Turns out I forgot to solder the ground and power pins! So they worked, but very poorly.

Combined with switching to shielded cable and swapping the piezo input from +1.65v to ground, it's working well now!

I'm seeing 68ms in current Chrome Canary. If I use current stable the test page doesn't work because of this bug. Filed a bug!

I did decide to redo it for the Teensy 4.1, and I hooked up all 18 inputs:

I also added mounting holes, and a bit of writing.

When you get deeper in you will hit the issue that almost every modern part is smd with no through hole equivalent.

I'm not currently planning to get deeper into this, but we'll see!

Audio science review forums will have domain experts who are much more knowledgeable than I am about this, it's very hard to make "perfect" analog acoustic circuits where any design compromises are no longer audible. But it can be done.

One nice thing about this project is that I'm not trying to capture high-quality audio: I only need it to be good enough to work as a sensor.

Testing with a breadboard the 3.3v digital seems to be good enough, and the noise I'm getting seems to be RF on the piezo lines which is hard to avoid.

Note that if you have a hot air soldering iron and paste it's not difficult to use smd parts of you order the big ones or have a microscope.

I don't, and haven't used one. I suspect it's not worth getting into it for this project?

I silkscreened the actual values not "r1...rn" and the same for capacitance. This makes hand building easier.

My current draft (as pictured here) does both, which is the KiCad default.

The post isn't trying to cover all cases of harmful careers, just ones where the career seems to be clearly net positive when approached from a costs-and-benefits framework, but still involves some harms. Trying to think about your class of objections, all the ones I can think of are covered by "that's actually net negative" and not "that's clearly positive, but you shouldn't do it anyway"?

For example, say someone cares a lot about animals and thought their best altruistic option might be working in their family's ranch. They'd (a) they'd earn a bunch of money (hypothetical!) that they'd donate to ACE recommendations, (b) they'd have some influence in the direction of better treatment of animals, but (c) they'd be complicit in raising animals for food. [1] It seems to me that the question here is whether (a) and (b) outweigh (c)? Or do you want to give additional weight to farms like this being incompatible with the stricter moral standard you think is correct?

[1] If the movement were working to outlaw ranches like this I see how working at one could undermine that, and so be another harm in addition to (c).

That's right: if it were free to include then sure, even if only 5% of attendees can read it. But it's actually quite a lot of work.

I can't tell if you're joking? But at the risk of missing the joke, where do you see this in EA philosophy?

Twilio has extended this by two years: https://www.twilio.com/en-us/changelog/Extension-of-Twilio-Programmable-Video-End-of-Life-to-December-5-2026

Speak up if you want me to keep this running until the new EOL date?

It could be fun to see how much of this is automatable: I have a camera roll that goes back to early 2012 combined with my selections for each year. That's a decent amount of annotated data!