 Thank you It's funny Gabe said the same thing next year, you know, yeah, I will replace him and I said, you know He's better. He's going to be better looking for a while So I think it might happen and thank you very much to the to the presenters before this Gaurav Jerry Jim their hard acts to follow But before I get started, I know there's a lot of live streaming happening Can I just get a sense not that I can see properly but get a sense of people how many people for the AI? So how many people here for the Cassandra summit raise your hands? Wow, great. I'll try to make sure that I talk about AI just a bit just joking so Thank you very much for this opportunity. It's been a long long long time Since we actually had a Cassandra summit and the last time we had it here was 2016 so it is awesome to get everything back everybody back together and Thank you very very much to everybody that contributes and uses Cassandra because it does Actually run the world in so many different places, right? I mean there's a phenomenal stuff going on so Absolute phenomenal innovation my journey. I'm just going to spend a couple of minutes my journey started It's funny. I see somebody in the audience who actually was was in it and My journey with two people actually Cassandra started in 2012 I was at a little company called Apogee and we were running about 45 percent of Black Friday traffic and whatever database we were using was not working and we had these massive fights between H base and Cassandra And guess who the two people were who were talking about Cassandra is going to rule the world The no-sequel Jesus Nate for many of you that know him as well as Ed and Oof and they obviously won And we've never had it down ever since and have actually done the done that piece really well It's been it's been four years at data stacks and it is It has been awesome and again I want to thank the community if to actually give us an opportunity to come back in and kind of energize and embrace the Community and work together for all the innovation that we've done in for oh and all the great stuff that we've done in 50 which is in beta now and What we are going to do when we release it as well as what we're doing with a cord and other things for five One so it's super excited about that. I'm going to talk a little bit about leading the future with nai and Cassandra So let me start with the Vinod Kosler quote for those of you that don't know Vinod Kosler He was a founder of Sun Microsystems. Most people may not remember Sun Microsystems. They were actually a big part of the client server piece We know that actually went off and started he start not started joined a company called client of Perkins One of the legendary VC firms and then has something called Kosler Ventures. It's a big investor in open AI probably one of the best investors in the history of investing and I find him to be very very bold And this was a quote he did at a Wall Street Journal event down in Southern, California Which is in 10 or 20 years 80% of 80% of jobs will get done by machines. I actually think he said five years personally I don't think he said 10 to 20 years But the best part about this quote is not whether it's five years or eight years or ten years is that we The AI deaf folks as well as a Cassandra folks have a chance to actually see this happen and have a front row seat We will not be on the receiving end of this Right, we will actually be the folks that actually help with this. This is not the first time This is not the first time that no our civilization has gone through a massive change We've gone through many of these before and we will sort through Sort through this one just like we have done in the past So here's how I get a chance to spend a lot of time with enterprises And we have some great customers just phenomenal customers and I get a chance to talk to Their view on it and this wave that's called gen AI is going to be faster than any Frickin thing we've seen Any frickin thing why you know like all other waves before of technology waves? This one actually is additive it builds on the web It actually builds on mobile and it builds on cloud without those three gen AI would not have taken off We wouldn't have gotten open AI would not have gotten a hundred million users without all three of those being in place So these time frames that you see may look a little for those of you that are in it might actually look like it's too far out These are super aggressive. So let's talk about them really quickly Incremental and efficiency based use cases are basically things like co-pilot Right and and people are dotting like crazy We have a great partnership with github and we're working very very closely with them And it's about go off and doing chatbots right making a chatbots more aggressive And I think SMB companies and enterprises will actually use them will actually adopt them in a very different rate Then because just because enterprises actually have to worry about regulatory pressures being wrong things like that And then the real fun starts with transformative and revenue use cases, right? When you start talking about not just you're not just talking about human beings using, you know Recommendations and this and that you're talking about human beings with agents right people with agents That's when it gets really really interesting That's when it actually has an impact on PNL And that's when you start figuring out when you have new revenue models, right? I mean just think about retail banking. When did the web happen? It depends on who you ask Some would say 92, you know Some would say 94 with Netscape it started taking off and when did retail banking happen retail banking happened in 2011 2010 right and the reason was because you needed the web and you needed mobile this time around it'll go a lot faster But the transformative use cases are coming and people are starting to think through it Here's what the view of the CIO looks like they have to there is no AI without data You can do all the LLM stuff that you want But at the end of the day you're not going to fine-tune the enterprise data out in the cloud Right and if you are you're going to do it in your Particular instance right and so the CIO's view is I want to consolidate my data real estate and thanks to all the Cassandra folks And all the great work we've done in 4.0 and what we are doing in 5.0 with Accord We have the opportunity to actually consolidate a bunch of databases into Cassandra, right? That is an awesome awesome opportunity The other thing that CIOs are doing is they're letting developers drive really really hard They will all decide what the AI stack looks like over the next 12 months It doesn't mean they will implement it But they are definitely doing that these are the people that write the checks for most of the people in the room Please realize this is how they are thinking about this right and they are very convinced and thank you for the open AI drama They are very convinced that It is going to be pure plays as well as hyperscalers It is not a winner-take-all with open AI with Google. I think open AI is far ahead, but I still think it's early innings Just having been through a few of these this is still early It's going to play out and Gemini is going to start getting better and better and better and Anthropics are going to pick up momentum They have access to a lot of data from the small company called Amazon and so they will actually get better We think they'll actually do really really well the developer view is they are back They're back in a massive massive way and they are the ones there is no CIO out there that is making a decision They're talking to the data science people. They're talking to the operators They're talking to everybody else, but they're asking the developers how to build genai apps They're definitely back and it always happens the beginning of a technology wave Everything everybody is experimenting with everything. They're getting a little tired I'm glad the holidays are coming up because they're looking at every tool possible And then they but they are building the future of genai apps as you go forward How do we how do we as a genai company that is completely built on Cassandra think about our success The only way we think about it is through production How many customers are going into production take a look at all these companies and they've all gone into production in 60 days or less So I'm going to walk you through a couple of examples, right? These are smaller companies that are moving fast This is in the smb category of what we talked about. So let's talk about physics. Walla love physics. Walla Walla in Hindi by the way actually means Prevair of of technology or anything you can have vegetables. You can you can also say physics any which way you look at it But it's a very interesting name They are They are actually doing serving six million users six million users Multimodal ai bot and they did it entirely in 55 days. What a phenomenal customer and we love love This is this is what we get up in the morning to do What we get up in the morning to do is actually build cool technology that that actually developers love Developers love that change the trajectory of the companies they work for and physics. Walla is one that excites us a lot The next one is skypoint. That's both based in portland healthcare company in the senior living space Tissen is here actually And so he's gonna he's gonna I think he has a chat. He has a talk They save 10 hours per week and imagine the burden they have Can you imagine their app actually hallucinating that would kind of suck Right and so they really have to work hard at it and they have Astra Cassandra right in the middle of it and it's really really good to see that make it happen I would have loved to have gone through all all All nine of the customers. So here is the journey They are multiple every company in the world has a version of this slide Mine is based primarily on what my conversation with customers are and how they view this they start with the chunking strategy They have an embedding strategy. They figure out how they do semantic search They do llm selection and by the way, this may happen out of order Not necessarily linear praise and then you they figure out what the drift is They have to then figure out how they do things responsibly Governance and other things and ethics that everybody else talked about. This is a part of their responsibilities Definitely a top-down view and then it has to be affordable We actually land up forgetting that, you know, a lot of people have racked up There's everybody has gone through cloud optimization and they do not want to go through a geni optimization cycle again Right. So so absolutely financial responsibility becomes a big part of it. So what are the components? I see this big These these are the number of minutes you left. So, uh, what what gen what gen i what what does gen i What does gen ai apps need they need ease of use? They need relevance and they need production and i'm going to talk through all three of them So ease of use you you've actually seen jerry come come up and talk about how he talked about frameworks By the way, you start the apps and you say i'm going to use python javascript and go off and build the app Which is awesome use versell. You can use root tool retool all of that that's awesome The frameworks will be things like, you know lama index you can use lang chain and others We have a great partnership with jerry and at lama index. They're doing great work And then you need data The new data types models and queries and that's what this community is going to be working on beyond 50 and above Right, you have to you have to have new data types You have to work making sure you focus on new models and obviously you have to focus on queries as well So this is going to be something that we work on as a community Now relevance is interesting for all the ai dev people you probably are very used to it right having done this for a little while I think about ease of use and scale and it's easy It actually works for me and a prize ready all that stuff relevance is absolutely A new metric to think about in an app It's not something i'm used to and i'm sure it's something that a lot of people are getting used to in the audience Or maybe you're just way ahead We use f1 because it's a combination of recall and precision and so we did a recent test Pine cone it seems to be a leader in the vendor in the in the in the vector db space So we've done a bunch of benchmarking against them J vector which is now part of kassandra and we we talk whenever you see the word astra think kassandra Right because that's what that's what it's all based on apache kassandra But here is what the results look like like 16 percent more relevant and now you may look at this and say 16 percent that doesn't seem like a lot that's right I mean that's like much much more much much more relevant and not getting the wrong answers right is really really important So we've we and we are just getting started with them. I have some other numbers as well This is just the beginning. We're just going to get better and better and better with this something that we talk a lot about inside the company production J vector is now part of kassandra 50 beta It is blazing fast index. Who is going to start clapping? Come on Blazing fast 10x faster than lucine It is no pre-source. You can go off and use it and it's available to all of you Absolutely kick ass. I love the innovation that we land up doing as a community superior performance On throughput against pinecone again some results we I was I was I was going to do what I normally do Which is take something that we were going to announce on thursday or friday and bring it into this keynote I didn't do it this time because it's a third party We actually have a third party that has done some awesome performance tests And they have gone off and done this and they're going to be releasing a report at the end of the week So stay tuned to actually see that We also did some performance testing against mongo. We obviously kicked their ass for people in the kassandra community You're used to that to seeing that right? I mean, it's just it's just something we do really easily And I think we should continue to do that and make a big deal about it. That's not funny. That's true Come on like, you know, you're used to kicking their ass, right? I mean, this is like this is the niners against cowboys No, I'm just kidding. Just a joke. Just a joke. Just a joke. All right, so um Again, um, here's really awesome the fact that we can do instantaneous indexing And actually do reads and writes at the same time and not have to wait for it Like other databases is something that we get thanks to all of you based on what we've done with SAI Secondary index is what we've done with kassandra as architecture. Uh, none of that would have been possible Yes, obviously it's available in the cloud as a service through astra But it's based all on on on open source technology that we've built as a community You can have all these effused you can have all the relevance you can have all the three put you can have the latency You can have all of that you can have the scale But if you do not think about the production filter, this is something that we've been talking about Something that we are actually educating wall street on right is that if you do not think about the production filter now It's going to get out of hand You absolutely you at the costs are going to go up You will have to decide what part of your vector about part of your database or your data state That you're going to vectorize and so price performance is going to be something that is really really important If I go back to the chart I talked about it is something that people will have to do You obviously have to worry about relevance. You have to do all of that But please be aware that Magically, you're not going to have four percent of your total budget, which is five billion dollars become ten billion dollars because of gen ai Right, it's not going to magically happen And this is going this has to be something that price performance has to be a category that you have to think through What's really interesting and I and what's interesting is you always think about ease of use we talked about this We talked about relevance right 16 percent better immediate indexing We talked about production ai you really need to make sure your HIPAA PCI SOC2 compliant things like that But the most important thing the absolute most important thing and a lot of people in the audience who build really Awesomely scalable Cassandra, right Is going to think they're going to think about why developer experience because in the beginning of every wave It is about getting things into production as quickly as possible and go through experimenting And that does not happen by making it hard for people to do You have to focus on focus on developer experience because it's about getting production apps Quickly you will not know which apps will take off you might and it might take two years But there are a lot of small apps are going to take off as well that are going to affect Your reputation as well as the companies that you work for and so ease of use becomes something that we as a community You just start thinking about as well So thank you i'll end by saying thank you very much again to all the people in the Cassandra community for the Database we built together It is absolutely the best database for gen ai All right, let's just be very clear. It is the best database for gen ai, right? I mean there's no if then buts about it And it starts with data Everybody in the audience has been working with it. It starts with you. There is no reason There is absolutely no reason Why any other gen ai database should be used in the companies you work for No, absolutely no reason. We have the performance data. We have everything. We are absolutely there You should all take it on as your responsive make sure that you kill everything every other gen AI database that comes in your company. There shouldn't be why shouldn't Cassandra rule the gen ai world Yeah, it should absolutely right why shouldn't it and so I think the responsibility is with you and with everybody else in the community And it starts now. Thank you very much