May 7, 2026 · 22m · latent-space
Senior Dev: This "Grill Me" Prompt Is Going Viral Among Top Engineers
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this interview at AI Engineer Europe, educator Matt Pocock joins Swyx to discuss why classical software engineering principles, structured prompt techniques like the viral 'Grill Me' prompt, and disciplined technical pedagogy are essential for building reliable AI systems amidst rapid tooling churn.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The hosts as informed peer 5.0 Guest teaching 4.6 Guest disagreement 1.3 The hosts pushing back 2.1
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 17:05 Evals are unsexy contrarian take
Matt dismisses the hype around AI evaluation frameworks, claiming that evals are not sexy and audience engagement on the topic falls flat.
Hardest push from the hosts ▶ 18:18 Pushback on agent control preferencesShawn directly interrupts Matt's premise on agent architecture, questioning whether developers truly desire more granular control over black-box harnesses.
Biggest teaching moment ▶ 7:21 Embedding DDD latent representations into promptsMatt explains how classical Domain-Driven Design constructs allow LLMs to reliably communicate using pre-trained architectural concepts.
The host holds their own ▶ 4:24 Narrow waist networking model for AI systemsShawn showcases deep systems knowledge by mapping internet protocol layering and deep modularity to AI workflow boundaries.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Welcome to AI Engineer Europe and Rapid Tooling Shifts | 4 | 2 | 2 | 1 | Shawn welcomes Matt to AI Engineer Europe and explores his transition from TypeScript educator to AI tooling teacher. Matt lightly rejects the pundit label, clarifying his role purely as a teacher focused on fundamentals. | |
| Applying Classical Software Design Principles to AI Coding | 7 | 4 | 1 | 1 | Matt explains his thesis of adapting classic software design texts to AI coding. Shawn displays strong technical expertise by introducing the internet architecture concept of the 'narrow waist' and MVC analogies to contain AI slop. | |
| Domain-Driven Design and the Ubiquitous Language Prompting Skill | 5 | 6 | 1 | 2 | Shawn asks about the viability of Domain-Driven Design in modern software. Matt educates the audience on how DDD concepts like ubiquitous language exist in the latent space of LLMs and describes his automated skill repo. | |
| Pedagogical Workflow and Structured Course Development | 3 | 6 | 0 | 1 | Shawn inquires into Matt's rapid content production and teaching methods. Matt details his prior career as a voice coach, his Zettelkasten Obsidian pipeline, and strict single-objective lesson structures. | |
| Educational Frameworks: Knowledge, Skills, and Wisdom in AI | 5 | 4 | 2 | 4 | Matt breaks education into knowledge, skills, and wisdom, remarking that users still prefer traditional lectures. Shawn playfully pushes back that passive lectures are the proven optimal medium, prompting Matt to defend mandatory hands-on exercises. | |
| TypeScript's Ascendance Over Python in AI Engineering | 6 | 5 | 3 | 6 | Shawn and Matt discuss TypeScript overtaking Python in AI engineering. Shawn challenges Matt when Matt claims evals are unexciting and queries whether developers actually want inverted control over agent harnesses. | |
| The Viral 'Grill Me' Prompt and Cross-Domain Generalization | 5 | 5 | 0 | 0 | Shawn shares surprising stats about Obsidian's lean four-person team. Matt explains his viral 'Grill Me' prompt pattern, which instructs the model to interrogate user assumptions to establish shared understanding. |
Statements from this episode (13)
Pocock: AI courses require short cohorts rather than self-paced formats
“I run courses in two-week cohorts these days because AI moves so quickly. Hard to release a self-paced course and guarantee updates forever.”
Pocock: Codebases that are easy for humans are easier for AI
“If you have a code base that's easy to change for humans, it's going to be easy for AI to change too. And it's going to mean that you can work a lot faster because A code base that's difficult to change for humans is going to be even worse for Aeon.”
Swyx: Introducing narrow waists in codebases helps developers contain AI slop
“And I think introducing narrow waste in a code base Allows you to contain slop, right? You can say, I don't care what goes on in here, but what goes in is defined by me. What goes out is defined by me. Do whatever you want inside.”
Pocock: LLMs already understand Domain-Driven Design vocabulary
“Because it's been going around so long, it's kind of in the latent space of these models already, is if you say, you don't need to invent this whole new set of terms, you can just bolt on this framework, this system that's very, very flexible, very composable,…”
Pocock: Students prefer traditional lectures over experimental AI teaching methods
“What I've noticed is the more I lean into the kind of AI experimental stuff, the more it actually turns people off my materials. I'm finding that actually a traditional approach is what- Give me the lectures, yes. Yeah, what still works for most people. Most p…”
Swyx: TypeScript overtook Python in the annual GitHub developer survey
“I think this year is the year that TypeScript overtook Python in the GitHub survey, and I didn't really foresee that coming.”
Pocock: Developers building AI chat apps and UX mostly use TypeScript
“Everyone who's building this stuff, everyone who's building chat applications is mostly doing it. If you're concerned about UX, you're concerned about shipping great stuff, you're mostly doing it in TypeScript. At least that's what I see.”
Swyx: TypeScript could win the AI engineering ecosystem over Python
“I think it could be that TypeScript is going to win AI engineering, and that's something I haven't anticipated or seen, or I don't even know how to sort of Position around this, because I think it really does mean a lot of different things for what frameworks …”
Pocock: Software developers are generally not interested in AI evaluations
“People are not really interested in evals, you know, like evals are not sexy. Like no one's excited to do evals these days, right?”
Pocock: Coding agents are converging on file-based systems and simple tools
“I think that I think that the coding agents kind of debate is sort of ending really like they're mostly coagulating towards a similar approach, which is, you know, file based systems, a simple set of tools.”
Pocock: Inversion of control to developers will become critical in AI tooling
“And I think that coming up for the next couple of, maybe the next few months, let's say, I'm not a clairvoyant, that the inversion of control is going to be really important, where you put more control in the hands of the developer and less in the harness.”