Alex Bartoletti

Building at the intersection of AI + hardware engineering

[Alex Bartoletti @AlexBartoletti1]

For the last few years I was a hardware engineer in aerospace, watching AI go from a novelty to something that quietly saved me hours.

[Alex Bartoletti @AlexBartoletti1]

@deedydas @steipete I’d love to learn more - I’m building agents for manufacturing workflows in mechanical hardware

[Alex Bartoletti @AlexBartoletti1]

The highest-leverage AI dev skill nobody teaches: deleting extra code the model wrote.

It'll happily hand you 200 lines for something that needs 20. The people who ship fast aren't the ones who generate the most - they're the ones who cut the most.

[Alex Bartoletti @AlexBartoletti1]

One of my biggest learnings in hardware was knowing which tolerances actually matter.

I used to tighten everything until mentors and experience taught me that not every dimension is critical.

Safe to say my parts started to cost a lot less after that.

[Alex Bartoletti @AlexBartoletti1]

@Brennan_Lup Maybe if you have a software background. But trying to vibe code a project past MVP to a scaled solution is a different story

[Alex Bartoletti @AlexBartoletti1]

@aaron_epstein What if the individual contributors love it but the higher ups don’t want to pay for it?

[Alex Bartoletti @AlexBartoletti1]

@JunaidAckroyd A lot of planning and yelling at your computer

[Alex Bartoletti @AlexBartoletti1]

We overrate AI models getting smarter, but the plumbing around them that does the real work is still very underrated and under appreciated

[Alex Bartoletti @AlexBartoletti1]

@MicroLaunchHQ Going well. I’m still early doing customer validation and prototyping so no live product yet. Stay tuned

[Alex Bartoletti @AlexBartoletti1]

@keithcrowwd CTO

[Alex Bartoletti @AlexBartoletti1]

@rackSpreader1 Unless your parents basement is in SF

[Alex Bartoletti @AlexBartoletti1]

@TTrimoreau Someone’s crazy enough to make it happen

[Alex Bartoletti @AlexBartoletti1]

Two hardware conferences, dozens of unscripted conversations. I never once asked anyone about "design intent."

But, 2 execs, 7 engineers, different companies, different roles brought it up on their own: there's no system that tells you who actually understood why a part was designed the way it was, before someone goes and builds it wrong.

[Alex Bartoletti @AlexBartoletti1]

@rackSpreader1 What are your best strategies here? I’m used to LinkedIn and just started exploring X

[Alex Bartoletti @AlexBartoletti1]

"Design intent" usually means clean parametric CAD - models the next engineer can edit without breaking things.

But that's just the half that gets written down.

The other half lives in the designer's head: how the part was meant to be made, assembled, inspected - the reasoning behind every tolerance. None of it makes the model.

So manufacturing and quality reverse-engineer it from chats, emails, and hallway conversations. That's where production problems start.

[Alex Bartoletti @AlexBartoletti1]

@dolylupec @Figure_robot @AGIBOTofficial Excited to follow these deployments this year. I wonder how long it will be before they have full adoption and widespread use though.

[Alex Bartoletti @AlexBartoletti1]

@levie File systems in different industries can mean very different agent interactions. Many of these file system humans use are not really that great - they’re just what people are used to. Agents provide a great opportunity to improve them.

[Alex Bartoletti @AlexBartoletti1]

@pmarca Would like to see the location distribution on this

[Alex Bartoletti @AlexBartoletti1]

@Trace_Cohen And they they have equity to stay motivated during that process