Based in Sydney, Australia, Foundry is a blog by Rebecca Thao. Her posts explore modern architecture through photos and quotes by influential architects, engineers, and artists.

Episode 298 - Veto Powers, Programming Languages, and Terminals

Episode 298 - Veto Powers, Programming Languages, and Terminals

Max and Aaron discuss recent works at Local Maximum Labs, including feedback on the new Great Compromise constitution, and the programming language for newmap.ai

Links

The New York Times - A Founder of the Crypto Hedge Fund Three Arrows Capital Is Arrested

The New York Times - A Silicon Valley Supergroup Is Coming Together to Create an A.I. Device

TechCrunch - Ray-Ban Meta smart glasses livestream to Instagram and Facebook

Related Episodes

Episode 296 - Government vs Google and Amazon Anti-Trust

Episode 297 - Stephan Kinsella: The Fallacy of Intellectual Property

Episode 172 - The Fourth Turning Part I - Cycles of History

Episode 295 - Rewriting the Constitution: Did the Founders Screw up the Senate

Episode 100 - Milestones, Questions, and Wisdom from the First Hundred Episodes

Episode 200 - Milestone Metachat

Episode 193 - New Beginnings

Transcript

Max: You're listening to the Local Maximum, Episode 298

Narration: Time to expand your perspective. Welcome to the Local Maximum. Now here's your host, Max Sklar.

Max: Welcome everyone. Welcome. You have reached another Local Maximum. It is my pleasure today to be joined again by Aaron. Aaron, how you doing?

Aaron: Good. It's good to be back. It's been a couple episodes.

Max: It has been a couple episodes. Not too many episodes, but it's good to have a nice string of interviews that I did. And so I have a bunch of interviews in the can. And I know I just did two interviews recently. The first one was Adam Kovacevich again, on the the comings and goings in DC, not like the not the the partisan politics that we're used to but the kind of, what the heck is the government doing suing Amazon and and was it Facebook? No, Google? I can't keep them all straight.

And then the second one was a lot of fun. That was with- I should just pull these up. But that was with Stefan Kinsella. Now I like Stefan Kinsella, I follow him on Twitter. He's, he's really interesting. He has some really interesting ideas when it comes to patents and copyright. And so that was a lot of fun as well.

Aaron:Yeah, that was a fun listen.

Max: Yeah. And we have a few more fun listens coming up. But I don't know if I should ruin it yet. And then I have two more scheduled so I literally have four in the can. So that is good to have finally. And then I'm reading some books I'm trying to get these authors on, we'll see if I can, we'll see if I can swing that as well.

Aaron: Exciting. And I don't know if you'll be able to get our friend is it Strauss or Howe who's still with us

Max: That's one of the books. 

Aaron: Yeah, the amount of talk about the recent book that came out and some of their older writings and their relevance to the moment.

Max: I am in the middle of it. This is not a fourth turning episode. But I just, that stuff changes the way, I think we should do…not only should I try to get Neil Howe on the show, but I almost want to do like a whole show on like literary analysis of like different movies and music that we know about and trying to face, trying to put it into the different archetypes, and different like story structures that he has based on the four different four different types of generations.

There's the Hero, the Artist, the Prophet and the Nomad. So that and of course, one of those movies has to be Grandma's Boy from the Adam Sandler crew. But also I was thinking like — we shouldn't get into this, should we? — we have a whole list of things you should get into.

Aaron: Now, let me mention one idea to that and then we can move on. Yes, it's and I don't know how many properties franchises this would work for. But there are some — Star Wars is the one that comes to mind — where there are characters who there were movies or stories told about those characters one or two generations ago, and then modern tellings of of those same characters stories in the current generation, and how have they changed.

Is Han Solo of episodes, one, two, and three, being portrayed in a dramatically different way from Han Solo in the solo movie, which is his own origin story. And given that it was written and produced, what, 40 years apart or something?

Max: If we were gonna do that, then I'd have to watch the solo film, which I don't think I'm willing to do just yet.

Aaron: I would do that for the show.

Max: I guess, well, one thing that's interesting about the I'd rather have you watch grandma's boy for the show. One interesting thing about Star Wars, the way I put Star Wars into this is that it's about it clearly portrays a fourth turning, but it is written, it is clearly written and discussed for a second turning audience, for an awakening audience, which is obviously the whole like Jedi mystical kind of aspect of it too.

Aaron: And that raises the idea of just how history is not necessarily told but recorded. Are the people who live through the fourth turning the ones who write the history of the fourth turning? And I would have ventured no, they are not. That history of fourth turnings are written by people in first and second turnings because there needs to be a generation or two between the events and the authoritative histories of them.

Max: Yeah, yeah, I'm watching. I'm watching a variety of lectures now in great courses, like those things you can get on Amazon. And one of them is on Ancient Egypt, which because I read this book, Pharaoh — that's one of the interviews coming up with Alexander Hool, who wants to, like, change the entire timeline of ancient Egypt.

So I'm like, okay, if I'm going to do this, I better learn a little bit about the standard timeline. So I'm taking that, and I'm also watching the Islamic Golden Age, and I’d definitely fit it in. I'm like, Okay, well, this happened in a second turning, then this was the first then this was the fourth turning, then this is the next fourth turning, when, like the dynasty changed and that whole thing. But for those of you are interested in Strauss-Howe generational theory, one that I kind of had trouble and I think I figured it out fitting into something: Wizard of Oz, that's a good one to try.

Maybe try not to answer that right away, Aaron, unless you have something on the tip of your tongue. And so it is because it took me a while, but I think I think I figured it out, like when it has to take place. So that's, we could talk about that another time. Great.

We're gonna have tons of episodes on this, just like we have tons of episodes on my constitutional plan. So for those of you who have been following the show, you know that I wrote a plan called the New Great Compromise where I rewrote the Constitution of the United States.

The main thing that it does is it changes the Senate breaks it up into the Senate, which represents the states and also the Executive Council, which takes care of some of the kind of parts, some of the executive branch aspects of the Senate. Variety of reasons why those need to be broken up. We covered them. But first of all, Aaron, you posted this online? Did you get any feedback online from this?

Aaron: I got a little bit. Yeah. And so there's-

Max: Localmaxradio.com/labs, by the way, if you guys want to check that out.

Aaron: There were kind of two major points that were raised. One was: this sounds like it adds a lot of complexity. And complexity is generally more complexity is an opportunity for more more corruption or, or unknown unintended consequences, which is not something that I can project, so there’s a potential issue there.

Max: Hang on a second. There's a problem where like, so I know this as a, as a software developer, people when they come into a new codebase always say it's complex. And then when they've been working with it for a while, it's simple.

So how do you measure complexity? Is that just I would say to that person, like, how so? How's it complex? What parts do you find complex? And is it the Condorcet method for voting? Because while that might be somewhat dense to figure out, I don't think it's a vector for corruption.

So what exactly are they talking about? And how do they measure complexity? I would try to get in. I would try to go there.

Aaron: Yeah. And I think that makes sense that you have to separate out when you're talking about complexity, your “objective measure” of complexity versus a status quo bias that just because this is changing, and in a way, it's adding more parts and that you're adding the Executive Council, you're not adding any, well, you aren’t adding limited new powers and responsibilities. It's more that they are being shifted where they lie.

Max: Shift into a more rational way.

Aaron: It's not a line item veto but there's some override capabilities, which are new — that was the main objection. I think it was more that you're adding this new body and you're changing how everything's being elected.

There was also concern about moving the power away from the federal level, down to the state level, with changing how the senators are elected/appointed, that in smaller groups, it is easier to exercise corruption.

I'd be interested to see an actual, rather than philosophical, more rigorous scientific breakdown of that because it sounds truthy. I have some gut questioning, but I don't have a strong counter to it. It certainly does seem like that the fewer people you have to bribe, the easier it is to gain control. And ostensibly by moving it back to the state legislatures having that power means you've got a smaller group of people you have to control.

Max: Right, of course, it's not state legislatures. It's state law. So like, ultimately they would have laws they have to follow like, look during COVID, I think they broke, I think state broke some laws. But that's already the case for state law in terms of elections.

Aaron: You could make a weak to medium strength counter argument that states already differ dramatically on how they deal with filling these seats between elections. That some states like California, and maybe it's not universal, maybe it has to do with when it occurs, the governor appoints a replacement.

Max: Just happened today as we speak.

Aaron: Oh, I haven't been following that closely enough. I knew it was pending.

Max: He appointed someone who has on their Twitter bio, that they're from Maryland. That's, that's all I saw. Unless it was a troll.

Aaron: You can have the governor appoint someone, you can have the state legislature appoint someone, you can have temporary appointments until there can be a special election. There are many different approaches. And the federal government does not define how that works. It is left to the states.

And there are several different approaches the states take. So I think that's kind of a proof of concept that you could extend that diversity, that power, beyond just filling vacancies to the standard process. And it doesn't necessarily introduce a, it's not guaranteed to break the system by any means.

Max: Yeah. All right. Yeah. It's interesting. Well, I'm glad those people are reading it. And I hope, I hope some of the listeners today are going to finally break down and give it a read, let me know what they think. Interestingly, like in politics, people tend to form opinions vary quickly and say, Oh, this is why it will never work and, or that people tend to make much more snap judgments than I'm used to, like in a technical concept, which is, which is really interesting.

Aaron: It's very tempting to read it with the eye of, okay, as soon as I find something that doesn't sound quite right to me, I'm gonna say, oh, that's the problem with it. And I can just disregard the rest. It's hard to always put yourself in the quote, unquote, open mindset to give things a fair shake.

Max: Yeah, yeah. All right. So there is a potential change I am thinking of making to it. It's a tweak. But it's this provision with the Executive Council that I have. And I want to read the provision that I think is probably superfluous.

I'll read it right now, word for word. It says, “With a supermajority, the council may declare any legislation unconstitutional. And so shall include this reasoning and any such declaration, the legislation is immediately null and void. However, any member of Congress has standing to challenge the reasoning of a Supreme Court, which has the authority to elevate and overrule the decision.”

So, first of all, it's quite a complicated process there. But the more I think about it, the more I think like this, this is a power that will never get used. Because if it's an old law, the council has rules, has powers to go through and edit all laws, that sort of a, the bookkeeping kind of repealing old stuff that the legislature can't get around to type rules. And if it's a new law, well, if it's a new law, first of all, the president probably signed it. And so already, they're gonna have a big problem getting a supermajority because the President's on the council, and secondly-

Aaron: So that means that they need everybody else to go against the president.

Max: Oh, well, five of them.

Aaron: So five, five of the seven with the president. Okay. Yeah, just just clarifying what the supermajority requires.

Max: But we have to remember, these are still politicians, and if the if both houses of Congress pass this, and then the President signed it, how likely is it that something like something sails through those three things? And then you have five counselors who are like, oh, my God, oh, my God, oh, my God, we can't do this. I just don't think that's that's going to happen.

And so I mean, I don't I don't know. And so I'm thinking COVID, for example, a lot of that suffering, because, of course, I don't think that there was much in the way of legislation during COVID. That was questionable. It was more executive actions on the state level, and partially on the federal level.

So I can't think of a time in history where this would have applied. And so I was thinking that instead, I should do what I didn't do the first time, which was to allow the council to have a role in the veto process the President can veto a bill have it passed the council, it's not just to say it's because I didn't want to make the argument like, well, there's already three traffic lights, you have to go through, let's add a fourth. I don't think that's gonna save that much.

But it kind of maintains the balance of power in that the council and the Senate, were supposed to be like, split offs from the current Senate. So they should have a say, in legislation as well. So that's kind of what I'm thinking right now. Does it make any sense to you?

Aaron: So I’m curious if I can follow your logic on the ability to declare legislation unconstitutional, being unlikely to actually be executed to be utilized. But my question, though, would be, does the presence of that ability create a chilling effect, which will somewhat constrain Congress to not push the bounds on that front? And so even if it is a rarely, if never exercised power, that it has a useful purpose? Probably grasping at straws a little bit.

Max: If they have, if they have to consent to the laws to begin with, like if you have to get a majority of the council, otherwise, it's vetoed, then A) they don't need a supermajority? And B) like, this could only happen then if like, oh, maybe four years ago, under a different Congress and a different president, there was some kind of a outrage that now they have a supermajority overturn? That just seems very unlikely. I think that chilling effects that you want would be accomplished by that via that veto power.

Aaron: So the other way to interpret that clause or that that power is, is that it creates a fast lane to the Supreme Court. Because if Congress passes something, and then the council immediately comes in and declares it unconstitutional, what are the odds that Congress is going to sit back and say, yeah you're right, we oopsie and just let it die.

No. Somebody's gonna object to it. And so it goes to Supreme Court. So I would say that functionally, this, this is a way to cut through some red tape and bring things to the Supreme Court more quickly. Now. If that's your goal, there might be a better way to streamline cases to the Supreme Court, than specifically going about this, this change in powers.

I don't know exactly what it would be. But it would seem, if that's really all you're trying to do here, if that's the ultimate purpose, that there's, there's other ways you could accomplish that.

Max: Right, I think I would need to study how the court system works a lot more. And I'd like to talk to someone because I the bulk of the research that I was I was doing was on kind of the structure of governance. And that's why I didn't touch the Supreme Court. There's a lot of very reasonable amendments to the Constitution that you could make with with regards to the Supreme Court. But I just I didn't focus on that here.

So maybe you could leave that up to a different one. And I think the veto power is very well understood. And I think it's, I think people would be able to grok more what that's going to do. So the question is, so the President still needs to sign it. So the President needs to sign it and then like three counselors, probably. So you need to have some kind of-

Aaron: Just a naked majority of the council? Well a naked majority of the council, if that includes the president. If it doesn't include the president, okay, so the president can still veto anything, but he needs to get at least three counselors to go with him.

Max: So four counselors disagree, then they are the ones who send the veto back and say, hey, here are reasonings. Here's our reasoning. Why don't you try to override it? Or change it? Usually you don't goad them and to be like, oh, yeah, override me. No. But usually you say, Hey, I'd like these changes, or you say, Don't do this at all. But,

Aaron: I mean, sometimes Congress will do that. We haven't seen that in a while. But sometimes Congress will do that, where they'll, I guess, you, you really need to have a Congress where they control both houses, where they can send a piece of legislation to the President and say, Go ahead, veto it. And then we can run against you on the fact that, you know what, we've passed these laws and you refuse to sign them.

Max: There actually have been veto overrides. And this usually we don't hear about it, because it's usually stuff that's not very partisan, and not very hot button II. And I'd love to get like a list of veto overrides that happened.

Aaron: Yeah, that immediately makes me think, A) did the President when they vetoed it know that this was going to result in an override? And B) If so, why did they veto it because you see that sometimes where, I know, at the state level, there are some states where if a law passes, like with a supermajority the governor has X number of days to sign it. And if they don't sign it, then it automatically becomes law. But then they could say, well, I didn't sign it.

But it's not the same as vetoing it. Because if they veto it, then they know it's gonna go back to the legislature, and they'll be overridden. So they're, they're kind of saving themselves that embarrassing step, but still preserving the ability to say, Well, I didn't sign it. It's not my fault. I just didn't stand in the way of it.

Max: Yeah, I'm having trouble finding successful overrides. All right. Well, I don't have them available. But I know they do happen occasionally. But yeah. All right. So I think that's a go. I think maybe I'll come out with a new one in a few weeks, and we'll see what is. I do want to send it out more broadly. Okay.

Aaron: So I shared some of the feedback that I'd seen. Have you gotten any other broad feedback in the last, what has it been about three weeks? Since you published it?

Max: Ah, no, not as much as I would like. I did see. I did have a few people read it. A few more people read it. But yeah, I'm busy with a lot of stuff. And so I'm not like, necessarily emailing it out to lots and lots of people. Well,

Aaron: speaking of which, so other stuff you've been working on?

Max: Yes, my new programming language.

Aaron: What have you been working on? 

Max: Well, I have a bit of time off right now, a bit of time off, or maybe starting my next thing, because so now I am sort of, I've moved to kind of a consulting role. And so now I am working with a bunch of people on their companies and startups and ideas, including my own newmap.AI which, as you'll remember, I worked on full time for a bit back in 2022. And that's like a new programming language/data layer, that I think is somewhat innovative.

But unlike a lot of other projects, a lot of startups where you could like build things like really scrappy and get them off the ground. This is my project, where I'm trying to build something in really a new way. And it's taking me a long time to do it. If I can kind of give my advice, and I've been working all weekend, and that's why I want to talk to you on it.

Aaron: Before you get to the elevator pitch — when did you first start working on newmap?

Max: Five years ago?

Aaron: Okay. I was, I was almost certain it was pre COVID. Because we've been kicking these ideas around for a while. Okay. But it wasn't really in full gear until, like you said, you were working on it full time in 2022 for a while.

Max: Yes. Yeah, exactly, exactly. And what I found was, it's easy to make a basic programming language. But once you want, like a really sophisticated type system, and I've books on type theory and all that. And I've been listening to lots of podcasts on type theory and reading academic papers and stuff.

But really, that doesn't, that's not helpful. You really have to dive in yourself and do it, the type system becomes very difficult. And the organizational system, how do you organize data that becomes very difficult. And when you're coding this thing up, it's not like when you're coding I do a lot of server work, where it's okay, we're building these server API endpoints. You're gonna send the server this and it's gonna send that back. Let's build it, boom.

Now it's like, you're thinking at a much higher level of abstraction, and that kind of hurts my brain sometimes. So I had a friend of mine come down here a couple of weeks ago who I used to work with at Foursquare. And we busted through a lot. He also helped me do a lot of things that were big improvements to just my workflow, for example, something as simple as, oh, make it a real command line that has history and where you could use backspace.

And it's like, I actually tried to pay someone on Fiverr to do that previously, and they weren't able to do it. It was some guy in Pakistan or something. It was like, we just asked ChatGPT, and we got it done quickly. And I was like, You genius. And he's like, I asked ChatGPT. And I'm like, Yeah, but that's what I need people to do. But alright, so elevator pitch, what is it.

Aaron: What is the use case? Why should we care? Okay, well, so many software, programming languages, why do we need another one? 

Max: Okay, so I want to combine several different ideas. And I think when I combine these ideas that are going to be some really interesting applications of it, that are possible, or perhaps more natural, that can't be done otherwise. So the first idea is the idea of functional programming and static data types. So there's the idea.

So there's functional programming, and then there's like, object oriented or imperative programming, this is a little object oriented too what I'm doing. But essentially in one paradigm, if you declare something as a piece of data, it does not change. And if you want to update it, then you create an entirely new piece of data. Okay, so I'm going to kind of combine that idea into a versioning system.

And so now, yes, if I'm pointed at certain objects, it won't change. But I could like find the next version. And so that allows you to do a lot of things that you can't do if everything's being overwritten on the fly, because there's a lot of optimizations that can be made. And there's a lot of problems that don't come up. So that's good.

And that's kind of a good basis, an interesting basis for a database, as it says. So we have this versioning. And we have this very hard type system, which is going to allow you to, which essentially proves that your code is correct in many ways. And so we're going to have a versioning on the type system. And so anytime you have data and newmap, the idea is that you have the data, but then you also have the interpretation of the data. And the data that represents the interpretation of the data, that's the type, is static, it doesn't change.

So you can like, store that file elsewhere, come to it in a few years. And it's like, okay, I still know how to interpret this data. So it kind of helps you make sure that things don't get out of date. And then another thing that you could do with versioning your metadata is like, okay, if I have an API, or I have a phone that hasn't been updated, that's still calling, calling my server the old way, well, every time it calls the server, it says what way it's calling it, it's using that code. And so it should always get back the thing that it expects.

And so I think I'm going to try to eliminate a whole class of problems that way. And then the third, interesting part of it is, okay, I want everything in this system, across the entire internet to have a unique ID. And then I have a certain way to do that. And then okay, so now you have like, kind of a decentralized, kind of Web3 type database, sort of like the web of data. And so the idea is you could sort of, I could say, okay, like, I want, I am running a decentralized Twitter. Here's what I expect a tweet to look like. Okay, I can receive from other sources tweets in that format, because they know that format.

So those are three, three ideas. I know, it's a little hand wavy now, right now I am sort of stuck. And I've been stuck for a while in the programming language phase. But why don't why don't I stop for now and let you respond?

Aaron: Well, so I don't know if this makes sense. But the first thing that popped to mind when you were talking through that is, could this be used for something like climate models, because I

know one of the controversies has been that so much of the original data is lost or inaccessible. And all that's available now is the corrected or there's another more nuanced term for it for how they've adjusted data. Yeah, but something like this would allow you to start with the raw data. And then you could as you apply corrections and adjustments and interpretations on top of it, you're preserving all of those intermediate steps. And it's got full traceability there.

So you never have had the concern of how did you get from A to D here? It's all encapsulated in that system.

Max: Yeah, if this system works, it's going to be very good at figuring out the provenance, the data provenance, which is like where it all came from, you still have the problem of getting all the data into the system. So you still need to do that. And you still need to write models to interpret it.

But hopefully, with versioning the models, you can be a little bit more organized with that. But that kind of goes to one of the first demos that I hope to create, which is oftentimes, when you're running a machine learning model, when you're when you're doing AI, essentially, you're learning through small increments, small updates. And so what does this system do for an update?

Basically, we're going to be going in and we're going to tell the system, hey, update, update, update, update, all day long, it's going to be updating. And every time it updates, rather than doing what a system does now, which is simply compute some new values and overwrite the old values, it's going to save the old state and then construct the new state elsewhere. Okay, fine. But what's the benefit of that?

The benefit of that is, now you can also communicate with the system from another process. So I can say, hey, I want you to change some of my hyper parameters, I want you to make the learning rate smaller, I want you to make the prior tighter, and then you could see how that affects your model. And I think it'd be cool to have some kind of visual representation of a model.

That is, and I have some kind of hand-wavy idea about using WebGL with WebSockets. So WebGL is a vector library for doing graphics through JavaScript on your web page. And then the idea is you use the WebSocket to be in constant communication with the newmap server that's running the machine learning algorithm.

So every second, you’re asking, what's the state, what's the state, what's the state, and then when you get the state, you render what it looks like. And then you could get the state again, and render what it looked like, but you're not the one actually updating the state, like maybe the server's updated it, maybe it maybe it's updated 100 times since the last time you've seen it, but you're constantly just your constantly just displaying the current state, and then maybe could like run those in reverse and forward at a certain speed, I don't know.

Or you could fork them all sorts of interesting things that you could do with them. So all of this is great. So long as I complete my programming language, which unfortunately, it always seems like, oh, I have like two or three features left, I have to build and then everything will become possible. And it's always just my graphs. And then it's like, oh, I need to do a reorg. And then I redo the whole thing, which is what I did, this weekend, which wasn't the first time.

Aaron: So was this literally a refactoring of the core code or something slightly different?

Max: Yeah, well, what I really realized is, so I've been through this code a lot. And the change that I made was that I wanted each type to actually be kind of like, a full description of what the type is. So for example, let's say I have a data type called. I mean, I hate to use this example, because it's so contrived, but this is like one of the first examples you learn when you learn Haskell, which is like, okay, let's say I have a type called shape. And I have two different types of shapes. One is a square with a side and another is rectangle. And that takes two sides. Oh, but then later on, I add a circle with a radius. Okay.

So, I think the idea is that those types can change over time. And when I create a value of that type, I not only say that it's type shape, but rather than just saying, Hey, this is type shape, I have to give a bunch of more data that tells me, specifies exactly what state what state of the world the world was in when I said shape. And so it knows, okay when the person tagged this as shape, this is what shape looked like at the time. So that's essentially what I did. And it took a while, because I had a bunch of.

Aaron: So it will know that shape version 1.0 supported rectangles and squares and shape version 1.1 also supported circles. And when you call for shape, you have to designate my shape 1.0 or 1.1.

Max: Right, right. And also look, I'm gonna have a lot more work to do, we're gonna have a lot more work to do, because that's exactly what I did, right? I have sequential versions. First of all sequential versioning: not always a good idea, but at least it solves the problem for now, right?

And then the second problem is, okay, I'm going to need more than just sequential versioning, we're going to have to know okay, well, what if what if the whole environment was forked? So I need an environment ID. And then ultimately, I'm going to need the Instance ID. And an instance ID is a single node in the system. So I have an instance where I'm running newmap, you have an instance where you're running newmap, and anything that's on my instance, I can prove that it was on my instance, because I signed something cryptographically. And so can you.

And we could run multiple instances there's no rule that's one person per, per instance. But that's how to ultimately make it global. But still, I think I'm going to try to solve the one person game for now, the one player game.

Aaron: Start simple.

Max: Yeah, exactly. Well, that's one of the things that I unfortunately did not do. But look, there's kind of a benefit to it. Because at work and whatever, you always have to start simple. You always have to start with the least abstract level, I want to try to do something where I was, I was trying to start more abstract trying to be more long run- ish by doing this a long time. I learned a lot from it. But now I'm ready to kind of pull it back a little bit and actually get something out the door.

Aaron: Yeah, there's, there's a constant tension between simplicity that has limited utility and complexity that provides a lot of use, but with complexity being a dirty word.

Max: Right, right. And usually, when you put out your first product, you should find, I think, what I wanted to do here is have a complex base or an abstract base and be like, Okay, this is, this is a great outline of, of what I want to build. But then maybe you have a simple use case to show. I'm kind of leaning towards the tools for data scientists right now. There are a lot of tools for data scientists.

Because I think even if I show, hey, this is a better way to do a logistic regression. Well, a lot of people are doing logistic regressions. And so they might be interested in that. And if it's dead simple, then then it might work out. So I don't know, that's one idea to approach it. But we go in a lot of different ways on this.

Aaron: Very cool.

Aaron: All right. So this is episode 298. We are approaching another tech retreat, which we've talked about previously. We're approaching episode 300. In a couple of weeks, I hope you'll join me for that. Maybe we'll do…what do we want to do? Another look back? I don't know what we want to do here. Do you have any ideas? What do we do for the other hundreds?

Aaron: Yeah, it's tricky because they usually also fall pretty close to our year end episodes. I guess they're they're drifting further and further apart over time, but.

Max: Yeah, I want to know, I mean, what's my all time, what's our all time listens? Are we at half a million yet? No, not quite. We will be soon. That'll be pretty cool. Let's actually see what we did here. Episode 100. We did a look back on the first 100 episodes. Okay. So bad. Episode 200 is lookback. Milestone meta chat.

Some interesting stats from the first 200 episodes. Um, I don't know. Maybe we could do a look back of episodes, because we haven't done that in a while. Seems reasonable. I find it's if you have if any one in the audience wants to make another suggestion we are all ears.

Aaron: So going back and doing a highlights clip show and Greatest Hits is interesting. But that sounds like it requires work. And so I kind of push against that.

Max: No, no, it'll be more like, Hey, what are your favorite episodes from the past? We're not gonna we're not gonna make clips. Look, we're gonna have AIs that are going to be able to do this. We're going to have AIs that can read all of our all of our conversations, going back to the beginning and pull out clips appropriately. So we don't have to do it manually. That's just waste work at this point, I think.

Aaron: Yeah, I mean, I don't want to depend on ChatGPT. But I am mildly curious. If we fit in the full contents of the Local Maximum website, if it would tell us what it thinks the highlights are? 

Max: Well, I plan on feeding it the transcripts once I get my act together, we'll see what we can do with that. That'll be very interesting. I think my first trick there, and I'm gonna, I'm gonna give it away. Firstly, what if I fed it my documents, including with the Constitution, as well as our discussions here? What if it could pull out the best quotes for tweets?

Aaron: Just a Local Maximum bot?

Max: No, no, it'll just go out on my own Twitter feed. So long as it's interesting enough, like I'd approve them, but I don't know what could possibly go wrong?

Aaron: Well, when Twitter or should I say X is involved, so much can go wrong?

Max: Yeah. Well, yes, but so long as I approve the tweets, I think it's pretty safe. So that's where we are today. All right now.

Aaron: Is ChatGTP still your go to, for the large language model AI assistants, or if you've been tinkering with some of the other options out there?

Max: Yeah, yeah, ChatGPT is still my main one. Don't call it ChatGTP. That's, that's not anything different. But it's just, it just hurts my brain. Just like when you sit when someone says instead, we're going to stop and shop, we're going to shop and stop. And all of a sudden, my brain starts to be like, argh.

Aaron: There’s still a chance that I'll get it wrong. So I rolled poorly,

Max: I feel very uneasy. Okay, sounds good. I'm sure I made that mistake. Several times in the podcast as well. I get it. It's generative, pre trained transformers. Okay. Yeah, exactly. But I don't know if they should have named the GPT. Because that doesn't make a whole lot of sense. You don't think generative pre trained transformers.

Aaron: It certainly wasn't a name that they came up with for marketing purposes. I can't remember what Microsoft says. Are they just calling it Bing? Or do they have a specific? Oh, I know Google has Bard

Max: Yeah. I don't know.

Aaron: But anyway, I've only tinkered with them a little bit of late. That tab has gone largely unused in my browser.

Max: Yeah. All right. Well, which one, Bing, or OpenAI?

Aaron: All of them. Although I guess I did flip the switch on my phone recently. So that I'm getting whenever I do a Google search, because I've got an Android phone. So that's my default search on the phone. It is adding in the AI generated results and analysis, which I don't know how I feel about yet. I usually disregard it.

Max: Yeah, exactly. When Google starts typing to me when I did a search, I'm like, No, I'm here. Because I didn't want to go to OpenAI. So yeah, it's very interesting how this is turning out for me as a consumer. I think that it has to be integrated in search in a way that makes sense.

Aaron: Not just AI, but I think we've seen a should-have-ication, trademark term there, of search, especially with the big players in that area in the last few years. Yeah, mostly driven by the advertising piece. But yeah, but I think AI is potentially going to make it worse, not better.

Max: But you get a bunch of ads, a link to reddit, a Wikipedia link if you're, if you're lucky. And maybe some like news articles from, like, very few approved media sources. Like that's what you get these days. Which do you remember back in the day when you used to search for a news item? And then Google had like a streaming list of things that people were saying, like in the last hour about it? They had that like ten years ago. They ripped it out.

Aaron: I do remember that news.google or google.news, whatever it was used to actually be a a decent resource. doesn't feel like it today.

Max: Oh, yeah, you’re better off going right to the New York Times.

Aaron: I don't know whether to blame that purely on the search engine or on some of the other news producers and aggregators and how they operate. But, either way, the result is not desirable.

Max: Yeah. All right. Well, on that note, Aaron, let’s see if we wrap up today. Any last thoughts on everything we discussed?

Aaron: Well, the only other thing kicking around in my head is that I did recently uninstall Reddit from my phone. So maybe that will make my life more efficient.

Max: That is your last connection with the outside world, I believe.

Aaron: So not quite, but I was definitely more of I haven't deleted my user account. But it's no longer easy for me to go down those rabbit holes on my phone.

Max: You know, if I announced on the podcast every time I deleted Reddit on my phone, we would have probably had 40 different announcements by now. And then somehow it ends up right back there on my phone.

Aaron: They do prompt you every time you go through there for the browser. Would you like to do this in the app instead?

Max: Yeah. And then it's, yeah. Because then I'm always like, why? What do I have better to do right now? I always have some justification. And it's I don't know. I don't know if it's making my life better. Probably not. There's nothing I need to check on Reddit. So I gotta remember that. Gotta remember that, folks. All right. So let's call it a day. Hope to see you for episode 300. Just a few episodes away. Have a great week, everyone.

That's the show to support a Local Maximum. Sign up for exclusive content and their online community at maximum.locals.com a Local Maximum is available wherever podcasts are found. If you want to keep up. Remember to subscribe on your podcast app. Also, check out the website with show notes and additional materials at localmaxradio.com. If you want to contact me, the host, send an email to localmaxradio@gmail.com. Have a great week.

Episode 299 - Sam Kamani with an Intro to Web 3

Episode 299 - Sam Kamani with an Intro to Web 3

Episode 297 - Stephan Kinsella: The Fallacy of Intellectual Property

Episode 297 - Stephan Kinsella: The Fallacy of Intellectual Property