That's Girl Code

In this solo season-three opener, Ellen recaps Databricks' 2026 Data + AI Summit, unpacking the keynote's core framework — Context, Control, Cost, and Choice — and CEO Ali Ghodsi's argument that AGI already exists but lacks the enterprise access needed to be useful. She walks through the expanded Genie product family, including Genie Code, Genie One, Genie Agents, Genie ZeroOps, and Genie App Builder, along with the underlying Genie Ontology layer that powers them. The episode also covers two major infrastructure announcements: the Rayden engine for real-time analytics, and LTAP, which unifies transactional and analytical data, eliminating traditional data pipelines. The episode closes with a brief look at Omnigent, a Day 2 open-source tool Ellen has adopted into her own workflow.

Why does representation in tech (especially in AI) actually matter? Because we've seen what happens when it's missing. Eden and Ellen walk through a history of products built without women in the room: crash-test dummies modeled on the average man that left women 47% more likely to be seriously injured, drugs dosed for male bodies, heart attacks misdiagnosed because the textbook symptoms are men's, hiring algorithms that penalized the word "women," and facial recognition that couldn't see darker-skinned women. The through-line is simple. When women aren't in the room, tools get built without them in mind, and the harm is real. Now AI is being shaped in exactly those rooms. This episode is about why that can't happen again, what the numbers at the top labs actually look like, and why getting women in early isn't optional.

It's an off-season check-in because some things just can't be an email. Ellen and Eden are back at the mic to catch up on everything since production wrapped — TGC Media is officially an LLC (yes, there's a prenup), merch is incoming, and the new website just went live thanks to Eden's Claude Design and Claude Code build.

Both Ellen and Eden came back from their work trips with the same question on their minds: how do you actually vibe code as a team? From merge conflicts and over-documented AI-generated code to whether any of it is production-ready, this one's for anyone navigating the messier side of building with AI. Plus — AI slop horror stories, Eden's surprising change of heart on in-person work, and a TikTok about Gen Z IQ decline that makes the whole "AI native" thing feel a little more complicated,

In our season two finale, we’re talking all things AISPIRATION: (AI + Inspiration) the AI tools, automations, and workflows helping us work smarter, create faster, and make life a little less manual.

From streamlining podcast production and prepping for high-stakes presentations to building custom Notion automations, organizing shopping decisions, cleaning up chaotic inboxes, and using AI as a real tool for self-improvement, this episode is one big brainstorm on how to make AI actually useful in everyday life.

This week I'm diving into one of the most fascinating infrastructure stories happening right now — data centers. We're going full 101 on what they actually are, how they work, and why AI has completely broken everything the industry thought it knew.

We cover the four pillars every data center is built on, why a random suburb in Northern Virginia became the backbone of the global internet, and what happened when drone strikes took out three AWS data centers in the Middle East earlier this year — and what that means for how we think about digital infrastructure as a target of war.

Then we get into the AI shift — why a single Nvidia GPU rack draws 42 times the power of a standard server rack, why Meta scrapped two years of construction mid-build because their facility was already obsolete, and why Google, Microsoft, and Amazon are all racing to secure nuclear power plants just to keep the lights on.

And then we go to the frontier. Space-based data centers — and why the cooling problem doesn't disappear just because you're in orbit. The moon — and why 14 days of complete darkness makes it nearly impossible. And the ocean floor off the coast of Denmark — which is actually working right now.

Sources & further reading:

• Middle East AWS attack highlights data center vulnerabilities — TechTarget: https://www.techtarget.com/searchcio/feature/Middle-East-AWS-attack-highlights-data-center-vulnerabilities
• Space-based data centers — Real Engineering: https://www.youtube.com/watch?v=osMooAvpSSs
• Underwater data centers — Denmark / North Sea: https://www.youtube.com/watch?v=eyklsgOClJA
• Equinix data center infrastructure: https://www.youtube.com/watch?v=v477fvbj3rk

If you've ever copy-pasted the same giant prompt into Claude over and over again... this episode is for you.

Eden breaks down Claude Skills: what they are, how they work, and how to set them up. Whether you're a developer or someone just trying to get more out of AI in your everyday life, Skills are one of those things that once you get it, you can't go back.

We cover:

• Why LLMs are stateless (and why that matters more than you think)
• What a Claude Skill actually is
• The brain aka SKILL.md
• How to build your first skill in 5 steps
• How to refine skills over time to avoid bloating them with tokens
• Where to find inspiration / pre-built skills you can download and use today → clawhub.ai

If you're new here or want to go deeper, check out these two episodes Eden referenced:

🎙️ Agentic Systems: Details for the Developers — the technical patterns behind how agentic AI actually works. You can use these principles to write better SKILL.md files.

🎙️ AI Basics for the Baddies — the foundational episode on how LLMs work under the hood

This week, we're covering four companies sitting at the bleeding edge of AI and robotics.

We start with Runway, the $5.3B NYC startup quietly reshaping Hollywood from the inside out. Founded by three NYU art students, Runway's text-to-video tools are already in Oscar-winning films and our very own episode promos ;) Their bet on world models could put them at the center of enterprise AI in ways nobody expected.

From there, we get into Fauna Robotics and their robot Sprout, a 3.5-foot foam-covered humanoid that is Ellen's new obsession and might be the most interesting developer platform in robotics. Sprout is a safe, expressive platform designed to live next to people.

Then we look at Daydream, the fashion AI agent built by the woman who launched Nordstrom.com and scaled Stitch Fix to a billion in revenue. It's ad-free, hyper-personalized, and asking a hard question: can a vertical AI shopping layer become a durable business before Amazon or Google absorbs the idea?

We close with Neko Health, Spotify Founder Daniel Ek's proactive health company that puts you in a chamber of 70 sensors and hands you a clinical picture of your body. Same day, no MRI, no radiation… for $370. In their first year alone, 1% of people scanned received potentially life-saving interventions for conditions they had no idea they had.

Sit with us as we chat through four companies with four very different bets on what comes next.

In this solo episode of That’s Girl Code, Ellen breaks down Databricks: why it exists, what problem it solved, and how data actually flows through it.

She explains how companies got stuck between messy data lakes and expensive data warehouses, and how Databricks’ lakehouse approach keeps data in object storage like S3 while adding reliability and performance through Delta Lake (Parquet files plus a transaction log), powered by Spark for scale and Unity Catalog for governance and lineage.

Then she maps real-world ingestion into three patterns—files landing in storage (Auto Loader), SaaS and apps pulling via connectors/APIs, and real-time streams like Kafka/Kinesis (Structured Streaming)—all landing first as Bronze Delta tables, before being cleaned and joined into Silver tables and shaped into Gold tables for dashboards or machine learning. She closes with two concrete examples: a near-real-time retail dashboard combining foot-traffic sensors, transactions, and store metadata, and a customer-segmentation ML workflow using feature engineering, Feature Store, MLflow model tracking, and even vector search to turn clusters into meaningful segment descriptions for marketing.

This week’s tech tea covers everything from China’s jaw-dropping humanoid robot showcase to Meta’s rumored facial recognition rollout in smart glasses. We unpack the Amazon Ring and Flock Safety controversy, surveillance creep, and what happens when facial data meets federal subpoenas. Then we shift gears to Elon merging Elon, breaking down the SpaceX–xAI acquisition, trillion-dollar valuations, and why the future of AI compute might live in low Earth orbit.

OpenClaw is the open-source AI agent that exploded across GitHub almost overnight. It runs locally, connects to your email, calendar, social accounts, and developer tools, and can take action on your behalf without you in the loop. It’s the personal AI assistant many frontier labs have hinted at but haven’t released.

In this solo episode, Eden breaks down how OpenClaw actually works, the origin story behind its rapid rise, the naming drama with Anthropic, and why companies like Anthropic and OpenAI have been hesitant to ship something like this themselves. From prompt injection to exposed API keys and massive attack surfaces, we unpack the security trade-offs behind giving AI “hands and feet.”