Databricks podcast mentions

Quote

When MLT is a very important thing to do, you Teams try to run complex workflows through traditional orchestration tools they hit walls. AshApp discovered this with their fraud detection models. They needed flexible compute, isolated and environments and seamless data exchange between workflows, but their existing tools couldn't deliver. That's why Cash App relies on Prefect. Now their ML workflows run on whatever infrastructure each model needs across Google Cloud, AWS, and Databricks. Custom packages stay isolated, model outputs flow seamlessly between workflows. Companies like Whoop and One Password also trust Prefect for their critical workflows, but Prefects Didn't stop there. They just launched FastMCP, production ready infrastructure for the first time. AI tools. You get prefects orchestration plus instant OAuth serverless scaling. Claude, cursor, or any MCP client, no more building auth flows or managing servers. Prefect orchestrates your ML pipeline. Fast MCP handles your AI tool infrastructure. Structure. See what Prefect and FastMCP can do for your AI workflows at AI Engine. Podcast dot com slash prefect today. Bruin is an

Quote

fully understand it, fully test it and knowing which of these uh which of these chemicals, which of these drugs to test is a big problem. So everyone in the AI biotech conversation is basically talking About the generative side, and almost nobody is building the kind of picks and shovels layer that's underneath of it. And so this is why I think 10x science is an interesting company. The next thing I want to talk about is. is neocognition. So this is an AI research lab that came out of stealth with about $40 million in their seed round. Uh you sue an Ohio State professor who runs an AI agent lab there is the founder. And the round was led by Cambium Capital and Walden Catalyst Ventures, Intel CEO Lip Bu Tang, and also Databricks co-founder. Aion Stoke both wrote angel checks into this, so I think that's a really strong signal, right? When you when you have the CEO of Intel and when you have uh kind of these high profile co-founders of something like Databricks. This is phenomenal. But essentially the thesis on this company that I think is important is that the current AI agents succeed maybe fifty percent of the time because they're basically unreported. reliable generalists. And so what they're arguing is that humans aren't great at doing tasks just because we know everything. We're great because we specialize fast when we're dropped into a new domain. So Neocognition is tryin…

Quote

fully understand it, fully test it and knowing which of these uh which of these chemicals, which of these drugs to test is a big problem. So everyone in the AI biotech conversation is basically talking About the generative side, and almost nobody is building the kind of picks and shovels layer that's underneath of it. And so this is why I think 10x science is an interesting company. The next thing I want to talk about is. is neocognition. So this is an AI research lab that came out of stealth with about $40 million in their seed round. Uh you sue an Ohio State professor who runs an AI agent lab there is the founder. And the round was led by Cambium Capital and Walden Catalyst Ventures, Intel CEO Lip Bu Tang, and also Databricks co-founder. Aion Stoke both wrote angel checks into this, so I think that's a really strong signal, right? When you when you have the CEO of Intel and when you have uh kind of these high profile co-founders of something like Databricks. This is phenomenal. But essentially the thesis on this company that I think is important is that the current AI agents succeed maybe fifty percent of the time because they're basically unreported. reliable generalists. And so what they're arguing is that humans aren't great at doing tasks just because we know everything. We're great because we specialize fast when we're dropped into a new domain. So Neocognition is tryin…

Quote

But on the defense side Army, US Air Force, um, Pentagon broadly. They are really focused on this and uh almost a parody with some of like the most mature Really? Hmm. Yeah, you think about the amount of paper that moves through the Pentagon on a daily basis. Um they've been out Okay, that makes sense. Yeah. There's been huge contracts with Databricks for a Ivana, um Salesforce is doing a lot there. Snowflake's doing an increasing amount across the federal government. government this last twelve months you've seen hundred million plus contracts awarded to Most of the big um uh model labs. Um this is all in response to a huge demand within the department for efficiencies, but also um But also for innovation. Okay. What do you think is the Uh where you guys are at versus where you guys are going, um kind of Making it a a bigger and better uh product. Like what we have been focused on will continue to focus on this is being an excellent partner for the

Quote

fully understand it, fully test it and knowing which of these uh which of these chemicals, which of these drugs to test is a big problem. So everyone in the AI biotech conversation is basically talking About the generative side, and almost nobody is building the kind of picks and shovels layer that's underneath of it. And so this is why I think 10x science is an interesting company. The next thing I want to talk about is is neocognition. So this is an AI research lab that came out of stealth with about forty million dollars in their seed round. Uh you sue an Ohio State professor who runs an AI AI agent lab there is the founder. And the round was led by Cambium Capital and Walden Catalyst Ventures, Intel CEO Lip Bu Tang, and also Databricks co-founder. Aion Stoke both wrote angel checks into this. So I think that's a really strong signal, right? When you when you have the CEO of Intel and when you have uh kind of these high-profile co-founders of something like Databricks. This is phenomenal. But essentially the thesis on this company that I think is important is that the current AI agents succeed maybe 50% of the time because they're basically unrelated. reliable generalists. And so what they're arguing is that humans aren't great at doing tasks just because we know everything. We're great because we specialize fast when we're dropped into a new domain. So Neocognition is tryi…

Quote

which are these drugs to test is a big problem. So everyone in the AI biotech conversation is basically talking about the generative side and almost nobody is building the kind of picks and shovels layer that's under. underneath of it. And so this is why I think TenX Science is an interesting company. The next thing I want to talk about is neocognition. So this is an AI research lab that came out of stealth with about $40 million in their seed round. Uh Yu Su, an Ohio State professor who runs an AI agent lab there, is the founder. And the round was led by Cambium Capital. and Walden Catalyst Ventures, Intel CEO Lip Bu Tang, and also Databricks co-founder Ion Stoker both wrote angel checks into this. So I think that's a really strong signal, right? When you when you have the CEO of Intel and when you have uh kind of these high profile co founders of something like Databricks, this is phenomenal. But essentially the thesis on this company that I think is important is that the current AI agents succeed maybe 50% of the time because they're basically unreliable generalists. And so what they're arguing is that humans aren't great at doing tasks. just because we know everything, we're great because we specialize fast when we're dropped into a new domain. So neocognition is trying to build agents that self-specialize the same way instead of kind of the current model where you hand, y…

Quote

All data teams, even in the most regulated industries, use a cloud data platform as their core center of operation, right? Whether it's data breaks. or Snowflake or GCP. And each one of those platforms offers their own LM and points that are governed by the same terms of service as the rest of the platform. And And you can use those LM endpoints for agentic coding, like you can use your even favorite agents like Club Code with the LM points that are hosted within Databricks or Snowflake for um coding. And that means that none of the data leaves your security perimeter. So LLM and data Are all within the same security perimeter from the data flow perspective, but also from the legal perspective. And furthermore, uh we've seen data platforms like Snowflake. Databricks very aggressively roll out their own agents that are even more, I would say, out of the box ready and security compliant because they just by defect. definition work within the the same environment and not using data for kind of training something that is completely outside the the environment, as far as I know. But obviously ask your lawyer. So I think I think that the maturity of those solutions has evolved for enterprise to be able to adopt those tools pretty aggressively. And I think that the adoption is lugging way behind the capabilities right now. Now digging into some of the second and third order impacts o…

Quote

layer. So we have observability and governance. And we can run this compute to our AI deployment layer all on top of Kubernetes. Right. So we start because Becoming this one platform where agents are developed and orchestrated, run. Okay. And this I'm still talking about the Next few years where agents become a real thing. But in the long term, the vision of the company is that we want to become the central compute orchestration. platform for organizations. So reasonable way of mapping True Foundry is what data warehouse companies like Databricks or Snowflake did. For data of an organization. They unleashed magic by centralizing the data of an organization. True Foundry intends to do the same thing for Because once you start bringing all the compute layers in one central control plane, you will notice that a lot of the other things start falling. falling into place and this vision that we started with in the beginning is started to become more and more apparent with agents taking control of this compute layer through MCP layer and skill sets. Skills of the of the agents basically. Let's switch to you Nikaj. Who influences the way that you work? Name a person or many persons or something. You look up to and why. Maybe like I will mention two people here, one from my personal life, one from my family.

Quote

a leg to hire your first sales leader? Do you have to only work with Kleiner to to do Uh I'll give you some anti-patterns. The first is do not just go on their LinkedIn. and look at all the fancy logos that they have gone and worked at. And immediately assume that because they were at Snowflake or because they were at Databricks, they must be good for your AI company. It just Doesn't work that way. In fact, in many cases, it's the inverse of stress. True where if you had to sell the number three product in a market. And you had to fight tooth and nail, and you were still successful there. You're probably like if you go to a great company gonna have a much higher procrastination. proclivity to do well, right? Whereas if you were I don't know, if you joined Snowflake at a hundred million of ARR and you join like their enterprise team in the Bay Area. It's like yeah, I get it, but like that's not that impressive. No offense to anybody that joins So I think that's that's one. Well it's like more like They're a fit for exactly that scenario if you're in that scenario. That's right. But you're not. That's right. That's right. Especially for startups, right? The problem with that is that you have to

Quote

connect to the engine and then you're going to be able to scale horizontally. And so uh as a as a as a company. Um our bet is that people will build more and more with rivet actors. And as you invest more into it than there are certain enterprise use cases that a lot of people don't need. And so we're technically an open company if you were to to get to technicalities. As in we feel strongly that our core should be Apache So permiss permissively licensed and be usable for like a huge amount of use cases. But we can make money off more traditional enterprise deployments, things like Databricks, for example, or like comparison. able to data bridge with Spark, for example, and with our cloud we can provide a managed solution, able to to run without having to self-host. But I strongly like I I like this company a lot because I think it's a company where we can have a very strong open core and a strong open source. velocity and a strong investment in libraries as opposed to trying to kind of walk a fine line. And then like not really being open source, but calling themselves open source uh or following a Hosh Torp and I I I hope I don't buy that boards in in in a couple of years, but I'm fairly certain that like this is gonna be this is actually going to work if you look at companies Yeah. There's a lot of companies I can list who have made this word were a lot quieter than the peo…

Quote

Aaron Powell If you fail there, you basically block your ability to adopt future innovation. You're stuck. And that clean approach has to be mirrored in the data layer. The second focus area, data management and governance. AI models are absolutely useless without the data. out high quality data, availability, quality, accessibility, these have to be mandated through robust governance. And this is where SAP Business Data Cloud or BDC. provides that unified foundation. The paper highlights a really modern feature here: collaboration with platforms like Databricks for seamless zero-copy data experience. Zero copy is a game changer. It means you are referencing the data where it lives, eliminating data silos and the massive redundancies and compliance risks that come with duplicating Sensitive corporate data across various platforms just to train a model. So the data doesn't actually move. It just stays put. Exactly. Imagine an energy company merging data from Field sensors, financial systems, and public weather reports, all without ever creating copies. That enables real-time asset management and predictive maintenance. C, using tools like SAP datasphere and master data governance, processes both structured and unstructured data while ensuring absolute compliance.

Quote

Because even if they were creating models all the time, by the time you tray them and then finish them and release them, they're already out of date. So I think um concentrating more on search was direct learnings with customers. on on this release cycle. Yeah, that's really interesting. Do you think ultimately we see both. I mean I've had this idea for a long time and it doesn't seem to be really Happening. Um in fact Databricks, you know, acquired mosaic and then kind of killed. this offering in the market as far as I know. But I've I've had this idea that if you're G E or three M that you could imagine having a model that was trained on all of your historical in-house proprietary data, which is vast, right? And you would love it if your model knew kind of on an intuitive world model basis. As much about your company and what it does and all its history as they obviously do about the broader world. Do you think that you can get there with pure search or is Is there still something to be said for kind of continued pre training or mid training? or whatever you want to call it that would try to bake in a sort of corporate world model. that presumably would complement a search, but I I don't know if it's necessary. It sounds like you maybe think it isn't.