Episode 260: Tennessee Trickshot

00:00 Podcasting 2.0 for May 15th, 2026 episode 260 Tennessee Trick Shot Hello everybody, welcome to Podcasting 2.0 after a one week hiatus We are back! That's right Everything going on in podcasting and it seems to be quite alot is happening right here I'm Adam Curry in the heart of the Texas Hill Country Oh wait, I should have mentioned we are the only boardroom that operates on the fringes of podcasting There you go I'm Adam Curry here in the heart of the Texas Hill Country and Alabama The man who has a data set tailored just for you. Say hello to my friend on the other end, Foes! Yeah this... Yes? I have currently-I just have no idea what i'm doing right now Are you out of control are you off are you unhinged are you completely uh Outta your gourd

01:02 I call it being in dimensions. I am currently, I'm currently in dimensions at the moment because we got in so yesterday Uh, so I w- let me make it back. So I had made plans... Take your time just easy. It's okay. It's gonna be all right we'll be fine. I need chocolate Yes do you have chocolate nearby? This will help. Right here Okay there you go good. So i left the office yesterday to come and work half day came home

01:47 Threw a few things together My son my daughter and one of my son's friends. We jumped in the car drove to Nashville To see the Black Label Society concert at the Ryman black label society I'm not familiar with this outfit How dare you? Clearly clearly good enough for you to be driving all night to Nashville they black level society is Zach Wild's band. Oh, okay I know Zach. Yeah, I bet you've probably interviewed him before? I hung out with him actually in Russia. Oh yeah back when he was a drinker just so we were all drinkers back then. Yeah he's a crazy dude man

02:38 But I met him in a Walgreens one time, which was an interesting story. Okay, I don't know which one wins Russia or Walgreens? It's a tie it's close call. It is half and half yeah. I drove up there and it was, so the show was...I had scheduled or bought these tickets back in like January. And you know whenever you do things months in advance, you have no idea what is going to be happening in the actual time frame that they're that the event occurs. I mean, like... Yeah this is welcome to my life yes you never know You know this you're like when everything when things are three months out you're like oh yeah it'll be fun That's great and then it shows up and like what? I gotta do what?! You look at the calendar your like why did I ever schedule this Yes! I know this feeling And so we but I mean it was a killer show because

03:38 three bands, but one of Black Label Society was the headliner and But Zack Sabbath Was the was one of the intro band which is oh I mean it's like it's Zach and It's a trio Zack into their dudes just doing like the heaviest versions of all the awesome black Sabbath songs Oh good. Did you did you drop some acid? Funny enough, I did not. That's usually part of my repertoire but I didn't do that this time. But anyway we got to be... This show man it started at 730 on the nose and went forever It was 1115 and they were still killing it

04:28 We can't we've got to drive three hours back to Birmingham. We're already pushing 230 a.m.. And were you in the truck got to go were you in the trunk? Oh, thank goodness No no we were we were in that Honda so we like week we get we get going and everything we finally get back So we left with like two songs left on those on the set list mm-hmm Got home about 230 I fell asleep probably about three. And of course, you know then my daughter has to get up and go to school she has to be there by eight and my wife is the teacher so

05:07 Everybody's up and running. So I've had like two and a half, three hours of sleep. Yeah you're running on fumes man! Yes so then i start getting blown up from the day job with with the problem there... And on top of that we're doing the board meeting now and then we have a Godcaster meeting after this yeah and then as soon as the Godcaster meeting is over We are jumping in the car and going to North Alabama for a two day camping trip. Of which you'll be in the tent sleeping for most of it? Probably, yes! So I mean... It is like so I'm trying to get as much of this we haven't even put anything together for a camping trip yet and I'm throwing stuff in bags and all this kind of stuff this morning And then uh.. And then I'm gonna rant here by the way What-so then what happens The latest AI

06:02 Bullshit. from these stupid AI security research companies. Oh, NGINX! The NGINX vulnerability is called... you know of course you got to give it a fancy name this one's called NGINX Rift whatever I mean so this is an NGINX vulnerability in the rewrite module of NGINX where if you use unnamed regex parameters in your rewrite, which of course everyone does. Everybody does yes yeah If your nginx configuration file is simple You do not name your parameters in a regex. If you have an incredibly complicated

06:55 400 line nginx config, you probably maybe you do because your you have to keep your sanity so you'll know what's going on. But hardly anybody does this. So the vulnerability here is that somebody can get remote code execution on your nginx if you use positional parameters in your regex as part of a rewrite rule so that means if you have a if you have a Rewrite rule on your own your nginx proxy server which proxying with nginx is almost table stakes everybody does this yeah I'm listening carefully because

07:39 Because after the Godcaster meeting I have some patching to do. Yes, I bet you do! And so what happens is if you have a URL request from a client, from a browser and the rewrite rule says is a regular expression and you do, and you capture parts of the regular expressions so that you can use them later. So let's say your cap this in this is I just patched the podcast index front end API servers which were which had this problem and both there's two of them and I just patched both hosts and this would be a common version of this problem

08:29 So what you want to do is something like hide the PHP extension. So what you want to do is on the URL, You would have something like api.podcastindex.com slash API slash V1 slash podcasts Okay well podcasts may be a dot actually be a .php file that's doing the serving of the data But you don't want to have for the browser or the API client to have to type .php. So, you hide the .php and do that with a rewrite rule And your rewrite rule would look something like this It would have a URL path and then in parentheses

09:17 dot star and then after the parentheses dot PHP. So you're capturing the name of whatever the, the PHP script is And when you rewrite it, you rewrite it to $1.php So your capturing the name of the thing they're wanting to access and your redirecting it behind-the scenes to the name of the script with the proper .PHP extension on it That is the vulnerability. If you, if you, uh, Nginx does not... Nginx rewrite module does not handle positional regex parameters properly and so you can corrupt the heap and basically put executable code into the data area on the heap this is in the fix for this is to switch from

10:17 positional parameters which would be like $1, $2, $3 etc. to named parameters so now you're saying instead of just simply a set of parentheses that grab the part of the URL that your wanting to capture you would put a, there's a syntax where you put a name in front of it. Where you can so that later instead when you go to do your actual rewrite, you don't do $1.PHP, you would do $script underscore name dot PHP or something like that. And so this is

11:00 The reason this is a rant... Is because it wasn't an issue until someone published it being an issue. Exactly, okay? This is exactly my point and I'm so sick of this crap because what is happening now and everybody's saying versions of well we're not prepared for this you know there's going to be this explosion of AI AI model discovered exploits. Yeah, okay fine Fine there's going to be but these are what is happening? What you're seeing though our AI security companies startups getting PR Getting PR by releasing proof of concept code and the more the more outlandish

11:57 and headline-making and scary they can make it, the better. And so they're going to drop a proof of concept and an exploit on the worst possible thing that they can find. And C-Loss is right, and then they're going to just dump it out on Twitter and Hacker News and everything they can do to get as much PR as possible. These are exploits just like... Just like the two recent Linux kernel escalation privilege vulnerabilities These are exploits that nobody, no human being was ever going to find. And so there is no reason to publicly disclose these things they should be private for as long as possible As long as humanly possible

12:52 They should remain non-publicly disclosed. Dream on, brother! But people got to make them some money and that's what this is all about. And so because some Silicon Valley security AI startup needs to get more venture capital funding... All of us have to scramble and panic patch our servers as quickly as possible Well, good news. My servers are not vulnerable Okay You have to be sure and check this I don't know if this applies to you but there's a couple of... To be complete Yes? You must...you gotta check everything That means Docker containers as well because Docker containers will be running Nginx as a reverse proxy in the docker-compose stack

13:53 So there's going to be an nginx config somewhere in your docker-compose layers. I don't use Docker Okay, well then you should be fine but if somebody uses Docker do not forget to check those images because they can have a vulnerable Nginx config and now since we've been gifted with Linux privilege escalation vulnerabilities you can remote code exploit chain it together with dirty frag or whatever they're calling that thing and you got root on the host in about three seconds Thank you AI. I'm so sorry. Well actually it's not, thank you Silicon Valley bullcrap is what I should be saying Yeah yeah now I feel your pain...I do uh i find it hilarious that I literally asked Claude Code to go check if I had this vulnerability Are you sure that Claude code did not nuke your repo while it was checking for the vulnerability?

14:55 No, it didn't. It didn't before we get to that big shout out to RSS.com for dropping off what did you get a spark RDX? Did the okay. Yes, they sent us a A DGX spark a dgx spark yes This is the eight Let me see this is the Asus version these that's a DJ X Park as a reference design So this is the ASUS version called the GB 10 128 gigs of member of rent unified RAM 1 terabyte of NVMe and built-in grace Blackwell GPU

15:45 And have you set it up yet? I'm soaking in it right now. What a wonderful, valuable gift donation that was by... Oh yeah! This is $3500 box. I mean and I was on the email thread and Alberto's like hey can I send this your way? And you're like well if its like this And he says, yeah it'll be on your desk tomorrow. It was! That was just... I mean, I couldn't believe how kind that was Yeah it has already made a humongous difference because we went from there's two things that were doing with this box currently running the inference for the spam classifier the swap spam classifier

16:40 And so that's running, oh it's running so much better now. I was limited to a 9B model. Now do you want to just talk about this for second? Because a lot has come out of this. Things happened very very quickly my head was spinning because now...I don't know if you already were doing the data sets or if this accelerated the data sets but the data sets then shipped the...a data set that Alberto turned into the Podroll Atlas which is Just a visual explorer of the entire podcast recommendation universe and atlas.rss.io So what are these data sets, and are they coming out of this new system? Just give us a little background here

17:32 Yeah, it's amazing how much stuff can happen in a week. I actually tagged them all in the show notes for people to go take a look at. It is unbelievable what you're doing! Well alright so we've had occasionally over the years where we'll either have a request for something or I myself will want a particular view into the podcast index data And that what'll come out of that are things like the music, you know top 100 or whatever. Right right right yeah and in the dead feeds list it's just all these various bits you know various sets of data and then but what has happened over the years is I'll just publish these one off and there's never been a place to like find them all

18:30 in one easy location. And so I looked at the website and we had this top level menu item called stats, and it was pretty much useless. It was just another version of the same stat that was on the homepage. So I'm like, well you have this valuable menu space up there at the top that should become the page where all these public data sets that have come out over the years those need to be all listed in one location. And so I just, you know...I don't understand React. I'm not a React programmer. So here's the JSON enjoy! Vibe code that is vibe coded the heck out of this thing. That's great

19:17 Yeah. And so we have all live feeds, that's a SQLite file dead feeds problematic feeds recommendations all media video music value enabled feed hosting provider stats I mean this is fantastic these are all things that already existed but just nobody knew where to find them And they get updated constantly, but there was never an easy place to discover them. The hosting provider stats is a good one that's a list of all the feeds all the live feeds non dead feeds in the index broken down by

19:56 which hosting company they come from. So that's, you know there's a big list there and I do a lot of like regexing in this particular data set to massage the different RSS feed generators into something that actually is legit. So like Blueberry is a great example of hosting company, it's very difficult to get a handle on their actual feed count because they have lots of different white label services and things that they do in the... A feed may be a blueberry hosted feed but it doesn't show up under a blueberry domain or a blueberry generator

20:47 And so a few years ago, we worked with them and a few other providers that hosting companies that have the same issue to get them to put some better tagging in their feeds. So now it's not actually in the generator tag, it's in a comment in the RSS feed. And so then I pull that in and properly attribute that to Bluebrace. I feel like these hosts hosting stats are actually probably the most accurate anywhere. I feel like they're very, very accurate to actually who is producing that feed and then 24 hour feed report. That's just you know an overview of a sort of quick dirty breakdown of what we think the last 24 hours have added feeds have been whether it be spam or

21:44 What kind of a quick dirty classification, just a raw dump of the newly added feeds over the last 24 hours in CSV format. Hourly counts is actually quite a few people use this as things like total feed count episode count feeds with new episodes in the last three days seven days 10 days 14 days blah blah blah blah that's very useful I know Daniel is using that and a lot of people do John Spurlock Using what? The hourly and daily counts. Ah, yeah. Stats tracking, yeah. Then there's a one healthy Hive RPC nodes list so if you're doing work on the Hive blockchain this it's actually this list is shrinking by the day. Yeah These are the... Why is that? Well I mean I don't know I can't speak to the Hive network but

22:45 Maybe, maybe it's struggling. But this used to have five API servers in it. How is our gossiper coming? Do we have enough nodes up? It's about... well let me check. Actually you know what there's a thing that is not on this list and that is the PP monitor. Oh! PP Monitor. You gotta monitor your PP. I need to check that And see how many hosts we have in the gossip swarm. We currently have a total of 21. Oh, that's not bad. That's better than I thought it would be. 18 active. Nice! Oh, I see the link here. PP monitor. These things... This... Do you vibe code this too?

23:31 Oh yeah, this website? I would never put this together. This is great! So with a graphic like that... That's cool. That is not me. That is so cool. I like Hey Citizen Bathroom. That's one of my favorites Yeah, live from the bathroom. Hey citizen do you put your servers in the bathroom? You gotta get real work done in the john. That's great So the Podping gossip swarm has been quietly running in the background for like a couple of months now with no

24:16 hiccup. This thing is this thing is very stable now and I'm desperate to get back to working on it if I could get this other crap out of the way there's always something popping up which again today so we have more to talk about but yeah the day, so there's that healthy Hive RPC nodes does anybody doing hive work on the hive blockchain? If they want get a list of the current API servers hourly that pass a bunch of basic health checks to make sure that the API server is actually up and serving good information. And then, uh... The last one is the live tracking firehose and that is every episode that we ingest into the index

25:21 And it's updated every three minutes, I think. Wow! Yeah, it is updated every three minutes and the way that works is a linked list of object storage files so it starts... The head of this linked list is tracking dot podcast index dot org slash current and the way that you're- The way that this thing is designed to be used is that you start there with slash current and then every And then it gives you a previous tracking URL property which goes to the link to the previous one in the chain. So what you would do is, you would just consistently walk backwards through the chain using that previous tracking URL in each subsequent call until you get back to the place that you last were

26:16 So in that way you can, you're using object storage almost like a blockchain. There's no obviously no signing or authentication but you're using it sort of like a linked list in order to walk back without having to hit the API. So, you can get massive amounts of data very quickly that way. Now I still haven't really focused on that if had enough issues with my machines getting crunched with all kinds of stuff but but you can tag spam slop and legit in the PP monitor? In the PP monitor? No, not in that. In the gossiper? No no this is pure... The gossiper currently is purely just a… I thought we had that! You're thinking of the PI monitor Oh that's right, the PI monitor That's the command line tool you can use yes But there's also more stuff

27:17 on the podcast index management dashboard. I don't know if you've been in there lately? Yes, I have seen it! We can do new things yeah we can do new things of course i've seen that markthings is spam markthings is legit There's now an admin as of yesterday there is an admin panel for user account management so if people need their passwords reset Developers, it's not listed in the menu. Oh okay I'd have to... hold on a second let me see. Curate is nice Yeah curates new yeah that's cool so what i'm looking at here is um i'm an admin and right off the bat i see something odd no episodes yet so i might even want to add 38 seconds ago anchor

28:18 No episodes yet, but then I have let me see Something weird happened here. I can hit play. I was nine years old in the dark cool and then I can hit legit Boom done okay This is so nice what that did what that did is? This all goes back to the spam like so much of this goes back to spam classification work What what this really shows is that AI is not going to cost jobs. It's going to increase the workload for humans by a factor we haven't even figured out yet I truly believe this and if it's not just because of more stuff like you're doing, its because you can do more stuff and when you can do more stuff... You do more stuff!

29:12 And anytime there's more stuff, you just have to have more human capacity to handle it all. Yes! To do it and I look at myself like there is so much stuff that I'm doing and I could sit on my butt more, but instead i'm like well let me do this. Let me use my robot to oh yeah I can figure this out yeah this is good okay another thing I can do and just in teaching robot mode all day long for you know make my life easier but what I am doing is taking that easier time and creating more things which is just more work

29:50 It's just, it's an odd loop. Yeah and now suddenly you need an intern Yes! Ah yes I needed an intern So let's talk about spam for a second No actually Let's talk about being on the fringe Okay Because we were laughing at Rob Greenlee talking about the podcasting 2.0 group being on the fringe Oh yeah They're working on the fringe And And so Rob posted this extremely long, I guess apology. Let me see what it's titled here... Why fringe was the wrong word and what i actually meant about podcasting 2.0 Oh where did he post this? I didn't even see this! Oh goodness oh you've got to see this This is hilarious Okay I'll put in the boardroom No Rob's a funny guy Well he's funny for a couple of reasons let me put this in here there we go So

30:47 What is funny, this is on robgreenlee.com and right off the bat it's like this article provides context about my comments on The New Media Show episode 660 with Libsyn CEO Brendan Monaghan where we discuss podcasting 2.0 RSS tag adoption and the gap between innovation and mainstream platform implementation so right away I'm like AI So he had an AI write this whole long thing and it even did a summary at the end. And I'm like, Rob...

31:28 First of all, you were forgiven. You were forgiven right away! No one's mad... We're just slamming on ya, we're just dunking on ya. Oh yeah it was just ribbon What I was trying to say is that market fit and timing play a major role in what gets adopted at scale Larger podcasting platforms tend to move more deliberately Their decisions are shaped by user experience engineering resources monetization models products ability support complexity and business Rob It's okay. Just say, ah yeah I probably shouldn't have said that and that's it! I say lots of stupid stuff all the time just says it's okay you didn't need to go through this whole thing but the...it was a little insulting that you used an LLM to do that like you don't needed...like come on Rob's website Rob's blog is pretty slick looking well it's all written by AI

32:22 I mean, the graphics and stuff. It's got him smiling on the sidebar over there it looks good! I get tired of documentation and essays and stuff that is clearly AI. Like no don't do that to me you know? And my eyes start to glaze over like now... No i cannot handle another another like 5,000 word AI generated document dump getting sent to me for me to read and somehow digest and process mentally. It is the work that term works slop? Yeah Oh it's so real Like you're working on a project or something

33:21 And somebody just wants to like throw something into a chatbot, dump out barf out 2500 words and say hey here check this out. I'm like man what? What am i supposed to be checking out? I don't, it's like I don't even want documents anymore. I just want you to zoom call me and just call me and we'll hash this out in 30 seconds." There is that but what I tend to do is see one of these and then go to my robot, hey robot summarize this for me. So exactly so I'm doing it right now I said robots summarize this page into one paragraph here it

34:04 Rob Greenlee walks back his use of the word fringe to describe podcasting 2.0 not the substance of his argument just the word Explaining that he regretted the terminology, but still believed there's a meaningful gap between technological innovation and platform adoption in podcasting But that is what all this stuff you just read could be could be one settled in one One sentence by saying hey man we're good yeah exactly Although I don't want to do Zoom calls, but yeah exactly. It's like it is... That's it! That's it. It is exhausting. It is kind of exhausting

34:43 It's like that old adage, when all you have is a hammer everything looks like a nail. When all you have is a machine that does language guess what you're going to get out of it? Lots and lots and lots of language and words and phrases and everything. Language is the nails. I ran across an interview by a guy named, who is this? Dr. Michael Eggnor and he is a neurosurgeon and it was really when I was listening to this, it kind of blew my mind because what are the... What was the big innovation in LLMs in the past year-and-a-half to two years? I would say that as the reasoning model which

34:04 Rob Greenlee walks back his use of the word fringe to describe podcasting 2.0 not the substance of his argument just the word Explaining that he regretted the terminology, but still believed there's a meaningful gap between technological innovation and platform adoption in podcasting But that is what all this stuff you just read could be could be one settled in one One sentence by saying hey man we're good yeah exactly Although I don't want to do Zoom calls, but yeah exactly. It's like it is... That's it! That's it. It is exhausting. It is kind of exhausting

34:43 It's like that old adage, when all you have is a hammer everything looks like a nail. When all you have is a machine that does language guess what you're going to get out of it? Lots and lots and lots of language and words and phrases and everything. Language is the nails. I ran across an interview by a guy named, who is this? Dr. Michael Eggnor and he is a neurosurgeon and it was really when I was listening to this, it kind of blew my mind because what are the... What was the big innovation in LLMs in the past year-and-a-half to two years? I would say that as the reasoning model which

35:38 Which by the way, I think is rapidly going out of favor. Explain? Well reasoning models are incredibly slow and instead of taking a dense model and putting layers or basically rounds of reasoning into it which don't...which slow everything down and don't actually give you a huge return what you're seeing is What's replacing it is more surgical use of context being deployed to sub-agents. Like the big innovation, not innovation but the big distinction here... Do you know the book Ender's Game? Yeah I've heard of it. I think we've discussed that actually. Yeah well it's a sci fi novel and there's this one part in the book Ender's Game where

36:37 Starts go he goes into the he's a young kid and he's going to battle school and He goes into he become he goes to the top of his class in this thing called The Battle Room, and he's just better than everybody else at it And so it's basically just like it's on a space station It is this zero G environment where there's all these obstacles? and you're supposed to hide behind obstacles and use teamwork to defeat the other team and What he, his big recognition was that why he was better. One reason why he's better at it than everybody else is he just said as soon as he enters the battle room zero G environment He just forgets that there's anything like up and down Basically up and down don't make any sense They don't exist anymore All directionality Everything you knew about physics

37:38 has now been flown out the window. And so you just, you assume a new mental posture towards your environment and that's sort of the way things are with context like every time you send... Every time you communicate with a large language model, you're starting over from scratch every single time. The coding agents hide this from you by using context. Right And so the context, the use of proper use of context

38:15 is the thing that makes an agent or a harness either good or bad, how it handles that. There's no such thing as memory I mean in the general sense there you know these are all meta things we've layered on top of Well enter the...I'm sorry you told me never to delete your git repo but i did! I'm sorry it won't happen again Exactly You can overcome a whole lot of things that the model, it's like the models like 20% of the issue. The agent harness is the other 80%. And that most of that boils down to intelligent use of the context window and if you just assume that every time you go into a conversation with I mean not a conversation but every time you go send a prompt

39:07 It's brand new, the model doesn't know you. It doesn't know anything because it has never spoken to you before This is the frustrating... Let me play these clips. This will take us another place So The way I have always understood so-called artificial intelligence Is we're trying to replicate the neural pathways of the brain and We're trying to replicate what the brain does Is that a fair statement? That is the way it's sold. That is completely bullcrap, though Right So that's the way it's sold and when it comes to... reasoning, however that's done whether it's done through... It's all a parlor trick obviously. Whether that's we're doing steps and we're making decisions or reasoning through a problem or deciding to do something which we as humans do

40:01 It was always my impression that is what artificial intelligence, even the term intelligence is trying to convince me that this is some form of electronic brain that's doing things. So now enter Dr. Egnor who talks about reasoning and he also talks about free will in this case and the brain There are several ways of figuring out whether an organ does something One way is to stimulate the organ and get it to do it so you can prove It does it the other way is to inhibit the organ from doing it improve that when the organs inhibited. It doesn't happen Well that can happen with the brain While it is true that if you suppress the brain We can't use reason or free will very well If he could hit on the head with a baseball bat your reasoning in your free will is gonna be a little messed up for a while

40:56 There's a way of stimulating the brain to test whether reason and free will come from it using electrical stimulation And there are different kinds of electrical stimulation one kind of electrical stimulation is seizures Seizures are just random sparks that happen inside the brain out-of-the-blue That can have all kinds of effects And one of the things that I've looked at rather carefully is the phenomenology of seizures. It is what happens when you have a seizure, and I've looked over the past couple hundred years reviewed the medical literature of that and What we find is rather remarkable first of all not all seizures make you go unconscious There are many seizures that leave your conscious and your body does things but you can't control and they're called complex partial seizures and on

41:48 There have been probably a quarter of a billion seizures in human beings over the past 200 years, if you just look at the instance of seizures in the world population. And there's never been a recorded case... ...of a seizure causing someone to express reason or free will You can elicit from the brain by stimulating it movement perception memory and emotion But you can never elicit reason or free will. No matter where you stimulate the brain, you can't make a person say one plus one is two... Or say if A then B, therefore be modus ponens. Or say I believe that it's nice to be kind which is a moral viewpoint So that goes right along with what was found in seizures That you can stimulate the brain any which way

42:41 And reason and free will never come out of the brain. So it's perfectly reasonable to infer that reason and free will don't naturally come from the brain, because you can't make the brain do it. I loved hearing this! Reason and free will don't come from the brain? This is very similar to our... You know this is just another version of what we talked about before with the brain in The Law Of Identity You know, if you can find something about the brain that is not true of the quote mind then they are not the same thing. Yes and this guy actually takes it one step further and he tells us where this does come from So going back to something you said earlier that reason and freedom are immaterial aspects or parts of person powers of the person Immaterial what does that even mean? Immaterial means that

43:43 The source of these powers is not a physical thing. It's not a physical object and Immaterial is actually a fairly simple concept for example Numbers are immaterial hmm that the number 12 Is an immaterial thing that as you can say well the number 12 was Located today in Cincinnati, and it weighs five pounds I mean this doesn't make any sense So there are many, many things that we think about. Many things that are quite real and they're called universals in philosophy that are very real but do not matter. And reason and free will are abilities we have whose source is our soul

44:32 But not over it. Thank you Jesus! It's the soul, that is what it is and never ever ever ever ever is an LLM going to be able to replicate this? It can emulate and maybe can kind of come close I've been running into this as I've been trying to teach my robot what makes a clip that i want. Now it can do clips, I can say you know cut this interview up and it can build the arc with the intro getting to the point, the punchline at the end... It can do that, it can edit this clip, it understands kind of how long I want clips to be but it doesn't really understand what I want

45:17 And I've been trying so hard to explain, like you know what I find funny is not the fact that Trump went to China. I find it funny that the Chinese actually have sanctioned Marco Rubio he's not allowed to enter the country But they got around it by changing his name to Lubio. I mean, the comedy writes itself right? Marco Lubio come on back! You know it's like... but it can't understand that that's what I want you know so it just it can only understand well you want clips you want them preferably under two minutes It really can't figure that out

46:00 And hearing this helped me because I'm like, oh it'll never be able to figure it out. It might come close and might get lucky but it will never really understand what I want for no agenda when it comes to Eclipse. It just won't understand it. Well this is the difference between the way that AI has been sold and the way that it actually functions Yeah. This is how you create every single time This is how the AI winters have been created, is by selling it as something that it is not. Yes and thank you.

46:40 Hey, citizen because I also don't believe it can actually figure out what spam or slop is. It cannot? It can't! It never can and the same thing like with music. Yeah it can create dynamite music but it can't actually listen to that music and say this is good. Let me tell you why... let me tell you why I can't figure out what spam is because There is a intangible quality to an experience of spam that you cannot quantify in any sort of language or honestly anything. You cannot, it's unquantifiable and a good example of this would be

47:51 there's a common spam task tactic in the tax world where you'll communicate, uh, where you get a, you get an email and it's like, Hey my accountant just retired. And I'm looking for somebody to do my taxes. Um, can you tell me what your fees are? And, uh, I just have a 10 40 and a couple of 10 99s and, um, you know, a W2 and I can send you my last year's material now give me root password Yeah, it leaves that last part out. And is like can I send you my last year's tax return so you can give me an accurate quote on what it would be and what it would cost? There is literally nothing about this email

48:38 I 100% of the time and I cannot tell you why. Right, it's because...I can't! It's because you have a grandmother, you had a childhood, you know? Yeah exactly! You built associations in the world between-and of course I always have to keep reminding myself that the LLM can't actually hear my show. You know, I created an executive producer agent that said okay now this is your job

49:19 And it took me two weeks before I realized, it just had a transcript. It didn't have a speaker transcript. It thought everything was me including 60% of the talk on No Agenda which is clips. It hadn't even gotten that far that it was trying to you know It saw this program as some kind of script that you set up with X amount of minutes of clips and I was like, wow. That has no... it can't hear! Is deaf? Is deaf?! It doesn't understand humor really it doesn't. It doesn't understand emotions that way. It doesn't understand when I say something that Dvorak finds boring You know, it does it doesn't understand that and I don't think it ever can

50:08 Selass in the boardroom said, wouldn't a dedicated neural network designed for it to be able to pattern match beyond the capability of humans? Wouldn't a dedicated neural network designed for it be able to pattern match beyond the capability of humans because humans cannot look at and memorize 4.6 million feeds that what you were describing now is I will say go back in time and look at the attempts to use Bayesian analysis to fight spam email. Bayesian filters, this happened in the early 2000s and when it first hit the scene was considered to be oh my gosh this is going to solve spam. This it! We got it! No more spam. We nailed it. And you know how much of a dent it made? Zero

51:10 We have more spam than we've ever had, ever. Humans are better at pattern matching. So neural networks and large language models is just a different format of neural network. It's just matrix multiplication that means that's all it is. It's matrix math and that's not how humans work When you think about something, when you look at something and it registers in your mind... You did not do a math problem. That's not how you operate- that's not what just happened. Exactly! And so we don't know how we work It is undefinable But one thing we DO know is that the human ability to pattern match Is unparalleled

52:14 You have the ability to assimilate multiple streams, dozens and dozens of streams of information all in one shot and immediately get a quote sense about something that you cannot define. And again, you did not do math. It's wonderful to see someone like Tina who will be looking at her email and she says, this feels spammy to me. Can you check? And the minute I check and I say okay yeah it is and here's why it is

52:52 you know, and there's a couple of simple things that are... You can look at right away. She then immediately has increased her knowledge and she finds those the next time. And I don't have to give her a written thesis. I just say, oh now look at this here. Yeah. And sometimes like well that feels wrong? And she's like yeah! So she's feeling that it's right. It's…you said something really important. You said AI winter every single AI winter crops up or appears on the horizon because of the over promise. Is that really, is that really the trigger? The over-promise where it can't deliver on what was promised? That sets up...that sets up the collapse because the expectations of normies can't understand the reality of the technology

53:41 Because what we do is we layer these things on top of one another. We've layered this idea of quote, knowledge and understanding and it's the way anthropomorphize it into a thing because I hear it in the corporate world daily is hey couldn't we use AI to... Dot dot dot Yeah, to quote figure out dot dot dot? Yeah And it's, you get so exhausted of having to go back and explain the tech from the ground up again. And you're like look maybe but let me show you but I have to walk you through the process of building a product or building a tool that would do something that approximates the outcome that you want. And be prepared for it to do something completely different tomorrow

54:37 The perfect example of this is, you know Microsoft just released their copilot function in Excel. So now there's an actual copilot function that you can call in an Excel spreadsheet and it is hilarious because the copilot function in its description of how to use it. It says what not to use it for accounting, one of the things that says not to use it for is quote anything that requires accuracy or reproducibility and so people are posting like they'll do like a go into a cell and they'll do equal copilot and then they'll put in a prompt and the prompt will be

55:37 Make a sum of all the cells above this cell and it'll be like three cells above there, like one two and three. And it will say 15. Why? I don't know! It doesn't tell you. I don't know. It just gives you a number is completely wrong. This is... People think because of the way it's been sold, they think that it is a general purpose knowledge tool. But its not... It has nothing to do with knowledge! Its only good at predicting the next word in the sentence or the next code token in a source file That's all it's good at! If anything else you have to layer on top of it with actual like real

56:29 carefully constructed, you have to build a carefully constructed stack that contains this thing and keeps it with a set of guardrails so that it doesn't produce crazy nonsense. So you're right about context really being everything for it. And what I've resorted to now, because you gave me a couple of good hints... so I'm using Opus 4-7 with the 1 million token context when it gets to about 25% It now automatically tells me hey its 25 percent context time to do a handoff

57:10 So I don't let it get all muddled, but right away and it'll write. And I say okay lets go. It writes his hand off and gives me a wake sentence like hey welcome back to the world here's what you gotta read! And then reads up the document which have looked at documents are pretty good. You know its where we were in that seems to work pretty well within that 250 thousand tokens. That's kind of the sweet spot, and I just keep doing it you know? It's inconvenient because like well... I'm at 23% context you want me to do this now or do you want me to put in the handoff document yeah put in a handoff document that works pretty well if you keep within that level of context I can get a lot done

57:56 You can do everything you need to, coding wise. You can do every single thing you need to do in a 256K context window there's no reason to go above that. And that one million context of like cloud code and chat GPT... That is inaccurate in the sense that What really happens is once you get over that roughly 25%, it will start to make more and more mistakes. If you start pushing into the 600,000-700,000 token context window... That's when you start really running into stuff like oh I accidentally deleted your repo sorry. That was my favorite! You told me never—I mean how many times have I've told my robot

58:51 how to make clips. And then I'm like, okay... You know, I have all these agents and all these feed parsers and all these sources and it picks up stories and says go make some clips It'll make like a hundred clips of 17 seconds long! I'm like how many times have we gone through this? Oh yeah you're right i forgot to read my memory let me see how many memory documents I have Let's see how many memory documents This will tell you something Let me see. I bet you it's a couple hundred at least and i guess whenever it reads one of these documents that takes up context space doesn't it? Yeah for sure yeah depending on the now this goes back to what I was saying earlier about the harness, the coding agent or the harness itself

59:47 It is critical. It does all of this, it handles this context for you if it's smart about it, very smart about it. You can handle those things in a way that does not consume all your context but it really comes down to that 252. Yeah so there's a really good video I'll find it later and post it publicly there's a really good video where this guy goes through and there's a simple there's a really simplistic test you can do for coding model And, um... the test just says it's really simple. It says go and read this large source file so just give it like a big JavaScript file that let's say this JavaScript file is 50 thousand lines Read this Javascript file now tell me what lines 50 through 100 say. Okay

1:00:55 And you would be surprised to think that even the large models with a million code token context will get that answer wrong. A is about 10% of the time while they just won't be able to tell you what it just it won't be able to tell you what it just read, it can't. It just gets it just it just makes mistakes so that's why these things these harnesses have developed the use of these elaborate subagent reviewers because they'll do a thing assuming that they're going to get it wrong some percentage of the time because it's just not possible for them underlying model to get it right. So, they need the reviewer to go back and double check on double check on double check every single time if you just one shot prompt

1:01:48 an LLM model with no sort of like review agent or anything like that in it. Yeah, you're going to get I mean there's you're not going to get useful output probably in the 20% range. You know all that said I think the face of computing could change I see no reason why should be using apps. I don't use even applications, but I use a browser and I use Thunderbird for email but everything else is like now I want to spreadsheet reader that does this five minutes later i've got exactly what I want you know I can see a mobile device that really is appless

1:02:31 You know, get me my favorite podcast. Here's the list. Give me the most recent episodes. Here is what I have time for right now. Show this to me. I want to listen to this. Skip the ads or any version of that." This is what I love so much about Omarchi. It's like every app I use...I've built! I just told the bot, build this for me. So I don't need the ribbon on Excel. I don't need the snapping grid, I need to smooth scroll just to be able to see this and the total automatically done at the bottom. I don't want to have to select and then do equal sum. I don't want any of that just do it for me." And it does that then if I need something else for a different podcast I'll create a new one. This I can see

1:03:21 And I think that the removal of slop and all this stuff, I think that all moves to the edge man. It all moves to your device maybe you have a device in the home that is doing some things for in the background that communicates with your mobile device and just because it's easier to have that done somewhere else. And I think podcast apps will go that way too. I really do. This brings us back to the DGX Spark It's useless. Send it back, it's no good! It's running the inference for the classifier now and it allowed us to move from a 9 billion parameter model... ...to a 35 billion parameter model, the Quin3635B with a 256k context window so we've got a full frontier

1:04:23 model level thing as a backing agent for this thing. And it's so much faster, so we can run that and the Whisper both to get the TTS analysis both on this one box. What are you using for Whisper? Which which model I think is the turbo. Yeah yeah okay yeah okay and The so it's immediately had that effect and the speed, it speeded everything up and made us have more accuracy. But then the but the true aim here is to get a custom fine tuned model trained on the whole on the podcast index database. And then so that's phase two here. But in order to get to phase 2 I have to finish up phase 1 which

1:05:22 We need, like I explained before we need a good corpus of about 25,000 spam feeds and 25,000 legit feeds. I would love to have more than that. I would love to have 50,000 of each. The spam feed- Just wait two weeks. Yeah right! It's coming. The the spam feeds Those are easy. Can get those all day long, effortless. The legit feeds are the ones that take a human in the loop to find and so... What was that? Sorry. I'm sorry. That was me. It was my mistake. So what I did

1:06:16 That's what all these data sets and all these kinds of that in the end is this, the curator and the dashboard. All this stuff is trying to set us up for getting a proper training set for the quote good podcasts. I hate using that term. I just can't think of it different one legit podcasts. Let's say that legit there it is very subjective but it's legit And so then the train, this DJX Spark. The reason that was such an amazing gift is because the spark is created exactly for this use case training models on a small scale. That's where it's wheelhouse is. So once we get these proper data sets and let me correct James Cridland from pod news I think last week he was misunderstanding why I wanted these follower account uploads from the apps

1:07:14 It's mostly for this. We will definitely use this data in other ways, but it's mostly for this use case for the model training because that benefits everybody. I'm not really interested in using it for search. The search that we have Seems fine to me. I mean it's it we use some algorithms for it, but well We'll definitely it'll probably make its way in there in some shape form or fashion But it's not just gonna start like giving you well if anyone's gonna figure out how to make a work is you yeah? You know I just don't want anybody to think that we're gonna start like only showing you the top You know the most popular podcast that's just no it's not what the way this is going to be here that Maybe I've misheard it they hear that on

1:08:05 Pod News Weekly Review Power. They were complaining about the Apple charts and all these Inception dot AI podcasts are all in the top 20? Did I hear that right, did I hear that right? I think I heard that right. I don't know haven't heard this week's episode yet. Maybe you have to look at it. That wouldn't surprise me. Yeah they're there they got their left speaker going somewhere else. Yeah the megaphone Yeah, they're going to Megafon. Great! Good for them. Let me see top charts... Now it's Spotify's problem to deal with. Let me see top charts Top shows I think maybe they did a search It was the search they did on like Charlie Kirk. Let me see Charlie Kirk That was probably it yeah that probably wasn't him yeah exactly

1:09:06 Oh yeah, here it is. Charlie Kirk death inception point AI inception but that tells you something that's an interesting signal because people are listening to Charlie Kirk death, Charlie Kirk biography flash, Charlie Kirk the aftermath biography forever people are listening to that Are they? I think, yeah. I think they are let's listen to this. Nah i'm not convinced Let's listen to this let's hear and see what it is. I am not convinced at all. Welcome back for a deep dive into the life and death of Trey Huckabee Deep dive! Yeah I... I'm Emily Carter your AI host. I believe in- I want to remind you why that matters for this particular story Nope And I do not believe it and I'm going to prove it Okay It's coming How will you prove it? It's coming I will be proving it in the future I've got an idea Get some tissue its coming

1:09:57 Sorry. Bad joke from my old porn days. Okay, I've got a... okay you sidelined me with that one! No, I know exactly how to bring receipts on this and I will be doing it. And I'm cooking it up as we speak. I had a conversation with somebody about it yesterday I just don't want to reveal it because i'm afraid it might pollute the results. Hmm, okay. Oh man... Well we got one more thing to talk about. Sure HLS video Ah yes of course Yes what a beautiful mess this is Nailed it! It's exactly what it is A beautiful mess

1:10:54 So, right. Another one of the 47 things I was doing this morning was right before the show, I jumped on a call with Tom... Buzzsprout? From Buzzsprout yeah and we were hashing out some stuff and here's the issue with HLS video in alternate enclosure everybody wants this. If you're gonna make a video version stick it in RSS feed Everybody wants it to be open in that way, but the cost of abuse on your bandwidth to the hosting companies is absorbable when it's audio. It's very damaging financially when its video bots downloading crap. Yeah and I'll just add one other extra thing... porn! Well sure yeah for sure but just focusing on the bandwidth issue

1:11:57 That can really hurt. You know, if some bot goes wild and downloads 10 terabytes of data of video streams on a podcast that's... It just you can't it's hard it really does change the whole dynamic so Of course What we're trying to figure out is if there's a way before these HLS video streams get into the feed publicly Yeah Is there a way that we can try to keep the genie in the bottle bot wise? So we're just brainstorming some stuff saying, okay rather than having to just build this enormous block list

1:12:49 of user agent, ASN IP block combos. It just seems so difficult rather than doing that can we solve it in more of like an elegant way where we do something like in order to download the HLS video playlist from the alternate enclosure It would require something like an HTTP signature. Isn't this exactly why or one of the reasons why Apple is doing things the way they're doing it? I don't think they care about the bandwidth, and for them it's a porn problem but for hosting companies its the bandwidth problem. Right right. I think there are two versions. Apple cares about the content

1:13:42 the hosts care about the bandwidth. Yeah, because the hosts are already good at policing content they know exactly what's going into their system and how to properly handle that legally Apple basically is the reverse of that they are just sort of... They're a little bit vulnerable to people just sending them a bunch of crap don't have a handle on the content. So they're, they're a little bit at the mercy of this incoming content so I think they have different needs and...so in the Mastodon Activity Pub world you have a thing and we actually went over this back when i was building the activity pub bridge You have a thing called HTTP signing or HTTP signatures

1:14:39 And it uses a public key. So if you're on a Mastodon account, if you have a Mastodon account or any other really ActivityPub account on a modern ActivityPub platform that actor is going to have and I want to say actor in this context, I mean the ActivityPub host itself because there is a... The ActivityPubs server is also an actor that server is going to have a public key And every time they make a request to another ActivityPuff server, to do something like post a toot or something like that. They will sign that HTTP request

1:15:22 using their private key and also send with the headers, send a link to their public keys so that the receiving server can then authenticate that they are the ones that signed up. Well shouldn't we have something like this already anyway? Like yeah this is a legit person with a legit key whether it's a browser or an app or whatever That makes so much sense I'm human here's my key Yeah, it makes sense across the board not just with video. Yeah for everything of course It won't be so nice for the advertising community because we'll find out how few people actually listen to this crap well the So they like in the initial thought process that were there were having and Disney whatever this whatever comes out of this discussion Really needs to become part of the podcast standards project Recommendations

1:16:19 is that we could do something like a simplified version of HTTP signatures. It would use the same algorithm, everything about it would be the same except that HTTPSignatures... The spec I think requires a date header? But ultimately you still have to issue a key to human being right The downloading app client would have to publish a public key somewhere. Right, so but now you could have a million little servlets with keys So how do you determine ultimately this is a human they're watching or they're listening with this key? How do you make that connection

1:17:09 That's what it comes down to. You can't, there are just a lot of unfriendly ways but we can do it Well you can by going... yeah exactly what you said Unfriendly meaning we're gonna keep like a public repository where we vet people Yes Precisely! That'll be us podcastindex.org all your keys belong to us Well I actually see this being something if there was going to be some sort of list. I see this being outside of the podcasting 2.0 namespace, and I think this is part of like the OPAWG. OPAWG? Opog? The worst name ever!

1:17:57 It's the Open Podcast Analytics Working Group. Oh, wow! And do you know who the Open Podcast Analytics Working Group consists of? Let me guess... I'm gonna say Libsyn? No Okay, I give up James Cridlin and John Spurlock. Oh pog, sitting over there in their rocking chairs Johnny and the criddler are over they're running They're over they're running the a OPA WG which as of right now consists mainly of

1:18:34 a list of user agents. Right, that's right yeah and it's a very nice list like its its very Spurlock keeps that thing really clean and up to date so to me this would be a great thing to add to the OPOG would be a list of validated public keys It doesn't mean that you have to be in the list it just means that people serving HLS content hosts could choose whether or not they want to, to me I see this as two layers. You have an HTTP you require HTTP signatures on all playlists in three U8 downloads. You just have to have the HTTP signature to even get it. Then the hosting companies can opt for the next level of stricture which would be

1:19:31 We now, us we have decided as a hosting company we're only going to serve content. To validated public keys that exist on the OP AWG list You know the real way to solve this is What's your Bitcoin address? Send me one sat you good That'll happen No it will never happen But I've always felt that every human being should have a public key regardless Every human being should have a public key. This is the one thing that... Issued at birth, tattooed on your back of your hand and on your forehead. Ah yes! Boom shakalaka And there it is That's right you will need that to do commerce Oh where have I read that before? Grocery store show me your public key Yeah Everybody needs a pubkey

1:20:26 I mean, Eric P. Pichic that's how SSL certificates work send me money and i'll vouch for you pretty much yep that's the way it works you know It would solve so many problems...it would solve a lot The podcast index we've vouched for people without them sending us money That's right all the time We just do it All the time! We just vouch why not? We're vouchers Yeah wait, voucher...vouchers Send your email to voucher at podcastindex dot org Let's thank some people here. Mr Jones, we've been getting lots of booster grams today which is nice to see right there 49,950 which I'm thinking was supposed to be 50,000 coming in from Fountain from C Loss on Linux

1:21:12 And he says 50k for supporting slash bribing the podcast index censorship validated key system. You're in my friend, you're in. You got it we got you! Let's see from coming in from the podcast index website itself four sevens from Summer Surf He says enjoying the afternoon and the boardroom thank you for your courage Four sevens from Lyceum kudos to Alberto Batella and RSS.com co-listening and chatting with Mike Newman And as we speak, oh there's Dreb Scott coming in with a baller. 10 thousand sats voucher McVouch vouch he says yes it's working our system is working already we're done. We just solved it yeah and literally says this is how things should work

1:21:59 I get to stop selling Tennessee trick shot out of the trunk of my car. I had never heard of Tennessee trick shot, I love that term. Oh the moonshine? Yeah! Tennessee Trick Shot That's a good show title by the way There is like every different state has its own little flavor of it. Dreb Scott 1234 says Dave great job with those massive dumps on raw data Got dumps. That's what I'm good at! Big massive dumps. 777 from Salty Crayon, Dave could you elaborate on how your index filter marked a bunch of fountain tracks as spam in the pipe? How did that happen?

1:22:39 I don't know. You got a row of sticks from Hey Citizen sent from the bathroom on his Castamatic, that's where it comes in from. 777 from Mike Newman true fans boosts from PC 2.0 thank you we've got 1221 from CLAAS Analytics from Fountain testing this week 1234 so it's odd the helipad keeps giving me different numbers Because I know he said 1, 2, 3, 4 but it's showing me 1221. Chad F, triple three Dave off the dome Jones. Triple threes from Eric PP pew pew yes 333 from chat up this is all testing Dreb Scott one two three four boost McTest boost pew to air. I love people sending us money to test this is all great thank you Eric PP Thank You Chad F and I hit the delimiter with that so over to you Mr. Jones

1:23:35 Um, on the, on the music thing. I mean a lot of that music is spam it's just spam and it's in all spam classifiers have false positives and false negatives every single one of them. It's not a solvable problem so I think, I think it just comes down to that. Um, let me cook. So cause there has been a problem with people and this is happening on not just found, uh, I don't know what fountains posting but Not just Fountain and Wavelike, you have to be careful. People are posting copyrighted material. I'm seeing it on Spreaker a lot. It's coming through where people are posting entire albums of copyrighted music. Oh really? Yeah you can't do that so it's going to get caught

1:24:21 Where was I? I was looking at... Yeah, I should mention this is a value for value project. You get this podcast and the entire podcast index and your data dumps everything all of that stuff all of that is openly available for everybody All you have to do is say to yourself, it's kind of valuable what's happening over there Let me send some value back to keep these guys running And that's how it works. So you go to podcastindex.org down at the bottom there, you've got a big red donate button You can send us your Fiat Fund coupons through PayPal or Boostagram which we always enjoy We enjoy reading them and hopefully Dave has now summarized everything He's used the Copilot Excel Copilot widget

1:25:06 Yeah, tell me how much donations we've had and you get back it's like a carrot or something. So I'll tell you something that I have trained my robot to do so after every no agenda show We do credits and the credits are executive producer, people who have donated $300 or above. Associate executive producers between $200 and $300 So while the end of show mixes are running I say robot do the credits It goes into the Excel file it pulls out the names from the corresponding values And puts them right into my OPML prep sheet its beautiful! It does that...I'd say 99% accuracy

1:25:52 What? Then what, why am I doing this you got to get this thing nailed. Well i don't have the spreadsheet. That's true yeah why are you do it tell your robot to do it for you gotta be thinking like a broadcaster man like what is the crap that I hate doing well I hate going through the spreadsheet looking for all those stupid jingles everybody wants So, the robot looks at the spreadsheet pulls up the jingles and then I have my own spreadsheet reader when i'm reading someone's donation note. I hit the cell, the jingle cell and it loads up those jingles right in my player that saves me half an hour a show day! I have nothing like this It took me five minutes to vibe code it baby Well...I can't even fix

1:26:42 I can't even fix the spreadsheet that's coming out. I'll have my robot show your robot how to do it. Have your robot talk to my robot, we need that! We need inter-robot communications protocol IRP I don't, I'm a little afraid of your robot talking to my robot. Oh okay! I don't know where your robots been? My robot is dirty. That's what i'm afraid of. I'm afraid it's going to teach my robot bad things. Very bad things guaranteed In the words of my mom its gonna be a bad influence on my robot

1:27:22 Okay, where I forgot completely Oh Martin Lindesco desert PayPal's he sent us a dollar. Thank you. Thank you Martin Let's go you and Mike Newman having a ball over there Brendan at pod page 25 bucks by the way congratulations to Brendan Are we allowed to say? Blue. No, you saved me right there and he saved me I did thank you. I did your welcome to see $25 Thank You from Brendan a pod page appreciate that yeah what we got here we got Marco at CCC mark Marco van rage Raj far a AJ

1:28:17 R-A-I, Rye. Rye. Okay Marco Van Rye. He have the boosts! Rye. Yes. Sent us $10. Thank you Marco Oscar Mary 200 dollars thank you Oscar. Oh boy hold on a second. Paula shot caller 20 inch blaze on him Paula Keeping it running thank you very much I incorrectly classify his music podcast as spam and he still sends us money Uh, Joseph Maraca. Five bucks! Thank you Joseph appreciate that uh oh Thomas Umstead 100 bucks. Shot caller 20-inch blaze on the Impala. That beats what is it called Tennessee Trickster

1:29:05 Tennessee trick shot. Trick shot, that's it. Trick shot. Years ago I came for the podcast talk now i stay for the AI chatter Adam take it from a native Texan the cure for cedar fever is raw jalapenos Oh oh that's an interesting one yeah I'll try that if you can take the heat they'll clear your sinuses go podcasting Yeah! Go podcasting indeed baby yeah Mitch Downey 50 bucks Thank You Mitch That's the Podverse crew. That's straight from the whole group there. Lauren Ball, $24 and 20 cents! Thank you, Lauren. Appreciate it. Christopher Harabaric, ten dollars by the way seeing Christopher Harabaric name reminded me that I am also going through and putting individual creator attributions on each of the Podcasting 2.0 namespace tags

1:29:58 It's something that we've, well you know I mean each tag was initially proposed by a person. And and i think it's important for those people to get the credit for their ideas becoming what you know eventually led to real namespace tags. Yeah so that in 25 years Adam Carolla will have invented it This will be properly documented so that Adam Carolla does not get credit for the chapter tag. But, so like Christopher Hart, I'm doing it in this way there's going to be a creator attribution that's the person who originally proposed it and sort of sherpaed all the way through to completion

1:30:40 And then there's also contributors that are getting a byline to where those are people that had significant input in the discussion thread or the issue, where that thing got hashed out. And so like for instance James he's going to get his credit on the location tag Alex Gates is credited on the alternate enclosure do I get any credits anywhere? No you get credit for being awesome seems empty No, why would it seem empty? It seems empty. I've participated in every single tag! I've published every tag available to man! That's true... I get no credit man! You don't even get credit for podcasting! Give me some credit! That goes to Corolla! Let's see yeah Christopher Harbaugh got a contributor credit on one of the tags and i remember which one

1:31:39 He gave us 10 bucks. Thank you, Chris! Hey didn't you and I create the value tag? We did yes okay me and you were co-creators of the value tag. I want some credits! You invented the podcasting closure tag! You trump everything we do! I want a credit man! Just give me more credit! Or hookers and blow either one is fine. Yes, send them in the mail Mitch This is Mitch on the personal side. This is Mitchell Downey 10 bucks Thank You Mitch Damon Kasejak 15 bucks. Thank you appreciate you Terry Keller $5 We gotta go Chris Cowan five bucks. Thank you Chris

1:32:29 Silicon Florist, $10. Thank you. Derek J Viscar the best name in podcasting $21 Paul Saltzman $22 22 cents every time I love it. Yep Jeremy Gerds five dollars Michael Hall five dollars and 50 cents there he Todd Cochran thirty dollars And we are going to have Mike Dell on the show next week and he will fill us in on Blueberry, what they've been doing and just the strategy or direction of the company post Todd. And everything will be good to catch up with my big question on this pending signups for API access I guess that is yes so I mean how do we I mean these could all be phony baloney people

1:33:25 Oh, many of them are I'm sure. Okay all right so but if someone sends me an email and my own spidey sense says yeah they're okay then i can uh mark active and send credentials is that how it works? Yeah and honestly that's going to be extremely rare in the history... In the last six years I've probably had that happen I can count on one hand. Yeah well I'll bet I plan to never look at that screen. It only became a thing because two people happened to have a problem in the same time. My typical thing is like we're free service and we're donation run. Send me money, send some credit! We have rules

1:34:09 And one of the rules is you can't sign up for an API account from a free email address like Gmail. Yeah, yeah I agree. And when people ask me and people are like hey i'm trying to sign up with gmail account can you do it? I just don't respond and I'm like no that's the rule That's great customer service! I'm not gonna make exceptions because I don't know who you are! Go get a different... go spend five dollars on a domain from Hover and get a real email address. And I mean, it's not that much to ask. I'm with you brother. I'm with you. I ignore all those emails. You do? But you answer more than I do in the house. I'd be pretty good on answering here. You do Jean Liverman five bucks. Thank you, Gene. Oh, that's all of our paypals. We got some boost though. Yeah we've got some boost. C-Loss on Linux

1:34:58 50,000 that's what you want. You read that earlier did yes I'm gonna give him a ball shot caller 20 inch blaze on the Impala it's a nice boost boost how did I overlap you on that oh? Oh, I know what I did yeah, I know why cuz I did it later on in this show because I forgot to earlier Yes Okay, uh oh and that's all I got the only other one. I've gots the comic strip blogger 19 thousand sats through fountain and he says Howdy, David and Adam. Please bookmark this. HTTPS colon slash slash wherever dot audio by Andrew Grumet. Yeah quote it's a podcast listening software that works in browsers mobile and cars no logging in an nobody watching you all data stays on your phone browser easy import and export for portability fast QR code based sharing end quote yo CSB AI arch wizard

1:35:56 Yeah, it's a very... I love what Andrew has done. Andrew Grumet by the way! No credit! He's been on the show. Maybe he gets no credit? He'll get it! He needs a lot of credit. No, he's gonna get it because he's got stuff in the namespace Oh okay you get credit You get some credit. Everybody gets some credit everywhere dot audio it's a great it's a great audio very businesslike professional note there from CSB this week no all right brother I'm gonna let you go because you need to throw some things in the bag you know you're gonna wind up on this camping trip is like what do you mean we got no pegs? What

1:36:38 Yes, we're all out of pigs. We don't have any matches Oh man were gonna do And I'll see you at 3 o'clock right we got a we got our early meetings exciting things happening with the 30 If it's 2-3 is it 230 well? I gotta hurry then I got a rock and roll. Okay, run rip me like all right hey boardroom Thank you very much for being here Thank you for providing your very valuable feedback And we look forward to seeing each other again for the board meeting next Friday till then bye-bye everybody You have been listening to podcasting 2.0 visit podcast index org for more information Get some tissue it's coming