Simplifying the AI Stack | IBM's Dr. Maryam Ashoori
Why the future of engineering is about managing AI-driven decisions.
This week, Andrew Zigler sits down with Dr. Maryam Ashoori, Senior Director of Product Management for IBM watsonx. Together they discuss the evolving AI stack for enterprise and the growing skill gap challenging developers. Dr. Ashoori shares insights from a recent survey of 1,000 developers, highlighting the need for better tools and strategies to manage the growing AI tool sprawl.
The conversation also explores the rise of AI agents, the potential of no-code AI development, and the future of software engineering in an AI-powered world.
But first, co-host Dan Lines (COO of LinearB) sets the stakes for engineering leaders everywhere: the future of technical work will evolve with agentic capabilities. Must we all become “AI managers” now? Check out the full episode below for the scoop!
“ I've seen a lot of developers that chase after technologies as a solution to figure out how they can just integrate their stack with what pops up or is popular in the market. I would say that the shift shouldn't be chasing the technologies. The shift should be focusing on what is the problem that you're trying to solve.” - Dr. Maryam Ashoori
The Download
The Download turns everyday work into something worth talking about. 🧲
1. You’re probably using Cursor wrong, so here’s a primer 🔥
Last month, Geoffrey Huntley published “You are using Cursor AI incorrectly...” and it quickly became one of our favorite reads of the month. If you thought AI coding assistants are just autocomplete, think again. This week he published an in-depth follow-up guide that teaches developers how to push Cursor to its limits with prompting rules. By using his /specs method, Huntley claims you can crank out weeks of engineering work in hours.
“Instead of approaching Cursor from the angle of ‘implement XYZ of code’ you should instead be thinking of building out a ‘stdlib’ (standard library) of thousands of prompting rules and then composing them together like unix pipes.”
📌 Start with the first post: You’re Using Cursor AI Incorrectly
📌 Then read the follow-up: From Design Doc to Code: The Groundhog AI Assistant
2. How Duolingo builds a better app by using it themselves 🦉
Dogfooding… Drinking your own champagne… or whatever your company calls it! — It isn’t anything new!
At Duolingo, dogfooding isn’t just about testing——it’s about learning. Since their product teaches, every engineer can engage with it as a real user (ie. everyone can learn), experiencing firsthand what works and what doesn’t. By embracing this mindset, they’re living the learner experience, uncovering insights that fuel engagement and retention.
For engineering leaders, the real question is: What part of your product applies to your engineers? Find that, and you’ll unlock a whole new level of meaningful dogfooding/champagne-ing.
Read: How dogfooding helps us build a better Duolingo
3. What AI agents can actually do vs. what we wish they could 🧞♂️
If you’re like me and couldn’t get enough of Dr. Maryam Ashoori’s wisdom in today’s episode, I’ve got great news for you: there’s more to check out! She’s one of the experts behind IBM’s latest report on AI agents, breaking down where they’re thriving, where they’re struggling, and what engineering leaders need to focus on right now. It’s a detailed follow-up to today’s guest, so be sure to check it out. (psst—— later this week, more is coming on Dev Interrupted 😉)
Read: AI agents in 2025: Expectations vs. reality
4. Get your guide to go beyond DORA 📈
Are your metrics driving real results, or just ticking boxes? A new LinearB guide shares insights from engineering communities (like Dev Interrupted and
) on how to go beyond DORA metrics to measure what really matters:🛶 Flow efficiency
🤩 Developer satisfaction
💥 Measurable business impact
Want to learn how top teams are driving results with this approach?
