From 1c99d11292e844375c525753572a3f14942fb2de Mon Sep 17 00:00:00 2001 From: Logbot <40303173+changelogbot@users.noreply.github.com> Date: Tue, 5 Nov 2024 12:09:48 +0000 Subject: [PATCH] Apply standardised formatter to the-changelog-614.md This commit was automatically generated by the formatter github action which ran the src/format.js script Files changed: podcast/the-changelog-614.md --- podcast/the-changelog-614.md | 34 +++++++++++++++++----------------- 1 file changed, 17 insertions(+), 17 deletions(-) diff --git a/podcast/the-changelog-614.md b/podcast/the-changelog-614.md index 2f6c1df6..d015b42f 100644 --- a/podcast/the-changelog-614.md +++ b/podcast/the-changelog-614.md @@ -26,7 +26,7 @@ So that was a really tough decision. And also, it took time to get to it, because you always hope that somehow you won't have to do it. And hopefully, I don't know, Amazon will change the name of Amazon Elasticsearch, or somehow we'll manage to work some sort of like an agreement with Amazon, or something along those lines. But when it gets to a point where so many users just confuse, either because they think you work with Amazon, either because they think that Amazon Elasticsearch is Elasticsearch, or your cloud service, or something along those lines, that we felt like we had to just go and change the license so it will force a change of name and a clear distinction between the two products that was not there. I wish that we didn't have to do it. I wish that there would have been another way to solve it, and I know a lot of people have a lot of ideas on Twitter on how this could have been solved... -**Adam Stacoviak:** \[00:08:13.05\] Oh, yes. +**Adam Stacoviak:** \[08:13\] Oh, yes. **Shay Banon:** ...but I can tell you that was one of the hardest decisions, if not the hardest decision that I've ever made. We tried a lot of them, and we felt like we had to go and do something about it. @@ -40,7 +40,7 @@ And at least for a while, people respected trademarks. You know, Apache, and Red **Shay Banon:** Definitely. I mean, like any good open source community member, I've spent a lot of time thinking about licenses since 2005, and LGPL, and Apache, and the difference between GPL and Apache, even within the open source licensing world... But yeah, the fun stuff is to build products, to talk to users and figure out what their problem is, and try to go and figure out how do you solve them, or sometimes things slightly ahead of where they're heading, and build it. And yeah, it's fun. That's where the fun is. -**Jerod Santo:** \[00:12:09.22\] What is different about a license versus a trademark when it comes to the law? Because isn't a license also a legal mechanism? Couldn't they just violate your license and you'd also have to go to court, similar to a trademark? +**Jerod Santo:** \[12:09\] What is different about a license versus a trademark when it comes to the law? Because isn't a license also a legal mechanism? Couldn't they just violate your license and you'd also have to go to court, similar to a trademark? **Shay Banon:** Totally, potentially. But I think the thing is that Elasticsearch was under the Apache license. Amazon was totally within their rights to take Elasticsearch and provide it as a service. That was never the issue. I think that sometimes companies have a problem with strip mining open source by cloud providers. There's that saying, and it's not fair to take open source software and provide it as a service, and then it's like there's a company like Elastic that invests so much, and other companies that don't invest as much in the open source end up ripping the benefits of it. But to be honest, it's legal. It's like, if you have an Apache license, then what they do is what they do, and that's totally fair. The part that bugged us is just the confusion. That's the thing that was really weird for us. @@ -56,7 +56,7 @@ And at least for a while, people respected trademarks. You know, Apache, and Red **Adam Stacoviak:** But... What you're saying, Jerod, is that it may have just been - and I don't want to use the word 'simply' to make it seem simple, but simply a trademark issue. Because that was the issue, right? AWS and Amazon Elasticsearch service was the quote from the original blog post back in the day. "Why relicense this?" It was "Amazon not okay. Why we had to change Elastic licensing." -**Shay Banon:** \[00:16:09.08\] Right. +**Shay Banon:** \[16:09\] Right. **Adam Stacoviak:** And what you're saying, Shay, is in retrospect things were different then. Maybe a license change was not the right potential way, which - obviously we get to reverse rug pulls. We're back here again discussing this. We're trying to understand why it happened in the first place. Could it have been just simply a court ruling or a court dispute over trademark law, and a trademark misuse over a community-disruptive license change? @@ -84,7 +84,7 @@ And nothing bad with Amazon as well. They're a great partner of us now, and we w **Jerod Santo:** Yes, that makes total sense. And so, in so doing, because SSPL, and I assume Elastic License V2 - neither of these are Open Source Initiative approved, according to the open source definition, which Adam and I both think that's an important definition, and they do hold a line in the sand which is important for the brand of open source itself to continue to mean what it's meant for so long... Because those don't live up to that, literally the thing that you lost, even though they are very permissive licenses and allow a lot of different uses - they are not that. And so you could no longer call Elasticsearch open source. What did that do to you, to the brand of Elastic, to you personally, to the company? What was the knock-on effects of that change? -**Shay Banon:** \[00:20:19.25\] First of all, I was very sad, and it was painful. And as someone that loves open source, and believes in it, it was a sad thing not to be able to call it open source. And I agree, by the way... I think if I understand things correctly, we could have called ourselves open source, because there's nothing illegal by calling ourselves open source. But I believe in the social contract of open source and OSI. And we stood by it. We search and replaced all open source with free and open, and other stuff. And that was not great. +**Shay Banon:** \[20:19\] First of all, I was very sad, and it was painful. And as someone that loves open source, and believes in it, it was a sad thing not to be able to call it open source. And I agree, by the way... I think if I understand things correctly, we could have called ourselves open source, because there's nothing illegal by calling ourselves open source. But I believe in the social contract of open source and OSI. And we stood by it. We search and replaced all open source with free and open, and other stuff. And that was not great. Our usage, interestingly - we live in our bubble of interest, and things like that... Our usage didn't change that much, to be completely honest. So the usage of Elasticsearch continued to grow. People just like downloaded Elasticsearch and used it and ran it, and it was free, and using it for search, and in all the other use cases that people use it for... But we were still not called open source. And I think open source matters a lot, and we always hoped that we will be able to get back to open source. @@ -100,7 +100,7 @@ I do think also that open source is also great for when new markets get created, **Shay Banon:** Free and open. It's not as strong. There's so much in open source. There's so much beauty and romantic aspects and history just in these two words; they encapsulate so much, and I love it. -**Break**: \[00:24:05.26\] +**Break**: \[24:05\] **Adam Stacoviak:** I think it's important to dig into that. I think your question is spot on, Jerod, because there was that study from Red Monk in terms of "Does it really help or hurt the company that does--" and I don't want to say rug-pull badly, but... @@ -112,7 +112,7 @@ I do think also that open source is also great for when new markets get created, **Adam Stacoviak:** Rug pull... Change license. Let's just say change license. I won't be negative. I'll be neutral at least. To change this license, to change a license of open source to non-open source, or a non-OSI-approved open source permissive license... You know, because we assume from the outside there might be a detriment to the business. But the true detriment really was not so much the misuse, but the muddying of the water between what truly is Elastic's Elasticsearch, versus Amazon's Elasticsearch. And even Werner had tweeted - I think this is back in your original post, too - that it was done in partnership. So there was no blurred of lines available anymore. So the community could then choose to use the technology; whether it was licensed open source or not, they can choose the technology that best represented their problem set, versus this other route, which was not there. The license didn't really impact that. It's important to know that, because a lot of companies struggle with this same struggle you have there, and we see rug pull after rug pull, whether it's because of an IPO company that wants to protect its moat, or some other route that over years things change. There's always a reason to do a relicense. And regardless, the community feels the pull of the rug. -**Shay Banon:** I think it's fair that there's more companies changing the license today, because they're really worried about cloud vendors. And it's funny to me that there's not more of "Hey, this is real pain that you're feeling." Startups... You know, I developed Elasticsearch in my free time, in my living room. You know what I mean? And somehow -- I don't know, maybe it's human nature. I find it funny that people take the side of Amazon in this case. I've found it hilarious. But I think there's real fear, and we need to acknowledge the fear. And if the outcome that we want is more open source in the world, why do these companies change it? It's not because they IPO'd. I'm talking to tons of startups that go and say "I'll never open source my code" that would have open-sourced the code, because they say "We don't want the cloud provider to come in and take all the stuff that we've built, and pooled something around it." I think there's a few companies that say "We need to bring trademarks back, and enforcement", and things along those lines... All of these are good discussions. Those are the discussions that we need to have. +**Shay Banon:** I think it's fair that there's more companies changing the license today, because they're really worried about cloud vendors. And it's funny to me that there's not more of "Hey, this is real pain that you're feeling." Startups... You know, I developed Elasticsearch in my free time, in my living room. You know what I mean? And somehow -- I don't know, maybe it's human nature. I find it funny that people take the side of Amazon in this case. I've found it hilarious. But I think there's real fear, and we need to acknowledge the fear. And if the outcome that we want is more open source in the world, why do these companies change it? It's not because they IPO'd. I'm talking to tons of startups that go and say "I'll never open source my code" that would have open sourced the code, because they say "We don't want the cloud provider to come in and take all the stuff that we've built, and pooled something around it." I think there's a few companies that say "We need to bring trademarks back, and enforcement", and things along those lines... All of these are good discussions. Those are the discussions that we need to have. What I can say, at least with our change, and go back to what we spoke about in the beginning... I think that enough time has passed from the hectic days of early cloud that AGPL is potentially good enough protection. AGPL is the license that we chose to be open source, and it's potentially a good enough protection. So when I talk to companies today, I say "Go open source, build things in the open, choose AGPL. It's probably good enough protection, because we're choosing AGPL." You know what I mean? Because we're showing that that's the case. So hopefully that changes their trajectory of choosing an open source license and which one ends up being chosen. @@ -122,7 +122,7 @@ What I can say, at least with our change, and go back to what we spoke about in **Jerod Santo:** Yes. And of course, AGPL - cool with OSI, or OSI cool with AGPL, so it is officially open source... What about slightly modifying the definition of what open source means in order to account for the change in the world that we've seen? Because while I believe that open source definition needs to exist, and there needs to be people that protect it and all that stuff, I'm not hardline on exactly that definition being written in stone tablets. Like, you could slightly modify it in order to broaden the tent slightly. Is this something that you've approached OSI with? Like "Hey, here's a license." Like, "Why isn't SSPL good enough? And can we change the definition slightly, because the world has changed?" -**Shay Banon:** \[00:29:50.17\] Yeah, I think it's a good question. MongoDB, that created the SSPL license... So MongoDB changed their license from AGPL to SSPL being concerned that cloud providers will take AGPL software, will test AGPL, basically. So they changed it to SSPL. So even from then I think enough time has passed that I think when you put AGPL out there, it's a good sign for cloud providers that this is the contract that you expect to have with them. I think SSPL tries to solve it. MongoDB went through a whole process, and you can read the mailing list discussions and things like that, around trying to get SSPL approved by OSI. It didn't happen. I'm not pointing fingers at one side or another. There's a lot of emotions, and the devil is in the details, and what have you... +**Shay Banon:** \[29:50\] Yeah, I think it's a good question. MongoDB, that created the SSPL license... So MongoDB changed their license from AGPL to SSPL being concerned that cloud providers will take AGPL software, will test AGPL, basically. So they changed it to SSPL. So even from then I think enough time has passed that I think when you put AGPL out there, it's a good sign for cloud providers that this is the contract that you expect to have with them. I think SSPL tries to solve it. MongoDB went through a whole process, and you can read the mailing list discussions and things like that, around trying to get SSPL approved by OSI. It didn't happen. I'm not pointing fingers at one side or another. There's a lot of emotions, and the devil is in the details, and what have you... I think there's a good question. I think there's a balance that OSI tries to put around what is open source and how do you define open source, which is very important to maintain. That social contract is important. There's a lot riding on the term of open source that we should cherish it and treat it very dreadfully, if that makes sense. And that's like rushing to adding like five licenses to it and it's fine, and then we do it, it becomes meaningless. @@ -148,7 +148,7 @@ So I can conceptually take the very, very minuscule risk of just going with AGPL I actually also think that big companies care much less about -- like enterprises. We used to say Apache is much easier for enterprises to adopt... But look at MongoDB. It's all over the place in large enterprises, and it's AGPL, you know? So obviously, that's not hurting adoption; or it was AGPL. So it's not hurting adoption. -\[00:34:16.06\] And I think a similar thing might happen in AI, and that might be the missed opportunity. We see open models, we see open this, we're saying open source AI, open source models when they don't strictly match the definition of open source. It doesn't matter about OSI or not, I just don't want the term open source to get diluted. And I think that's the responsibility of OSI, and us, by the way, to help, to try to figure out -- because we obviously play a big role in gen AI as well, and building models, and things like that... To figure out how do we build a system that allows for it. Because you just don't want it to be diluted to become a term that means everything. +\[34:16\] And I think a similar thing might happen in AI, and that might be the missed opportunity. We see open models, we see open this, we're saying open source AI, open source models when they don't strictly match the definition of open source. It doesn't matter about OSI or not, I just don't want the term open source to get diluted. And I think that's the responsibility of OSI, and us, by the way, to help, to try to figure out -- because we obviously play a big role in gen AI as well, and building models, and things like that... To figure out how do we build a system that allows for it. Because you just don't want it to be diluted to become a term that means everything. **Jerod Santo:** Right. I actually think that what Meta is doing with LLaMA and its license - which is incredibly permissive, borderline open source, but not, because of that one clause in there, that if you are operating at however many million monthly active users, then it's not for you. Like, that one little thing, which makes it not open source according to any open source definition, is similar to saying you just can't rehost it as a service, right? It's like similar to that kind of a clause, and compete with us. But they're calling it open source... And because it's so stinking awesome - what they're putting out is hugely valuable. I mean, just the raw cash value they put into training that thing over and over again... And it's great. I use it every day. And regular people now, non open source nerds like us, getting into this stuff, and they're just "Mark Zuckerberg calls it open source." It's pretty much open source. And so I think the OSI, maybe they're already missing the opportunity to define that sucker, because I think Mark Zuckerberg might be defining it for the next era. @@ -180,7 +180,7 @@ I actually also think that big companies care much less about -- like enterprise **Shay Banon:** ...whether it's monetarily, where we can contribute to OSI so they have more people and more money to help figure this out, or by working with them to try to figure this out together. I do think in such a new space, that moves so fast, like Gen AI models, open source has been hijacked, the term, from OSI. -**Jerod Santo:** \[00:38:17.10\] I think so. +**Jerod Santo:** \[38:17\] I think so. **Shay Banon:** And I think it's important to try to figure out how do we consolidate that fact that is happening as we speak. @@ -202,11 +202,11 @@ I actually also think that big companies care much less about -- like enterprise **Jerod Santo:** Yeah, 100%. I mean, I think that it's so compelling of a piece of software, slash data, slash whatever it is, that the value it brings is immense, and almost incomprehensible to everybody, except for those like seven companies that happen to hit that one clause... And it's like, we don't care about those companies necessarily. It's like, okay, they're Meta's competitors. It's like a handful of orgs. Everybody else, have at it. It's so close to open source, and so valuable, that I think it can actually completely hijack the term, and it won't mean what it used to mean... And that might just be something that we have to accept at some point. -**Adam Stacoviak:** Well, they were very forthcoming with the usage of the word open source. They used it in the hyphened version of it on llama.com. It says, "The open-source AI model you can fine-tune." So it's used there in the -- and then in the announcement \[unintelligible 00:41:09.08\] "See how LLaMA is the leading open source" - and in this case they don't use the hyphen version. They use just open, space, source, model family. They link up to a blog post that is titled -- I won't title it. It's a long one. I'm not reading that. I'm not reading that. \[laughter\] But then all throughout this linked up post, which is how LLaMA is used in this calendar year, 2024, is just open source all through it. "Open source promotes. This is a leading open source." It's again and again the usage of the word open source, so it's pervasive; pervasive usage of the word open source, hyphened and not hyphened. And so it's not like there's any dispute they're trying to say it's open source, is the point. +**Adam Stacoviak:** Well, they were very forthcoming with the usage of the word open source. They used it in the hyphened version of it on llama.com. It says, "The open source AI model you can fine-tune." So it's used there in the -- and then in the announcement \[unintelligible 00:41:09.08\] "See how LLaMA is the leading open source" - and in this case they don't use the hyphen version. They use just open, space, source, model family. They link up to a blog post that is titled -- I won't title it. It's a long one. I'm not reading that. I'm not reading that. \[laughter\] But then all throughout this linked up post, which is how LLaMA is used in this calendar year, 2024, is just open source all through it. "Open source promotes. This is a leading open source." It's again and again the usage of the word open source, so it's pervasive; pervasive usage of the word open source, hyphened and not hyphened. And so it's not like there's any dispute they're trying to say it's open source, is the point. **Shay Banon:** Yeah. And I think the interesting part is that, for example, they made the license more permissive just recently, if I remember correctly. -**Jerod Santo:** \[00:41:58.16\] Yeah, the newest version is even is even less restrictive than the last one. +**Jerod Santo:** \[41:58\] Yeah, the newest version is even is even less restrictive than the last one. **Shay Banon:** And I think this is the important work that we need to do... Because open source within models might have a lot of fidelity. There's the weights that might be open or open source, there's the retraining opportunity that you might be allowed to do or not, there's the data sources that you use to train the model... They don't have to be open source, but are you opening - which data sources? Like, just listing them, so you can go and certify them or not. @@ -222,7 +222,7 @@ So there's a lot of areas in these places that if we have a standard way to thin **Jerod Santo:** Listen. Mark, if you're out there, just take that restriction off. It'll be fine. You'll still be rich, Meta will still succeed... Just take that one little restriction off. -**Shay Banon:** Well, he can do what we do. It's like, he can just AGPL it. If you end up using this LLM, then you need to open-source the rest of your infrastructure as a service. \[laughs\] +**Shay Banon:** Well, he can do what we do. It's like, he can just AGPL it. If you end up using this LLM, then you need to open source the rest of your infrastructure as a service. \[laughs\] **Jerod Santo:** Yeah, exactly. @@ -232,7 +232,7 @@ So there's a lot of areas in these places that if we have a standard way to thin **Adam Stacoviak:** Sure. Yeah. -**Break**: \[00:44:08.07\] +**Break**: \[44:08\] **Jerod Santo:** Well, Shay, how well did this news land? Four years later you're back, you're fully open source, you have an AGPL license now... You feel great... Did the community welcome Elasticsearch AGPL with open arms? What has the response been? @@ -240,7 +240,7 @@ So there's a lot of areas in these places that if we have a standard way to thin Obviously, there's still people that try to find the bug in something good... But I think there always are. I try to focus on the fact that this is just a net positive progress for Elastic in bringing back open source, and it's hopefully an even bigger net positive for the open source world, because hopefully it will convince more companies to do open source versus less. -**Adam Stacoviak:** Why? That's the question I have. Why open-source again? I get it, but truly why? +**Adam Stacoviak:** Why? That's the question I have. Why open source again? I get it, but truly why? **Shay Banon:** I think there's a few reasons. The first one is we never stopped behaving like open source, right? All our code is still on GitHub, all of our issues are on GitHub issues, pull requests, reviews... People send us pull requests... It's never stopped. And if you go back to ask yourself why you changed the license, we changed the license because of the trademark. But the trademark is no longer an issue, because Amazon changed the license to OpenSearch. So really, the only question is why not, in that case... Because if we're not afraid of Amazon calling it back Elasticsearch - which we are not, because we work well together now, and they went with OpenSearch, and it's just totally fine... Then I think the question is sometimes like "Why not?" Sometimes the easy one is "Why not?" @@ -254,7 +254,7 @@ I also totally believe in the value of open source, as I mentioned. There's so m **Shay Banon:** I think so. -**Jerod Santo:** \[00:52:10.26\] Or is the story not over yet? +**Jerod Santo:** \[52:10\] Or is the story not over yet? **Shay Banon:** I still would like to take back the years that we were not open source. It worked out, but not in the way that I would have loved it to work out. You know what I mean? @@ -278,7 +278,7 @@ I also totally believe in the value of open source, as I mentioned. There's so m **Shay Banon:** Yes. So some people say it's great, but it doesn't align with the norms of open source, or something along those lines... Which is fine. It's still legal to do it. Our case specifically was just around the trademark. -**Adam Stacoviak:** \[00:53:59.09\] Yeah. In that case - I mean, it was a massive blip then, really. I mean, because we haven't really called out the blog post yet, Jerod, but I think this is a fantastically written blog post, and it was written by Shay himself. I'm sure you probably had some feedback from other team members that said yes or no, or - I don't know if you wrote this in isolation, but I love the way you wrote it. You wrote "DNA, love, not like us"... I'm phrasing some of the things you sort of earmarked in here. And the big part of this, the way you opened it, was the DNA. The DNA for Elastic was and has been and is to be open source, and what a pure joy it was. As I read that - and I'm sure, Jerod, we've had conversations as you read it as well... You can hear the excitement, proverbially, in your written words. We can hear -- +**Adam Stacoviak:** \[53:59\] Yeah. In that case - I mean, it was a massive blip then, really. I mean, because we haven't really called out the blog post yet, Jerod, but I think this is a fantastically written blog post, and it was written by Shay himself. I'm sure you probably had some feedback from other team members that said yes or no, or - I don't know if you wrote this in isolation, but I love the way you wrote it. You wrote "DNA, love, not like us"... I'm phrasing some of the things you sort of earmarked in here. And the big part of this, the way you opened it, was the DNA. The DNA for Elastic was and has been and is to be open source, and what a pure joy it was. As I read that - and I'm sure, Jerod, we've had conversations as you read it as well... You can hear the excitement, proverbially, in your written words. We can hear -- **Jerod Santo:** Also, can't you see... Is this Kendrick Lamar? You can see some other passion here... @@ -328,7 +328,7 @@ I also totally believe in the value of open source, as I mentioned. There's so m **Shay Banon:** Totally. That's already interesting. And that's fun. That's why we're having fun discussions. That's what we should be having. Not heavy, depressing, accusational discussions. You know what I mean? It's like, that's the fun that we need to get back to having. -**Adam Stacoviak:** \[00:57:46.05\] Well, I'm excited to see Elasticsearch be open source again. It's kind of a strange thing, because you've kind of had to be something you weren't for a little bit, you know? Not as a product, as a team, as a company, but as little as a license change as it is, it's such a core - as you've said, a DNA component to who you are. It's the beginning of what you've created and built... To have to change that for this protection. Which we have investigated, disseminated, discussed, etc, to its nth degree, more than I think we probably should ever again... But it's good to know. It's good to know, I think, the people of open source... And this will be transcribed at some point, so this will become part of the zeitgeist of what we consider as open source or not. So when all the LLMs eventually forge together and mine us for information, they'll have this conversation to look back on and say "This is the reason why they made the change, and this is the reason why they came back." I'm just happy you did. +**Adam Stacoviak:** \[57:46\] Well, I'm excited to see Elasticsearch be open source again. It's kind of a strange thing, because you've kind of had to be something you weren't for a little bit, you know? Not as a product, as a team, as a company, but as little as a license change as it is, it's such a core - as you've said, a DNA component to who you are. It's the beginning of what you've created and built... To have to change that for this protection. Which we have investigated, disseminated, discussed, etc, to its nth degree, more than I think we probably should ever again... But it's good to know. It's good to know, I think, the people of open source... And this will be transcribed at some point, so this will become part of the zeitgeist of what we consider as open source or not. So when all the LLMs eventually forge together and mine us for information, they'll have this conversation to look back on and say "This is the reason why they made the change, and this is the reason why they came back." I'm just happy you did. **Shay Banon:** Yeah. I think details matter, as you said, and for people that are interested in it, it matters. I think the end result of having more open source in the world, it's much easier to just all agree on. And if the arc of it is where we're heading, I think it's goodness. And to be honest, I actually think that -- like, if you build enterprise software, that's my recommendation. You can pick AGPL if you want to go open source, and not worry about the cloud stuff. As we discussed, I'm more worried about the large language models and AI area, which we need to figure out. That's an area that we also need to invest in. And hopefully, this podcast also helps push it in the right direction, if that makes sense.