Blog
Insights on hardware, AI, and building for the physical world.
Why LLMs Won't Be Your CAD Designer
Apr 15, 2026 · 4 min read
Text-to-CAD tools are getting a lot of attention, but they're missing something fundamental about why CAD exists in the first place.

You Don't Have a PLM Problem. You Have a Much Bigger One.
Apr 9, 2026 · 5 min read
Hardware startups are flying blind and most don't know it yet. The culprit isn't your CAD tool or your spreadsheet — it's the missing layer between them.
Why Every AI Agent Needs a Feedback Loop
Mar 20, 2026 · 4 min read
Without feedback, a system cannot correct itself. The same principle that drives control systems applies directly to AI agents — and most deployments are ignoring it.

Hardware Engineers Are Clowning on AI Tools. Sound Familiar?
Mar 17, 2026 · 3 min read
The Reddit comments on hardware AI tools sound a lot like what software engineers were saying two years ago. That pattern is worth paying attention to.
FPGAs and LLMs: From CERN to Your Desk
Mar 5, 2026 · 12 min read
The TPU rabbit hole led somewhere unexpected. A particle accelerator, 40 million collisions per second, and a chip that just might change how we think about local AI inference.
TPUs and LLMs: A Hardware Engineer's Research Rabbit Hole
Mar 3, 2026 · 7 min read
What are Tensor Processing Units, how do they power AI at scale, and why can't the Coral Edge TPU run local LLMs? A hardware engineer's research notes.
Why Hardware Founders Need Better PLM Tools
Feb 25, 2026 · 3 min read
Enterprise PLM was built for Boeing, not a 4-person startup. Here is what went wrong and how a new generation of tools is fixing it.

BSD Systems in IoT and Embedded: A Hardware Engineer's Perspective
Jan 26, 2026 · 5 min read
A hardware engineer explores whether BSD systems offer advantages over Linux for embedded and IoT work, and what "hardening" really means in practice.

Redefining TRL for IoT Startups
Jan 19, 2026 · 3 min read
The TRL framework was built for aerospace and defense, not IoT startups. Here's how to apply it to hardware products before you rush to production.