Lorenz Verzosa

Cross Posting #Databricks CEO Just Dropped The Most Honest Advice About The Future Of AI The next wave of AI success is going to be completely dependent on structured data. Maybe that’s a no-brainer to you. Maybe I’m telling you that the next time you take a drink of water, that water will be wet. They don’t say things like all software developers will be replaced by AI (the deadline for which has since passed) or all tech employees should become baristas and bartenders (the dream is over, nerds) or don’t worry, we’ve hammered out all the unintended consequences for letting users create their own AI At the same time, AI is, theoretically anyway, offering executives and leaders actionable insights, options, and answers. AI Is Motion, Data Is Experience AI is fire, while data is fuel. https://lnkd.in/gqcRq6vY

2

Databricks just shipped Zerobus Ingest to GA. The pitch: stream data directly into Delta tables, skip the Kafka cluster entirely. Sub-5-second latency. 10+ GB/sec throughput. For teams doing high-volume sensor and telemetry ingest, this is interesting. Not because Kafka is bad (it's not), but because most IoT data pipelines don't actually need pub/sub fanout. They need a fast, reliable pipe from source to storage. And operating Kafka for that is like renting a warehouse to store a bicycle. https://lnkd.in/guvXmJTJ We see this pattern constantly at Tiger Data. Teams running industrial IoT workloads with millions of sensor readings per second, and half their operational headaches aren't the database. They're the middleware. The message bus. The glue code between systems that shouldn't need glue code. This is why we built Tiger Data Cloud around direct ingest into TimescaleDB. The less stuff between your sensors and your database, the fewer things break at 2am. Streaming architecture is getting simpler. That's a good thing. #IoT #TimeSeries #DataEngineering #Streaming #TigerData

14

You don’t leave Databricks after 9 years unless something real pulls you — and Sameer Paranjpye isn’t someone who needs to go anywhere. He helped build one of the most important software companies of the last decade, and that kind of experience just changes your lens on everything… what scale actually feels like, what breaks when things move fast, what’s actually worth doing next. Somewhere in the middle of this, Baseten hit a different gear. Revenue started to inflect, demand picked up in a real way — you could feel the company changing while we were in it. So we stopped and reset. What we were originally looking for just wasn’t good enough anymore. When a company crosses that line, you don’t hire for upside or potential — you go get someone who’s already lived it, who knows how to scale without slowing things down or overcorrecting. That’s why this took time — close to 10 months, which isn’t exactly something you brag about, but every once in a while it’s worth it to just stay in it until the right person shows up. And that ended up being Sameer. People at his level don’t take calls unless it’s obvious — big problem, right moment, founders fully in it — and credit to Tuhin Srivastava, Amir Haghighat, Nicole Kaylor, Alana North, and Sofia Dzialga who just didn’t let up on this… nights, weekends, whatever it took, no shortcuts. This isn’t really about leaving Databricks — it’s about recognizing when something is worth running toward. Baseten is very clearly in that category. https://lnkd.in/gWvK44xR

129

4 Comments

Databricks Lakebase just dropped! It finally bridges operational and analytical databases - something data pros have wanted for years. By bringing separation of compute and storage to Postgres-like operational systems, it enables AI agents to spin up sandboxed, versioned databases in seconds. Imagine your coding agent testing new features on a full product DB clone - safely, instantly. Total game changer.

25

1 Comment

"The hard problem was governance at the access layer. This, absolutely this...60-90% of projects fail due to data not being AI ready...

3

When two $100B gorillas connect, the junior partner has to suck its data throw a very thin straw. Not a good look for either side - Databricks should provide ADBC or ArrowFlight, while Palantir looks like an add-on that Databricks haven't gotten around to implementing yet. JDBC is not how to connect data between big data platforms. Arrow, with ADBC or ArrowFlight, gives columnar to columnar transfers without data serialization/deserialization and without columnar to row-oriented pivots. Hopsworks has had ArrowFlight support since 2023 - https://lnkd.in/dyAAcUSv Databricks and Palantir discuss their new integration in this video - https://lnkd.in/d_c_jSqD

50

8 Comments

Carlos Guzman’s #OpenMed walkthrough is a fantastic proof point: bring a healthcare-tuned model to your data, run it at scale, and only pay while it’s running. 1M clinical patient notes, 4 GPU’s, 5m30s, ~$2–3. This is how we vault healthcare forward with AI —by unlocking real use cases at low cost. #Snowflake #CortexAI

12

Excited to be partnering with Databricks and several of their customers as we build out both contract pods and permanent teams of top-tier Databricks professionals. What we’re seeing across the market is clear: The new wave of AI Ops / MLOps engineers is becoming one of the most in‑demand skill sets - experts who can deploy models into live environments, operationalize AI at scale, and fully leverage the power of the Databricks platform. Across multiple projects, this capability gap is growing fast. Are you seeing the same on your side? If AI model deployment, platform operations, or scaling your Databricks talent is becoming a bottleneck for your team or your customer projects, let’s connect. Always happy to share what we’re learning and help where I can. Barney Green Nathan Miller #Databricks #AIOps #MLOps #AIEngineering #DataEngineering

13

5 Comments

Big news: Snowflake Ventures is leading a strategic equity financing in AtScale. This investment validates the critical role of semantics in the modern enterprise data stack. As enterprises consolidate data on Snowflake AI Data Cloud, a unified, governed semantic layer is essential for consistent metrics and business logic. A semantic layer like AtScale makes accessing that data more accurate, accessible, and efficient by bridging the gap between business language and data language. AtScale’s Universal Semantic Layer provides this single source of semantic truth for customers like Fidelity Investments and The Home Depot. The company remains independent and interoperable, supporting a broad ecosystem of data platforms, BI tools, and AI applications. This is a good moment to reflect on the future of data and AI. New to semantic layers? I'll post a tutorial link below. And here are five big topics from the 2025 Semantic Layer Summit (link to Substack articles on these below): 1️⃣ Semantic layer standardization is coming 2️⃣ Semantic layers are BI’s Gift to AI 3️⃣ Semantic layers make AI 3X more accurate 4️⃣ Long live SQL (with semantic layers to help) 5️⃣ The great debate: Should semantics be in your database? This is a pivotal moment for the data industry, and I’m thrilled to be a part of it! https://lnkd.in/eVJ276Pp

88

10 Comments

It’s not that clinical trials need more dashboards, what it needs is less friction. Because right now, trials are still moving at the speed of paperwork, while the rest of the world is operating at the speed of light with AI. That gap is the real bottleneck. Here’s the paradox healthcare is living in today: we can design drug candidates in weeks, but it still takes a decade to get those treatments into patients’ hands. Sometimes it’s not because the science isn’t ready but because the systems around it aren’t. That’s the problem we’re focused on at Ryght AI. We’re streamlining site selection and feasibility at the very start of the trial lifecycle, helping sponsors and CROs identify the right sites faster, forecast enrollment with more confidence, and move from protocol to activation without the usual drag. And yes, building AI agents in an industry that’s historically… let’s call it “cautious” with technology hasn’t been the smoothest ride. It’s been bumpy. Sometimes very bumpy. But it’s happening. Traditionally, site selection can take 6 to 12+ months. We’ve brought that down to weeks and in some cases, within a day. Feasibility cycles that used to take months now take days. Documentation that dragged on for weeks now under an hour. Our goal is simple: make clinical trials faster and more accurate, so life-changing treatments reach the people who need them, sooner. #RyghtAI #ClinicalTrials #HealthcareInnovation #ResearchSites #InnovationWithPurpose

42

4 Comments

It’s time to be real about what HTAP actually means and how to implement it in a way that supports agents and other similar workloads…

11