🏭️ Why America Can Make Semiconductors But Not Swabs

Dan Wang on American industrial production:

Learning to build again will take more than a resurgence of will, as Andreessen would have it. And the U.S. should think of bolder proposals than sensible but long-proposed tweaks to R&D policies, re-training programs and STEM education.

What the U.S. really needs to do is reconstitute its communities of engineering practice. That will require treating manufacturing work, even in low-margin goods, as fundamentally valuable. Technological sophisticates in Silicon Valley would be wise to drop their dismissive attitude towards manufacturing as a “commoditized” activity and treat it as being as valuable as R&D work. And corporate America should start viewing workers not purely as costs to be slashed, but as practitioners keeping alive knowledge essential to the production process.

🚲️ Why We Focus on Trivial Things: The Bikeshed Effect

“Bikeshedding” is a common term in tech circles. When starting on a big new software project, start by asking a design team for opinions on which programming language to use and you’ll get to see it in action. It applies all over; humans love an opportunity to look like they’re contributing meaningfully, especially when they perceive that they should know something about the subject:

Bike-shedding happens because the simpler a topic is, the more people will have an opinion on it and thus more to say about it. When something is outside of our circle of competence, like a nuclear power plant, we don’t even try to articulate an opinion.

But when something is just about comprehensible to us, even if we don’t have anything of genuine value to add, we feel compelled to say something, lest we look stupid. What idiot doesn’t have anything to say about a bike shed? Everyone wants to show that they know about the topic at hand and have something to contribute.

⌘ Glyphfinder

Hat-tip to Julian Lehr’s recent post for the referral to this one. It’s a simple menubar app that gives you a search interface to unicode symbol sets. The speed here is phenomenal; so much faster than the built-in emoji keyboard (plus it has a much larger library).