Matt Mullenweg started out as a jazz saxophonist, and went on to create WordPress, which is the platform behind an astonishing 42% of the websites in the world.
We chat with Matt about his journey from musician to developer to entrepreneur, his perspective on distributed work, and his thoughts on the transformative capabilities of the latest generation of Generative Artificial Intelligence.
We also come back to Matt’s roots in jazz and his continued love for music and musicians.
Bio
(via Wikipedia)
Matthew Mullenweg is an American entrepreneur and web developer living in Houston. He is known for developing and founding the free and open-source web software WordPress, and its parent company Automattic.
After dropping out of the University of Houston, he worked at CNET Networks from 2004 to 2006 until he quit and founded Automattic, an internet company whose brands include WordPress.com, Akismet, Gravatar, VaultPress, IntenseDebate, Crowdsignal, and Tumblr.
Notes
Moravec's paradox: https://en.wikipedia.org/wiki/Moravec%27s_paradox
Jazzy Claire de Lune: https://ma.tt/2016/02/jazzy-claire-de-lune/
C Blues Jam E x p l a i n e d: Jacob Collier and Cory Henry Go Crazy:
Summary (via ChatGPT 🤖)
Matt Mullenweg, the founder of WordPress and Automattic, discusses his journey from a jazz saxophonist to a web developer and entrepreneur. He talks about his perspective on distributed work, the transformative capabilities of the latest generation of Generative Artificial Intelligence, and the success of WordPress, which powers 42% of the websites in the world. Mullenweg explains how open-source software and community-driven development are key to WordPress's success.
Highlights
📈 WordPress powers 42% of the websites in the world.
🎶 Matt Mullenweg started out as a jazz saxophonist before becoming a web developer and entrepreneur.
🤝 Open-source software and community-driven development are key to WordPress's success.
🏢 Automattic is the parent company of WordPress.com, Akismet, Gravatar, VaultPress, IntenseDebate, Crowdsignal, and Tumblr.
💻 Mullenweg dropped out of the University of Houston and worked at CNET Networks before founding Automattic.
🧑💼 Mullenweg believes that community-driven development creates a more responsive and agile product than a top-down design system.
🧠 Mullenweg takes a long-term view and is informed by political science and macroeconomics.
Help us make the show even better by taking a short survey: www.dbtr.co/survey
If you're interested in sponsoring the show, please contact us at: sponsors@thecuriositydepartment.com
If you'd like to submit a guest idea, please contact us at: contact@thecuriositydepartment.com
This episode is brought to you by:
Fable: Build inclusive products: https://makeitfable.com/designbetter/
Freehand by InVision: The intelligent whiteboard that's half the price of Miro and Mural:
Methodical Coffee: Roasted, blended, brewed, served and perfected by verified coffee nerds:
(use code "designbetter" for 10% off of your order).
Cruise: We’re a team of researchers and designers creating a self-driving transportation service for the people and cities we love. Visit design.getcruise.com to learn more about how you can help design the future of transportation!
Transcript
00:46.90
Eli Woolery
I'm curious at what point WordPress actually started and then maybe we can walk into the present, because the last interview I was listened to was maybe about a decade ago with Tim Ferriss, and at that point, you're saying that around 20% of the web was built on WordPress, fast forward to now and it's like 42% — holy cow that's crazy! How did you get there from here?
01:33.85
Matt Mullenweg
I was really learning a lot particularly by participating in online forums whether that was news groupsups or bbss or um, you know, using early forum software um and the equivalent of github at that time was this site called Source Fourge which is where it was kind of like a code repository and bug tracker and everything else for open source projects much like github is today.
I was learning so much from blogs I decided to start my own and I went through some of the different dominant software at the time which was, moveable type or livejournal, blogger basically trying them all out and eventually ended up on something called b two which was pretty early at the time but it was written in php which was my favorite language maybe still my favorite language and um and was fully open source.
So it was under what's called the gpl license where some of these other things like movable type were quite good but they weren't open source. They were under proprietary licenses and so at this time I'd already become radicalized to believe that like open-source software was very important and what I want to spend all my time and development effort on.
I just began contributing to b 2 as one of these things where again total amateur knew nothing but eventually after my twentieth time visiting the forum I saw someone else ask a question that I already knew the answer to and so you know you learn some. You always get back. So as part of giving back I started to answer other people's questions.
03:09.90
Matt
And that led to writing code to help people which led to contributing to b two and um, sort of the interesting twist in the story was that ah b two is really driven by a single developer and he disappeared just without any messages or anything and so the community. Which I was part of at the time wasn't sure exactly what to do? Um I posted about this to my blog and one of the other contributors a guy named Mike little who lived in the Uk left a comment saying hey I was basically saying like hey you know b two is not developed or what do we do. We can take something continue the b two code base with you know the best parts of these other systems like text pattern which is done by Dean Allen Moveal type other things and he said yeah do you want to do that I want to work with you and so we just started collaborating together from across the atlantic ocean me in Houston again a kid as like a freshman at college. And um, no real qualifications him is like a real developer who's a professional developer at that point but I had a lot of time on my hands you know youth is blessed with the gift of of lot of time and so um, yeah, we worked together. We eventually started calling that WordPress.
04:23.99
Matt
And we did our first release. Um later I forget the exact timeline but maybe a few months later Michelle who had developed b two kind of returned. You know he had just gone through some personal stuff. That's why he kind of logged off for a little while and um, he didn't ah really want to work on b two as much anymore.
04:43.79
Matt
And you know we were what was called a fork of b two meaning we had taken the existing codebase and started developing on top of it which is something you're allowed and encourage to do in open source. There were 4 or 5 other forks so 4 or 5 other projects that had kind of done the same like groups of developers and individuals and um. Michelle looked at all those and said hey this WordPress thing seems like the best so this is now the official continuation. Um, so we were kind of blessed as the official successor to be to which again also started to bring over some people um to how that led to this success that WordPress has had um. There's not a single thing I can point to it's except being extremely responsive to um, our customers and community you know and also something else I think is different from WordPress about most software projects is that it really is community developed. Even though I have a company which contributes a lot to WordPress. We're typically only like 10% of the contributions. So 90% is coming from either other companies or other volunteers and that sort of community-driven development when it's done right? I think creates ah very responsive and agile products. And also just allows us to move faster. You know ourrie tear competitors like ah like a Wix or Squarespace might have you know 500 or even a thousand engineers doing r and d to develop their system. Um, but it's no match for the tens of thousands or hundreds of thousands of people.
06:16.20
Matt
Develop WordPress and plugins and themes and everything so just over time assuming again, you're you're being responsive and iterating well you're just going to develop something better much like the Wikipedia when it launched was much worse than in carta or Encyclopedia Britannica like it was measurably way worse but then over time. It got better and better and better and better and that's the same thing that happens with open source and why I think open source is somewhat inevitable to you know dominate every area that it enters right.
06:46.46
Aarron Walter
Maybe we could dig into that philosophy a little bit further because there's there's sort of like ah ah, some might might categorize it as a compromise that comes with that open source many contributors that. Things can get a little messy. You know there's a lot of different contributions and is there like a grand vision that's uniting all of it. But somehow that does happen with WordPress where things hold together into like ah the sum being greater of all of all of its parts.
And then there's kind of the closed ecosystem as you described. Ah you know Wix or a Squarespace where it's like it can be perfectly controlled but its pace of iteration is very limited. Could you just maybe unpack those philosophies like there's 2 different directions. Ah. You talked about being indoctrinated into open source that that was something that you believed in very deeply but as you think about this from a commercial perspective. How do you decide to take a right turn instead of a left turn and what were the concerns that emerged from that open source. Philosophy.
07:54.73
Matt
And yeah, this is an interesting question because people have a really hard time imagining how a chaotic open develop system can outcomp compettes sort of like ah a top-down design system. Um. So perhaps to be instructive. It might be good to Zoom out from products or software to just capitalism versus kind or some other you know top-down economy or a control system. Um, you know human progress is incredibly chaotic and includes you know thousands of uncorted.
09:11.14
Matt
Um, a great you know quote to illustrate this is actually from Adam Smith and now the wealth of nations from 17 76 around the time America was started and he said it is not from the benevolence of the butcher the brewer or the baker that we expect our dinner. From their regard to their own interest. We address ourselves not to their humanity but to their self-love and never talk to them of our own necessities but of their advantages. Okay, now we've seen this work now over past four hundred years of of human development. And in cases even 1 with a lot of really good intentions. We've tried to create economies that were kind of top down or controlled by a single entity or a group of people a polypbu. Whatever it might be um, that just hasn't been as effective. Even though theoretically I I am also drawn to those models right? It's just makes a lot of sense but in practice it just hasn't worked. Um, so think of a healthy open source ecosystem as being an ecosystem that is much like an open free market. So now. When you look at Wordpress and say okay, it's a little bit chaotic. It's sometimes not as organized. Um, sometimes our first versions are really you know junky and we have to iterate away to something that's good, but it's really driven by common philosophy.
10:41.31
Matt
Ah, common set of beliefs in our you know ecosystem or country. That's a philosophy around the freedom imbued an open source software being really important. It's a a desire for user control you know and including privacy and customizability something really key to our tenets. And you know some cultural stuff. We're really in the jazz so we name releases after jazz we have talked about Otis Poetry. So We're always trying to refactor our code to be more beautiful but pragmatic that what matters to us more than anything is not beautiful code but a beautiful user experience and that not just and you know.
11:00.48
Aarron Walter
How nice.
11:19.23
Matt
How nice a button looks or how good a single interface is but the entire totality of that user experience including how easy is it to find a developer how well does your site indexed by Google how easy is it to like you know, customize your site by installing a theme or plugin like. Those types of things aren't necessarily captured if you were just put a screenshot of Squarespace in a screenshot of WordPress side by side but those were the things we obsess over what are the long-term incentives not just over the following year. But the following ten or twenty years of the community. We're trying to build.
And so did the extent. We've been successful I Think that's that we take a very longterm view a view really informed by political science and macroeconomics which were two of my other passions and um and just making the into ton mistakes but always coming back and ah.
12:11.46
Matt
And trying again. So I guess a persistence we haven't given up over many many now 2 decades of doing this we turn twenty this year and honestly I had personally been to work on WordPress the rest of my life. You know as long as I am intellectually and physically able to I plan to keep working on this mission to democratize publishing and um.
12:21.37
Aarron Walter
That's incredible.
12:32.99
Matt
Looking forward to doing so alongside Hopefully some of the current people that also do it but perhaps also new generations that want to contribute and be part of this this mission and you might if we take a quick pause right? there we're at back.
Okay, um so Matt that's this is a fascinating philosophical approach to a product to a way of building to a way of building a company and I wonder if we could. Maybe unpack that a little bit more um because I don't think a lot of business leaders think the way that you're thinking you know you're describing. It is basically connecting a lot of what seem like disconnected concepts but ultimately are they inform 1 another. And there's a thread in your in your personal growth as someone who studied music which feels a world away from being a developer and you know building computers for people and then ultimately building companies. Ah, but they these. These things are connected I wonder um, ah and Brian can edit this down to a more concise question. Um, could you just share a little bit about how you think about the world and systems. It seems different and unique.
15:00.44
Matt
Um, thank you I also am not known for my concise answers or my grav so I apologize if this is already going longer. Um I would actually disagree that not many businesses are run this way I think that if you look at. Businesses that have gotten to scale or um, been at the top of the totem pole so competitive in an open market for decades. It would you would be hard pressed to find one that doesn't think in larger systems, economies and and standards and so. Even when I've never met you know like I've never met bill gates or something like that I've studied Microsoft deeply because look at what they did over the 80 s and 90 s um wow so much to learn from there. They created one of the first true computing platforms and did it like but. Somehow behind the bag of ivm like the other dominant companies at the time I think it's also good to study what happened to them in their early two thousand s and today we're getting another master class from the same company Microsoft on success planning you know with Sauty and Adella there. Performance that he's had over his tenure and also just how they're making Google dance as he said with their investment in openai. Um chat chebt all these sorts of things are quite interesting and bringing back things people assumed to be dead like you know internet explorer and it's now called edge and.
16:34.88
Matt
You know, bing are now part of the conversation again. Um, and then if you look if you dive into that story. You get a little bit behind the scenes and see that oh wow Reid Hoffman was at the center of a lot of this at the center of you know social network duration at the. Have board of open Ai you know created Linkedin obviously which is sold to Microsoft he joins a board of Microsoft like okay, who's this for gumflike character behind the scenes and check out his podcast and check out his writings and so I think that there's so much to study from and if there's anything that I learned.
At Hspva and in that arts high school. It's that um, studying areas outside your own is a great place for inspiration and also that there's something to learn from almost everything including things you strongly disagree with or you know have a negative reaction to. For example. I love apple and Apple products and you know think how for a long time. They wouldn't allow you to set defaults or install apps and everyone was jailbreaking their phones things like that like so that was also I think informative um, in terms of you know, even giving their position.
17:37.70
Aarron Walter
Matt
And their incredible powers and you know at one point they were the most valuable company in the world trillions of dollars more cash from their balance sheet than most countries and yet they're still kind of behaving in like an underdog way that to me feels kind of consumer hostile I get if they were.
17:57.89
Aarron Walter
Um, right.
18:07.39
Matt
You know, early two thousand S and about to go out of business again. Why they might behave that way but from the point of view of their marker position. It feels like they in my opinion should move with a different level of grace and sort of of responsibility to their position in the Market. They've been. Rewarded by consumers with an incredible position and now I feel like they have a bit of a duty to also give back to those same consumers.
18:36.67
Eli Woolery
Matt a few minutes ago you were talking a little bit about you know the potential future for WordPress and bringing in new people and then you also touched on openai and I'm curious about your perspective of the sort of interface of these new technologies. Um machine learning and artificial intelligence and then. You know people that are kind of growing up with these technologies. What's what's going to look like in the next few years as these become more commonplace and more powerful.
19:04.75
Matt
It's a great question. Um I think it's interesting to look at what's been truly society changing technology of the past twenty years and also where we maybe had some detours that got a lot of Hype but warrant. As useful so when I look at the current chat interfaces to ai I think the easiest line to draw to is what's been the most successful ai of the past twenty years which is Google search engine. What. You could ask it anything you like you could structure the queries as words or sentences and its interface in returning it back was a list of links and more recently also providing some answers at the top of that you know that little box that sometimes answers things. They also created a whole ecosystem around webmasters and seo and everything which got millions of people around the world creating answers for them optimizing it so that Google can the googlebot can read these easier and then um, they were kind of give them a little cut. You know they'd send some traffic to the website which by the way. Probably running Google Analytics and Google adsense so they were kind of making money in this distributed ecosystem in a way that um was very very successful for a while. But then the incentives of this ecosystem. Um, kind of got worked.
20:34.85
Matt
And so I'm sure a lot of us have experienced in the past few years a degrading quality of Google search results one around they are doing a thing that they said they would never do which was have a lot of ads at the top and in fact, are mobile now for certain search terms especially related to WordPress because those terms are very valuable. You only see ads the organic results are not visible Second those ads look just like the organic results more and more every year they used to have different color backgrounds and everything and they've suddenly evolved that ui to where so an average consumer at least when I've observed people do searches. They have no idea what's a paid result versus that organic one and you know there's kind of a fig leapaf of respectability there where Google says oh it says sponsor there and you know they can kind of get by it. But I think an actual user testing at least in my anecdotal experience like people are not distinguishing. There. Paid results versus organic and then finally even in the organic results. Many of these websites are so driven to monetize that they create a really bad user experience with bad ads that are slow to load popups everywhere or content which was essentially a content farm. Generated to optimize for Google search and isn't necessarily like the best organic it's not like an expert like you are someone posting their answer to this question. It's like a website created by someone trying to optimize for Google so the vary system they created and the incentives there and ended up being the yeah.
22:09.75
Matt
Seeds are their downfall and that's often true often contained in every success is a seed of the thing that will eventually make that success a failure if you don't change and adapt. Ah so now look at the sidebar I think the sidebar was these voice assistants. And everyone created these smart speakers Alexa was very very popular and we were told this would be a conversational interface. However, in the reality of how it was really used and again you could just go to your friend's house and watch how people use this they were using for very very simple queries like setting timers. Alarms and really playing music is at least what I observed I'm sure Amazon has better data on this but like again, just go to your friend's house and and ask them to use things and observe and always try to have that like learning beginners mindset and watching other people use things. Um, and if you tried to do something more complex. It didn't work. Well. Also because you didn't have really a screen they started to add screens to these things like the echo devices started to get screens which is really nice. You could kind of see what was playing or or see answers to questions in a visual way but that kind of like iterative conversation impression was not good and it wasn't until chat gdp. Gpt which to be honest is not really new technology from an Ai point of view but was an interface and openness you know is freed to everyone for anyone to try innovation unlocked right and inspired like an entire generation of things now being built around this.
23:46.60
Matt
And they made the apis open and they just lowered the prices by 10 x and like openai is really nailing it on the execution in such amazing ways and they raised funding in a way that allows them a competitive moat to be able to do it. You know they just raised I think $10000000000 so they're really executing on a number of areas including business strategy. Which I think is igniting this um new ecosystem look at what else is happening now there's there's like 500 other ai startups with check hitting funding because all the vcs are like oh wow now we can pour money into something and like there's a lot of overcapitalization in the venture. Feels right now and so we're going to get like 4 out of that 500 there's gonna be 450 failures 40 that are probably like good home. You know home runs and 10 that are probably less probably like 1 or 2 that maybe have an open a I level of success but you know what? that's okay. That's messy. That's chaotic, but it's also like how open markets should should function and everyone doing that is doing it. You know with open eyes all the investors. Expect many of those things to fail all the you know the self-interest of all the entrepreneurs in that space is this its see so that's cool. You know that's kind of the system working. Well.
25:06.89
Aarron Walter
How do you think Ai is potentially going to change the way that WordPress operates. It seems like there's just a myriad ways that it could make ah you know using WordPress setting it up publishing just much richer.
25:23.56
Matt
Yeah, um, in ways, it's hard to articulate and that are massive. So I mean think of it because just like so they say like people don't want to drill. They want a hole in the wall. When people Google they they don't want a list of 20 links that they have to sort through. They want an answer um and these conversational ais even chat Gpt today you can have it write WordPress plugins with code that works you can copy and paste it and execute it. Um, you can ask it like what's the best.
25:50.63
Aarron Walter
Law.
25:57.61
Matt
Like it's a very and there's already like a ton of startups start doing something that essentially like creates an entire website just from a few conversational questions and you pair that with like generative Ais around images and how all of a sudden you can have like really customized photography or easy editing of your currents. Imagery like maybe have a bunch of pictures but they're not very high quality. You'll want them all a certain illustration style again things that you know I can't remember how many dozens of hours I spent like learning Adobe photoshop and illustrator. Ah and now with some of those same effects.
I used to have to take long tutorials on and read like bid site point you know articles on now just being a single you can ask the computer do it and it magically happens. Um wow like and that's you know Wordpress has always been about the democratization of publishing since we started saying that.Matt
Everyone and their mother has adopted the democratization phrase in ways that I don't think are really true to the intention. But I think this conversational Ai is actually truly doing it because it is going to allow literally millions of people who are able to type a prompt. Um.
To do things that used to require you know tens or more hundreds of hours of learning to be able to accomplish on your own and to finish the analogy of Google search to these new. Let's call chat gbt is just a a stand. In example for all the conversational ais. Um. It's remarkable like all of us have probably gotten pretty good at crafting search terms and this is a skill. We didn't even realize we were doing but you talked to anyone who's like good at the internet and how we search is very different from how ah like a novice or less technical person uses Google.
We know how to craft the queries how to do site searches how to remove certain terms from the results how to scan through and go through results really quickly to find the right answer. This is a learned skill. No one's born with this but we've all learned it the same way I think that prompt engineering or asking the ai. Right? Question instructing it in the correct way is going to become a skill just like it like searching Google was a good skill because think about it. It's a computer interface and we've evolved from command lines like on linux where you have to type certain specific commands have. Specific argument orders and pipe the results between different sets of commands. You know piping my grock to an and nc command and then piping that to like a file which then I I parse of the script like think of all those steps now I'm able to command. Ah.
28:41.54
Matt
And Ai basically I do think of these things as new intelligences to do the same thing. Um people who have run businesses or managed contractors are going to be naturally pretty good at this because if you've ever worked with someone else who you're directing their work how you ask them? what to do. Is actually really really important to them be successful at accomplishing the task I'm sure we've all experienced where like we hired a designer or developer and we didn't ask them quite right? How to do something or didn't clarify it and you end up with a bad result and then you have to iterate and everything. Um.
29:02.19
Aarron Walter
Absolutely.
29:17.37
Matt
It's the same way asking these Ais and and so a lot of people we're starting to adopt Ai a ton inside automatic and we've had developers that report that it's 3 x to productivity 3 x that's ridiculous. Um, and these are good developers. Ah yeah.
29:25.81
Aarron Walter
Wow.
29:32.95
Matt
And a lot of people who've tried it out are like oh I tried it and it just didn't work for me like it seemed kind of dumb and I think that experience isn't driven by the ai being dumb. It's being driven by not asking it what to do properly and knowing what to do with those outputs and feeding them back in in a proper way and other things. So. Now of course as these technologies advance and you know it should get easier and easier and better at guessing what you want to do or better at iterating on some like right now there's at a limit to its responses as limits on like um, the big version of chat and how many questions you can ask before it resets but all of that's going to evolve and go away. And the rate of change here in all of my experience thus far as a technologist you know, call it the past twenty years um I've seen major changes happen on the order of usually like 6 to twelve months and often taking years to really iterate and get good. Google used to shift a major algorithm. Update kind of every 12 to 18 months sometimes a little longer WordPress does major releases. You know every four months squarespace will do a major release or drupal will do a major release every two or three years squarespace kind of similar. Like so there's like different cadences. The advancements of these Ai Technologies is happening on the order of weeks I've never seen anything like it and so as good as these things are that's why I say it's hard to predict exactly the impact the stuff that is going to launch this year ah
30:53.30
Aarron Walter
Wow.
31:02.43
Aarron Walter
Um.
31:07.62
Matt
Blow your mind if this current stuff hasn't already blown your mind the stuff that's coming is not just a little bit better but an order of magnitude and so I would just encourage anyone listening to this to really spend time with these tools and play with them because it's. Potentially um, obsoletes all of your current knowledge of how to work with technology in the same way that being like the most amazing buggy whip maker wasn't useful for like working on cars. Ah. There's perhaps some meta skills like learning or communication or other things which were transferable but the actual kind of applied skills um are completely different and I would just encourage people to spend a lot of time with these new areas.
31:58.39
Eli Woolery
Now you seem to have ah a pretty optimistic take on on where this is headed and I think overall Id too. But you also read you know the potential for some darker things and as these systems become more advanced and sort of more of a black box. How do you think we can prevent. You know some of the potential negatives like Bias or um, you know, just just the the machine itself having its own intent that that maybe contradicts what we as humans want have you thought about that much.
32:28.31
Matt
Ah, yeah, it's called open source. Ah so you called you called it when you said this is could be a black box or we can't really see how these things operate um I think open source and open source has fast followed with things like stable diffusion again. Things I'd spent to to take years happen in months which is wild now. The bigger problem isn't that we can't see how these things work it's that we don't fully understand how they're evolving including the top researchers at Google and Openai. We're seeing emergent behavior for example, training something in one language and it's somehow learning another language along the way without being trained for that. Um wow and even the the people developing things these things. Don't really know and there's actually a paper about this where I forget which model it was I think it was one of the Google ones. Um they found that it could answer things in spanish with about a 15 decrease in accuracy but it was wasn't trained on spanish at all. It has picked it up from um, essentially in its own. How these systems work as they develop kind of a vector based space for certain concepts part of what we don't fully understand but what they seem to do is they put concepts near each other and near is a relative term because it's essentially not spatial but is a good metaphor to think of it what much like in the human mind. Um.
33:57.21
Matt
Lets say the concept of a chair you and I can look at things different chairs that could look radically different. You know, think of an ames chair versus you know what's the ghost chair the louis chair or something versus like a cafeteria chair versus like. A chair in a classroom that has a desk attached or you know you know think of all the different things and it would take us milliseconds to know like that's a chair all of these different concepts I just mentioned in our brain are stored in a similar area. The neurons for these things when we like. Functional Mris and watch parts of the brains light up are near each other like literally spatially in our brains and that appears to be happening in the databases and storage structures of these Ais that they're storing lion and Tiger and Jaguar and cat near each other. And sort of our emergent taxonomies that are like the taxonomies that we might explicitly create. But again, they're not being explicitly trained on this. No one is telling them here is the taxonomy of like all evolution of felines and here's the big ones here's the little ones here's the best you know it's just kind of figuring that out on its own. So that's actually the the black box that I worry about more um and then just on the final thing you know you said how do we eliminate bias and other things from these systems again, really good question and especially given some of our experience not just with technology but with humanity over the past.
35:32.74
Matt
2000 years but let's call it even particularly like the past like 60 um, we should be thinking about this. But also I think a good way to think about it is not eliminating bias because we appear to not have been able to do that in any system including our own. Um, even the most enlightened human. Like ah I don't know so people might pick that differently but assume that you have very progressive views yourselves or let's say the dally lahoma or someone like that. Um, still when we do brain measurements will have differential reactions including like how many milliseconds it takes to respond to things based on. And color and other things. So to me, it's not about eliminating the bias. Let's assume that that's an emergent property of just any sort of neural network learning system. Um, but instead having a step afterwards which says okay, this was maybe a biased response. Let me observe that. And then do a second response based on that who cares? Honestly, if the Dalai Lama or something like that responds to something in their brain in seventy five milliseconds versus ninety milliseconds based on the color of their skin. Let's just assume that that happens and say okay well what what's the next thought. The next thought in my opinion should be that doesn't matter I am now going to make a decision not based on that sort of built in neural network response rate but on some more enlightened principle um or something that's derived from like a moral framework believing.
37:06.36
Matt
As I do you do like that's those surface things don't matter that it's you know the depth of someone's character or their actions which we should judge them for and respond to them based on or some other thing some person far smarter than me some philosopher some you know civil rights advocate has like taught us all. Um, that's what we want to make our decisions on. So really, it's about taking these systems and not just responding and whatever the base layer responses are but then building in levels above that that are essentially like must like us have multiple levels of reasoning the Ais can and do actually multiple levels of reasoning. Where they check their own initial responses and then iterate based on it before they send a response this is much like the sort of safeguards quote unquote safeguards that are built into these systems to not send like you know, angry or bad words to you? Maybe that's their initial response but then they have something else like another program that runs and says oh. Maybe I shouldn't curse out Aaron let me then create a new response based on like you know something else which is again very similar to how our brains work and how humans work. Yeah.
38:11.48
Aarron Walter
Yeah I mean this is one of those areas that if if we're going to be successful and avoid really um, you know some some negative situations. We can't even fathom just yet. We really need to think laterally and and kind of. Ah, adjacent topics philosophy political Science economics. All all kinds of different things need to inform the way that we approach ai and how we integrate it into our technology into our design philosophies et cetera curious like where. What are you reading? and um, like how do you think about this like bringing a broader view to what is to come.
38:54.54
Matt
It's good. Question. Um, again, the best metaphor here is that what we appear to have been able to do with this modern ai is replicate how the human brain learns. So or it's informs by how the human brain. Learns and works and we're getting really cool um results there including from capable things. Computers are good at that humans are bad at like reading the entire internet you know like all of us could maybe hope to read you know, call it. 30000000 words in our lifetime and it can do that in a minute so it's just going to be able to answer certain queries or or tie together certain concepts or knowledge better than we can. Um, now there's certain things that we're way better at so and this is I forget the name of the paradox here. But there's basically something that says things we thought ai would be fast at like driving a car are actually taking way longer than even the experts in the field predicted and things we thought it would take a long time like creating illustrations. Ah. Or like what we consider creative work writing poetry. It's actually gotten really good at really quickly by way a fun thing we chat shebt if you want to play with it. ask her to write a poem or ask her to write a haiku or a limerick. It's actually incredibly good at these and then once it writes a limerick say like hey.
Um I don't like this rhyme could you change it or you know could you add more? could you make it funnier or could you add a joke at the end and just watch how it does that it'll kind of blow your minds. Um.
40:35.88
Eli Woolery
I've done that by kids and and then had it actually read in the voice of Snoop dog via this speech if I app so you can like take it to another another level too.
4041:24.27
Matt
That's close. It's not actually the one I'm thinking about but is a different paradox that describe the same thing. So the polai paradox which is named for british Hungarian Philosopher is a theory that human knowledge of how the world functions in our own capabilities. To a large extent beyond our explicit understanding another good concept to introduce here for your listeners is called the lindy effect the lindy effect basically says that something that's lasted a long time is more likely to last a longer time. So let's talk about this with books. A book published last year has I think a low percentage likelihood of being read 100 or or thousand years from now now let's take a stoic philosopher or a greek philosopher or some of the plays that thousands of years later were still not just reading but studying. And saying there's something very special here. The works of Shakespeare you know, think of art that's lasted the test of time I could say I would bet actually that in 1000 years we'll still be reading Epititis or Marcus Aurelius because there's something in the lindy effect. It's lasted a long time. So it's likely to continue lasting. And new things maybe have a lower likelihood so in terms of what to read and when I try to read is um, the works that have lasted a long time and like we referenced Adam Smith earlier you know in colleges they still study Adam Smith it's been 3 400 years three 3 hundred 300 years basically
42:53.19
Aarron Walter
Yeah.
42:55.37
Matt
And we're still studying it so there's some good stuff there. So um I think that you should kind of balance your reading if you can and I'm not great at this because I also love twitter and magazines and news. But like if ideally I were designing my my weeker month I would spend 60 to 80% of my time. On old stuff maybe new interpretations or new criticism of old stuff but like really based on ideas and things that have stood the test of time and maybe like that 20% on new stuff including Twitter can do cycles like. Again ideally I don't do this perfectly. But if I could design my life perfectly and I had perfect self-control. That's what I would do.
43:40.80
Eli Woolery
Matt Automatic has been a remote company and 1 of the larger remote companies from it I believe it's very beginning. Um I was I've been in invision number of years and for a while we were you know one of the bigger ones too and then the pandemic happened and this way of working shifted. And now it seems like a little bit of the pendulum is starting to swing the other way to some degree for some companies where they're asking people to come back. What's your what's your current take on remote work and what are you know and some of the successes and failures associated with it.
44:12.22
Matt
Um, so my experience has been that working in a distributed fashion I don't like the word remote because that implies essential and a remote distributed implies that every contributor is at equal footing. So I'll say working in a distributed fashion. Um, you can create world-class products experiences companies, everything. Um, that does not mean that everyone who works and distributed will do so nor does it mean that everyone who makes everyone go into the office every day is going to be more successful. It appears that there's a lot of other factors. Ah because gosh how many in-person companies or companies that force everyone to go into the office failed most of them actually that have ever been started have mostly failed. Um, so again like i. Think we need to be careful when people try to draw causal relationships from correlational data which I think folks are doing if I have to guess why some companies that worked in a distributed fashion are now asking people to come back. It's that many managers. Who have worked in a mostly in-office environment thus far feel more comfortable with their skills in that in-office environment or in their ability to monitor or control their employees. Um the past few years I would say that the balance of power was firmly in the.
45:44.19
Matt
Ah, beside of employees because the demand for talent far outstripped in in technology areas. Um you know great developers, great designers to demand for those people far outstripped um the supply and so. Guess what? that means that the employees can dictate when and how they work and everything like that and I I have experience and see that distributed work is way better as an employee you know, assuming that you're driven and still going to do a great job like gosh being able to choose when and how and where you work. Yeah, who wouldn't choose that It's awesome. By the way that also means that you could choose to be in an office. We don't tell people you can never go into an office tons of people go into coffee shops or co-working situations or anything. Maybe they really like that buzz of an office or something great. Do that? Well actually give you money to do that so you don't have to pay for it yourself. So again. We're not about a dogmatic way of saying one or the other is purely better. We're about putting the choice into the hands of the colleagues. My colleagues like our employees that automatic so it's not actually me saying dist distributed work is is better I think distributed as a default. And giving choice to employees to do one way or better is way better for the employees and I believe that what's better for employees ultimately dreams to lead to better outcomes for the company as a whole. Um, so what's changed in the past year while the economies crashed major tech companies are doing layoffs I think there's been.
47:20.81
Matt
You know over 100000 layoffs in supply and demand has shifted now. There's I think an over excess of incredible talent that is searching for an ever smaller number of jobs. This is now going to be itasturbated because of tors technology companies are going to be the first and earliest adopters Ai. Which might take things that used to take again. Let's say my developer is three x productivity. Um, do I need to hire? Let's and my business is growing. Let's say 50% per year before I used to have to hire you know between 50% and and 0 more people every year to keep up with that growth if now someone's doing a 3 x I might not need to hire for a year or 2 again to meet my current customer demand in a high growth business. So if productivity is being multiplied. Um, that changes. Now that means for people who I'm employing their salaries probably go up quite a bit ah because my revenue per employee which I think is the biggest sh driver's compensation like ultimately that's the only honest thing that drives compensation is going to go up quite a bit and so naturally through supply and demand that is going to drive higher compensation per employee. By the way for my current employees. That's great for the thousands of people that might want to be hired by automatic. That's not great because all of a sudden there's going to be more competition for each job. There's going to be more kind of productivity for my existing folks and um.
48:57.20
Matt
And economic environment which is uncertain so a lot of folks who previously were burning a lot of cash to grow their business are now trying to get the break even and profitable. So in that case I guess that's a long way of saying that the balance of power shifts to employers and so I think you're going to see some employee unfriendly. Moves from many businesses. We're obviously not going to do that because our our policies are trying to be driven not by what we can get away with but what we think is by first principles. The right thing to do as 1 example there, we pay similar salary ranges globally so we don't care what. State or country you're in if you're doing the same work. We try to do the same pay again. This is a principle which I think many companies follow locally right? like it's hard to find you'd be hard pressed to find a technology company in the us that would pay a man or woman differently or pay. someone who's from India differently from someone who's from America not only do we have laws against that. It's morally odious like who would who would do that and feel good about themselves who could sleep at night yet. There's a hypocrisy and that would they would have policies to do that. You know, cost of living adjustments based on where you live and morally if you really think about it why is it moral let's say I have 2 engineers. Um, that do the exact same work and create the exact same value from our customers.
50:28.12
Matt
Why should I pay 1 differently for being in San Francisco and a different one for being paid in Ottawa and the differences internationally were even further. So all of a sudden like let's say I have ah an engineer again doing this exact same work in Pakistan or Nigeria why should I pay them differently. Like give me the moral argument to pay them differently when I dug into it at least in my personal exploration. The arguments. There were all around Basically what you can get away with or 2 local norms. Oh.
51:00.61
Aarron Walter
Um.
51:01.56
Matt
And then I I find these amazing contorted things. So the people who've gotten most mad at me about this policy were actually entrepreneurs in other countries who would say hey you know these people if you pay this engineer 5 times more than they would make locally. It's going to create issues with their family. They're going to be ostracized from their community or whatever and like that actually might be true by the way you know if you're super successful in America and all your family you know makes 50 grand a year and you make 500 guess what that creates issues and you have to navigate that. But. As a company is that my place to then morally have a decision to pay you less I don't think so like I can't justify that I can't find like a first principles moral argument for that. So by the way we didn't start this way. We used to pay people differently based in different countries. But it was like I think around 15012 did this thought process and really examined our own hypocrisy and said wow I don't think that's right, ah and the change and happen overnight. We couldn't afford to change that overnight with the you know because our our kind of. Business model was built on our sort of our current cost of like employees and everything so it took us probably a year or 2 to get a very one to similar ranges. But um, yeah, that that was a decision we had to make and so I really encourage business leaders. Actually every person listening to this.
52:36.95
Matt
Sometimes examine your own beliefs because we're all huge hypocrites. There's a lot of areas. We've just never thought of and we start with kind of a default set of behavior or beliefs based on what was done and guess what when automatic started and in fact till today what businesses do is pay different people based on where they live. Or where they moved to and so it's really kind of like a and full credit this was driven by some of our employees who came to me and said hey I'm doing the same work with this other person. Why are you paying me less and we had some default answers and then they kept asking the question and I really had to be like wow. I think you're right? Ah I'm an idiot I've been wrong and so also just having some forgiveness and grace for yourself that hopefully there's beliefs that you and I might have today that twenty years from now we might consider morally odious or or wrong and just knowing that. Hey we're all going to be wrong all the time humans are fallible and let's just try to be heading towards correct and also not be too guilty or beat ourselvesrself about when we've been wrong now. Let's try to get to right? but you know what? um we made the decisions based on the socials of the time and. That was paying people differently based on their location and once we realized that was probably not right. We tried to switch to the right thing as soon as possible but in the same way as like I don't believe in like super hirsch judgments of looking at you know, let's call it a greek philosopher from two thousand years ago
54:11.65
Matt
And judging them by today's morals like yeah let's let's look at that and it and inform that but should I not read Plato anymore because you know he did something bad with they had slaves in their society or something like that like yes, let me judge that by the say standards and say. I think they were wrong on that but let me not ignore all of Plato's work and sort of look at it in conversation with today's society and what we know to be right today.
54:42.80
Eli Woolery
Now this has been awesome and and we promise to get you out of your time but we're already a few minutes over do you have time for 1 more quick fire question to wrap things up. No, it's for us. It's great. Ah no, no.
54:46.64
Matt
I Sure and again I have very long answers. So again I apologize like feel free to edit this down like even tell actually include that part where I apologize for having long answers I it's it I I am not got to brevity. So.
55:01.60
Eli Woolery
If the thing is there interesting answers so people are gonna want to hear it. So um, okay, well we're gonna wrap we ra things up by kind of bring things back to the beginning and um I listen to a lot of jazz at home. But I'm you know, kind of a a newbie as far as like Noam, what's great I just kind of go back to the classics like.
55:06.13
Matt
Ah, thank you.
55:19.64
Eli Woolery
Coal train and Duke Ellington and a lot of stuff from the sixty s is there anything interesting in that space. You're listening to right now or playing.
55:26.84
Matt
oh yeah oh wow ah um first I would say for you and everyone else listening to this look up where there's live music in your locality and go there because you will not believe how much leaving five bucks in the bowl of a local musician. Changes their life and also how you showing up to that restaurant on those days also causes them to hire more alive musicians and things like that. So again, how's the system work support local musicians first and foremost and also what's cool is they'll hang out. You can talk to them become their friends like. Hire them for your next party. You'd be shocked at like if you're hosting something at your House. You can get some amazing musicians for your house for like two hundred bucks ah and you're like oh wow I'm spending that much on booze. Maybe I could also spend that on having like a cool band in the corner and that'll elevate your event like so much. So. First for local musicians and just find the cool local cats like um second in terms of like answering your question like what albums to buy or things. Um, recent stuff that's good I mentioned Robert Glasper earlier his stuff is super accessible in that he's combined if you like r and b or rap or anything like that. He basically combines super good jazz like harmonics and rhythm and stuff with essentially like modern day performers like um.
56:58.38
Matt
Gosh I think he has eica badu and like yeah um, is folks like that. That's why he run the r and b album. He's a jazz musician but he run the r and b grammy so that's I think cool stuff so jazz has always been about fusion. There's a section of the community that says like. We peaked in 1940 and like we should just copy that or 1960 like miles davis job cultureing. So like I think it's also cool to see the evolution of what I I consider that still jazz some people don't but second I would say kamasi will washington amazing tenure player I see aaron doing a thumbs up. Yeah. Um, in particular fact.
57:36.90
Aarron Walter
There's so what one one 1 song in particular like I would love to have played at my funeral the his version of claireoon. It's that's transformative. It's amazing.
57:43.39
Matt
I was just about to say that Aaron that is I I would check out that just that track actually start with that his album he came out with a 3 cd album as his first album which again very controversial in the jazz world. And so it's a little more intellectual or maybe hard to get into if you don't appreciate like all of jazz but that claiir du lom is incredible because it takes again a Lindy song a song that stood the test of time now for hundreds of years which by the way was based on an amazing poem. And I actually um Claire I blogged about this in 2014 so I'll send youall the link if you could include this on the notes where do I put the chat it turns out that um, they.
You say? ah I'm probably pronounced that wrong. Um da you see which claiy loom means light of the moon was actually based on a poem by Paul Verlaine check out the poem that inspired Deb Youy to compose the song and now listen to the original like listen to class version.
58:42.98
Aarron Walter
Debc.
58:58.67
Matt
And then listen to like Kamasi Washington's version from you know I think forget when that album come out like 20122013 or something version of it. It's a cool way to learn about music the same way that you can study art excuse me, um, by looking at the evolution. How artists evolve over time. What is it in Philadelphia the Barnes Museum I think you had a barnes museum in Philadelphia if anyone ever goes to Philadelphia visit this museum so cool for seeing how like a thousand year old or 2000 year old sculpture. That we discovered in the aztecs could inspire Picasso you know two thousand years later which then inspires some other artists like watching how these things evolve I think is how you learn the same way that if you're like want to learn about wine. You might do a horizontal tasting right? Let me taste a wine from. 5 different vineyards in a certain region. Maybe a napa or in bordeaux or something the same year but different vineyards taste how those taste differently. That's a horizontal tasting then do a vertical tasting. Let's look at the exact same vineyard. So the exact same vines and look at one from different years. So that's called a vertical tasting. What you learn there will give you an appreciation for wine. That's very different in the same way if you want to appreciate any art area more do figure out their equivalent of vertical and horizontal. So maybe look at the top jows albums from 2012 and see like what's the best that we thought was that year and then look at the same artist.
01:00:32.86
Matt
Like maybe look at like amasey Washington over the years and hear how he's evolved or miles davis or like you know coal train. Oh my goodness his his career which by the way he didn't really break out until like he was 32 33 which is kind of cool right? because it's well I was in arts high school. Gosh there were so many prodigies folks who were like so good as like a high schooler I was not that I just had to work really hard to get the gigs I had to show up early I had to memorize all the music so I had to do like the stuff I could do because I was not the prodigy at all and I think that's actually. Ah, good metaphor for business as well like it's just working harder and doing some work that's led to I think a lot of my success. Um, finally I will say the musician that is blowing my mind the most. Um I again, it's hard to describe that this this prodigy this this. Kid now a 20 something year old is actually taking not just music but music theory and creating entirely new types of chords and stuff that like a Herbie Hancock is like really respecting um is this kick I can't call my kid anymore. He's 20 something.
01:01:46.79
Aarron Walter
Yeah, guys. Amazing.
01:01:47.28
Matt
Jacob Collier and um, check out his albums check out as well. He post youtubes of him putting together certain songs and he'll do this thing will actually take hundreds of tracks and compose them and like ah I forget if it's logic I forget the software program he uses but just watching him. I have no interest in doing that but watching him work and he has like these 2 hour videos where he walks you through all the tracks how he recorded things. Everything is inspiring. You learn so much from that and his dedication to his craft and then also just check out Youtube like there's lots of Youtubes of people analyzing. Some of his tracks some of his solos. Um I'll give you another link you can share this might only be good for like musicians in the audience. But um, there's basically this cool jazz musician also work checking out called Corey Henry
01:02:43.36
Matt
Canist and an organist actually plays the the hammond b three really? well. Um, he posted like a 1 minute clip um of him and Jacob Coley just doing the jam. On c jam blue is which is basically like the most basic blues. It's like the first blue is you learn as a jazz musician and let me actually make sure I have the right link here because I want to make sure I give you the right one by the way I find that I search. My Youtube history more than it owes anything else I would ah.
01:03:21.48
Aarron Walter
Same man I live in Youtube.
01:03:22.26
Eli Woolery
No.
01:03:32.49
Matt
I think it's this one and we just don't remember this video. Yeah, it's this one so this is a 14 minute analysis that goes pretty deep into like jazz and music theory and a transcription of this 1 minute Corey Henry and Jacob posted again just randomly to like an Instagram story or something. Um, again. Even I think if you're a musician this is mind blowing even if you're not watch it from the point of view of how'll to appreciate the craft. These 2 musicians responding to each other in real time That's my favorite thing about jazz is when you see live jazz or jazz performance. You might be hearing music that has never existed before and literally will never exist again. Especially if it's not being recorded that's special. That's valuable.
01:04:13.51
Aarron Walter
Um.
01:04:26.75
Matt
Something about that. That's pretty neat. So um, this was happened to be recorded and um, there's this cool cool video of someone recording it Youtube has the coolest stuff I'll share 1 more thing which is.
01:04:46.21
Matt
Um I saw maybe the best concert of my life. Um a few months ago and it was Jacob Collier and which I'm probably mispronounceed sorry I mispronoced everything colier coer and this manda linus named Ryan Teel playing a duo at the blue note.
01:05:05.78
Aarron Walter
Wow.
01:05:05.92
Matt
New York city um they're both perfect pitch ultimate command and mastery of their instruments. What I would call prodigies like kind of like modern day mozarts or something you know that like. Just are so incredible. They're doing this duo. They're just looking each other in the eyes and playing things and riffing off each other in a way that it's hard to describe it as anything like intimates I'm going to be a little. Um, pc right now is described it almost like watching like a romantic intercourse ah on stage with notes and music and watching just just to observe this was it was a special moment I feel like in the world. Normally this would be ephemeral and no one would ever see this again? Um, but there happened to be someone in the front row recording all of this who posted it to Youtube this has very little views right now I was lucky to find it like right afterwards because I was telling everyone about this concert I was like. I wish you could have seen it and then like I found this actually not me but someone else who I brought there found this in the background noise you can hear some people hootin and hollering. That's actually me responding like and doing something amazing me be like well or like colapping or something which is again part of a live performance.
01:06:34.38
Aarron Walter
Ah.
01:06:40.29
Matt
If you're a musician that's performed the audience actually responding and making noises or or like being an active audience is part of what gives you energy to play something cooler or like something like that because there's actually a feedback loop happening between the audience and live which is why live music.
01:06:48.20
Aarron Walter
Yeah.
01:06:57.98
Matt
Personally I always prefer to like recorded records because that feedback loop I can I think just elevates the performance. So um, let me give you that link and again if you're a jazz fan sounds like both of you might be just kind of putting this on. Headphones if you can or like a good sound system putting on a big screen. The video quality. So so the audio is actually quite good and check this out 1 cool thing about this. Um, this performance was they actually did a lot of more popular tunes. So they covered again. They're such amazing musicians that they just consume the entire western music repertoire and incorporate into their music and you know something I think is quite impressive but not widely appreciated from perhaps a musical point of view is a lot of Beatles music. Look at what the Beatles did and how they evolved over their career man. They did some really advanced stuff and they were the most popular pop group of their time and of all history. You know in some ways of all history which is kind of cool. So also cool to study that like dilemma. There is these like musicians who became like incredibly enlightened elevated their craft also like went to India and like meditated for a month and did all these things that like enlightened themselves and still would come back and perform concerts where literally you couldn't hear what they were playing because people were screaming so loud. Ah.
01:08:29.38
Aarron Walter
Yeah.
01:08:32.36
Matt
Ah, kind of wild and I love also they were marrying this artistic integrity with and ah, a popular commercial tech or popularity and also how that tore them apart you know and if you watch like the documentary like the recent Peter Jackson 1
Aarron Walter
Get back.
01:08:50.62
Matt
What it's called but get back. Yeah and just the evolution of the creativity there of watching Paul Mccartney create an amazing song over minutes. Basically the evolution of a song that I think is it get back that they talk about or um.
01:09:06.13
Aarron Walter
Yeah, yeah, and what's fascinating about that is just how they build play is built into how they operate everything is play and nothing is productive per se like let's do a productive thing and make.
01:09:14.67
Matt
And.
01:09:23.85
Aarron Walter
You know, intentionally direct our energy towards a thing that we want to commodify or put on an album or make money from it is We just keep playing. We just keep playing. We just keep playing and then something shakes out of that.
01:09:35.11
Matt
Yeah, and then also how they elevated their audience so they became popular on perhaps some more simpler tunes. Ah, and then when they start working in some. What.
01:09:40.88
Aarron Walter
Um, yeah.
01:09:46.93
Aarron Walter
Yeah.
01:09:52.41
Matt
You know and more advanced I don't like that terminology because I don't want to apply a hierarchy of this being better or worse. But maybe let's call it more complex actually is a perhaps a better way to put it more complex modalities more complex influences bringing in actually instruments that maybe weren't more popular.
01:10:07.58
Eli Woolery
Robby Shan car yeah
01:10:09.93
Matt
And the western canon or western popular music at that time and how that might lead someone on a different exploration like maybe some percentage of their audience is then like wow let me check out Ravi Shanor or let me check out some traditional um other music from India or something like cool right. Like again that that leads more learning and discovery so I sent you a link to this the the Youtube playlist of um, this which is essentially like every song from the concert I'm telling about that is thus far I think the best concert I've ever seen in my life and I go to a lot of live music.
01:10:46.78
Aarron Walter
That's scary.
01:10:48.33
Matt
Um, and or maybe the best concert that I've seen recorded. Ah so check that out both for both of y'all and perhaps for some percentage of your audience if anyone checks this out is just it's kind of cool to see and then. You know, maybe appreciate improvisation and live music in a different way that encourages you to get out locally.
01:11:07.74
Aarron Walter
I love it. Matt thank you so much. This has been a fascinating conversation.
\
