP server. I'm like, it's the same thing.
• 9:41
  So it's complicated. People don't really grasp it yet until
• 9:46
  you try it, until you throw it into your robot,
• 9:50
  and then you can kind of see what it does.
• 9:52
  So, yeah, I sent you into a tailspin, but that
• 9:56
  was your first reaction to it. was because I told
• 10:01
  you about it, showed it to you, and they were
• 10:03
  like, advertising is over. I'm like, wow, what is wrong
• 10:07
  with us? That that's all that we think about. I
• 10:10
  mean, discovery is here instead of advertising is over. That's
• 10:15
  all that I could think of. And I just found
• 10:17
  it interesting that that's where most people – In the
• 10:21
  group. went to is oh well this is going to
• 10:24
  kill advertising oh no it won't yes it will oh
• 10:27
  it's great no it's not look at the discovery mechanism
• 10:32
  So instead of – I hear what you're saying, but
• 10:36
  instead of thinking – like instead of viewing it as
• 10:40
  – there was the wrong takeaway I think that we
• 10:44
  can step back and analyze and just ask why that
• 10:48
  was the first reaction ooh okay alright alright Mr. Freud
• 10:52
  I'm here laughing I think it's because... Um, The discovery...
• 11:03
  Discovery never really was broken. You've been saying this for
• 11:07
  many years. The discovery was never broken. Discovery of podcasts
• 11:15
  happens through word of mouth. Right. And you have said
• 11:18
  for many, many years that there is not a discovery
• 11:23
  problem. Podcasts will get discovered organically on their own, and
• 11:27
  that's in fact the only way they can be discovered.
• 11:30
  And can we improve it? Yes. But I think that
• 11:37
  improvement is sort of similar to saying, well, I –
• 11:46
  I just sit on the couch and eat ice cream
• 11:48
  and potato chips all day, but I'm just going to
• 11:51
  start taking magnesium because that's going to make me more
• 11:53
  healthy. Like – It's going to make like a fraction
• 11:58
  of 1% more difference to the health of your life.
• 12:02
  This other thing is going to be the bigger deal,
• 12:05
  which is discovery is already a thing. Discovery through the
• 12:10
  MCP server and through pod roles and all that. Yes,
• 12:15
  it is. These are all improvements. But there are improvements
• 12:19
  on a thing that may not have already been broken
• 12:22
  to begin with. I agree. The difference is that people
• 12:26
  increasingly have an assistant, a robot that is doing things
• 12:32
  for them in their life all day long, suggesting –
• 12:38
  Things to do, things to schedule, things to respond to,
• 12:41
  et cetera, et cetera. True assistant. So to have that
• 12:46
  thing recommend things to you, I personally have found already
• 12:52
  incredibly helpful. Because it's finding things that I didn't even
• 12:57
  ask for. It's saying, hey, you know, you're kind of
• 13:02
  interested in flying cars. Well, here's a podcast you might
• 13:06
  like. I did not tell it, go look for podcasts
• 13:09
  with flying cars. Mm hmm. So that's. Yeah. Yeah, I
• 13:15
  think it's the same thing. It's someone else telling me
• 13:19
  I'm not I'm not bumping into a YouTube video that's
• 13:22
  in the right hand sidebar. Right. And so I think
• 13:25
  you've empowered an agent to do essentially word of mouth.
• 13:32
  Precisely. So, and that is, like I said, that's legit.
• 13:36
  But if the question is, why was the other thing
• 13:40
  the bigger topic, I think it's because podcast advertising actually
• 13:45
  is broken. Like fundamentally, and we all know it. And
• 13:49
  it's broken in some – it's broken in not just
• 13:52
  one way. It's broken in many ways. And it's not
• 13:55
  just the advertising itself that's broken. Everything up and down
• 13:58
  the chain is broken. Like it's just – it's a
• 14:01
  completely broken from end to end function. Nobody even knows
• 14:07
  what to do with it. The listeners don't know what
• 14:10
  to do with it. The advertisers don't know what to
• 14:12
  do with it. Everybody in between. So I think it
• 14:15
  makes sense that that was the... That that was sort
• 14:19
  of like the gut reaction to, Because you're taking some,
• 14:23
  it's a big change in a thing that's already really
• 14:26
  effed up. It's pretty effed up, man. I mean, look
• 14:30
  at James' lead headline this morning about Adload. Yeah. You
• 14:37
  know, it's it's it's off. It's off the chart. And
• 14:41
  so I did what I was going to talk about
• 14:45
  last week. When you rug pulled me, I'm glad you
• 14:49
  did because this has given me so much to think
• 14:53
  about. Can I just mention one? Can I give you
• 14:57
  one more data point before you launch into that? Okay.
• 15:00
  Would love you to. Yeah, please do. So I don't
• 15:01
  remember the exact wording, but on the show I said,
• 15:05
  give me the most recent episode of Better Offline at
• 15:07
  Zitron and remove the ads at the front. And it
• 15:10
  took, I don't know, what, seven, eight minutes. And then
• 15:13
  all of a sudden, oh, there it is. We played.
• 15:15
  It was perfect. I did not ask. my robot to
• 15:19
  follow it. I did not ask it to do that
• 15:21
  from now on. I didn't do any of that. In
• 15:24
  fact, I just moved on and forgot about it because
• 15:28
  I subscribe to Zitron sometime. I don't really remember to
• 15:33
  listen to a lot of things. And this morning... Right
• 15:39
  after a flying car review... It says, new better offline
• 15:47
  episode, colon, monologue, colon, Australia's dodgiest AI IPO is a
• 15:52
  pale horse. Front ads cut, end ads remain. And right
• 15:55
  there, a player, exactly what it says on the tin.
• 15:59
  I did not ask for it, but it did it
• 16:00
  itself. Just, I guess, thinking that's what Adam wants. Thank
• 16:07
  you for explaining that because this – is exactly what
• 16:12
  i know is going to happen that's what i've been
• 16:14
  trying to say this week and you're right because i
• 16:17
  saw that i'm like dave's right so i mentioned last
• 16:21
  week and i want to talk about this later i
• 16:24
  want to talk about so the party time the aggregator
• 16:28
  that runs on on uh on all of our aggregator
• 16:31
  nodes I've completely replaced it as brand new. And it's
• 16:37
  rewritten from Node into Rust. and it's running on all
• 16:43
  the aggregators right now. It's superior. It used to take
• 16:48
  you a month to do something like that. I'm just
• 16:50
  saying. I'm telling you, this would have taken me months
• 16:53
  of work. Months of, not just months of piddling here
• 16:57
  and there, months of just, that's nothing I do. That's
• 17:00
  the only thing I do. So... It's completely rewritten in
• 17:04
  Rust. Every single metric is better. It's 70% less memory
• 17:08
  usage. It's 50% faster. And that's not – that's just
• 17:13
  scratching the surface. That's every – I want to talk
• 17:17
  about the full recipe of this in it, but –
• 17:21
  I mentioned I've given up. AI is just – it's
• 17:26
  better at coding than I am. Yeah, you just said
• 17:28
  that. And this was not true three months ago. No.
• 17:30
  No. Three, four months ago, I was better than LLMs
• 17:35
  at coding. I no longer am. You were a frontier
• 17:39
  model, in fact. I was. I was. Yeah. And –
• 17:47
  That we have to recognize... this. And when I say
• 17:53
  that it's better than me at coding, I'm going to
• 17:55
  say that I want to define those parameters. I'm saying
• 18:00
  it's better at structuring a code building process. The code
• 18:05
  itself, line by line, may or may not be better
• 18:10
  than what I would have chosen to write. If you
• 18:14
  look at it, just if you break it down and
• 18:16
  say, okay, line 53, would you have written that that
• 18:18
  way? Maybe, maybe not. Every choice line by line is
• 18:24
  not going to necessarily or function by function may be
• 18:28
  better and may not be. But if what you end
• 18:31
  up with in something like porting from one language or
• 18:36
  runtime to another, if what you end up with is
• 18:39
  byte identical output. With a 50% speed improvement and a
• 18:44
  70% memory reduction, then what I'm saying is the line
• 18:49
  by line is irrelevant. It really doesn't even matter anymore.
• 18:54
  It's better. It just is. And this thing built a
• 19:01
  process roadmap that I simply could not have built. And
• 19:07
  I'll talk about it, but it's like this whole thing.
• 19:11
  So I'm combining this experience. with what you hit me
• 19:18
  with last week. And then I start seeing other... Patterns.
• 19:26
  Patterns. Patterns. Patterns. Patterns. Yes. Patterns. So it looks like
• 19:33
  the new trend – With LLM development because in the
• 19:38
  last three to four months, it has become incredibly good.
• 19:42
  I think that Mo guy had a video in the
• 19:46
  last week or so that just said they fixed AI.
• 19:50
  And his sentiment is it's just it's just it's excellent
• 19:54
  at coding. And so the new trend is no longer
• 20:00
  doing things like React Native or Flutter. The new trend
• 20:08
  is. native apps on every single platform and you build
• 20:12
  one reference app, let's just say you choose iOS. So
• 20:16
  you're going to build your iPhone app or a Mac
• 20:19
  app as your reference platform. And then you just say,
• 20:22
  go make it great for Android. Yeah. And you say,
• 20:25
  go match, go port this to Android, Windows, Linux, Android.
• 20:31
  You know, every other platform in their native in their
• 20:35
  native SDK in their native platforms, native controls and make
• 20:41
  sure that it is bite for bite identical in its
• 20:44
  behavior. And it will just do it. And so like
• 20:49
  that, that's because it will build it. And the reason
• 20:53
  it can do that It's because it will build before
• 20:56
  it even begins the porting process. And this is what
• 20:59
  happened with party time. Before it even begins, it will
• 21:03
  create thousands of integration tests. And you just simply can't
• 21:12
  do that. You do not have the bandwidth. as a
• 21:16
  single programmer or even as a small team to build
• 21:22
  Let's say 7,000 integration and tests in a 2,000 unit
• 21:29
  tests. in order to automate fidelity with a reference app.
• 21:35
  Like you just can't do it. And so – and
• 21:38
  then a thing came up in the last couple days.
• 21:42
  vibe cloning vibe cloning there you go yeah vibe cloning
• 21:48
  wow show have you seen this stuff no oh yeah
• 21:51
  the adobe stuff amazing yeah yeah oh no so if
• 21:54
  you haven't heard about this it's pdf uh like pdf
• 21:59
  craft what's the name of this thing pdf create craft
• 22:04
  It's Artcraft. Artcraft.com. Yes. Let me copy this into the
• 22:10
  board. That was... I mean, I saw a kid, like
• 22:14
  a YouTuber photography kid, and he uses RAW, and he
• 22:19
  was showing how everything works. Some of it's even better
• 22:22
  in the Vibe clone. Runs on Linux. Hello? Hello? Amazing.
• 22:28
  Yeah, what this guy has done... Is it a guy
• 22:33
  or a team? It seems like a team. No, it
• 22:36
  seems to be one dude. One dude. It's an open
• 22:38
  source. I know. It's all open source. I know. This
• 22:41
  is the best part. So he just cloned every single
• 22:45
  Adobe app from Premiere down to Acrobat in Rust. Yeah.
• 22:54
  And it's like a thousand times faster. No, no, no
• 22:58
  scamware or or adware. And yeah, every it's everything is
• 23:04
  it's just perfect. And he did it all clean room
• 23:07
  to avoid copyright infringement and things like that. So it's
• 23:11
  all he's never he just said. Yeah. Clone this. Look
• 23:14
  at this app and build it. Well, I think Adobe,
• 23:18
  I had quite a conversation with a couple of friends
• 23:21
  of mine about this. Adobe will try and sue whatever
• 23:25
  they're going to do over proprietary layout and other things.
• 23:30
  They may have patented trademarks. You don't know. But that
• 23:33
  is completely irrelevant. It's open source. Everybody has a copy.
• 23:38
  You're done. And the money's gone. And the money's gone.
• 23:42
  And the money's gone. Yeah, right. And so then it
• 23:44
  comes out today or late yesterday that this same guy
• 23:48
  has also done the same thing for Excel, Word, and
• 23:51
  PowerPoint. Right. Did I not say short Microsoft software is
• 23:56
  over? I'm pretty sure I said that. Yeah. And so
• 24:01
  if you – and this is the thing. you can
• 24:08
  now As a company, You don't have to use this
• 24:13
  guy's thing. That's just an example. You just vibe clone
• 24:16
  his vibe clone. You do it yourself. Yep. You just
• 24:19
  do it yourself. And suddenly you say, Adobe Acrobat for
• 24:24
  like an enterprise – It's expensive. It's very expensive. It's
• 24:31
  as expensive as Microsoft Office. So 99 bucks a seat
• 24:35
  per year or something like that? Yeah, there's different volume
• 24:38
  licensing levels, but it's all subscription now. And so, you
• 24:42
  know, if you had like, I don't know, if you
• 24:44
  had like 300 seats of it, you're looking at. $200,000.
• 24:49
  Well, I mean, I was paying... Combined licenses for M-Air
• 24:57
  and... Hindenburg Pro, this is all when I was still
• 25:03
  on Windows, and CleanFeed, probably $300 a year. You know,
• 25:09
  not not bank breaking, but none of it worked exactly
• 25:13
  the way I wanted it to. And now I just
• 25:16
  built something in four days. That's not that's not entirely
• 25:20
  fair because I had I had a previous version. So,
• 25:23
  you know, I already gotten all the basics, mapping MIDI
• 25:27
  and all that stuff. But I just went to town.
• 25:31
  I just said, okay, I want this. And it came
• 25:32
  up with its own layout and everything. It said, no,
• 25:35
  that sucks. Change this. But there's a bigger thing going
• 25:41
  on here. It's not just, okay, let me step back.
• 25:46
  For decades, we have been using applications that were built
• 25:52
  for our computers and very restrictive. See all smartphones and
• 25:58
  app stores, very, very restrictive applications. I mean, this is
• 26:02
  why Microsoft Office is such a mess with just customize
• 26:06
  your ribbon. OK, you know, like you're working, you're the
• 26:11
  software is training you to work with it. And now
• 26:15
  we can just create our own software. That does exactly
• 26:20
  what we want it to do. It may not look
• 26:21
  great. It may be funky, but it's how I want
• 26:24
  it to work. And the only thing left of value
• 26:30
  is, is hardware. Transport. mass storage. And and data sources.
• 26:43
  If you have a data source that people want access
• 26:46
  to, you know, then it's up to you whether you're
• 26:49
  going to let them have access to that and how
• 26:51
  they can access it. And if they have to pay
• 26:53
  for it, because that's the next that's the next step
• 26:56
  that's coming because that anything this government regulated. Yeah. Come
• 27:01
  again. You listed out the only things that have value
• 27:05
  anymore in that chain, and I'm saying – Oh, yes.
• 27:09
  Okay. Things that are government-regulated behind it, and I think
• 27:14
  that's why – The AI frontier labs are trying to
• 27:17
  get themselves regulated because they want protection. Not just –
• 27:22
  in fact, so I had this epiphany myself. I'm like,
• 27:26
  no wonder people are propagating the AI is going to
• 27:29
  kill us all. It's going to kill them. It's going
• 27:32
  to kill their businesses. All their businesses are useless. It
• 27:37
  literally became useless overnight. And it's going to take a
• 27:41
  little bit because not everybody can concept things. And vibe
• 27:46
  coding. But I got a note from a – did
• 27:51
  I tell you about this? The homeschooling lady? Shelly? I
• 27:54
  don't think so. She said – she's been hearing me
• 27:57
  bang the drum about AI and Muse in particular. I've
• 28:00
  been saying, hey, Muse is a good place to start.
• 28:02
  We get to do this. And she says, hey, I'm
• 28:04
  a homeschooling mom. Ben, I've been kind of worried about
• 28:07
  this, but – I trust you. So I installed it.
• 28:10
  And right off the bat, three hours a week, I
• 28:14
  got back just immediately by letting Muse do the syllabus
• 28:18
  and whatever the homework is for these kids. So I
• 28:22
  got that back right away. And this is before the
• 28:24
  vibe cloning news came out. She said, I also have
• 28:27
  this business on the side that's kind of been dormant
• 28:30
  for about a year. She makes props and puppets for,
• 28:32
  you know, homeschool theaters or something like that. And she...
• 28:39
  She says really difficult to use Adobe. She just said
• 28:42
  Adobe. I don't know if she's using Photoshop or Illustrator
• 28:45
  or whatever she's using. It's been really debilitating to use
• 28:49
  Adobe to do my patterns and stuff like that. So
• 28:52
  I threw it into into Muse. It's about 20 minutes.
• 28:56
  It gave me a perfect pattern that I could sell
• 28:59
  right away. Which is interesting by itself. But it also
• 29:02
  built a little program for me that does exactly the
• 29:06
  things that Adobe was doing for me, but better. So
• 29:11
  when I hear this as a homeschooling mom, I'm like,
• 29:14
  OK, this changes everything. So people who are afraid of
• 29:20
  AI, particularly in Silicon Valley, it's because it's going to
• 29:26
  kill their business. All of it. All of it, except
• 29:30
  for hardware. But they come out with, ah, here's the
• 29:34
  new Surface. Notice they're not doing Dell, no Surface. Microsoft
• 29:38
  Surface, which we know is a pretty crappy platform to
• 29:40
  start with. But it now runs a local LLM and
• 29:45
  128 gigs of unified memory. So they see the writing
• 29:49
  on the wall. This is why Zuckerberg holds up that
• 29:52
  little fake and gay Tamagotchi. It's like, look, we have
• 29:55
  hardware for Muse. He knows. He understands.
• 30:00
  It's going to happen. If you don't have a hardware
• 30:02
  device, you're out of business. Yeah, well... Even social media.
• 30:12
  example. Well, newsletters is a good one. So Tina writes
• 30:17
  the church newsletter that, and she really, she's like, I
• 30:20
  hate this. Hey, I hate this robot stuff. I said,
• 30:22
  man, but a couple of her friends are using it.
• 30:24
  She said, no, I installed Muse. And, uh, no, it
• 30:27
  was kind of easy to, uh, It got me the
• 30:29
  newsletter items that usually take me a little while to
• 30:33
  find the promo reel and all this stuff. And I'm
• 30:36
  thinking, so then you go and you send this out
• 30:38
  via email. The next step is obvious. People will subscribe
• 30:42
  to your robot's communication, and it will be a direct
• 30:45
  communication. It may use... SMTP, it may use FTP, it
• 30:51
  may use distributed hash tables. I don't know. But why
• 30:55
  would we have to create something, send it through the
• 30:59
  horrible channel that is email? No, I want to subscribe
• 31:02
  to your robots newsletter. Boom. I can see it. So
• 31:11
  that That brings me to the ad thing. And I
• 31:17
  think that this – Advertising – has the same problem
• 31:24
  that Silicon Valley has. It hit a plateau in growth,
• 31:31
  and the answer, as we all know now, is they
• 31:34
  started to inshittify their products. Yeah. Absolutely. They started to
• 31:39
  make their products harder and more – they piled more
• 31:43
  and more friction into their products in order to get
• 31:47
  – More dollars per user extracted. And this is exactly
• 31:52
  the same problem that advertising has had over the last
• 31:57
  40 years. It has just slowly ramped up the friction
• 32:01
  of listening or watching. Anything that has ads in it
• 32:06
  has become ridiculous. Virtually unlistenable. We used to call that
• 32:10
  the interruption model versus the attention model. Okay. Well, it
• 32:17
  is absolutely an interruption model these days for sure. And
• 32:22
  because advertising could not resist in shitifying – The process.
• 32:29
  It is also going to be subjected to the same.
• 32:34
  End around. That software is going to be. People are
• 32:39
  going to just. Opt out. Yeah. They're going to opt
• 32:44
  out. Forget all for a minute. Yeah. Hang up all
• 32:50
  the ideas of whether or not it's moral or if
• 32:53
  it's – or any of those. It's completely irrelevant. It
• 32:58
  is irrelevant what the ethics of stripping out ads is.
• 33:03
  Ears. If it can happen, it will. It's going to
• 33:07
  happen. And and so it's going to happen because of
• 33:12
  this. At first, so if you're an app developer, a
• 33:16
  podcast app developer, let's say that you're Pocket Casts. No,
• 33:22
  that's too much. Let's say that you're the Xavier at
• 33:25
  Podcast Addict. At first, it feels a little bit wrong
• 33:30
  to put that button in there that says take all
• 33:33
  the ads, take out the dynamic ads. But then you
• 33:38
  start seeing other people sort of edging it. And you
• 33:42
  yourself as a developer start to get pretty annoyed at
• 33:45
  the ads. And you start to see how – you
• 33:48
  start talking to yourself a little bit from time to
• 33:51
  time about how you get zero benefit. Self? Yeah, self.
• 33:58
  Who's – am I benefiting from these – from not
• 34:00
  removing these ads? Am I in the flow at all?
• 34:04
  Is anybody paying me to keep these ads in the
• 34:06
  podcast, the ones that I don't like? Yeah. And then
• 34:09
  a few smaller apps with less listeners or less users,
• 34:13
  a couple of them start doing it. And you don't
• 34:17
  want to lose market share. So now you have to
• 34:19
  do it. But you have to do it, but you're
• 34:22
  also willing to do it. Yeah, I'm going to have
• 34:25
  to. Phoebe! It's a board meeting. I'm going to have
• 34:33
  to leapfrog you. This is what I was talking about
• 34:37
  with Sam. And it pains me, but why do I
• 34:41
  need podcast addict at all? Well, That is true. I'm
• 34:52
  not sure what I think about that yet because there
• 34:57
  are – Um, There are apps that... that I love
• 35:08
  That already worked so well for my life that I
• 35:11
  don't I don't see the pain of going through trying
• 35:14
  to like recreate it just as some little one little
• 35:17
  feature that's one percent, you know. And it's not in
• 35:22
  the... I don't know yet. I mean, like. Okay. Try
• 35:27
  this. Okay. And by the way, this is for our
• 35:31
  own business, too. I mean, we have an app, you
• 35:34
  know, so these are very valid questions. Okay. My example,
• 35:41
  you know, I really just want a Rachel Maddow app.
• 35:45
  I want Rachel, Rachel's adjacent shows, the people she likes.
• 35:50
  I want it to show Rachel news. You want that
• 35:52
  to be your desktop theme, you know, aren't you? It
• 35:56
  already is. It's Rachel. It's Rachel Maddow. It already is.
• 35:59
  Okay. So... Why wouldn't I just say, make it, make,
• 36:03
  you know, I mean, I don't even have to have
• 36:07
  an app. I have an app. It's called my robot.
• 36:13
  This is my this is my point. Like, just make
• 36:16
  it work for me. Just do just show it to
• 36:18
  me. I'm already seeing it now. Now, currently, I don't.
• 36:22
  I'm a podcast guru guy. I love podcast guru. It's
• 36:26
  interface works well for me. so I told my robot
• 36:34
  when you show me something to play just make it
• 36:37
  a podcast guru link doesn't go straight into the episode
• 36:40
  it goes to the show how long before I'm so
• 36:43
  annoyed by that that I just say if I've cloned
• 36:47
  this So I have an app that does it the
• 36:50
  way I want it to. Well, okay. See, I think
• 36:53
  you described a barrier, though. Once you rebuild the entire
• 36:58
  app, you're now on the hook for maintenance. For yourself.
• 37:02
  For yourself. For yourself. Sure, for yourself, but then now
• 37:06
  you have to do all of that yourself. Right, but
• 37:08
  I'm taking it beyond the app and app stores. I
• 37:13
  just have a slab. A slab? I just have a
• 37:16
  computer slab with a screen, and it's running whatever it's
• 37:21
  running, and it's running an agent. And the agent does
• 37:25
  all this for me. I don't need to deal with
• 37:28
  the App Store. I don't have to think about Apple.
• 37:30
  And I have to think all I need is the
• 37:34
  SMTP, RSS, HTTP, even questionable if I need that, DHT.
• 37:40
  I just need protocols. I just don't think we're there
• 37:45
  yet. No, but it's going very fast. Sure, but it
• 37:50
  may – that's a lot, though. From here to there,
• 37:53
  that's quite a big jump. I'm already there because I
• 37:56
  – But you don't want to do – like, you
• 37:59
  don't want to have to say – There's a usability
• 38:03
  thing that's in this. You don't want to have to
• 38:06
  say, hey, agent, play the latest – prick up where
• 38:09
  I left off from the latest episode of Rachel Maddow.
• 38:12
  Like, you'd rather just swipe over and hit the play
• 38:14
  button. Yeah, but of course. I'm not – I don't
• 38:18
  – your interaction is wrong. I tell the agent what
• 38:23
  functionality I want, and this is exactly the process I
• 38:26
  just went through with Currycaster 2. I said, well, I
• 38:30
  don't like that. Fix that. Okay, here's a new version.
• 38:33
  Yeah, that's not quite. Fix that. So I iterate with
• 38:36
  the robot. I am controlling the interface. I'm not saying,
• 38:40
  hey, robot, get Dave on the show or, hey, robot,
• 38:43
  record. No, I click record. But the interface I wanted,
• 38:47
  no one, no one, no one has ever built because
• 38:50
  it's the one I want. And maybe no one else
• 38:52
  wants it. You also hold a position that most people
• 38:56
  do not. Currently. Which is you have systems thinking. Yes.
• 39:03
  Most people seem like you can do that. My dad
• 39:05
  is not going to do that. Please. He needs an
• 39:08
  app. Stop with the old people. No, it's not the
• 39:12
  old people. That's not fair. I'm telling you. I know
• 39:14
  my dad. It's not old people. Your dad's old. Your
• 39:17
  dad is old. And your mom wears army boots. Come
• 39:19
  on. Yeah, he's old. Okay, let me give you another.
• 39:22
  My daughter can't do that. My wife can't do that.
• 39:24
  These are not old people. Hold on. Hold on. Hints.
• 39:27
  Hints. Hints. The need for systems, storage, transport, simple things.
• 39:35
  I mean, once you give your robot a token for
• 39:40
  Linode, I don't have to know what Linode is. I
• 39:43
  just say... Okay, I need a server for this. All
• 39:46
  right, would you like the $5 a month? Yes. Okay,
• 39:49
  it's done. You've already surpassed anything that my 16-year-old daughter
• 39:55
  even knows what. It's coming. It's coming. Hey, we used
• 39:59
  to type in HTTP colon slash slash. Sure, but that's
• 40:06
  – I think you take this stuff a little bit
• 40:10
  for granted. No, no, no. Because you think in systems.
• 40:16
  It comes naturally to you where other people are. No,
• 40:21
  I'm just years ahead of the curve and someone else
• 40:24
  will make a billion dollars. That's this is what I
• 40:26
  do. Well, that's true. So it's coming in our lifetime
• 40:29
  and it will be your daughter. It is already it
• 40:32
  is already Gen Z. I'm getting emails from people every
• 40:35
  day. Just look at what my kid did. And it
• 40:38
  did all this stuff. And by the way, the hosting
• 40:41
  will soon be built into the Frontier Labs package. You'll
• 40:45
  get hosting, domain names, all that stuff is going to
• 40:47
  be in there. A dollar a month. It's coming. Well,
• 40:55
  I am prototyping a spec document that's going to show
• 41:00
  exactly how to remove all these dynamic ads. And I'm
• 41:06
  going to put it out there, and it's going to
• 41:08
  be – and I'm not going to publish it until
• 41:11
  it is 100% reliable. Where you can just throw any
• 41:16
  podcast at it and it will – it's not going
• 41:18
  to be a piece of software. It's just going to
• 41:21
  be a specification document. Here are the steps that an
• 41:25
  agent will – that a harnessed model will take to
• 41:31
  remove – Every dynamically inserted ad that is not the
• 41:35
  host themselves speaking out of a podcast episode. And you're
• 41:40
  doing this – Because... You can't do it because I'm
• 41:46
  doing it because this is going to happen. And it
• 41:50
  might as well get done right. But no, no, no.
• 41:55
  It's going to happen, but the advertising industry is fantastic
• 41:59
  at ignoring things. So unless somebody knocks on their door,
• 42:05
  throws a bag of dung onto their doorstep. They're never
• 42:09
  going to do anything. They're not going to move. They're
• 42:11
  not going to budge. They're not going to do a
• 42:13
  single thing. And so this is really not about saving
• 42:16
  them. This is about getting the attention of actual individual
• 42:20
  podcasters. To say, this is coming. Make sure that you
• 42:27
  go ahead and move to either a ad-free membership model.
• 42:32
  Give that as an alternative or value for value. That's
• 42:36
  pretty much it. That's what's going to be left. Make
• 42:39
  sure that you give your audience a way to support
• 42:43
  you. A way to support you. A real way. A
• 42:49
  way that they can donate money. They can give them
• 42:52
  a path to pay you money right now. Because if
• 42:56
  your entire livelihood is built on dynamic ad insertion. Yeah,
• 43:02
  you're gone. You're toast. Yeah. You're toast. And rather than...
• 43:08
  Just wait for a bunch of checkboxes to show up.
• 43:11
  I'd rather go ahead and just show exactly how –
• 43:13
  write a document that shows exactly how it's done. You're
• 43:18
  so radical. Yeah. This rebel. You are a rebel. Well,
• 43:26
  that's – but I think that's – so the reason
• 43:30
  I brought up all those vibe clonings and those other
• 43:34
  issues before is – Is because I think this is
• 43:39
  what we are – this is the era that we're
• 43:42
  in. Um, It's... Like each individual. Go ahead. No, go
• 43:51
  ahead. After you. Each individual actor in the podcast advertising
• 43:56
  ecosystem is. Probably feels powerless to change anything on their
• 44:03
  own. Of course. And that was the impetus behind like
• 44:08
  Oxford Road bringing the amp cords and all that kind
• 44:11
  of stuff together because I'm pretty sure because Dan Granger
• 44:14
  and others probably thought, well, we – We just feel
• 44:17
  like we're getting nowhere trying to change things. But the
• 44:21
  fault there was that all the big players got together
• 44:26
  in a room, but they never intended to actually go
• 44:29
  along with any of it. So, you know... And they
• 44:36
  thought they agreed to a bunch of stuff, but then
• 44:39
  all the big players promptly left the room and did
• 44:41
  whatever the hell they wanted to. YouTube left and then
• 44:43
  redefined what a play is again for the 15th time.
• 44:47
  Spotify left the room and created a skip ads button.
• 44:50
  So there is right now. There's what comes next, and
• 44:57
  then there's what comes after.
• 45:00
  We don't know what comes after, but we do know
• 45:03
  what comes next. And what comes next is all the
• 45:08
  ads are going to disappear. This goes right along with
• 45:12
  massive disruption in the software industry. Just the advertising industry
• 45:17
  is going to be disrupted in the same way that
• 45:21
  people are vibe cloning Adobe Acrobat. Your ads are going
• 45:25
  to disappear. What is happening here is Dave Weiner actually
• 45:31
  talked about this a lot back when we still talked.
• 45:35
  He would say Silicon Valley, bunch of a-holes. They always
• 45:39
  force us to fit within whatever they come up with,
• 45:43
  whatever their vision is. And it's the same with. Everything.
• 45:51
  We are about to be given so much power as
• 45:54
  users, consumers, people, humans. That... So the first thing that
• 46:04
  shows up is, well, there's dynamic pricing at McDonald's. Okay,
• 46:10
  there is. It's already here. There's dynamic pricing at McDonald's.
• 46:15
  That's horrible. And they're going to do this at an
• 46:17
  individual level eventually. But it's based upon location, time of
• 46:23
  day, like Surge Big Mac. It's busy. People will pay
• 46:27
  more. But the way I see this going is I
• 46:31
  am going to tell my robot I'm going to hit
• 46:34
  McDonald's negotiate a price for me. And it will. And
• 46:38
  McDonald's will say, well, it's five bucks for a Big
• 46:41
  Mac meal. And I'll say, I'm going to pay you
• 46:43
  $450. No, we won't do that. Yeah, but I'm also
• 46:46
  ordering three Happy Meals. Oh, okay. Otherwise, I'm going to
• 46:48
  go to Burger King because here's the price I got
• 46:50
  from them. This is coming. It's already being done with
• 46:54
  Uber and Lyft where you tell your bot, I need
• 46:57
  a ride from here to home. Find the best price.
• 47:01
  And so right now it's just comparing. But the next
• 47:04
  step is the bot is going to negotiate that price.
• 47:06
  It is going to bring prices down. It is going
• 47:11
  to – and the open and free market is going
• 47:14
  to be amazing. It's going to be fantastic. Because it'll
• 47:20
  work on supply and demand. And you will be attractive
• 47:23
  because of your money. And your robot will negotiate everything.
• 47:32
  Airline tickets, theater tickets, concert tickets. Book prices, all of
• 47:37
  this stuff. This is why Amazon immediately went, no, no,
• 47:40
  no, you can't touch us. You can't come to Muse.
• 47:42
  Okay. All right. That's going to change very quickly. Margins
• 47:48
  will get thinner. And, of course, there's a floor for
• 47:50
  everything. But this is the free market. We are going
• 47:53
  to, as human beings, we are going to have so
• 47:56
  much power to, Over this token prices, if these guys
• 48:00
  even survive this apocalypse that is coming, all of this
• 48:04
  is going to be the power of the individual. And
• 48:07
  it's going to be amazing. I see such a and
• 48:11
  we're going to be able to use our creative minds.
• 48:17
  and create things for ourselves, for our friends, all of
• 48:21
  this stuff. I had an interesting experience. Of course, my
• 48:24
  node went down again. I'm like, crap, restart the thing.
• 48:29
  They go in, like, okay, check AlbiHub, make sure it's
• 48:32
  unlocked. And it says... connect your robot to albi hub
• 48:36
  oh okay all right yolo so i can sign me
• 48:40
  up and it connected and it's connected and now my
• 48:45
  robot and i gave it a you know on the
• 48:47
  albi hub side i gave it a budget of 10
• 48:50
  000 sats and now it can now it can uh
• 48:52
  pay for things And I tested it right away. I
• 48:55
  said, okay, I want to deposit some money into my
• 48:58
  wallet. Give me a QR code. Robot shows me a
• 49:01
  QR code. I pay for it with my strike wallet.
• 49:06
  Boom, accepted, done. What's my balance? Oh, now you have
• 49:08
  14,000 sats. It was 4,000 sats in that wallet. So
• 49:12
  now I can easily, easily do value for value payments,
• 49:17
  splits, everything. The whole problem is solved. I just need
• 49:21
  the information. Where do I get the information? From the
• 49:23
  value tag. The whole problem is solved. And where do
• 49:27
  you get the value tags? From RSS. From podcast index.
• 49:30
  From RSS. And where do you get the RSS? From
• 49:31
  MCP server around podcast index or wherever. Exactly. This all
• 49:35
  comes back to the issue. It's not an issue. It's
• 49:41
  joy. Joy has come in the morning. It is here.
• 49:45
  Rejoice and be glad in it. This is what I
• 49:48
  was saying earlier. Make sure you have a way for
• 49:52
  people to pay you. Mm-hmm. Do it right now. Don't
• 49:56
  stop. Don't go. Don't go past. Go do it right
• 50:00
  now. Go make sure you need to ask yourself as
• 50:03
  a podcast, as a podcaster. Is there anything in my
• 50:08
  podcast feed that links back to a way that somebody
• 50:12
  can give me money? And if the answer is no.
• 50:16
  Fix it. Because the only way... To... deal with Vibe
• 50:25
  cloning things away and that kind of thing is to...
• 50:31
  Give is to give the aid, give the agents what
• 50:35
  they want and interface. Give them something to latch on
• 50:40
  to. That was the other thing about these Vibe-cloned Adobe
• 50:45
  apps, the PDF craft and all them. Every single one
• 50:48
  of – they're not only faster and have no crapware
• 50:52
  in them, but they also – every single one of
• 50:54
  them has an MCP interface. So that you can drive
• 50:59
  the app from an agent. The only way, and that's
• 51:04
  something that Excel and Word and Adobe Premiere, they don't
• 51:09
  have any of that. Can I just try something for
• 51:11
  a second? Sure. Okay. Robot, send a Boostergram to Podcasting
• 51:16
  2.0. We're live right now. and make sure that send
• 51:22
  a thousand sats and make sure you split up payments
• 51:27
  between the recipients according to the value tag. I have
• 51:30
  no idea if that's going to work. Interesting. But let's
• 51:33
  just see. Let's see if that works. I haven't heard
• 51:36
  any Instagrams come in, actually, now I think about it.
• 51:40
  So if we hear the harp. Oh, I didn't. I
• 51:43
  should have done a triple seven. Oh, to get the
• 51:47
  jingle. Yeah, to get the harp. It'll give me a
• 51:49
  pew-pew. So it knows. So it has the MCP server.
• 51:55
  It has a wallet. So I just gave it a
• 51:59
  little extra instruction for the first time, and let's see
• 52:01
  if you can do it. Let me just add to
• 52:04
  it. The boostagram should say it worked. Okay. I just
• 52:08
  added that. Anyway, continue. Oh, no. Well, okay. So, I've
• 52:16
  said this phrase in the past, and I think that
• 52:19
  it's more true now maybe than when I first said
• 52:22
  it, was that AI, LLMs is not the beginning of
• 52:25
  something new. It's the end of something old. Mm-hmm. And
• 52:30
  I think that's true. In the sense that. There's so
• 52:36
  much about what had existed before was just broken. And
• 52:41
  did not work well and in shitified. The world is
• 52:45
  in shitified. Everything's in my car is in shitified. Right.
• 52:50
  Piece of crap. I get in the car. Good morning,
• 52:52
  Adam Curry. The other day I was singing a song.
• 52:55
  I was singing along to the radio, which is how
• 52:58
  was the song? Um. for King and Country. I was
• 53:03
  like, I'm looking for a place in this world, a
• 53:07
  place in this world. And my car goes... Finding a
• 53:11
  place in this world. I'm like, no, no. Stop. Don't
• 53:15
  do that. What does it even mean? Stop. Here's a
• 53:18
  place in this world you might want to visit. It's
• 53:21
  5,000 miles away. Oh, my gosh. This is what I
• 53:26
  mean. This is the world. The. The world and the
• 53:30
  platforms and the technology that had built up over decades
• 53:34
  has become so inshittified and so broken and so strung
• 53:41
  together with chewing gum and wire that something had to
• 53:46
  just sweep it away. And this is like a clean
• 53:51
  what is what is happening is a clean slate is
• 53:54
  forming. I find it so exciting. I can't tell you
• 53:57
  how exciting this is to me. I love this. I
• 54:01
  wish I knew a way to make money on it.
• 54:03
  That's the thing. You're right. Value tag, funding tag. That's
• 54:07
  it. That's it. We don't screen here. What are you
• 54:13
  doing? Trying to take a screenshot. Uh-oh. This is my
• 54:18
  job. You're doing screenshots now? I'm doing screenshots. Let's see.
• 54:22
  Okay. All right. Can you put a... You can just
• 54:25
  copy... Yeah, just paste it in. Just paste it into
• 54:27
  the browser. It'll immediately... I don't know what you're using.
• 54:32
  Paste it into your – What are you doing? Oh,
• 54:34
  you want me to paste something? What are you doing,
• 54:36
  Dave? What are you doing? I'm trying to get –
• 54:39
  What are you trying to do? I was thinking, is
• 54:43
  there a way to put an image into the IRC
• 54:46
  chat, into the boardroom? Yeah. Are you on the browser?
• 54:49
  What do you – I mean – No, I'm just
• 54:51
  in – Just put it on Signal. I'll put it
• 54:52
  in. Oh, okay. All right, let's see. So my robot
• 54:56
  is doing something. It says it's sending payments. This would
• 55:00
  be very exciting. Because once it knows how to do
• 55:02
  it, I don't have to train it again. Oh, man,
• 55:06
  it's going to be so badass. Okay, where are you?
• 55:10
  Are you putting it in? What are you sending? A
• 55:12
  screenshot? Yeah, there it is. Oh, got it. Okay. Hold
• 55:17
  on. Download. to download Okay. Where'd it go? Yeah. Hold
• 55:28
  on. How do I do that now? I'm sorry. Uh,
• 55:34
  six, hold on. Files. Open up. There's an easier way
• 55:39
  to do this, no doubt. Download. Probably. Here we go.
• 55:43
  Here we go. Here we go. There it is. Okay.
• 55:46
  Got it. What are we looking at here? Yeah, so
• 55:49
  this is... This is my dashboard for this show. So
• 55:55
  when we do the boardroom, it's live transcribing everything. Showrunner.
• 56:04
  Ooh, showrunner. It's live transcribing everything, and then it's when
• 56:09
  it thinks that I said something that needs to be
• 56:13
  – checked or needs more or needs like some more
• 56:15
  research or something it'll it'll give me a heads up
• 56:19
  hey dave you said this wow this is good yeah
• 56:22
  that's what i'm talking about look at your creative genius
• 56:28
  Booster grams went out. 991 of the 1,000 sats went
• 56:32
  through, split across the value tag just like the feed
• 56:35
  says, each one carrying it worked. One snag, Dave's nine
• 56:38
  sats wouldn't go. His lightning address keeps returning a bad
• 56:41
  invoice. That's on his end, not ours. Everything else landed,
• 56:44
  including the key sends to the two node recipients. holy
• 56:49
  crap let me see if i got it did we
• 56:51
  get is that true did we get it well my
• 56:54
  my node has been offline so so i don't know
• 56:57
  oh i can look on our albi hub um i
• 57:00
  can look in my i think i still have a
• 57:03
  split on my yeah look in our albi hub for
• 57:05
  a second Okay, let me see if I got it.
• 57:08
  I think I might have gotten it on my strike.
• 57:11
  No, not yet. It says it went out. I'm on
• 57:15
  the wrong page. Okay, here we go. Previous. One minute
• 57:20
  ago, it says we got 406 sats from something. Does
• 57:26
  it have a boost? No, four minutes ago. A boost-o-gram
• 57:30
  thing? I didn't get anything here. Wait, I got a
• 57:34
  boost that says it works. Ha ha! Okay. It says
• 57:42
  comment. It worked, but there's, it doesn't show like where
• 57:46
  it came from or what. Well, I just, it doesn't
• 57:49
  know. I mean, I just told it to do this
• 57:51
  thing. Here, let me send it to you. How about
• 57:55
  that? Yeah, here it is. How badass is that? Come
• 58:02
  on. Open that up. Copy. So you can see this.
• 58:08
  I'm going to put in the signal. Wow. Paste. Yeah,
• 58:13
  it doesn't give an app name that I can see.
• 58:15
  Well, I'll tell it to give an app name. Next
• 58:19
  time, make sure the app name, which is required, says
• 58:24
  Adam's Robot. Okay. Adam. Let me go see. Yeah, came
• 58:32
  in. It worked. Comment, it worked. Holy, holy, holy, holy,
• 58:37
  holy, holy. So, I think that This being the end
• 58:45
  of something old. Capitalize the R. Sorry. it also coincides
• 58:54
  with Whenever there's some new – the introduction of a
• 59:03
  new technology or a new social phenomenon or any time
• 59:09
  – we've talked about this before with – With Charles
• 59:11
  Taylor. And that Charles, you always Charles Taylor. Who's Charles?
• 59:18
  He wrote a book called a secular age. Oh yeah.
• 59:21
  Okay. I remember. Yes. I remember. And, uh, One of
• 59:25
  the main points that he makes in there repeatedly is
• 59:29
  that Whenever you have a new social movement, That movement
• 59:35
  always contains the seed for its own demise or its
• 59:39
  own reversal, whatever word you choose to put there. It
• 59:45
  is going to – it carries along the things within
• 59:50
  it. That will eventually result in that thing being surpassed
• 59:57
  or removed.
• 1:00:00
  And so we have to be conscious of that throughout
• 1:00:03
  the process of whatever this thing looks like as AI
• 1:00:07
  marches forward and takes and blows itself through podcasting and
• 1:00:11
  through software and that kind of thing. It is going
• 1:00:14
  to look at times unstoppable and it is going to
• 1:00:18
  be a world-changing event. It will also end up resulting
• 1:00:25
  in itself being replaced with something that we don't understand
• 1:00:30
  yet. Have you ever seen the movie Twister? Yes, with
• 1:00:35
  Bill Paxton. It's coming. It's coming. It's already here. Game
• 1:00:42
  over, man. Game over. Fabulous. So what I'm saying is
• 1:00:53
  I still hate AI. I hate every bit of it.
• 1:00:59
  I wish I wanted to pull a clip because my
• 1:01:02
  experience, your experience with AI, you focus a lot on
• 1:01:06
  the time savings that has brought you. Yes. My experience
• 1:01:11
  with AI has been that I am working more than
• 1:01:15
  I ever have. And as a – in my day
• 1:01:21
  job, I now have – I don't stop. I'm mentally
• 1:01:28
  slammed from the moment I wake up to the moment
• 1:01:30
  I leave the office. I'm very sorry to hear that.
• 1:01:35
  That is not what it's supposed to be. Well, I've
• 1:01:38
  heard people – I should have brought a clip, and
• 1:01:41
  I'll try to bring it for next weekend, but there
• 1:01:42
  was a clip of an accounting podcast I was listening
• 1:01:45
  to recently. Riveting. No, it's just as exciting as it
• 1:01:48
  sounds. But they were talking – they talk a lot
• 1:01:52
  about AI, of course, because it's – Yeah, because what
• 1:01:56
  else are you going to do? Blowing everything – yeah,
• 1:01:57
  blowing everything up. Yeah. And – One of the things
• 1:02:01
  that the guy said, and this guy's a big AI
• 1:02:04
  booster. I mean, he's like singing his praises all the
• 1:02:06
  time. Every Claude co-worked this. It's the best thing ever.
• 1:02:11
  Well, on this last episode, he was like, man, I
• 1:02:14
  just got to tell you, he's like, I am exhausted.
• 1:02:17
  He was like, I've got. I'm doing so many things
• 1:02:21
  with these agents, but I have – he's like, I
• 1:02:24
  can't stop having to manage them because if you let
• 1:02:28
  them do too much by themselves, they make mistakes. And
• 1:02:33
  he's like, all I'm doing all day long from the
• 1:02:36
  time I get to the office till the time I
• 1:02:38
  leave – It's just managing agents and saying yes and
• 1:02:42
  redirecting them and going and fixing the fixing their direction
• 1:02:45
  on this thing that they took a wrong turn on.
• 1:02:48
  And all this, he said, he's like, he's like, I
• 1:02:50
  am mentally fried. And that's the way I feel, too,
• 1:02:54
  because what we have this thing where. When you're given
• 1:02:59
  a magic box that will do everything, suddenly, guess what?
• 1:03:02
  You're doing everything. Everything becomes equally important and equally motivating.
• 1:03:12
  And you can't stop managing the hoarding. Well, this goes
• 1:03:17
  back to – in a way, it goes back to
• 1:03:19
  what you said about it being like playing a video
• 1:03:22
  game because it's just, oh, just one more. I can
• 1:03:24
  get to the next level. I can do this next
• 1:03:26
  thing. You do need – You do need the fruit
• 1:03:31
  of the spirit. You need self-control. You really need to
• 1:03:34
  have that. Well, when your company tells you, you know,
• 1:03:40
  if somebody walks in and your company is like, hey,
• 1:03:44
  can you pull out this – I'm just trying to
• 1:03:48
  make up an example – Can you give us a
• 1:03:49
  new dashboard for the billing system that shows us this
• 1:03:53
  information? In your mind, you're like, well, yeah, I could
• 1:03:55
  probably do that in a couple hours. I could vibe
• 1:03:57
  code that out. You do it. Now the next person
• 1:04:02
  sees that. Oh, hey, can you add this? Can you
• 1:04:05
  add that? Can you add this? You take that same
• 1:04:07
  scenario and you multiply it times 10. Now you have
• 1:04:10
  10 threads of vibe coding going all at once that
• 1:04:15
  never stop. It's sort of like that old thing of
• 1:04:17
  like, be careful what you ship because you're signing up
• 1:04:20
  to maintain it for the rest of your life. Don't
• 1:04:23
  ever offer to create a service or an app because
• 1:04:28
  now you'll never be able to get rid of it.
• 1:04:31
  And that's the thing here. You have to be very
• 1:04:34
  selective because you vibe code something in a few hours,
• 1:04:37
  and now you've just signed up for a lifetime of
• 1:04:40
  maintenance. maintenance on the thing you vibe coded or correct
• 1:04:45
  correct And it's not because you can't just – well,
• 1:04:49
  I mean it's like party time. So the new party
• 1:04:52
  time rusts. is wonderful. But it's a lot of work.
• 1:05:01
  So what the agent did – is it created an
• 1:05:07
  entire rollout and process pipeline. So the first thing it
• 1:05:11
  did was it created – it was like, okay, what
• 1:05:13
  is your goal for this first release of the Rust
• 1:05:17
  binary? And my goal was just byte parity with the
• 1:05:22
  old node aggregator. I'm like, I just want it to
• 1:05:25
  be the same. You know, for the beginning, it has
• 1:05:27
  to be the same. So it cloned every bug, every
• 1:05:32
  quirk, every mistake that was in the node stuff. It
• 1:05:38
  cloned it directly. And it had to build up an
• 1:05:41
  entire framework of unit testing to make it to where
• 1:05:46
  every – even the flaws were also reproduced. Once that
• 1:05:51
  was in place, and this took about 48 hours of
• 1:05:56
  solid time. agent work of it just plowing through and
• 1:06:02
  building all this stuff and then it took a couple
• 1:06:06
  of days to refine some things it took Then we
• 1:06:12
  rolled it out to one of the aggregators as a
• 1:06:14
  test, and it built an entire framework for – so
• 1:06:18
  the node aggregator and the Rust binary were both running
• 1:06:22
  together at the same time. And there was a third
• 1:06:26
  process that was checking that the Rust output was identical
• 1:06:31
  to what the node aggregator. one was doing. So then
• 1:06:35
  after three days of parody checking, I demoted the node
• 1:06:41
  and upgraded and went live with the Rust binary. Three
• 1:06:46
  more days of no bugs on that, and I rolled
• 1:06:50
  it out to the rest of the fleet. But now
• 1:06:52
  there's this entire thing that's in place, this entire process,
• 1:06:56
  where now that all of that is working... So now
• 1:07:01
  we've got to start fixing the bugs and actually improve
• 1:07:05
  and not just get parity but actually improving the code
• 1:07:09
  itself so that it gives us better results. Like I
• 1:07:13
  said, the speed and the memory is already a million
• 1:07:16
  times better. So now there's an entire process that happens.
• 1:07:21
  Right. You make one change, it gets promoted, it gets
• 1:07:25
  uploaded and then shadowed for a couple of days. If
• 1:07:29
  there's no errors and no crashes and everything looks good,
• 1:07:34
  then we'll promote it to be the live version. And
• 1:07:38
  I'm going to go through that process for every feature
• 1:07:41
  change. This is great, and it's a wonderful way to
• 1:07:47
  make sure that production stays bug-free. But it's a lot
• 1:07:51
  of work. And that's why I feel really bad now.
• 1:07:58
  But but you see what I mean? Like there's a
• 1:08:02
  there's there are parts of a person's life where I
• 1:08:06
  will very much make things easier and save you time.
• 1:08:12
  But if you're a full-time system admin or developer or
• 1:08:15
  something like that, AI has a different side to it
• 1:08:19
  where it will also increase your workload 5x. And that's
• 1:08:24
  what a lot that's what I was hearing him say
• 1:08:28
  in that thing was like, I'm doing so much more
• 1:08:31
  and now I'm exhausted. So I think that that may
• 1:08:38
  be the seed – of of how this thing also...
• 1:08:48
  destroys itself. Is that there's a dark side that's in
• 1:08:54
  this as well. And maybe that dark side – Of
• 1:09:00
  technological exhaustion is actually what humanity needs right now. Oh,
• 1:09:07
  well, I completely agree. What we need is... is to
• 1:09:14
  put down our screens and pick up our Bibles. And
• 1:09:18
  I'm not joking about that. I really mean that. People
• 1:09:22
  need something else in their life besides all this. I
• 1:09:26
  am wholly convinced of that. I'm making life decisions based
• 1:09:31
  upon that conviction. Maybe we never would have gotten there
• 1:09:36
  without something like this driving us hard over the edge.
• 1:09:40
  Yeah, very possible. Wow. That's deep. I like it. I
• 1:09:43
  like it. Well, that's what Eric PP in the boardroom
• 1:09:46
  just made a good point. He's like, it's applying large
• 1:09:49
  scale development to a single developer. And that's true. Yeah.
• 1:09:53
  Like, I think. We're going to see... That's a really
• 1:09:57
  good way to describe it. AI puts too much on
• 1:10:01
  a single person. It makes you responsible for too many
• 1:10:06
  things. Well... Yeah, it depends. I think it really depends
• 1:10:13
  on what you're doing. You know, you I mean, you're
• 1:10:17
  you're you've gone up this horrible curve with this stuff
• 1:10:20
  in your workplace of people coming to you and showing
• 1:10:23
  you chat GPT and saying, how do I get it
• 1:10:25
  to do it better? All the way down to you're
• 1:10:28
  doing MCP servers, you're supporting people who are using it,
• 1:10:34
  you're trying to figure it out. I have some experience
• 1:10:36
  with that, but it's not my livelihood. My job doesn't
• 1:10:39
  depend on it working or not. But when those accountants
• 1:10:44
  and people in your office... I mean, it almost might
• 1:10:49
  behoove you to train them out of Vibe code. Yeah,
• 1:10:53
  do this. Figure it out yourself. You might be surprised.
• 1:10:58
  Well, you can't do that because of security regulation. I
• 1:11:01
  understand. But, you know, I don't have the same results
• 1:11:06
  you have with… This is a pretty big software package
• 1:11:10
  I've built. Look, hey, there's 15 radio stations running on
• 1:11:14
  crap I've vibe-coded. And I've got customers, and I think
• 1:11:19
  this would have been three, four developers to do all
• 1:11:22
  of the stuff that's running. In fact, people from the
• 1:11:24
  radio industry are going, what? What? I want that. What?
• 1:11:29
  And I'm just like, well, I hope they don't try
• 1:11:30
  and vibe clone it. Um... So I haven't gotten to
• 1:11:36
  that point, and people do depend on these things working.
• 1:11:40
  It's not quite the same as your industry. I think
• 1:11:44
  you will get to this point. I might. I think
• 1:11:47
  we all – that's kind of my point. I think
• 1:11:50
  we all will get to this point. I think this
• 1:11:52
  – I think the dark side of AI, of agentic
• 1:11:57
  – of the agentic world we live in and we're
• 1:12:00
  heading into we're all gonna spin out we're gonna spin
• 1:12:04
  out yeah we're gonna spin out man yeah man we're
• 1:12:07
  gonna spin out man i'm gonna spin out and i'm
• 1:12:10
  gonna have a good time doing it yeah well that
• 1:12:14
  that's what's gonna happen i'm i'm i'm convinced Because you
• 1:12:19
  can't moderate. It takes... What kept us from over committing
• 1:12:31
  before was the fact that we had done it before.
• 1:12:34
  That gate is gone. You've become load bearing in the
• 1:12:37
  office. Load bearing. Load bearing. Wow. Yeah. It's going to
• 1:12:45
  happen. Yeah. Well, I really don't want to see that
• 1:12:50
  happen to you. I don't want to see it happen
• 1:12:52
  to you either, brother. Yeah, but I'm… I'm afraid it's
• 1:12:55
  going to. Maybe, but what just happened here where I
• 1:12:59
  just sent a boostagram and it did splits and… And
• 1:13:03
  I just said, hey, man, make sure that next time
• 1:13:05
  you ask me the show name and the episode number,
• 1:13:09
  if you don't have that information from whatever I'm doing.
• 1:13:12
  Okay, no problem. That was amazing what just happened. That
• 1:13:17
  was truly amazing. That blew my mind. Oh, no, it's
• 1:13:20
  impressive. And that's the intoxicating part. That's what drives you
• 1:13:26
  to go further and further and further. And before you
• 1:13:29
  know it, you got customers for an app and you're
• 1:13:32
  like, oh, crap. Yeah. But see, this is. Yes, maybe.
• 1:13:38
  But when do the customers just stop being customers and
• 1:13:41
  do it themselves? I just don't see. I have no
• 1:13:45
  hope that that will happen. It's going to happen. No,
• 1:13:48
  no. Yes, yes, yes, yes, yes. The people that would
• 1:13:54
  be the the hyperscalers that would need to get involved
• 1:13:58
  for that to happen will not let that happen because
• 1:14:02
  that is. Oh, destruction. That horse has left the barn.
• 1:14:07
  It's left the barn. You don't need the hyperscaler. At
• 1:14:10
  this point, we don't need hyperscalers anymore. We don't need
• 1:14:14
  them. The models are at a point where I'm like,
• 1:14:19
  oh, this is doable. It does it. Now, would I
• 1:14:22
  like it to be even better? Sure. Why else do
• 1:14:25
  you think they're doing? Oh, I got to slow this
• 1:14:27
  down. Of course, I got to slow it down. We
• 1:14:28
  can't have the open source models catching up to us.
• 1:14:31
  It's no good. You know, PewDiePie? Yeah. Yeah. biggest YouTuber
• 1:14:38
  probably. So he built all his own models. I don't
• 1:14:44
  know if he's going to make them available, but he's
• 1:14:46
  been, that's been his thing now. He probably lost a
• 1:14:48
  lot of audience. People like, I'm not interested in watching
• 1:14:50
  you distill open AI's models, but that's, that's basically what
• 1:14:54
  he's been doing. Even distilling all their models. And he's
• 1:14:58
  influencing
• 1:15:00
  An entire generation on this stuff. Way beyond what anyone
• 1:15:04
  else can do. And, you know, it's like. You know,
• 1:15:08
  you're a nerd today. You're Bill Gates tomorrow. But it
• 1:15:11
  won't be Bill Gates. You'll be Bill Gates of your
• 1:15:13
  own little universe, your own little home, your own little
• 1:15:16
  stuff. People are tinkering. In America, we are a nation
• 1:15:22
  of creative tinkerers. It's what we are. It's what we
• 1:15:26
  do. It's in our DNA. just like our sexuality. It's
• 1:15:31
  in our DNA. And so we're going to tinker, and
• 1:15:37
  then we're going to share stuff. And I think, I
• 1:15:39
  don't know, man. I just think it will be better.
• 1:15:45
  We're also a nation of capitalistic a-holes too, though, so
• 1:15:49
  don't forget that. Yeah, but it's not – we're not
• 1:15:54
  – yeah, there's – we're creative and liberty-driven, but we're
• 1:16:01
  also a bunch of rent-seeking a-holes too. Yeah, but this
• 1:16:06
  is what I think is ending, that there's no more
• 1:16:09
  – There's no more gates. There's no more barriers. There's
• 1:16:13
  no more barrier to entry. I can do it myself.
• 1:16:17
  They said that about that was the that was you're
• 1:16:20
  singing Aaron Schwartz's song and you see what happened to
• 1:16:23
  him. What did he mean? Killed himself? No, I'm saying
• 1:16:27
  that every the gatekeepers. Found a way to like we
• 1:16:32
  the promise of the Internet doesn't look nearly as good.
• 1:16:37
  Oh, I disagree. Look, look at podcasting. Podcasting has survived
• 1:16:41
  all the gatekeepers, even a hostile billion dollar takeover by
• 1:16:44
  Spotify. It survived. Open protocols. But you have to agree
• 1:16:49
  that it is more regulated now than it's ever been.
• 1:16:52
  The Internet? Yeah. No. My podcast, I do whatever I
• 1:16:57
  want. I say whatever I want. Where's the regulation? On
• 1:17:02
  your podcast. But then somebody goes to a porn site
• 1:17:06
  and it pops up and says, oh, you're in Texas.
• 1:17:08
  Sorry. It was supposed to be the Internet was supposed
• 1:17:12
  to be no. And then we created VPN around. They
• 1:17:15
  were supposed to get around that. And then we created
• 1:17:17
  VPNs. Yeah. And then they block it. Well, then we
• 1:17:22
  go make our own porn. I just tell them I
• 1:17:24
  just vibe codes and porn, porn code. I'm just saying
• 1:17:29
  that the Internet is more we we thought that the
• 1:17:32
  Internet was was it's still pretty great. It's still pretty
• 1:17:37
  great. Yeah, but my point is it's not untouchable. Tor.
• 1:17:42
  And this stuff is not going to be untouchable either.
• 1:17:46
  No. Whenever something is this – This is what they're
• 1:17:50
  – exactly. That's why the frontier models have hired all
• 1:17:55
  these lobbyists. That's why they're all in Washington. And for
• 1:17:59
  the next two and a half years at least, we
• 1:18:02
  probably will be okay. What comes after that, I don't
• 1:18:06
  know. You know, the psyop of you've got to regulate
• 1:18:10
  us is is failed. It's failed. It has failed. Not
• 1:18:14
  that it will fail in the future, but it has
• 1:18:16
  failed for now. I think it's failed against Trump. Yeah,
• 1:18:20
  there's other there's other administrations that would be very happy.
• 1:18:23
  Yeah, I agree. But I think I think the horses
• 1:18:26
  I think the horse has left the barn. I think
• 1:18:28
  it's. I think it's just too late. Prompt hub. Prompt
• 1:18:33
  hub. Oh, I like it. That's very good. Prompt hub.
• 1:18:41
  Somebody register that thing right now. Right away. Sam Santhi
• 1:18:48
  said he's been boosting, but he hasn't heard a harp.
• 1:18:50
  He says you closed the Albie Hub channel. No, there's
• 1:18:54
  something going on with Pipewire because now all of a
• 1:18:56
  sudden it works again. No, it was just an audio
• 1:18:59
  thing. Yeah, because I see his 777. He just wants
• 1:19:04
  to hear the harp. Send another one. Send another one.
• 1:19:08
  Keep doing it until you hear it. Yeah, keep doing
• 1:19:10
  it until you hear the harp. You've been unsuccessful until
• 1:19:14
  I hear the harp. Well, I am extremely excited about
• 1:19:21
  what is right in front of us. And I think
• 1:19:24
  it's a great... a great warning that you put out
• 1:19:28
  there for yourself, but for everyone, um, you, To make
• 1:19:33
  sure we don't get spun up and spun out on
• 1:19:36
  just trying, you know, you can be there's such a
• 1:19:38
  thing as being too creative. And probably we all just
• 1:19:42
  need to go back to being ranchers and farmers while
• 1:19:46
  we vibe code. We got to have some normalcy in
• 1:19:49
  our lives. And we'll still need ranchers, farmers, welders, electricians.
• 1:19:54
  The real world is still fabulous. There's so much wonderful
• 1:19:59
  stuff out in the world. And we just got to
• 1:20:02
  remember to do that. Well, the beautiful thing about humanity
• 1:20:06
  is that you can hold you can do something that
• 1:20:08
  an LLM cannot do. An LLM is just a is
• 1:20:11
  purely a left hemisphere in a jar. And a human
• 1:20:15
  being can hold. Two things in their mind at once
• 1:20:20
  that are actually paradoxical. You can hold opposing ideas in
• 1:20:26
  your mind at the same time, and they are both
• 1:20:30
  true. Yeah. And what those two ideas can be is
• 1:20:35
  that AI is better at coding. Yes, we are. And
• 1:20:40
  it's really good and it's really fast and it will
• 1:20:43
  just make incredible things possible that were never possible before.
• 1:20:48
  And it will also ruin your life if you let
• 1:20:50
  it. Those are both true. Yeah. So you can and
• 1:20:54
  we can recognize that two things are true at the
• 1:20:58
  same time that don't seem to jive. In a way
• 1:21:01
  that some AI will never understand this. True. There it
• 1:21:10
  is. There it is. No, that was Lindiskog. Oh, okay.
• 1:21:16
  Well, he's on True Fans, too, because him and Sam
• 1:21:18
  have a wine party every Friday. Oh, boy, oh, boy,
• 1:21:22
  oh, boy. Oh, wait, did you see PC sales dropped
• 1:21:25
  20%? No. What's up with that? Biggest decline since 2023.
• 1:21:31
  In PC sales? Yeah, because the component prices are so
• 1:21:35
  expensive. Too expensive. Yeah, too expensive. Did you see those
• 1:21:39
  19-year-old kids who have an AI home box that they're
• 1:21:44
  selling? No, what is that? Yeah, it's like some, let
• 1:21:49
  me see. Let me ask a showrunner. Find the name
• 1:21:56
  of the company of millennials that are selling a large
• 1:22:04
  language model computer. They're like 19 years old. I have
• 1:22:08
  this article somewhere. You know what the worst part of
• 1:22:12
  AI is now? Mm-hmm. Is that it's no longer an
• 1:22:16
  excuse for me to tell Gordon that I can't do
• 1:22:19
  something. Hey, that's too hard. No, the best win we
• 1:22:24
  had is, dude, you've got to tell your robot to
• 1:22:27
  shorten your communications. And he did. And he did. He
• 1:22:32
  did. Yeah, he did. Company is called Ghost, also known
• 1:22:35
  as Ghost AI. Its computer is Core. Zane Javed, 19,
• 1:22:39
  co-founder, Nicholas Chua, 19, two other... co-founders Yifei Chen and
• 1:22:46
  Guatam Sharda 24 and 23 they raised 11 million dollars
• 1:22:50
  in seed round product is the core of 3,499 brain
• 1:22:54
  in a box that runs ai agents around the clock
• 1:22:57
  on your own hardware not on someone else's cloud nvidia
• 1:23:00
  rtx pro 4000 blackwell graphics card ships with three open
• 1:23:04
  models from Quinn and Gemma. Their pitch, an AI agent
• 1:23:09
  should not live on infrastructure someone else owns. Got to
• 1:23:13
  agree with that. Yeah, of course I agree with that.
• 1:23:16
  It's fabulous. Why am I – see, this is my
• 1:23:18
  question. Why am I running the showrunner thing? Okay, it
• 1:23:21
  says, Dave, PC sales down 20% with biggest drop. Worth
• 1:23:25
  a quick check. Worth a quick check. It might be
• 1:23:29
  a lie. Yeah, it might be a lie. So it's
• 1:23:31
  giving me some links. But why am I doing that
• 1:23:33
  when Nathan G just always posts the – he says
• 1:23:37
  youngest millennials are 30 years old. He's the fact checker.
• 1:23:40
  I just let him do it. I'm burning CPU cycles
• 1:23:44
  when Nathan's doing it for free. Are you running Whisper?
• 1:23:48
  What are you doing with your showrunner? Running Whisper? Whisper,
• 1:23:51
  yeah. There's a new one that came out that is
• 1:23:56
  super fast. What is it? Give me this information. Nate,
• 1:24:03
  what didn't Nathan just tell you? Go ask Nathan. Don't
• 1:24:06
  ask me. I'm going to rename the thing to Nathan.
• 1:24:10
  What was it? I'm scrolling. I mean, I don't want
• 1:24:15
  to ask again, but I should. Well, let me tell
• 1:24:19
  you, the new thing I'm doing is the Brave search
• 1:24:24
  API is fantastic. Yes. You can just tell any... agent
• 1:24:32
  to like always uh use the brave search api to
• 1:24:36
  look up and and add like amend its information with
• 1:24:41
  that thing with uh search results and you get like
• 1:24:44
  i think it's like a thousand searches yeah i i
• 1:24:48
  use kagi k-a-k-a-g-i kagi um Cactus Whistle, 16.9 megabyte speech
• 1:24:57
  model for local CPUs. Open model bringing speech onto its
• 1:25:04
  on-device needle runtime, whatever that means. What's it called again?
• 1:25:10
  Cactus Compute, and the model is called Whistle. You're using
• 1:25:15
  it? No. Runtimewire.com. Oh, wait. That's the... That's the website.
• 1:25:21
  ActorsCompute.com. With speech to text in 16 megs. There you
• 1:25:25
  go. There you go. For mobiles, wearables, robots. Oh, that's
• 1:25:33
  tiny. Yeah, it's tiny. This is actually good for my
• 1:25:37
  ad removal spec doc. So where are you going to
• 1:25:42
  publish this? On your sub stack? I think I'm a
• 1:25:46
  blog. Oh, you blog? Yeah. Do you still blog? I
• 1:25:50
  have a blog. I haven't written anything in one time.
• 1:25:53
  Answer, no. You still have something about Albiehub. Todd Cochran.
• 1:26:00
  Todd Cochran. Man, we've got to thank some people because
• 1:26:03
  you have to go back and be the AI wizard
• 1:26:05
  at the office, which I'm so sad to hear. No,
• 1:26:08
  I'm actually off today. Oh. Oh, well, then I'm just...
• 1:26:12
  We have some stuff we had to... Well, I'm going
• 1:26:14
  to have a sip of water. So, conclusion, the future
• 1:26:19
  is going to be beautiful and you're going to die.
• 1:26:22
  Is that it? That pretty much sums it up. Well,
• 1:26:26
  I have news. We're all going to die. We're all
• 1:26:29
  going to die! You were faster than me. I couldn't
• 1:26:34
  get to it. It's the power of humanity. We're all
• 1:26:38
  going to die. Hey, where is it? Where is it?
• 1:26:42
  Why can't I find it? I have this one. Kill
• 1:26:45
  all humans. Kill all humans. Oh, yeah. Must kill all
• 1:26:50
  humans. I like the trail off. Where is that kid?
• 1:26:58
  Die kid? Why am I having a hard time finding
• 1:27:01
  this? I'm really happy watching the little meters on your
• 1:27:07
  curry caster. Oh, now you like it. Now you like
• 1:27:09
  my little meters. No, I love the meters. I just
• 1:27:12
  don't like how quick it is to tell me I'm
• 1:27:14
  doing something wrong. Yeah, it says like back off. Get
• 1:27:16
  closer. You're not talking. Your mic is – something's wrong.
• 1:27:19
  You're in touch high. Back off. Oh, you're going to
• 1:27:21
  – yeah. I didn't come up with any of that,
• 1:27:23
  by the way. It says nothing is reaching the microphone.
• 1:27:26
  Yeah. Except spittle. Is this the right one? I'm like,
• 1:27:29
  wait a second. I'm just not talking. I will. So
• 1:27:32
  just so you know, I didn't come up with any
• 1:27:34
  of that. I did not define the interface. I didn't
• 1:27:37
  say anything. People seem to like the little log scrolling.
• 1:27:42
  People seem to think that's kind of – I do
• 1:27:44
  like that, but I'm a – A lot of times
• 1:27:47
  in clean feed, when we're doing the show, I just
• 1:27:49
  keep it on that page. I just see all the
• 1:27:53
  stuff that's flopping by. It somehow is like listening to
• 1:27:59
  – like ambient music or something. What's that guy? Zamfir?
• 1:28:06
  Georges Zamfir? Yeah, it's like the visual version of Zamfir,
• 1:28:10
  Master of the Pan Flute. It just puts your brain
• 1:28:14
  in this weird state where you can just concentrate. Oh,
• 1:28:17
  man. Hold on a second. I've got to play some
• 1:28:20
  Jorz Zamfir now. It's been a while. Wasn't he the
• 1:28:23
  master of late-night pan? television infomercial sell you some albums
• 1:28:29
  zamphir was he i don't remember that oh he was
• 1:28:32
  man he was always on the uh oh i don't
• 1:28:35
  remember that at all okay late night infomercial okay Here
• 1:28:42
  he is. *Ding* Oh man, he's so good. Well, let's
• 1:28:51
  fast forward a bit where he's going nuts. Wow. Okay,
• 1:28:58
  we just got booted off of Spotify. Master. Master of
• 1:29:02
  the pan flute. He is the master of the pan
• 1:29:06
  flute. He's the master because nobody else gives a crap
• 1:29:10
  about the pan flute. Yeah, it's an interesting instrument for
• 1:29:13
  sure. Not a lot of us. Danny boy. I want
• 1:29:16
  to hear him play Danny boy. Does he do Danny
• 1:29:18
  boy? No, no, that's not true. He's got it right
• 1:29:22
  here. It was on the infomercial. Danny boy. Amazing grace.
• 1:29:25
  Okay. Now I got to look it up. 1988. Hold
• 1:29:28
  on a second. Is your some fear? Danny boy. Okay.
• 1:29:32
  Let's see if we can find this. Oh, he's got
• 1:29:34
  a hat and a gold chain running through that chest
• 1:29:37
  hair. Jorzom fear, Danny boy. 100% chance this guy's in
• 1:29:44
  the mob. There's Danny's knight. Danny boy. Oh, brother. Oh,
• 1:29:57
  brother. There you go.
• 1:30:00
  Oh, yeah. home, Xabby. He's dead, Jim. He's dead. So
• 1:30:11
  good. That guy sold 20 million records? Oh, my goodness.
• 1:30:15
  Oh, yeah. And you know what? He deserves it. All
• 1:30:18
  right. Martin Lindeskog, 777. Harp music. Harpo Marx is the
• 1:30:24
  guy. 420 from Martin Lindeskog, who's having his little party
• 1:30:29
  with Sam. PewDiePie equals Felix Arvid Ulf Kjellberg from Gothenburg,
• 1:30:34
  Sweden. Yes, we're aware. Sam, true fans. Triple seven on
• 1:30:40
  a different note, true fans has added a demo of
• 1:30:42
  the new contacts tag. This is good. I appreciate that
• 1:30:47
  since, uh, I think I started that one with, um,
• 1:30:52
  What's his name? Alex? Sanfilippo. Sanfilippo, yes. Good. You can
• 1:30:57
  see it working on power on True Fans. We use
• 1:31:00
  the at symbol for the icon. Very good. Very good.
• 1:31:03
  We have 2222, Martin Linduskog. Dave, I enjoyed your conversation
• 1:31:07
  with Sam Sethi. Yes, that was a great conversation. Thanks.
• 1:31:11
  On last week's power. It was good. It was really
• 1:31:15
  amazing. I love Sam. He didn't ask me the question.
• 1:31:19
  He's like, have you and Adam just like really gotten
• 1:31:21
  into it and gotten mad and said, what the hell
• 1:31:23
  are we doing? And you're like, no, we never have.
• 1:31:27
  No, ever. I look forward to the upcoming episode of
• 1:31:31
  Power. You guys have the ducks in a row. N.A.
• 1:31:34
  Millennial, false positive with 770 sats from Fountain. Dreb Scott,
• 1:31:39
  1234 from Cast-O-Matic with a boost. and Sam as triple
• 1:31:44
  seven again as an app developer I get no benefit
• 1:31:47
  by including ads on my platform James is talk about
• 1:31:49
  Apple getting paid via their proprietary API for every ad
• 1:31:53
  served on their app he says that other apps like
• 1:31:56
  true fans could slash should get a similar deal to
• 1:31:58
  play ads otherwise let's add a skip button like Spotify
• 1:32:01
  Sam's going to be the first one I've talked about
• 1:32:04
  VTS and splits, enabling everyone in the podcast ecosystem getting
• 1:32:09
  paid by advertisers, hosts, creators, and apps, but also listeners
• 1:32:12
  who get paid for their time and attention. Brother, you
• 1:32:14
  preach to the choir. I don't see it happening. Freedom
• 1:32:18
  Boost for a free market. Best premises from Martin Luther's
• 1:32:21
  Code with 1776. There's 4321. I see what you did
• 1:32:24
  there, Dreb Scott. He says, my dude named Ben Stuff
• 1:32:27
  today working through nine IDFs to label and straighten things
• 1:32:31
  up in our new 250,000 square foot lab slash office
• 1:32:35
  building. Making notes for the construction team on what they
• 1:32:38
  need to fix. Thanks for being here for me to
• 1:32:41
  listen to live. Go podcasting. And that's it. I don't
• 1:32:45
  have the delimiter because I guess my node was offline.
• 1:32:49
  uh let's see you may have some other some other
• 1:32:52
  ones that i missed as possible okay i think i
• 1:32:55
  may have uh we got paypal we got christopher harbark
• 1:32:58
  ten dollars thank you christopher uh mitchell downey that's mitch
• 1:33:02
  and the boys uh ten dollars thank you um we
• 1:33:05
  got lauren ball 24 24 cents thank you lauren uh
• 1:33:10
  podverse $50. Oh, wow. Thank you, Potverse. That actually deserves
• 1:33:15
  a – I think that deserves a – Baller! Shot
• 1:33:17
  caller, 20-inch blades on the Impala. Thank you very much.
• 1:33:22
  Yeah, these days, that's a baller. Joseph Marocca, $5. Thank
• 1:33:26
  you, Joseph. Goldstein Media, $18. Boy, it's Goldstein Media. $18.
• 1:33:32
  Was that Scottish? That was my Israeli voice. My Jew
• 1:33:38
  voice. Goldstein. That was so Scottish. I'm sorry. $18. Thank
• 1:33:45
  you for all you do. $18 is a lucky number
• 1:33:48
  in the Jewish faith. It means... Chai, chai, chai, chai,
• 1:33:52
  chai, chai, chai, chai. L'chaim. No, l'chaim is different from
• 1:34:01
  chai. It's got an M on the end of it,
• 1:34:04
  so you don't pronounce the M. Oh, I don't know.
• 1:34:06
  I got to animate the back. I have to brush
• 1:34:09
  up on my Hebrew. Keep up the good work. Thank
• 1:34:13
  you. Thank you, Goldstein Media. Appreciate you. Let's see what
• 1:34:17
  we got on Boostergrams. We got a bunch of true
• 1:34:21
  fans, Anonymous, 1186, 19. You're not resolving the boost? Well,
• 1:34:29
  it just says anonymous. I don't think there is any
• 1:34:31
  message. Oh, okay. I mean, any message. All right. It's
• 1:34:34
  possible. Magnolia Mayhem. Hello. Through Fountain says... 920 sats. It's
• 1:34:39
  over for the podcasting industry. Good. Burn it to the
• 1:34:41
  ground so I can piss on its ashes. Gee whiz.
• 1:34:46
  Woo! Angry much Magnolia Man? There we go. He's ahead
• 1:34:50
  of the curve. He's already passed AI into spin-up and
• 1:34:53
  spin-out. He says, burn it to the ground so I
• 1:34:56
  can piss on his ashes. Long live podcasting. Yay, go
• 1:34:58
  podcasting. Yes, beautiful. He follows it up with another 167
• 1:35:03
  sass. He says, also, I have to say that we
• 1:35:06
  already have this to a small degree. My pinch flat
• 1:35:09
  server. that downloads YouTube channels and re-serves them as podcasts,
• 1:35:14
  use SponsorBlock to set all of the ad reads as
• 1:35:16
  skippable chapters. Once the first person sees where the ads
• 1:35:20
  are and marks it, everyone else after that benefits. Yeah,
• 1:35:23
  just a quick note for anyone out there who is
• 1:35:25
  a web developer or app developer in the Christian –
• 1:35:29
  podcasting space, please stop replicating RSS feeds all over the
• 1:35:35
  place. Please stop this practice. Read up on what you're
• 1:35:39
  doing. Stop doing that. Do not take focus on the
• 1:35:45
  family and serve it again under a different URL. No.
• 1:35:50
  Don't. Even though you think it's easy and it's cool
• 1:35:53
  and it solves some work for you somehow, please don't.
• 1:35:58
  Mary Oscar, also known as Oscar. Mary, 2778 sats. He
• 1:36:02
  sent us a pair of eyeballs. Eyeballs. Thank you. Dick
• 1:36:07
  Rula at Fountain.fm, 3390. No message. Thank you. Anonymous user
• 1:36:15
  at Fountain, 1673 sats. We got some others here. We've
• 1:36:21
  got another anonymous, 5018. Dreb Scott. Did you read this
• 1:36:27
  one from Dreb Scott? My dude named Ben stuff today?
• 1:36:30
  Yeah, I read that one. Okay, that one. And we
• 1:36:32
  got some other anonymous. And then we've got... Comic strip
• 1:36:36
  blogger. There he is. The delimiter. 10, five, five, six
• 1:36:40
  sets. Howdy, Dave and Adam. Today, I recommend a podcast
• 1:36:44
  accidental tech podcast from www.atp.fm by Marco, the developer of
• 1:36:50
  overcast podcast player and his two friends, John and Casey.
• 1:36:54
  They talk about tech, mostly Apple and AI. Yo, CSB,
• 1:36:58
  Ozempic Taker, and still stonewalled by Adam Curry despite his
• 1:37:02
  promise. P.S. It was $15.69 when sent equals 19,000 sats.
• 1:37:10
  Ah, yes, comics for blogger. Maybe, maybe this, maybe, maybe,
• 1:37:15
  maybe. Maybe, you don't even know what, maybe, just maybe.
• 1:37:18
  He just wants me to do his AI tech wizard
• 1:37:21
  show. Do it, man, do it. It's a hoot. It's
• 1:37:25
  weird. No, I know. It's a strange experience. I know.
• 1:37:32
  I'm too busy vibe coding. you're spinning out i am
• 1:37:37
  i'm spinning out all right so what are you doing
• 1:37:40
  around the house that you're not working why are you
• 1:37:43
  playing hooky Oh, I mean, right now I'm about to
• 1:37:48
  just go and have a nice date with my wife.
• 1:37:52
  Oh, are you taking her out? I've not had. Yeah,
• 1:37:55
  we're just going to go out and eat somewhere. I
• 1:37:57
  haven't had much time off this year, so I'm just
• 1:38:00
  trying to pick a few days here and there and
• 1:38:02
  take a day off. Well, very good. You know, just
• 1:38:06
  as a final story. Speaking of dates with your wife.
• 1:38:12
  You've got to date them. Well, she's in Indiana, right?
• 1:38:14
  You do have to date them. So there's the Sonos
• 1:38:17
  system. The whole house music thing? Yeah. And so I've
• 1:38:22
  had a Sonos system 15 years. At a certain point,
• 1:38:27
  you had to upgrade. They basically bricked everyone's old system,
• 1:38:31
  long story. But then they came out with a new
• 1:38:33
  app, and the new app sucks. The CEO even got
• 1:38:36
  fired over it, and the new CEO, oh, we're going
• 1:38:39
  to fix the app. No. I mean, you can literally
• 1:38:40
  be in the bedroom, and then the bedroom won't show
• 1:38:43
  up on the app, and you have to go in
• 1:38:44
  the kitchen. And then it started in there. So I
• 1:38:48
  said, robot. Robot. and I connected the robot to my
• 1:38:52
  Sonos system. I had no idea how that works. And
• 1:38:55
  now I just say, sexy time. And it turns on
• 1:39:00
  Luther Vandross. Yes, I did. Yes, I did. A little
• 1:39:03
  Teddy Pendergrass. No, no, Luther Vandross, not Teddy. No, Luther.
• 1:39:07
  I just had a sexy time. Yeah. It's beautiful. This,
• 1:39:13
  by the way, this thing... My dashboard over here, it
• 1:39:17
  just fact-checked me on whether Zamfir had 20 million records.
• 1:39:22
  It's fact-checking you? Yes. Dave. You just stated that he
• 1:39:27
  sold 20 million records. That's worth a quick confirmation with
• 1:39:31
  a bunch of links. I'm like, man, I really don't
• 1:39:33
  get it. This is a useless thing to me. I
• 1:39:35
  hate I built this. No, what you need to do
• 1:39:37
  is you need the book of knowledge. That's what you're
• 1:39:40
  doing wrong. You don't want that thing doing it the
• 1:39:43
  whole time in real time. You just want to say,
• 1:39:44
  book of knowledge, how many records did George Samphir sell?
• 1:39:49
  That's what you want to do. No, it's slow. According
• 1:39:58
  to the book of knowledge, George Sampson, the Britain's Got
• 1:40:02
  Talent winner and street dancer, Well, I'm killing you. It
• 1:40:07
  just says – it also says, Dave, you stated that
• 1:40:11
  Zamfir is dead. And I never said that. You're in
• 1:40:15
  trouble now. All right, brother. Have a great day tonight.
• 1:40:20
  Thanks, man. I'm very happy. That's a good one. Hey,
• 1:40:22
  boardroom, thank you very much for being here. Oops, hold
• 1:40:24
  on a second. There we go. Oh. I cut off
• 1:40:28
  the book of knowledge and then cut off the end
• 1:40:30
  tune. We'll be back next Friday for another board meeting
• 1:40:34
  of podcasting 2.0. Thank you. Podcasts are cool. You have
• 1:40:51
  been listening to Podcasting 2.0. Visit podcastindex.org for more information.
• 1:40:59
  Go podcasting! Game over, man. Game over.