Zero Prime Podcast

Ethan Ding, co-founder and CEO of TextQL, joins ZeroPrime and AI Council Co-founder, Pete Soderling, to make an argument that most data leaders have been waiting years to hear: you don't have to clean everything, migrate everything, or centralize everything before AI can work on your data.

Ethan's path from Bessemer to a hypergrowth stint at Tackle, let him see how much time teams spent moving the same business logic from one database to another, over and over. TextQL was born from a different premise — that agents make data and business logic portable at near-zero cost, so the stack can be reimagined around data staying exactly where it lives. 

Today TextQL's agentic analyst, Ana, works across warehouses, BI tools, ERPs, and CRMs at enterprises like Dropbox.

In this conversation, Pete and Ethan discuss:

• Why enterprise data teams are still running a 1970s Detroit assembly line 
• Why dashboards become just-in-time, throwaway artifacts 
• How governance shifts from three-week committees to canonized metrics sourced from actual usage
• Where DIY coding agents on the warehouse break down: RBAC, semantic layer ownership, and runaway token costs
• Why warehouse bills are about to spike
• Why data leaders should own their storage, logic, and compute in an open stack

--

👉 Sign up for our "No BS" Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter

ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.

FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf

What's actually stopping enterprises from putting their AI to work? Bala Kasiviswanathan, who oversees Snowflake CoCo, Snowflake's AI coding agent, as VP of Developer and AI Experiences, says the real answer isn't access. It's fragmented context.

Pete Soderling sits down with Bala to make the case that democratizing access to data was last decade's challenge. The next one is democratizing who gets to build — without creating enterprise chaos in the process. They unpack why governance and speed are no longer in tension, and what it actually looks like when an entire organization, not just the data team, builds from the same trusted foundation.

In this conversation, Pete and Bala discuss: 

• Why the challenge is not accessing data anymore, it’s about actioning data
• Why governance and speed are no longer a tradeoff, and what changed architecturally to make that true
• What happens when marketing, finance, and ops all build from the same governed context — with examples from WHOOP and Thomson Reuters
• Why traditional engineers don't disappear in this shift — they move up the stack
• When everyone has a builder mindset it unlocks organizational productivity

--

👉 Sign up for our "No BS" Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter

ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.

FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf

Linus Lee, Head of AI at Thrive Capital, sits down with Pete Soderling to talk context engineering, monolithic agents, and the failure mode most teams aren't measuring yet.

Most teams building AI agents right now are reaching for bigger context windows. Linus thinks that's the wrong instinct. He runs AI at Thrive and ships an internal research agent called Puck that's processed billions of tokens in production — which means he's seen the failure modes most of us are still pretending don't exist.

In this conversation, we get into:
– Why context engineering is really a search problem
– Why monolithic agents get brittle the second you try to change anything
– The observability gap nobody's talking about — when your dashboards are green but your agent isn't actually doing what the user asked
– Why he's pulling 50-year-old IBM diagrams into modern AI work

Linus is speaking at AI Council SF, May 12–14, 2026.
🎟️ Grab your ticket: https://aicouncil.com/sf-2026

📖 Read the full Q&A: https://petesoder.substack.com/p/linus-lee-says-your-agent-is-too

Chapters:
00:00 Why a product person ends up at a VC firm
02:31 What working in investing taught him about data
08:49 The trap most engineers fall into with context windows
12:19 Context engineering as a search problem
15:37 The failure mode at scale nobody's talking about
17:31 What he's most excited about at AI Council

Subscribe for more conversations with the engineers and founders building the future of AI.

#AICouncil #ContextEngineering #AIAgents #LLMs #AIEngineering

--

👉 Sign up for our "No BS" Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter

ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.

FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf

For 50 years, the pattern has been the same: store the data in the database, keep the big files somewhere else, link them with a pointer. It's powered most production systems we've ever built. Chang She thinks AI is about to break it.

In this episode, Pete Soderling sits down with Chang She ahead of AI Council SF 2026 to talk about why the old data stack wasn't built for what's coming, what agents are doing to database throughput, and why anyone with a serious background in performance "starts to shake in their boots a little" when they think about agentic data access at scale.

About Chang

Chang She is CEO and co-founder of LanceDB, building modern data infrastructure for AI. Previously, he architected the ML and experimentation stack at TubiTV as VP of Engineering. In the mythical pre-pandemic epoch, Chang was the second major contributor to pandas, CTO/co-founder of DataPad, and a recovering financial quant.

Timestamps

• 00:00 — Storing blobs inline vs. as pointers: the trade-offs
• 02:34 — When you've blown past the bandwidth limit on object storage
• 04:03 — Six months trying to make Spark on Parquet work, and why it didn't
• 06:04 — The moment Chang decided to build something new from scratch
• 07:35 — Why Chang wasn't worried about adding another tool to the AI ecosystem
• 11:34 — Agents are firing 100,000 QPS, and most stacks weren't built for it
• 13:32 — Latency, scale, and the new ceiling for production AI workloads
• 14:54 — Pipelines written by agents, not humans
• 16:12 — From co-authoring pandas to rebuilding the stack on top of it
• 17:27 — Why Chang predicts "multimodal by default" within three to five years
• 19:48 — What Chang is most looking forward to at AI Council

Mentioned in this episode

• LanceDB Blob V2 API and multi-base feature
• Apache Arrow and the future of database integration
• The "hodgepodge tax" — what happens when one customer takes 24+ hours to process a single day of data
• Claude Code, Codex, and agent-driven data pipelines

--

👉 Sign up for our "No BS" Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter

ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.

FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf

On this episode of the Zero Prime podcast, Pete Soderling sits down with Vik Korrapati, Co-founder & CTO of Moondream — an open-source vision language model that lets any developer build vision AI apps through prompting, no PhD required.

Vik shares what he's learned building Vision AI in the trenches:

• Why the benchmark everyone optimizes for is the wrong one
• How an unglamorous engineering philosophy beats flashy research breakthroughs
• Why his team rewrote the parts of the stack everyone else ignores
• The hallucination problem nobody in vision AI is talking about

If you've ever wondered what it really takes to get vision AI out of the demo and into production — Vik has the receipts.

--

👉 Sign up for our "No BS" Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter

ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.

FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf

On this episode of the Zero Prime podcast, Pete Soderling sits down with Eno Reyes, Co-founder & CTO of Factory — a software development agent platform that has run missions lasting 14 days straight, with hundreds of agents executing continuously.

In this conversation Pete covers:

• Can you scale AI coding agents without things quietly breaking?
• Is there an equivalent of a PR for specs?
• What does developer trust actually require?
• What will happen to the product manager role?
• How does the quality story change for data pipelines and infrastructure, where failure modes can be silent?

Eno answers all of it from the ground up and shares a few takes:

• Agent readiness is as much a process investment as a tooling one
• The spec matters, but only as a starting point
• Winning developer trust is a UX problem as much as a technical one
• The PM role isn't dying — it's unbundling
• Data infrastructure is where most teams will get caught off guard first

If you're building with agents and want a practical, ground-level view of what quality and trust look like at scale — this one's worth your time.

--

👉 Sign up for our "No BS" Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter

ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.

FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf

On this episode of the Zero Prime podcast, tune in to our conversation with Scott Breitenother, Co-founder & CEO of Kilo Code. Scott built Kilo Code — an open source AI coding agent platform — after watching the engineering job description quietly transform around him. His take: the engineers pulling ahead aren't writing better code. They're running better agent teams.

In this conversation, Scott unpacks what that shift actually demands — why every developer at Kilo operates like a mini CEO, why he thinks the frontend/backend divide is already dead, and why the ad hoc agent is about to become always-on. If you work in software and haven't made this mental leap yet, this one's worth your eleven minutes.

Don't miss this episode if you want a practical, ground-level view of what AI coding looks like inside a company that's already built around it!

--

👉 Sign up for our "No BS" Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter

ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.

FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf

On this episode of the Zero Prime podcast, tune in to our conversation with Eldad Farkash, Co-founder and Chairman of Firebolt. With over two decades in data architecture, Eldad spent 14 years as CTO of Sisense before building Firebolt into a next-generation data warehouse.

Get a glimpse into Eldad’s approach as a rare founder who's built not just one, but two unicorns from the ground up. Eldad shares what life is like as a founder in the Tel Aviv tech scene, reveals how he applied hard-won lessons from Sisense to speed-running Firebolt to unicorn status and breaks down why he believes intelligent data ordering and storage management are the keys to guaranteeing lightning-fast database performance. 

Don't miss this episode if you want a glimpse into the future of data infrastructure with one of the industry's leading architects!

--

👉 Sign up for our "No BS" Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter

ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.

FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf

In this episode we talk to Dwarak Rajagopal – VP of AI Engineering and AI Research at Snowflake. Hear from Dwarak as he traces his journey from optimizing AMD processors to building AI systems at Google and now revolutionizing enterprise AI at Snowflake – discussing the real-world challenges and opportunities for bringing AI to business applications.

Dwarak offers compelling perspectives on why text-to-SQL shouldn’t be underestimated, how query logs can create unfair advantages, and why efficiency will be the next AI frontier. This technical discussion captures the evolution from cutting-edge research to practical enterprise value delivering real ROI.

You won't want to miss this!

--

👉 Sign up for our "No BS" Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter

ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.

FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf

In this episode we talk to Jake Moshenko – Co-Founder and CEO of AuthZed. Hear from Jake as he traces his journey from building flight simulators to creating cutting-edge authorization systems – not mincing words when discussing the security challenges in AI-driven applications.

Jake offers new perspectives on the future of permissions and access control in a technical discussion that captures the evolution of authorization from legacy systems to today’s AI revolution.

You won’t want to miss this! 

--

👉 Sign up for our "No BS" Newsletter to get the latest technical data & AI content: https://aicouncil.com/newsletter

ABOUT AI COUNCIL:
AI Council brings together the brightest minds in data to share industry knowledge, technical architectures and best practices in building cutting edge data & AI systems and tools.

FIND US:
Website: https://aicouncil.com/
LinkedIn: https://www.linkedin.com/company/aicouncilconf/
X: https://x.com/aicouncilconf