Skip to main content

Ducktober is here - 8 launches in 8 daysSee what's live

~/motherduck/podcast

EXPLAIN ANALYZE

Digest and explain AI and data news.With veterans from the field.Less hype. More signal.

$sudounlock--all-shows

Links and show notes on every episode.

$catlatest-episode

Building Reliable AI Agents with Seth Rosen: Checkpoints, Not Workflows

October 20, 202652:57

Hosted by Jacob Matson, Mehdi Ouazza · With Seth Rosen (Co-founder, ThruWire)

How do you give an AI agent freedom while knowing it did the right work? Seth Rosen joins Jacob Matson and Mehdi Ouazza to discuss checkpoints, structured receipts, analytics agents, changing data stacks, and why data modeling skills matter for reliable AI systems.

See show notes

$lsepisodes

Building Reliable AI Agents with Seth Rosen: Checkpoints, Not Workflows cover

Oct 20, 202652:57with Seth Rosen

Building Reliable AI Agents with Seth Rosen: Checkpoints, Not Workflows

How do you give an AI agent freedom while knowing it did the right work? Seth Rosen joins Jacob Matson and Mehdi Ouazza to discuss checkpoints, structured receipts, analytics agents, changing data stacks, and why data modeling skills matter for reliable AI systems.

AWS Bought DuckLabs and Users Want Their Own Agents cover

Sep 9, 202658:47

AWS Bought DuckLabs and Users Want Their Own Agents

AWS bought DuckLabs, but DuckDB stays open source—so what changes for MotherDuck and the data stack around it? Mehdi Ouazza and Dumky de Wilde unpack the acquisition, then connect Iceberg and Lance, SQL-first agent context, LLM eval funnels, AI lineage, stateless Git, query optimizers, and the fight over who owns the agent experience.

AGENTS.md vs Skills: MLOps, Evals & Agent Governance ft. Maria Vechtomova cover

Aug 24, 202651:05with Maria Vechtomova

AGENTS.md vs Skills: MLOps, Evals & Agent Governance ft. Maria Vechtomova

Agent code can look productive right up until a dependency changes, an eval misses the real failure mode, or an over-permissioned tool turns a routine task into a security incident. Maria Vechtomova joins Mehdi Ouazza and Dumky de Wilde to connect MLOps and LLMOps with agent evals, MCP governance, regenerated software, and AI security.

AI codes with outdated libraries. Why is this unsolved? cover

Aug 6, 202649:36

AI codes with outdated libraries. Why is this unsolved?

Python 3.15 is at release candidate, 3.10 goes end of life in two months, and your agent keeps pinning 3.11 because it does not know what day it is. Mehdi Ouazza and Dumky de Wilde trace that staleness through agent observability, the Stripe and OpenRouter talks, Google's leadership shuffle, and why a context layer still cannot replace a deterministic pipeline.

The Safest Bet in AI Is Hardware cover

Jul 28, 202650:30

The Safest Bet in AI Is Hardware

The model race keeps changing, but scarce compute still collects the rent. Mehdi Ouazza and Jacob Matson unpack why hardware may be AI's safest bet, then trace what graph context, reproducible environments, BI consolidation, and loop engineering mean for data and AI builders.

AI is relearning everything databases already knew cover

Jun 26, 202647:44with Stephanie Wang

AI is relearning everything databases already knew

AI is moving inside the database, and it's hitting the same problems databases already solved. Stephanie Wang (staff engineer at MongoDB, ex-MotherDuck founding engineer) joins Mehdi and Jacob to unpack Snowflake's Cortex AISQL, why AI inside SQL is really a query-optimization and cost problem, the compute crunch pushing RAM and inference prices up, multi-agent systems and model routing, and going slow to go fast when you build with AI.

AI is turning all data viz into JavaScript cover

Jun 19, 202649:57with Yan Holtz

AI is turning all data viz into JavaScript

Yan Holtz (data-to-viz.com and the R/Python/D3 graph galleries) joins Mehdi and Alex for a full episode on data visualization in the age of AI: picking the right chart, whether dashboards are dying, and why serious data viz is converging on JavaScript and D3 now that AI writes the hard part.

Data infra is printing money. Is it sustainable? cover

Jun 13, 202645:15

Data infra is printing money. Is it sustainable?

Infrastructure keeps winning the funding while SaaS bleeds, but is it sustainable when AI agents spin up databases that die in days? Mehdi and Jacob Matson dig into Supabase's Series F, local AI on personal chips, DuckDB turning up everywhere (MariaDB, MondayDB HTAP), Anthropic's "data is not software" claim, and a wave of people building software just for themselves.

DuckDB breaks the lakehouse? ft. Daniel Beach cover

Jun 5, 202647:06with Daniel Beach

DuckDB breaks the lakehouse? ft. Daniel Beach

Daniel Beach (Data Engineering Central) joins Mehdi and Dumky for a tour through the open lakehouse renaissance: catalog commits bring multi-writer Delta + Unity Catalog + DuckDB without Spark, Cloudflare ships a unified data platform with an AI agent called Skipper, ADBC quietly replaces JDBC/ODBC, Aaron Francis launches SoloTerm, Addy Osmani names the orchestration tax, and Mitchell Hashimoto gets a 40x speedup with an agent in a loop.

161k GitHub stars in 4 months with 65 lines of Markdown cover

May 29, 202652:27

161k GitHub stars in 4 months with 65 lines of Markdown

A 65-line Markdown skill file racks up 161k GitHub stars in four months — more than two-thirds of the Linux kernel. Mehdi and Dumky dig into GitHub turning into a social platform, GitHub Action supply-chain attacks, Bret Victor's ladder of abstraction, slow tech, evals for everyone, and Lenny's AI paradox.

AI tooling wars: Bun, Claude, and open source under fire cover

May 22, 202655:03

AI tooling wars: Bun, Claude, and open source under fire

Bun ditches Zig for Rust, Anthropic pulls the rug on pricing, agents are rewriting how databases work, and open source maintainers are drowning. Mehdi and Alex pick apart the AI tooling wars and why your brain feels fried.