Postgres at PlanetScale (Interview)

Well friends, we'll go back. Summer is over. The show is back. Some of you just can't get enough.

Let's just say you can't get enough And I'm here to bring it to you This week I'm doing by a good old friend Sam Lambert CEO of PlanetScale One of our first conversations happened back in January of 2022 So we had a lot to cover in this plot We talked about the new five dollar PlanetScale Postgres the story behind metal So they are not racking and stacking their own MVMEs They build on the hyperscalers bare metal machines the smart way the cursor demo They have where agents optimized a database and ship it safely through deploy requests while you can roll back a 500 terabyte schema change as fast as a one megabyte change and Sam's genuinely unresolved wrestling match over whether to open source Nikki their new sharded Postgres product a huge massive thank you to our friends and our partners over at fly.io That is the home of changelaw.com learn more at fly.io Well friends, I'm here with the CTO of build kite and one of the most challenging problems of modern era software development is continuous integration and continuous delivery And so Lachlan Donald build kite CTO what are you thinking about today's teams the challenges they face the speeds at which they're developing new features new code it is just overwhelming How do you all think about that? Such a good question it's the question everyone's asking right now All of our big customers are asking us at the minute like you know if we five or 10x our three put this year or a thousand exit what breaks and when and you know my answer is kind of same as it's being for the past 20 years which is that the bottleneck is still trying to integrate those code changes in and then deploy them and check they work and then keep them working as you keep throwing more and more code at it. I think a lot of the fundamentals are the same but we're just a thousand xing the speed of it and you know that changes nearly every variable. Yeah for sure okay so where does build kite thrive?

What particular tip of team or enterprise do you thrive in? The area that build kite has always thrived in is like this like fastest moving tech companies of the world. Like we've been disproportionately successful in that small niche that kind of Shopify class, Uber class, you know open AI class of folks that have this key problem around iterating really really fast and you know the thing about all those faces they all have subtly different needs, subtly different problems and so we've tended historically towards building like really well engineered LEGO blocks that scale like orders of magnitude more than what our nearest competitor does. So you know I think that puts our system in this tension where you know you've got to spend some time assembling those building blocks that it's those LEGO blocks to get the thing that you want but the end result is far and away more performant and scalable and the experience is better than what you get from something that's off the shelf.

So I think we've started from a position of a really well engineered LEGO blocks and then are kind of working backwards towards kind of creating the thing that scales down to a startup that starts with one person in 10 agents next week. My friends go to buildkite.com that's buildkite.com. You just have better CI engineer for the frontier we are all facing trusted by the teams setting the pace again buildkite.com once again buildkite.com. Friends I am back with an old friend it's been several years about four years Sam.

It's been the pod Sam Lambert CEO of Planet Skill one of my most favorite databases that I don't use I'm so sad saying I am on the list of try not metal because it doesn't love a good MVME right? Oh yeah go and have that performance especially with databases inside the server disk magic. Yeah you know I always felt like you know Planet Skill was out of my realm in terms of what I personally needed because you're so good is that a good thing or a bad thing what do you think Sam? Luckily for folks like yourself we now have a five dollar postcards option Planisco Postgres whoo very accessible the price of a kind of Diet Coke in the fine but expensive city of San Francisco and so like very well priced for people who just want to get going and have some simple postgres but all the great features of Planisco bundle in it's there we're ready come through.

Give me a landscape of what Planet Skill is today the last time we talked was in January 2022 now I know the audience pays attention to you but just for those who may not have caught up since then or they just want to lay the land what is the landscape of Planet Skill is I know you're in postgres world and you've got sharding you've got metal I mentioned already but give me a zoom in to what you're doing. Yeah so a lot has changed since 2022 that was like the very early days. Yeah I mean things have gone really well we well right now we have entered into the world of postgres which we announced mid last year we have announced our sharded postgres product Niki which is coming very soon and is already ramping up with customers and for tests keeps growing and hosts now some of the world's largest databases we run very very large sharded workloads for tons of very large AI companies. The AI era has been phenomenal for Planet Skill we've always presented ourselves especially in the latter years right as just a big engineering team is a company that kind of has a very thin abstraction over the top so our marketing is very technical we're very well known for our blog posts and our technical content and our incredible marketing team all engineers that spend the time kind of curating for our engineering audience and none of us really foresaw this kind of agent agent like revolution AI revolution but it turns out if you're just kind of hanging out and building a database that scales to extreme scales for customers AI companies need this and we've been very fortunate to have a load of them kind of come onto the platform and grow incredibly quickly and so it's been an amazing time here at the moment you get the kind of in that era of like you build you build you want to get successful and then you get all of the sort of success problems and stress and success and we're dealing with that and scaling ourselves and scaling our various products and just a very exciting time at the company right now and just running around 24 7 meeting customers getting customers migrated talking you know it's fun.

So you say that right but then I go on your website to your homepage and I just search command F AI and I don't see it anywhere what is going on how are you leveraging and building and growing so well in this era but yet you're not really seemingly positioned as an AI database how was that? I don't know if there is a specific definition of an AI database right now there's definitely AI database platforms and we certainly have that I just gave a talk at Kirsten's conference yesterday where I showed off 10 Kirsten agents working on the plant scale platform the whole point of the talk is they went through a bunch of performance optimizations deployed them very safely and you know this is the thing that's really cool about plant scale is we built this for humans and one of my slides that got shared quite a bit on twitter was that we built this for the flakiest agents of all humans if you remember when we spoke in 2022 we were all about allowing people to ship really really quickly safely like we have database branching online schema changes we have rollbacks for those schema changes we block agents from pushing schemas into prod that would break queries we have traffic control that isolates workloads and does QOS for databases so we have all these like incredible features to make databases extremely safe for humans or their agents the thing that communicates that we're there for AI is the fact that the best AI companies already work on our product and use it we use AI behind the scenes to deliver insights to our customers and really good things for our customers we don't need to necessarily screen the words AI over and over again you see these companies that don't really have a clue what they're doing and their marketing team has been told by the co we need like an AI narrative we need like just and they just slap AI on everything and it's just like not good it's just insulting to the users whether they're AI companies or not like things have not changed that dramatically but we do use AI and the AI companies use us and that's the important thing can you give me a glimpse into what's required behind the scenes to do what you said with cursor that you said you had several agents provisioning deploying to the plant scale can you give me a zoom into that yeah that's actually another part of the talk that was key like I didn't want to talk to be just one big demo of plant scale it wasn't a sales pitch right but the key points I was actually making is that infrastructure platforms are going to change now to give insights to AI right like you know there's so many database companies out there talking about the day one of oh wow we're the fastest spin up a database and then like put it to sleep and never use it because it's agentic or whatever that's like really not how software is built like day one is day one great like it's the point of creation it's fine you pick any stack it doesn't matter plant scale although you should start your startup on plant scale at five dollars because then you don't have to do a switch and like we have an unbelievable amount of startups in the new yc batch building on plant scale for this reason they know the ones that you used to default to just fall apart basically no scale and so we we really focus though on like day one you got to get going go and go to build but it's like day one thousand is where the real problems show up when it comes to maintaining right like any infrastructure is like this is like a big living city and it's complicated there is so many interacting services and like you'll get your most productivity from AI working with existing code bases that need constant pruning and evolution maintenance and care which humans have to do agents can go and do this and so what we what we do now is like we generate prompts and instructions for agents and it is us scaling our expertise through infrastructure services and like giving it to the customer very directly and giving it to their agents very directly and that takes an immense amount of infrastructure behind the scenes so we we take traces and log the performance characteristics of every query that hits our platform which is tens of millions a second and we have to stream those data warehouse house them we have to like apply our various algorithms to detect whether they're using indexes or not like we've had to edit postgres core my sequel core to go and do this all to like generate a three-line prompt for an agent and so you turn this like depth of infrastructure and expertise into this insight for agents and these emergent behaviors then get observed and become the next instruction for agents and it's an incredible amount of infrastructure and plumbing behind the scenes that's required as our platform to be very complicated and well managed and we've achieved that and it gets this very simple flow where I could just go on stage and say optimize this database and curse this agents went away and got recommendations from the system we are very early to this there's very few infrastructure providers that have built this way it's because we were early to understanding that humans break databases and we have to make them make sense our developer experience is superior to most companies out there especially in the database space and so it's not hard for people that have built developer experience for 20 years to figure out how to get agents that same experience yeah and so that's where we go take me to the technicals of this so you have cursor in this demo let's just say in your talk optimize this database what what happens is cursor talking to mcp is it an api is it cly installed into cursor help you understand what enabled that query in that question to happen yeah it loads up the plant scale skills and the plant scale mcp and asks plant scale what do you have for me what insights do you actually have for me rather than just like we're in the stage of most people are just like pointing agents at like massive streams of logs or just a code base and it's like very difficult to get the agents to sift through that and find the right thing and so we make sure that it gets very narrow instruction make this add this index remove this index um scale this cluster like all these areas things that the agents can actually act on and they will take the mcp so they're going to the code and then they will also hit the api to go and spin up a dev branch to go and do the work on and that's where the next chapter comes in i guess which is they get the instruction they figure out what to do with the code base and then they need a really safe isolated way to test their work and so they will use plant scale branches to have an isolated dev like environment sorry production like environment for dev or they can't break anything and then the thing that we truly know one else does is we get that change into production in an incredibly safe way so there's like a gate it's the deploy request right like it's like the pull request the same people that like designed to pull request work here and you know are doing the deploy request here so like there's various things there's like first of people we stop agents screwing things up before they get to so we um block you over here is upright like an agent has screwed up and they want to and i demoed like we poisoned some agent instructions to go and do the wrong thing um and they got rejected from doing it we may then let one through so we could show what happens even if the worst of the worst happened and i'll tell you that's not in a second but but basically like if you go to drop a column in prod which is a common thing like when you're doing a refactoring you want to get this like rid of this data you don't want it in the database if you have referenced that column within the last 24 hours we'll query we block you from deploying it and we show you what queries have referenced that column is a very common source of outages for people is just to think they are no longer accessing that column or table or whatever and they are and we will stop that happening and we prevent outages every hour for customers when they are like we block many many bad changes going in we then have the scenario we thought loud that to go through and it did drop the column and then we have another feature for fixing that which is we fire anomalies to say this is broken this if you have broken your database queries and are failing and we can roll that change back instantly and it brings the old version of the table back with a zero data loss like even if we've done like intermediate changes to the new table they are still going to be in the old table and we can bring it back and that is again nobody else does this in a inner platform we're the only platform that has this capability for databases and so you can see it's these like small sharp tools of a tool chain that get an agent from dev to production safely can undo changes these changes all attributed to the agent like you go into a plant scale graph and you'll see a line that goes down the graph that shows when it was deployed and you can then like see exactly what's gone wrong and then once you've gone and made like changes and it looks good and it's like fine we generate more for you to do if it needs to do more and so it's this constant loop that goes round and round and so now like agents are building kind of scale for us and deploying their own work for us in extremely safe environments against like we see customers doing schema changes on 500 terabyte tables and they just get it done and it can and by the way you can roll back a 500 terabyte schema change in exactly the same time as a one megabyte schema change like it will be exactly the same. That's pretty astounding to even think about this kind of changes and I think what I think about is the primitive you've probably built upon so you've been doing this for a while developer experience focused for so long and so that's not just magic wand waving that's literally hey we're developers we use this every single day we're improving it based upon our developer needs just even your your data warehouse where you pull logs into to gain insights is a clue to that for me but you have primitives you've built on so branching you've got these safeguards you call them gates help me help me understand how those gates come into play to give that insight but also block you if you're going the wrong direction how those gates appear to the developer when it comes to the x they look very tactile and they look like we expect from other products right like the deploy request is really meaningfully meant to look like the pull request it communicates it shows you a diff of your schema even though you didn't send us a diff you did schema changes on a branch by using the normal ddl that my ckolopostgres supports and then we show you a diff it should feel like your other tools and then for an agent the same thing just looks and feels native to them through mcp servers or api's and like we have past remote for us you can just keep an agent's can arbitrarily call api endpoints that we have set up for our own use so they get these like these abilities like in the talk I talk about you know the unix philosophy of these independent tool sets that humans or agents can gather together to get what they want and what they need and compose their own journey like you might not be doing a schema change but people spin up a planet scale branch every time they do ci so the ci runs against the branch and search basically running against a copy of production or a system that behaves just like production it's just about being getting out of the way but blocking you from doing the wrong things it is another way of thinking about this which is i've been in databases for a very long time on very large products and services and i basically took every outage i'd ever seen and we're undoing the ability for that to happen i think for me i'm curious what is the most exciting thing happening right now obviously you've got a boom and you've got a boom in the industry with ai and there's a lot of speculation of bubble bursting just a lot of cool call fear uncertainty and doubt fud obviously but aside from what might be the obvious answer what is the most exciting thing happening for you personally in planet scale so the thing i'm most excited about right now is watching the key get built which is our sharded post-cres solution like the test for post-cres in the last five years we have spent so much time and work like evolving for test from something that was a very raw open source project that was being used at very few companies so now something that's like a generally used to cloud product and people interact it's like impossible now to be a consumer in the u.s or a knowledge worker in the u.s about interacting with products that run on planets go we have cursor they use us to build to just run gigantic sharded databases that back all of the product interactions that cursor builds including origin then new products running on us already we have kick.com pokes interaction line scooters every square swipe whenever people are buying coffee or you know scon square terminals that's going through planet scale cash app just extremely large consumer products that need a very scalable database and they need charting run on planet scale and now we're bringing that to post-cres it's been phenomenal for what your team of experts like the world's foremost engineers on this subject and field that have the experience of running these gigantic workloads in production on our cloud now bringing the same to post-cres through nee-key not just how quickly they're building and who is rolling out like right now we're moving another set of just huge post-cres workloads at very very well known companies on to nee-key and then watching that go is just phenomenal so like bringing sharding and the sort of traditional practices of database scaling to the new audience it is the post-cres audience that is full of very dynamic startups post-cres is definitely the database of choice for new companies and AI companies and so i'm just extremely excited to see the demand like we actually there's so much demand for this product that we have companies just i had to just very gently tell the CEO of a startup like they told me it was unfair that they weren't being given first access to this product and i just had to say you know we have a queue and fortunately we're working with a lot of people right now and we try to rate limit the demand it's i've never seen anything like it and so it isn't it yeah it's an effort in like technical mastery and it's also incredible to see that we've landed in this part of the market and there's so much need for the thing that we are so good at so you have quite a queue for nee-key and would you uncover from the outside to help me shape a lens here but is nee-key synonymous with vatess in that vatess is my sequel nee-key is post-cres is that are they correct i would say that nee-key is more evolved in the sense that if you're doing a do over on a 50 like vatess is 15 years old now getting a do over yeah well i'm not too much of the architecture right now but like i mean we certainly we learned there was like at youtube there was youtube did not design it to be a general purpose system it was designed purely for youtube and so there was mistakes that were made or not even mistakes actually i will say they were not in any way mistakes no i think it's more pro i think it's actually i would have less respect if they'd have gone and done that work because at the time that would have been an over optimization they did things that worked for youtube and that was about it if that test did evolve over time to be used by a bunch of other companies but they were still hyper-scale and they were like pretty custom but nee-key is built with 15 years of a test knowledge five years of planet scale knowledge running it directly on the cloud and learning everything we'd learned and building general purpose systems and then just getting all of the different things we want to add new features new user experience on top and and built by you know the union of all those things it's an amazing piece of software where does i understand the test and my sequel in the relationship there help me understand nee-key is this something that is proprietary will we ever have our own hands on the software the the mxl open source is it we have not decided yet i'll be very transparent we have not yet settled on whether we will open source nee-key or not unfortunately we've seen some people some companies very popular ones in the database space be very very misbehaving when it comes to other companies open source projects claiming credit for them forking them like hiring someone that's 15th on the maintainer list and saying they now maintain that project just things that are very bad and that didn't really used to happen in the early days of open source but that you know there's a new generation of people that don't really understand these things and i think it's all cutthroat it's true cutthroat yeah they use their position and they use their leverage to try and do these things and their voice and you know i don't want to see that happening and then there's obviously amazon here in abysmal player in the world of open source no matter how much they try and whitewash that and a terrible partner and so working with them and seeing them um what i mean they just say they have a shot at the market called limitless um it's amazing name yes i saw this amazing tweet once where someone said we have had so much fun finding limitless is limits in production it was great so the largest database that ever ran on limitless runs on planet scale they had an average two outages a day on limitless and so now we have amazon telling people that the new version will be much more like for test and it's like well it's open source they can go and try and do that and so amazon the the bounds that which amazon will will try and screw over small companies or the world of open source is unlimited and so when we when it's been proven that we are better engineers in databases than they are now they wasn't the case back in the day we didn't we went around then but the early teams that built the early databases at amazon were phenomenal engineers they just don't work there anymore um also the largest database that ever ran on uncharted or all in our runs on planet scale so planet scale like it's just our bread and butter all day long um we move databases off of amazon's products um and they behave very poorly as partners with that they tell people they're gonna make their software like us and so we do not know what the future of open source veneake will be it will take some deliberation but yeah it is because of people like that and some other startups in the database space that it's causing you know it's just we're seeing everywhere people are moving moving tests or moving their stuff it's just because the same bad players are behaving badly and now you know it used to be much harder to copy pre-ai but now people can like slot fork but but even like they can just kind of move a lot faster in interpreting the code like back in the day you had defensibility because most people couldn't just read the code and understand really what it was doing and what the trade-offs were and how it could be done better now i can explain it to them so it's desperately sad that these people do this and that the companies they're even willing to work with them but they're the kind of casino you have to work with them and so it just means i don't have a good answer i'm not going to pretend to know what to do we may still open source it it's just a world where you have to kind of play with your hand closed nowadays because of you know if you are a rising company that's threatening others that are already more incumbent they will do dirty tricks and i would never want to give them any advantage yeah well you're touching on a very um i would say a very interesting hot button for me i'm not sure how much it's out there i think it is i'm actually oddly enough in this in this day and my age less uh slurping up what's on x and less slurping up on linkedin and places like that just sort of like in my own world in a way uh which is kind of good because i've needed a mental buffer for this reason but i've been in this in this thought process of like you know the the primitives you build on the seemingly crown jewels that you would normally open source because hey let's let's make this prim of available for others we built on it let's meet this let's be cool right it's just a good thing to do i just it's just a good thing to do earlier in this room i interviewed a friend of my jacob who runs a company called pia he created twitter bootstrap and we were kind of talking about how open source used to be used to just do it because it was a good thing to be and now it's got a fair you know i dedicated my lifetime so i worked to github for seven years right like i i supported the webin source i watched it from the inside of the nuances of it it's a fragile ecosystem as a collective the open source world is one of the greatest things humanity has ever pulled off right sharing our work for the betterment of humanity it's why we have ai today um it is very sad to see the cynicism that has overlaid open source because of the way commercial companies have behaved especially once like amazon you know i'm not sure i know that your lens and i'm curious if you feel it's just simply human nature i think it's really we're especially an american world capitalist for the most part really we're building companies we're oh that we can do that for a second while we're on capitalism i love capitalism i i'm captain is the greatest vehicle for him but i am so catholic that i think these large tech companies should be breaking up broken up i should not be competing with my own company by buying a rotis-riching of whole foods that should not be allowed like that that that is just fundamentally anti-competitive let's look at ai google saw the early versions of ai very early and pita teal is completely correct that google have run out of imagination based on how much money they have in the bank they've stopped innovating they use um they use the fact they have google search which is they earned that they earned and built from the greatest businesses of all time with google search but they saw the early version of ai and they had the research and they just sat doing nothing just wasting the opportunity and then others came up opening eye thank god for opening eye that woke up and built something and lit a fire under google and then google get to wake up in the world like the speed ai is moving right now an age later and there's almost no consequence for them because they still have a massive search business they still have so much capital to throw into ai this is not good for innovation i'm not i'm not this kind of socialist kind of break up tech companies anti-billionnet like i think it's phenomenal that people put rockets in space and get gigantic rewards for doing so i think we should be making more billionaires that get to make more things and build things and that should trickle down i think this is the true spirit of capitalism if you look at the forefathers of capitalism the people the way that in in the young america people talked about opportunity they did not foresee the kind of leverage that these giant tech companies get and it's hurting innovation these sleeping lazy giants sit on innovation and it doesn't hurt them that is not good for the general innovation of humanity i like that i want to coin that sleeping lazy giants so at the same time you say forefathers of of of capitalism they were cutthroat oh you should be cutthroat there's probably mirror even like this is back back original og business days right i can only speculate i'm not speaking from maybe uh you know a biography by any means but i imagine there was a lot of cutthroat and if there was a version of open source they would certainly glean from it like they're not benign right just because of it really right but also open source came from a cap like i mean they tried right it was like microsoft let's not let them let's not remember the october letters let's look at the october letters not so forget that they tried to kill in suppress open source and you know they bought github um to undo this and they have done a good job and they've done some great open source but capitalism big organizations like microsoft did try to stamp out open source and it did not work and they're still very rebellious almost political movement of open source again they take it too far in some areas but it did it throughout it did thrive in capitalism too there was a there was a there wasn't that good but i agree with you but i i don't the early forefathers did you know the crony capitalism is not what they wanted they some of the early decisions around contract law why america has such an incredible economy now because they because you could get a fair shaking court you can still sue and get your day in court um and win and that is extremely important so i don't think it's worth viewing capitalism is always having to end up into in a very cynical everyone's gonna screw but i do think you have to be cutthroat i do absolutely i think the say that i just mean that i think it's human nature and i'm with you so i'm in a way uh narratively lamenting yeah the scenario we're in right now from a position that the change like is by and large been supporting open source since the year after github has accepted like at the same year like i think github was born in 2008 we were born in 2009 and we've been riding the wave of open source moves fast keep up since the beginning well friends this episode brought to you by friends at coder.com secure environments where developers and agents work in parallel and i'm joined by nikki pike field cto forecoder nikki what is a field cto so i get that question a lot and it's you know half the people understand it half people don't so field cto i describe it very simply as we're dev rel for the c-suite so we provide a bridge between the customer voice between the c-suite and the managers and the leadership teams of our customers back into our product and then we go through and we help enable our teams to have the same message to make sure that the message is correct and that we're building on something that people actually want not just something that we think they want okay so we're taking the laptop away from the developer not really though we're putting them in a cloud development environment a secure environment where they can work with their agents in parallel these are blessed environments what's wrong with the laptop the laptop is the trap here and not only because of the fact that it could be stolen you could lose it it it breaks and you're out of work while you wait for a new one but there's also just the consistency you got there we all know developers developers are going to be looking for some of the latest and greatest and if you're not really controlling how they get out there that's where you get this it works on my machine it doesn't work on in production it doesn't work anywhere else because you don't have that consistency you don't have that ability to really standardize what that environment looks like and this is a problem not only for new people coming in you know the onboarding statement is it average i think it's like four to five weeks for a new employee to really get their their local laptop set up and ready to start doing the first time of code and you know the time to first commit is a metric that almost everybody knows and the reason they can't do that is because there's a lot of tribal knowledge out there they got to go talk to developers what do we use and where do we get our dependencies are we getting from public are we getting from private repositories but there's also the security and the supply chain aspect of this when you have local machines out there look at like the shaihele you know that virus that went out not long ago this was a compromise of the mpm public repositories they went and downloaded things mpm did what it did next thing you know you're compromised but when you use something like what we're doing with cloud development environments then you can mandate and you can put restrictions on there to say hey you can only go get your packages from our private repo those packages are expected to have been thoroughly vetted we know that they're clean now does this stop everything like shaihelele no if that compromised package gets into your private repo you can still have that but it really reduces the surface area of the attack and it also reduces the blast area of the compromise should it happen because if your laptop gets compromised and you have to kill the laptop for whatever reason that's weeks out of work while you're either fixing that or you're getting a new laptop in the cloud development environment allows you to kill that start back up fresh and you're back and running in five minutes you don't have to wait all that time well first step is to go to coder.com install coder self-host environments for your teams to enjoy to standardize around and it's open source so you try out today once again coder.com and so as that same person i'm sitting here today saying okay niki is it going to be open source i don't mean that like hey sam give all your crown jewels yeah how do you battle this struggle how do you deal with this struggle of like gosh we're in the scenario where my most fearless competitors that have more than me they can cibosh me maybe not technically but maybe with a purse potentially or a war chest how can i give them my 15 years of a test my five years of running planet scale and just hand that over yeah in open source it's a very challenging position to be in but yet we come from this past thread of lovers and supporters of open source yeah it's it's hard i don't i do love thinking about this way and like i said i used to say we never do postcards here we are i don't i just don't know you did say that when we talked in the first time i mean like um i might as go for life right the test for life right and i changed my mind on that and i'm still i i do not mind being insistent like i'm here i you know my public stance is i think about things very openly and i express my feelings i have absolutely like people when i tweet about stuff people love to dig up all tweets it's like some sort of gotcha um as if like i'm like but change their minds but yeah correct i will change my mind faster than all changes correct i get no embarrassment and no shame from not being logically insistent in fact my by-law one of my things i say in my kind of description on twitter is i'm making more mistakes than you and i'm actually very proud of that like people again they read that and they get it wrong like oh it's a guy's proud of making mistakes stupid um i genuinely deliberately make mistakes i move extremely quickly i believe life is a it's like a spread bet you have to make mistakes if i said to you there's a target down range i'll give you a million dollars if you hit it she's a sniper i for a shotgun what are you choosing you're just a shotgun because you're gonna get i don't know how many shot they go in putting a shotgun but you're gonna get more chances than the one sniper shot if you bundle everything into one big decision you better you better not miss well with a shotgun blast and i learned this in the early days of github that culture was incredibly iterative we put something small out learn iterate build i just will not i don't i refuse to to ever stick to anything in the past i have almost a fault zero respect for the past or time for the past i just want to see what's tomorrow and and if the most offensive horrible thing i can do to my competitors open source niki you bet you'll be done in days until then i think the best thing i can do is i'm watching some being out built in the open they are years behind us um i'm watching us take amazon's best swing and moving customers off constantly so right now i'm keeping that for ourselves but who knows we'll see yeah you know i would i would struggle just as much as you are right now uh with that choice and if i were you uh just speaking out loud i'd probably keep it on the inside for now just because what do you gain yeah like who do you help by open sourcing niki that's the question yeah it's not even with open source there's so much change maybe you get to speak to this too potentially as you um as you weigh these pros and cons as you see a lot of change in open source you mentioned it pulling tests back um all sorts of things not accepting pull requests not accepting contribution so we're open source sure you can use it it's still the free gift it's still usb stick on the front curve yeah thank you very much but we don't accept any any contributions it's just in-house only so there's no then you got uh the mitral hachimato direction which is like let my current contributors you know i think i'm sort of you know tailed during and away what they're doing but it's sort of like a merit-based if you've done some good things or somebody suggests you can kind of come in kind of thing and i could be wrong on that uh again i'm sort of out of the loop to some of your uncertain things but i feel like the the scoreboard and the way we participate in certain open source not open source but certainly high stakes open source like something like niki being open source for example there's a different game being played and it's it's just a little unfair to release your crown jewel that helps you do so much and then what do you really gain from giving it away i don't know yeah i think about this and i don't want to criticize anyone who's done like people that you have to find your own path in this world and people do amazing open source maintainers are saints like they're the things they put up where it's hard for now i just want to say that like you are all incredible people and i hope that we all find a better path forward for this maybe it's licensing i don't know but like i really hope so i never want to seem like i'm critical we're losing from this too right like we're still not the size of amazon so still when we're working with these gigantic businesses they're like well i know we know plant scale's profitable and there's lots of money in the banking is going to be around forever but you know it'd be easier if like we knew because you know we've never had an enterprise customer churn right and we never had a customer leave us so we've retained every dollar of enterprise business and they know that right and they know that it's like not incredibly easy to go from if plant scale is the best most scalable database if we hit them it's not like they know that they're not going to turn around and do it tomorrow but if a lot of the coders open source it does give them more of a sense of feeling that they have a plan if something was ever to happen or we got acquired or whatever right so we're losing from this too and maybe that's why i'm going to behave the way it's because it's easier for them to just tell me us if we know that we can't do that i don't know it's but it's sad and it's not an easy situation and i will always come back and update you as my thoughts change but right now i'm i speak like a lot of the people you've mentioned as being like shining stars open source i've had this conversation with people and we're all trying to find our way toward like figuring this out well i mean if you can give a demo with cursor have it do some cool stuff with insights and you can do the same thing with a code base you can look at some base code base whether it's even from a licensed perspective yeah you can graph intent you can oh yeah you know you don't have to copy code these days to copy the idea or even to i don't want to say copy is like a bad thing you know we've been all been copying or we've been staying on the shoulders of giants for a long time but it's really just an interesting space where you have a super intelligence available for most anybody with 20 bucks on most of the planet that has access to these tooling and to learn to examine to replicate you can do it afrously you can do it altruistically you know you can do it and from so many different angles but the point is the primitive the ability is uh is there yeah take me take me back to this unfair advantage of me i want to i want to blaze past that without even in the feelings you have on today's open source but what is this unfair advantage how long is this list and why do people feel take me further into that the remark well this two sides this and you just touched on AI one one unfair advantage is is that we have some of us have spent the last 15 years in our careers working on very large websites or very large database deployments and that is knowledge that has not made it into any single model right like the code that we wrote to solve these problems may do but the brains of that code was formulated are still in our heads and precisely none of that has gone in and the reason we have very very large AI companies building a tough bus is because they're buying for our expertise they're buying for what my engineers will do at two in the morning paged with a behavior that has never happened before that act of human brain instructing the agents to fix things and building the pipelines that the agents use is what people pay for so that gives one advantage which is yeah maybe i'm not 22 anymore doing all nighters all the time but agents can do all nighters yeah so we've got so not only has niki been the distillation of so many years of experience of incredibly talented people it's been built at a breakneck pace because those people turned out exceptional engineers can do exceptional engineering with AI and we're taking our experience and using it and so there is potential for a lot of tech companies that can you hear that yeah what in the world they're doing construction outside that is likely a drill it's okay let's let's see if it passes i hope so that was a good that was a great question you asked me want to make sure i don't forget that man you're so good at this by the way like i literally you know i i've never i asked you to come on this i've never asked i don't find hard to have me bad like nope wow you it stood out i start the t-shirt i also don't keep match either like i really thought you did an amazing job so like i would love to make sure there's seriously seriously i've been through some stuff man it's been it's been a journey i hope you're okay it's while we're waiting for this to go i'll give you a little snapshot so behind the scenes this last 2025 was a challenging year for me my business partner jared he's a great dude i love the death but we just had differing opinions on how the business should run and i just got tired of the partnership i really did and i'm sure he got tired of the partnership as well it's not like it's a one-sided scenario here but we were together as partners for a decade basically and that's a challenging change really and so right now in december of last year i asked him if i could buy him out he agreed and so in february of 26 this year we inked that deal and it's done so he's an only business so that's the change but at the same time you know we have this AI revolution happening i've been building a lot of cool stuff um all the while also podcasting and maintaining this business but it's been challenging because like the world has just changed drastically and i've had some imposter syndrome to be honest with you like what you should not show and you should not but say here you say that is what makes me tell you that seemingly tMI behind the scenes that i would not normally share so uh deeply with someone um i but so i appreciate that because i want you to wholeheartedly take my encouragement that you should keep doing going forward um for multiple reasons plant scale we did lay offs like we've been really dead at one point things are better than ever we are now like just more alive like ripping business like we got features like like brex did a report that we're in the top 25 of like things that start us spending on now like we are going it was so dark and if you don't push through that you will never get to that and it's really hard you're founded too and literally the story you just told me another creator told me the exact same thing business partner break up buying them out pain it's very common and like yeah but i'm serious like deadly serious ui p99 amazing what you do you should not feel it but all the best people from imposter mind you too like but you shouldn't it's also a long journey to having done this like my whole career's been a podcaster for the most part and i can honestly say i haven't built anything technologically that's worth really anything so being in the tech industry and not really having a tech to show for it is kind of embarrassing in a way no but like media now is me i i you know media is is is hotter than ever now right people want to tell because there's now like i remember back in the day and there was like two companies doing one the same thing right like you basically just like in 2013 you had to be the first person with the idea and you can make a big company now there's 10 people 10 startups in san francisco going after every bad idea and people are different shading now through storytelling craft design experiences i think you're about to enter an age where your skills are more important than ever generally but i look at how we build our marketing team and our buying behaviors are we're going in that direction yeah well i appreciate that i mean i think that the business is definitely changing i'm still excited about podcasting i'm just a little bum to be alone even though it's my choice yeah and so i'm just navigating some personal motivation i would say and then also i'd think a little bit of burnout and then also some apologize in terms of you mix all that together you get totally weird soup that just tastes kind of nasty sometimes well you should come out some for this kind of filling energy out here like it will really pick you up i think it will really make you feel like go to event see people like i think you'll really enjoy it i'll meet some new generation who found us an interview and chat to i think we really go man i got a ton of friends out there too i mean i would consider you as a friend as well but i would come through come through the office hang out you're more welcome i should just do like a tour yep go see Adam Jacob come see you exactly there's a few others that i would so reach out to uh my buddy at render on a rag great small handful folks i just like going hang out with if they would let me of course come through right i think we're in a drill this window so i'll carry on oh so i'll pick up the edges so we're looking at this very interesting time right now there's a bunch of companies being built by people who've got deep expertise from working in other places they know their domain really well and they can build for the next generation and using AI to do that really rapidly like AI scales expertise phenomenally and i do speak to startups but like they're kind of you there's an arc of time and this is how i view us for yamas and right which is like five years ago VCs used to tell me you're never going to be Amazon well five years later we've got the largest databases from multiple their products running on them and that people advocate for us over them constantly it's the kind of arc of time when it comes to startups there's an incumbent and there's the rising ones and the things that mean they are struggling right now to respond to a to demand they're doing great as a business right now they're the ones with GPUs but but like it's difficult for a 50 000 person organization and you know we are we have a 50 person engineering team they're 5 000 people running engineering on the competing products to us and we still beat them this is because we can move faster we are nimble we are involved we get there like large organizations thanks to AI although they are trying to keep everyone out like they are at risk because they can't move fast enough you just like the only massive people to do stuff anymore and so it's interesting like i'm seeing some young companies that come with a different angle of seeing things and they're going to be truly disruptive but i'm seeing some wash against the waves of just the facts of life and the facts of how some software dynamics are again that may or wash out in ten years time it might be completely irrelevant to have anything we built we might have a god completely can just do everything we're not there yet we've got one more cycle maybe before that maybe a few who knows we're going to be some great companies will be built by very experienced people scaling their expertise through AI so it's actually an amazing i'm never had more fun like we we have a customer what they are just phenomenal at what they do and they're young and they're building for their audience and they're just absolutely outstanding and they're scaling incredibly rapidly when it comes to consumer when all these places experience actually kills it like you don't like you can be too old for these things right but when it comes to infrastructure there are just ways of doing infrastructure that will persist for quite a while and there's rules to how that gets done and that's how it's going to be for a while i think help me understand your hardware story because this is something i'm personally challenged with is like how do you compete and innovate against amazon or even google whenever you deliver vatess or niki or planet scale to gcp or at ws how do you compete innovate in their playgrounds so to speak their hardware their world do you build your own machines what is your story we do not build our own machines i think when we named metal i think it was a good name at the time i think it was kind of obvious what it meant but people have confided it with us running our own data that's not true i thought you did right you bought some mvme's and right to mistake them and i'm glad you then thought that so i can correct that which is we use the bare metal machines that google and amazon give us and we build on top of that so that's how we compete we take the same machine they do and build our entire software stack on top of it which means the machine the profile that the cost profile the machine is low because they don't really put much margin on top because don't we do anything but the most impressive thing is those machines are femoral if we lose one of those machines or we give one back by accident and we misschedule something in kibinelli's or whatever we would lose customer's data and so are we differenti and no other startups are doing this um there's a few things there that's hilarious we can see the code and they're gonna lose data um that and so we differentiate by running databases in state and we run more state on kibinelli's than anyone on earth we just actually released a really amazing fat neon team released a great blog post about kibinelli's that it's just a wonderful masterful piece of writing i recommend anyone read it um and yes we build from the base up so it means we can perform as fast or better than their solutions with all of our dx agenta x on top of it um the the higher up time we are we compete with them by having higher up time um and just generally a better product while inside the cloud this is the thing there's a there's a few neo clouds that built really well outside of the of the major clouds but if you want to sell to large companies and companies that are far beyond startups you have to be in a ws and google because of latency issues and all these things um it's we have to be there and so the fact we found the perfect hybrid between on-prem performance and scalability well no we add the scalability on-prem performance inside the hype scares cloud is why we get to do what we do we're also running the customer's account too it is not on-prem you still get a plan to go on it's like bring your own cloud which i think you and i will have talked about in 2022 and we were early to that it was one of the greatest things we did like we have gigantic customers running it because they are customers of google and amazon so we run inside their accounts and so it means that they secure it they pay the bill they get like great discounts when you're committing like a billion dollars to um amazon you get a decent discount and so they can use that as well so we can get the net like the total cost like our cost plus the discount on those hardware costs we can often be amazon on price um and so that's why you know that's how we compete yeah it's interesting because you have to you know it's it's almost hypocritical but you can't help it right like you because you just said in one second they're telling terrible partners uh living to amazon but yet you know you have to play in their world so it's like this um like a mask party you know like maybe a masquerading kind of thing like a party where you have to go with a mask on you know but you're still kind of people it's just really weird scenario where you have to play nice but not quite so nice oh but even like i massively respect them for the position of leverage they put themselves in AWS is a phenomenal business built by phenomenal people and we should have nothing but respect for them at the same time they have to strike a fine balance of competing with the people that run on them right and like how long people will tolerate that now luckily there's that undoing as well because they've built like what 350 services and now there's 10 competitors for every one of those services and it hurts for them because they have to keep up and people want the best in class tools unfortunately the databases are some of the services that make the absolute most money for amazon like people i have other CEOs who have great partnerships with amazon because they compete with less important amazon services so you must i mean i just met one of your customers and they told me how much they spend on amazon how much they spend on you like amazon must love you and i was like we go out the terrible time because we're competing with the the software that they're trying to put their services they're trying to protect the most and so um yeah they behave really badly in that world um and so they just have to be careful and i hope we don't make the same with um some of the large model companies that constantly talk about how they're going to replace everyone it's like i would caution anyone from allowing that to happen again it was a day with a i it could be dangerous like it is really bad that you can be an amazon customer and that they can see you working with someone else and their sales and then they reach use that to reach out to that and they do do that and it is extremely bad and there's nothing you can just stop the thing it happens happens all the time our customers tell us constantly and so you have to be careful because like you don't get to rain forever there's other clouds and microsoft actually is not like that on azure their sales team the way they operate is much more friendly to the ecosystem of people that work on them amazon particularly um really do not do that they pretend to they suck you in like they have released it's it's a it's a company that i think most people think is compared to us not isn't really but amazon just released a competing product to them today after having that company on stage at reinvent last year so they will pull you in it's so funny actually uh founder i can't remember what company it was tweeted a screenshot of an email they got from amazon corp dev being like oh hey you know it looks like you're doing really well i would love to be challenged see how we can help that was sent like six months before amazon released a competitive product and i reached out to that founder and showed him an identical email from me i just pick up the phone because i'm entirely there so long there's no way i'm still containing any of them yeah it's like are you a fan of silicon valley bunny chance the tv show it's too close to i've watched i've enjoyed watching it and i've met the producers they came to github when they were doing research and i know a friend of mine is one of their advisors they did a great job of parading it to the degree where i find it a struggle to watch when i live in it in silicon valley um it's hard to watch oh and some of the some of the actual props were from github i didn't remember the scene where they get delivered merch on the doorstep github provided that as a as a um as a prop but yeah like it's hard to watch for me is a long-winding answer saying no i've seen some of it and it is good my judges are genius but when you kind of know people that are just exactly the same as the characters it's a bit much for you possibly watch tv so they go extremely long when do i have explained that to you the answer is no my friends i'm back with a good friend of mine michael grinich michael i know that i love workOS our audience may not know about workOS but what are the challenges developers face starting a new project choosing the right tools choosing the right database choosing the right path take you there when it develops starts a new project the decisions that they make at the very beginning end up having long lasting consequences what language you build in what platform you build on top of what database you choose these are things that are very hard to change later on so they have like major consequences and especially if they limit your ability to grow and scale at some point as the product starts to take off you're going to have to stop developing the product features and go re-architect or rebuild your system and that might be a killing blow right at the moment you need to accelerate so these decisions are really on are really really important and i think that's why developers gravitate towards solutions that are mature things that they know that will scale even things that are open source you're going to pick you know something like planet scale for your database provider not because it's the cheapest or because it's the you know most fun to use but because you know it's going to be a durable provider you can scale on for years and workOS is like that you know at the earliest earliest days if you look across all these different services they kind of look very similar but at day 1000 or day 2000 or day 10 000 you're going to want to have made sure that you pick a platform that could scale with you and today workOS is powering off and identity and security and permissions for all these AI companies literally fastest-growing companies the world like opening iron and drop it in for a certain complexity workOS is under the hood there so i think when people pick workOS early on really what they're doing is trying to pick the defaults to allow them to grow and rapidly scale and there's no platform other than us that's done at the same moment well friends the next step is to go to workos.com sign up today check it out free for a million active users try it today there's no excuse not to it is your default you should choose it so do so workOS.com once again workOS.com that's okay because i won't ruin the plot by any means but i'll just give a glimpse i think it was in season two it may have been season one they were brought in by a venture capitalist firm they being piper and they were innovating on this you know compression algorithm and they were hunting for money obviously because like any sort of you need money to grow and they were bringing republic guess is probably the proper term i don't know if it's the proper term or not so similar it's like you know let's email you let's get you into the room and ask you how your thing works with no intention to invest because they were not intending to invest and you know because they're naive and they're early in their careers they gave away information they were happy to like gosh these venture capitalist meetings no one ever asks us how the platform works let's just whiteboard the thing for you so they literally whiteboard uh you know they're crown jewels and now they have competitor which is why before they did not now middle out that was pretty much the you know the crux of piper was now given to a funded team who is capable of engineers and they go and build the same thing they'll have to better so this happens all of the time this is a real thing like you see a story once a week of someone on twitter being like oh yeah actually this company offered to try and buy us and do two judges and they use that really poorly and like it's very bad this is it it you know i chat to young founders and founders that are new in this and it's amazing how naive people can be to thinking that these giants are here to help and that it's not immensely and it's getting worse there's so much on the light like people are terrified by AI they are terrified by what it's doing to them even though they should not be um yeah it's causing this to happen much much more frequently and yeah you have to be very careful and very secretive about what you do we just don't we try like we have learned two lessons and we would we try not to fall for these things yeah it's it must be challenging to even come on a podcast to have me ask you questions that might go into an area i'm sure you're protective of it but nonetheless there's some anxiety to some degree like i i've got to be in the public because i want to promote what we're doing i'm the CEO and you have to be you have to do your job and there's a job one of the jobs of the ceo is to be some version of a figurehead to the company and the spokesperson and obviously speak eloquently and have clear thoughts and share your ideas well but then there's a part like well now i'm at risk because like if i share too much you know how much do i share how how do i navigate this this new world we're in and i think that's a how do you how do you personally handle that yourself with being in the public because you're pretty you're pretty well spoken uh in two ways well spoken but then also i hear you and i see you on x and i'm like dang there's sand going toe to toe with like benchmarking and like calling them out and like you're in the most positive way ruthless you know like i love that about you i love the way you come at folks in a way um but how do you how do you manage the the ability to sort of show up somewhere like this but then also not give it all away i think about this very often it is a constant evolving debate so we're very open we're known for our technical content is a great example right so how do you share really deep insightful things that have a meaning you can tell when tech content is just like fluff right like how do you talk about really deep things about giving away proprietary things i talk about this all the time at the beginning of everyone of our all hands there's a giant slide um that says um confidential on it and we're very clear and i remind the employees that we are incredibly open at plant scale like if you're an employee plant scale you know you can look up how our finances are if when we do board meetings we present to the board deck for the entire company i believe in hiring exception good people um i believe i mean i cannot self-assign this but it's something i'm aware of that people would be plant scales having quite a high talent tomorrow um and being a small engineering team that's clearly boxing above our weight um i do not believe in hiring incredible people and then hiding anything from them like every business problem can be engineered we are only engineers to engineer ourselves towards building a phenomenal business so i share very very liberally internally but we have to remind people that you you might not like it's very unless you're essentially myself or like the sea level at the company it's very hard to know whether the thing you are doing or working on or you might just say it plays an important role in a very big picture that we don't want to get out there so we guard very um you know we're very protective over what we do our road map what we work on like when we build postgres we have to work in secret like i have to have a code name so that when we can ship you front-end code because we're interested it wouldn't be referred to so that when we're in south bark we can talk about what we're like working on without them knowing like we are very secretive in that way but we have to do a fine line of like being open and like allowing our marketing team and myself and folks to go and speak on behalf of the company and not leave and mess things up and so it just takes refinement and takes culture and it takes thinking that through like a few bit like a minute suppose we did but fattie it's amazing like it goes very far into showing you how well we think about systems without exactly kind of giving you the keys to the kingdom and so it's this constant balance that will again it will change over time but i try when it comes to twitter twitter twitter is a bit of fun for me like half the time i don't take things seriously i just don't mind saying things that i know is going to make people use their minds and it's like it's just funny to me um how easy it is uh especially also when it doesn't impact me like i don't get anxiety about this i can have a fight with someone online and then just close my phone and never think about it i can sleep like a baby um so it's really bothering me but it's funny when it really lights up some of your competitors and they lose it and whatever and do some linkedin posting but also like partly the way i'm on social media is because i do not believe in letting in false truths or things come be said about your company and not jump on them immediately i've worked in places where i've watched other companies that just get drowned on social media with things that are not true about that company and they and they are too reserved and polite about it they don't squash these mis truths i go at that stuff incredibly quickly to correct it in the public record um not this i make it sound like this tons of drama around this company there really is not but you know people come up and say things and they try they do silly things and you know just call it out and just say it for it isn't it's just a it's a way of handling PR and public relations and sometimes companies misstep and you can get out there and correct them really quickly and i watch companies just getting absolutely destroyed in the public narrative and they see yoshos be out there correcting these things saying it confronting them whatever like i've had people say hateful mean things i've had the odd death threat to um i often meet the people uh you know like you just go and say hi and uh talk to them and make friends and i have a few people that i now talk to very very it's a bit of a meme internally that i've had people that have been real haters like just absolutely hated us and me and now are customers and you just i just reach out to them and have a conversation with them like i'm if someone if someone is mad and hateful at me and makes a good point i like you can really flip that by going yeah actually you're right i was completely wrong with what i said like i genuinely do try and put controversial things out there see how people react and then change my mind if i do believe and i very often change my mind i have zero fear change in my mind like it's my job to do that yeah a lot can change well a lot can happen in relationships whenever you can be transparent but then also be willing to have a conversation like so many people just don't have the conversation correct much rather sub-tweet you or they would much rather you know talk about you behind your back or right you know roundabout so to speak and i'm like negative or positive just sort of like just about you right but uh you know something change whenever you actually have a conversation with somebody and there's clarity and then there's also just transparency in terms of you know what i was not actually telling right here i've changed my mind here and i don't feel that way anymore or you were right i was wrong that is the way that happens there absolutely i completely agree that is the way to deal with subterranean by the way if people subterranean just immediately go at them just so i know that's about me and then just like react to it they stop because they sub-tweet because they're cowards and if you call it out but people to polite they won't call it out so it lets all this stuff happen it's silly for those that are listening that may be aware i'm not having zero expectation that a single person knows who i am that is listening to you and maybe they do now obviously that's why i'm here but you know maybe you've interacted or seen my Twitter or whatever and there might be people that really think this guy like this guy here's a piece of shan line or whatever i just would love to let you know that it is one of our largest sources of leads there is a drop down in our sales force to designate their source being my twitter one of the hottest AI companies in europe came inbound two weeks ago whenever i see i love meeting customers if you read blog posts from our customers yeah very much out there very many times that i will be the first person they spoken to and i will also babysit their migrations i'll be in the section i love it it's talking to customers one of my favorite i don't know people like why are you talking to me i'm like what else do you see i'll be doing and talk to customers work on the prime ones right but like they're very surprised anyway i joined i like i saw them come one of their engineers filled out a contest for and i was like that's a great company i wanted to talk to them i added myself to the meeting because i added them to me myself meeting the founder showed up and i said wow and this is a great company they're gonna pay us a few million dollars um i said uh how did you hear from that scale and he said i love your twitter and i was like that's great and i said you can't tell it i said unfortunately like that's gonna just make me worse but like it is it works and i get told all the time that people just want they don't hear genuine takes anymore tech has become so mealy mouth and there's fakers there it's never been full of more full of fakers i don't know how computers work i think people will feel refreshed tech got a little bit too friendly to the non-technical ones that can yap all day long we should drive them out and get back to talking about like actual engineering and i think there's a void of that and i think people miss it and i'm kind of i'm just saying my stuff it's not deliberate i might quit tomorrow i don't care anything so much about it i'm not obsessed with my father account like i just say the things i care about and believe and maybe don't believe i want to test the belief in and i just use it and it's that it does come up a lot though it's like yeah people mention my i get people mention my twitter to me constantly maybe i should change i don't know but we can't follow you so i pay attention to see uh i'm not involved in the conversations but like anybody else i lurk yeah you know and i'm here for the stuff i'm here for the good stuff sometimes it's drama sometimes it's not yeah and i can't really recall the ones that the posts that i'm talking about but i recall other database companies touting their superiority in terms of their speed and their migrations like i don't even remember what it was something like that and you were like nah not true like here's the real benchmarks that's how they're being not true with their their details because you can fudge that just like calling them out you're like dang you know i know i love that i mean like i guess because some people would be maybe too busy not that you're not too busy but like somebody might use the excuse of being too busy to come at them but that's what i think uh stood out to me was like that i liked about your personality was that you're not you're not unwilling to push back when you know uh you know maybe of an untruth or a fudging of a number so speak or some sort of benchmark detail it just didn't really fit that you have a reputation to maintain with plan of skill obviously but i would say that you are definitely edgy and i like that like it's not a bad thing it's not how i would respond because i'm just not quite that way but i respect it i like that yeah maybe it's just that i'm oddly broken i have brain damage i don't know i also don't but i don't believe that people are too busy if someone is comparing your product technically and lying about it you should stamp on that immediately so far yeah like if you care about your craft and like care massively about the craft of engineering i love the database space because it is just engineering there's so many companies that have out-taught us out-to-yapped us and now that they're the largest customers that are going to us because they use Postgres we are winning but it's fundamentally a game of engineering i take it very very seriously i love like we are not a very serious culture internally in sense we have a sense of humor i like to make jokes i like to be funny i like to push like be silly um but when it comes to engineering and that we take our job of engineering very seriously everyone accidentally lets down customers and things and now and then that happens with us and we take that very seriously as well it is we are here to be a database product and i feel like one of the best things i can do to attack my competitors out engineer them and then have a lot of fun doing it publicly and build an narrative around doing that and they can comfort like when we announced our Postgres product was very simple blog post it just benchmarked us against all the ones around the market we smashed some of them into the floor we had to give two of them um we had to give them double the resources just they could show up on the graph next to us um that was the way we showed up humble because we said we weren't going to Postgres because i we didn't go to Postgres i didn't know if we could do anything better i i did not i didn't want to be arrogant enough and believe that we can just ground the oceanally show up and do it a different database better once we'd shown that we could though we made that the point of what we're doing all of the companies that we benchmarked we got very triggered one one actually the cto came up to our people we released it while we had a booth at a conference that day they walked up to them and got so angry that spit was flat like someone else from another company a friend of mine whose company's booth is next door is like you need to check on your team they've just been screamed out by the cto of one of the people you benchmarked i'll say their name it would be so big for them to get mentioned on a podcast like for this so i won't um but uh they yeah they really made people extremely mad they all promised follow-ups we watched them all benchmarks it's been a year not single follow-ups they couldn't do it so they can sit back down and try and write code and all that yeah being they can hire that big brand designers to spend 90 grand on a website and we were just doing engineering and we'll take that very seriously and then we'll have fun messing with them publicly yeah that's wild um let's go to to maybe an angle i wasn't really unmentioned but you mentioned 90 000 websites and i'm not saying that your website is not 90 000 and i'm not going to try to offend you by telling you how much you did it and spend on your website but i did an a simple doc you're not going to offend me i did it in a google doc so so it was very cheap how do you mean it the first version of that website and Holly our incredible cmo she has a Twitter thread about this i was on a trip to see the red words and i was up late at night and i was like we need to go way simpler i just want to we've got great tech let's just communicate this like a read me um and so i opened a google doc and just started drawing askea into it and writing words and then Jason our phenomenal designer took it made it look like less like the let like sort of scribbling to a mental patient and put it up on the internet and it went viral immediately i got invited onto podcast i was like hundreds of thousands of listeners like marketing podcasts we i didn't go on those but like people were fascinated by this things that get spoken about all the time it will change we are going to work on a new version of the website like it is too simple now it's that format and that medium is too simple to like communicate the complexities of three separate products and all this sort of stuff so it will evolve but it was very fun to run in the other direction now by the way there is some beautiful 90 000 websites out there some companies produce have insane taste and produce absolutely incredible websites we just went very simple for now and i know we're gonna you know never gonna at least i also just don't like it it's bad engineering to me when you open a website and it's all flashy and like you can hear your MacBook spin up and chrome is stuttering like that's bad engineering that's not communicating for sure i want to communicate but it was just a little silly thing you know just again i just like to do try dumb stuff it works or it doesn't you can't do it whatever it you know it worked pretty well i think back in the day when we had you first on the pod and for several years after that you had closer to what would seemingly look like a 90 000 website and i think it's about a year or two ago you've gone the direction you've gone so you haven't always had this you know you've had this minimalistic opinion but it didn't seem like you were this um plane that's probably the way to say you know so this is very close to plain text it's not it's not a pejorative our previous website was also extremely good done by an extremely talented set of people um and it got copied there was actually people at open source libraries to copy our like gradient hero and i someone said to me recently like they remember that year when we had that website they saw copies of everywhere and we did like people ripped that old website off so now i'm saying is that what we're doing right now is figuring out the new one and i'm like when we tried really hard people loved it and like it was good when we didn't try at all people also loved it and copy both versions now i kind of feel like the pressure is on to do something else that's a bit crazy i'm not trying to pretend i'm some giant design genius and we're going to come up with anything amazing but you do feel that kind of like huh like we also didn't try to do those things we just did things that made sense to us but it gives you that pressure to like think what can we do this time that's special maybe we found it i don't know we'll see we'll see what people think i'm sure they won't be afraid to tell me online since you're a spicy natured fella that's probably the sweetest thing i could probably say that's cool i like that it's nice to me no i'm not gonna do it i'm all worth i'm sorry to say i'm say something you know i like to do it keep this in as well don't cut it's nice to me what do you think the worst we should no no i really meant that in that way i liked how it sounded i had a laugh it's like it's like a meeting on stage laughing their own joke that's like you shouldn't do that so i did that no it's good i like that you should enjoy it i maybe i'm spicy i don't know like you are spicy it's maybe i'm blind another all of that man listen i live in texas and it's just fine to be spicy okay just fine that's true um i want to go to this level because i this is something that i've had we've had podcasts on this we talked about multi-grace we've got the suga we've talked to i'm even an investor in subbase on the small angel investor in subbase well that was a very good example i couldn't be an angel investor well that was an exception investment because they built a very valuable business so yeah the it's my i've only had two angel investments that's my best one okay good solutions not not very many angel investments on my side but uh that one i think is paying off very well i'm happy about that i've got friends who use it based on like why do you use it and these are obviously uh well i shouldn't say that they're newer to software development aka vibe coding which is just fine you know everybody's gonna start somewhere in the audience and they've done a phenomenal job of building that audience they have millions of people that have like that foundation they're using subbase like i that's my database then is that it is that all you're using them for is just a post-grace database okay fine but you know whatever from this i mean they've written this live and they had great experience i'm i think they should feel very proud of themselves and they've created a amazing business most people have not done that it's yeah amazing what are your thoughts on you know what you're doing with niki and multi-grace how do those two rub i suppose in the marketplace it's their attempt there is like you've just referred to there is uh one of the people that created for tests at youtube and was a co-founder plan scale that is a subbase doing this in a couple of x employees um you know they are the competition is good they're welcome to go and do this it's great it's it's different i'll tell you how it's different well first of all i'll dress that super left the company before we had a single sharded customer in production and so the the thing that i don't like about this situation is the perception that someone who made plant scale what it is today has gone to super base that is completely untrue if you look at the version of the company pre um 2020 and 2021 that is the counterfactual to what plant scale could have been and it was not and now plant scale is what it is now he played no role in what it is now and so when people online make takes about oh you know they're gonna sue base of the shot becoming anything like plant scale that is just not true none of the plant scale dna has gone to super base so i can correct that like leaving that early in in a company that's achieved so much since that period of that person leaving is hardly the boom people think it is but that's fine and i wish him well i don't like talking about individuals in the histories but they've had no problem talking trying to claim the vatest name and claim an association of plant scale and i don't like that um so we what they're building i'm have no doubt will be of quality and the people that are building it are not unserious and they are not bad engineers however they're in a business that does different things it is not the union of the 15 years experience of a test of five years of building plant scale and all the other things that we do and so it'll be interesting to see how it goes i wish everyone well i think it's great building software however when nearly like there's customers that are still on their home page that have migrated to us already when i would do postgres the logos that are on our page and you consider the long-term sustainability of what it takes to build a sharded product and the companies that run on sharding a company that's really only service-fired coders is not going to compete with a company that has done the opposite of that and is running some of the world's largest databases and that will prove out but at the end of the day i would not try and go after their audience because i don't believe i could right i do not believe and until i maybe will believe that one day but i try and stay humble and say i i i would not try and compete with super bass right now on five coding on back ends of the service they've done a really good job at this they are welcome to try and compete on what we do which is what they think they're doing with multi-grads but as someone has been through this um there is many more things that are not just you don't just have an open-source source like sharding tool is not it is their record of security is very publicly quite bad like we do not have that right these all things that end-hives care about um at the end of the day when i've seen the largest database that they ever had that moved to us and it was like i mean minutes ago in comparison to it was far below our average customer size it's going to be a long battle when they have to like do you envision a world of opposite a giant sharded postgres customer versus us and what's your largest customer and ours is four orders of magnitude larger i mean i don't know we'll see i like the competition but i don't really think about like there's other ones that are trying other sharding ones it's like why if i do it i just don't think about it i we are on schedule we are on our schedule when it comes to database scaling we set that standard industry wide um you have to start be like you have to focus in other places if you want to win that way but i do wish them well and i wouldn't hate to communicate any disrespect for the incredible business that they have built yeah so you mentioned having extraordinary engineers you mentioned having uh the ability to give those extraordinary engineers artificial intelligence ai right models uh lm's anthropic cloud codex picker picker flavors i know what my flavors are i say use them all of course right use the ones that work for you um you've got niki out there you're not sure about open sourcing it for all the reasons we've already talked about but what was it like to take that exceptional engineering team 50 how many ever was necessary or how many ever used to light stuff from that larger swather folks to build out niki what was it like to kind of have a redo i would call it like i've been enjoying the redos in in this ai world because there's there's so much more possible i'm almost happy to fail at trying to implement something because i learned and that my next version would just be that much better and more closer to what i would actually want to deliver so what was it like having had that many years of experience with the test this brand new superpower some of the most amazing engineers out there anyways and then here you go and you're building niki what was i like phenomenal i like it's just been amazing like to watch people that are so good at what they do and they're craft to have superpowers to build now we are not just five coding this thing right i sure i'm not to be generous multi-grat team as well they did a really good post about how they use ai to build and people also interpreted that as five coding which i don't think that's what they were doing either right like some exceptional engineers are building really really quickly but i'm watching phenomenal people build in phenomenal ways with ai and it is like i thought niki would take two to three years right we're a year in and we're already moving customers now that is like and i thought three years with like earlier ai right like how was a year ago when we started this and then i also thought you know just doing 15 years of work that quickly right like now is it got featured power to you with the test no will it get there soon enough yes it is amazing it's actually a great kind of meditation on the fact that code was just a thing you had to get through and that taste and now that's a terribly overused word so whatever but like but your taste your biases your experience fed into things that will write the code is truly where it was like and it gives me like there's a lot of people feeling despair about ai we are ingenious creatures we have engineered our way off of this planet to the depths of the ocean we've engineered our environment around us ai will be a tool to do that and i when we start like when ai takes over more and our ingenuity gets flowed into those that machine we will gain more as a species it's very exciting to me to see this and see how we engineer using these tools and the productivity is there and it does truly gel with the belief that i have another people have that this stuff will scale the expertise that companies have we are proving that right now like the knowledge that things we know the proprietary knowledge we know about running databases at this scale can influence how we build and produce something very quickly and that shows that there is an incredible moat to experience and intelligence and your company's culture and the way you build and the targets you set for yourself and code is always just a syntax that you have to build and you have to be smart of reading it and you have to be really good at coding and there's a massive power law to the distribution like his ability to code has been unfairly distributed across the industry but if you're listening to this in your skip tool to ai i hate to tell you that it's here and that it's happening and the best engineers are engineering in a few ways and also happy to tell you it's nowhere near too late to adopt these tools the hype slop lords on twitter that are just saying random stupid things and making everyone get triggered and upset are wrong they've always been here they've existed throughout the industry young people have dumb takes on tech the good people that are good at engineering look at this as things engineer and they just engineer it and so it has been amazing to have that do over and it taught me so many lessons that i knew in an abstract way after building software for 20 years but now i really know it yeah do you think there's one part of my question that you're answering very well and i'm just curious about this that's fine is is the redo part of it do you feel like you know i don't want to say that necessarily i'm not sure that the actual code i haven't read it because it's not a source that's cool but when we look at niki and the opportunity to redo you mentioned it's not quite where the test is at will it be eventually of course yep but was it a joy for you and for others that were involved in writing this code or you know operating the ai that wrote the code and all the wisdom that came from that was it cool was it fun to do a redo oh yeah and take two on what the test was for my sequel to do what niki is for progress members of that team have told me that they're having the most fun of their career like just to start with an empty repo yeah like we had this meaning we all got together and we just talked about like how we're gonna build this like one of the main lessons we learned from taking the tests and like the test was built by very smart people at google right and they they did an incredible thing breaking out of all these things and like i said videssa's grown and become a real mature product in the last five years in the last five years since 2021 its usage has gone up 65 000 it's pretty good it's used by more people now than it ever has been but like there's nothing like that getting it on all of your like washing all your sins away and trying again and if your test is an amazing piece of software but of course if you can just start from like you do it some days you do it just anyway right you like take one crack at something and go okay i learned like i felt around i learned all the things i'm just refreshing to just like do the first shot again and it works and it goes well and so doing that decades go with massive customers running on it you just make things differently like vitez is still a phenomenal piece of engineering and it is far superior to niki is right now anything else out there in terms of database capability but over time niki may outdo it just from the sheer nature that it's like the architecture is different and we're very very excited about it and we are excited to keep pushing vitez but tessa's getting new things at it continually because of some of the things that my people can do like i wouldn't say one's gonna be better actually than the other it's gonna be different but taking a database system that's similar to my sequel in architecture this little sequel database and then doing it again incredible fun like the real do i was incredible fun everyone's enjoyed it and i feel very lucky to watch that teamwork yeah both sides of the major database coin my sequel and postgres will forever be in some version of a war um we've always been around here postgres for life so thank you for coming around sam even though you feel that postgres is in fact inferior to my sequel and we'll leave that for a different day oh can i hit that sure sure it is just different right my sequel lost the crown because they stopped listening to their users they stopped building for developers it is a very good database to operate and i think it is still easier to operate than postgres right like high availability is easier all of these things like we run into so many issues with customers and you just have to explain to them this is how postgres does it vacuuming is bad because it's a core thing like yeah it's a core decision so similar to the replication is not as how i would like it to be but it's different they have a different philosophy and they've achieved something phenomenal they it is the default database now because people have done a good job of helping you operate it hroku us the major clouds have taken that pain away and so now you're left with a dev interface and the extension ecosystem so much better so much better so i love postgres now like getting to work with that audience and see these brilliant and amazing companies and using it and then to bring something truly differentiated from our biases into my sequel world it's phenomenal so i wouldn't say ones better or worse they're just different but like they will get better over time or like they will win them i hope we can play a role in making both metta continue at one time yeah well you're a minute or two away from a customer call and i appreciate you going literally to the one yard line with me on this podcast is there anything you want to say in closing just as we tail out in the show anything whatsoever i loved the episode we recorded all those years ago it was one of my favorite i've ever done it's one of the ones that we've been people have referred to over the years and people still you know that was maybe how they heard a plant scale and they're very complimentary about how we both did on that show and so i'm really glad to be back and it's amazing to see these enduring things that go through this evidence the industry has changed a lot while you do what you do and so it's it's amazing to have something enduring that we we anchor on and i enjoy listening and i enjoy being part of it so i just want to say thank you and everything you've done for all of us they're listeners and for building something that you should be very proud of and that we like subscribing to and i really enjoyed being part of too and i'm very happy to be back with a company that's defaulted live with hopefully users listening now and potential users in the future so thank you to yourself and your audience and everyone involved so you've got something enduring and really amazing well thank you thank you very much it's been a lot of work as you know over the years i mentioned the pobby where our inception was our birthday it was november 2009 so we've been here for a very long time we're coming up on our next birthday here later this year and i forget how many years has been but as long as i care about this community which i think will be today i die i'll be producing some version of a podcast hopefully is always the change law because i feel like the name we chose change law dot com even is like you can't erase it it's an institution it may change and have different variations over the years but i feel like we're we've always struck the right balance of who to talk to what i talk to them etc and obviously having you a part of that has been amazing so thank you so thank you for all the work you do at plant scale and the whole team there and thank you it's not me it's our customers we're only here because our customers thank you to all of them um we love our customers you know they're very special people but like you have truly achieved the change law for the industry you've been doing this long you've tracked the change you know what you should do is you should go back and like reflect on old episodes i bet you've got some amazing people that honor like there's like you can do a single company yeah i mean like it has really been the change law for the industry and you should keep going because we're about to end a new era that will be fun and different and really cool i think so too this next day is gonna be interesting and i i sure hope we all endure it as you've said i sure hope we came and we do because things change seemingly every single day but sam thank you so much for being a friend coming back on the pod sharing all your wisdom and really just sharing the struggle of what it takes to last in this industry and what it takes to be a great player in this new year i appreciate your your perspective your wisdom in your team your part your customers thank you thank you thank you you better keep that glazing in if i hear that edited out i will be furious at you you better sit there and have other people go wow yeah actually now i think about it Adam is amazing and we're great oh man you better keep it in i will be so mad you'll get my twitter i if you don't well friends thanks for tuning in to this episode of the change law joined by sam lambert ceo a planet-scale big fan with a doing over there love the way sam has led this company you know love the almost died but not default dead now they are default alive and thriving which is cool databases have been all the rage for a very long time for august reasons and now in this agentic era we're in with agent scale being slung around everywhere we need planet scale we need planet scales to focus on my sequel to focus on post-grast to focus on niki to focus on metal to focus on mvme based databases and that's cool stuff and so i hope you enjoy the show a massive thank you to our friends who have sponsored this pod we got our good friends over at build kite our good friends obviously in partners over at fly and our good friends over at code.com and of course to the beat freak in residence break master cylinder all right friends this show's done we'll see you again very soon