AI Isn't Replacing Curious Developers
In this episode of the Data Engineering Central Podcast, Daniel Beach and Neil Roberts discuss how AI is changing software development, focusing on UX, agents, LLM workflows, and what developers should do to stay relevant.
Article intelligence
Key points
- AI is as much a UX problem as a backend problem
- 'Agents' in practice differ from demos
- LLM workflows have useful applications but also hard failures
- Junior developers should be both worried and excited
Why it matters
This matters because AI is as much a UX problem as a backend problem.
Technical impact
May affect model selection, inference cost, product capability, and evaluation benchmarks.
×
Share from 0:00
0:00
/
Transcript
Daniel Beach
May 06, 2026
AI isn’t just changing how we write code. It’s changing what it even means to build software.
In this episode of the Data Engineering Central Podcast, I sit down with Neil Roberts — a developer who’s been through every major wave of the web, from BASIC on an Atari to modern TypeScript, and now deep into LLMs and agentic workflows.
This is not another surface-level “AI will change everything” conversation. We get into what is actually happening right now, where it works, where it completely breaks, and what developers are getting wrong.
We talk about why front-end and UX matter more than ever in an AI world, why most people misunderstand agents, and what real day-to-day workflows with LLMs actually look like.
There’s also a hard look at who benefits from AI, who falls behind, and whether we are quietly building fragile systems that we don’t fully understand.
If you’re a developer trying to figure out where this is all going, this is one of those conversations worth paying attention to.
Expect to learn:
Why AI is as much a UX problem as it is a backend problem
What “agents” actually mean in practice, not in demos
Where LLM workflows are useful today and where they fail hard
Whether junior developers should be worried or excited
How building apps changes when AI is part of the system
What developers should actually be doing right now to stay relevant
Neil also has a podcast, The Skill Tree, on AI and agentic-specific topics.
We also get into a bigger question most people are avoiding:
Are we heading toward AI-assisted coding… or AI-orchestrated systems where developers become supervisors?
And maybe more importantly… which side of that shift do you want to be on?
Thanks for reading Data Engineering Central! This post is public so feel free to share it.
Share
Discussion about this video
Long Live the Data Engineer. No holds barred. Talking about Data Engineering news, topics, and general mayhem.
Long Live the Data Engineer. No holds barred. Talking about Data Engineering news, topics, and general mayhem.
Listen on
Substack App
Apple Podcasts
Spotify
YouTube Music
RSS Feed
Appears in episode
Daniel Beach
Recent Episodes
Data, AI, and DuckDB
May 27 • Daniel Beach
Why I Left Facebook to Work for Myself
May 20 • Daniel Beach
Academic → CTO: What Actually Matters in Data (Matthew Housley)
May 13 • Daniel Beach
AI Is Changing Data Engineering Fast
Apr 29 • Daniel Beach
Most Data Teams Are Doing It Wrong
Apr 22 • Daniel Beach
From Industrial Data at BASF to Delta Lake Committer
Apr 15 • Daniel Beach
He Quit Apple After 13 Years
Apr 1 • Daniel Beach