levpopov
Joined 16 years ago
Joined 16 years ago
Some suggestions (I'm sure you've already thought of most of these):
- It would be great to have notebook-like UI to isolate steps and allow faster partial rebuilds.
- live editing is great, but it'd be nice to automatically stop in-progress rebuild if another edit is made while it's computing. Right now it seems to always let a rebuild complete even if it's already out-of-date.
- Search on API/docs page would be really handy
- Face/vertex filtering API looks cleaner than CadQuery's, and I'm excited to try it out. In a lot of cases I would still really like to just be able to click on the model and get a reference in code. Even if it's just an index that would break if topology has changed. Auto-generating stable references/filters would be even better, when possible.
Whisper source is very readable, check out https://github.com/openai/whisper/blob/main/whisper/decoding...
However I don't quite see the Bluetooth use case - RP2040 is not really a low power chip, making it pretty hard to use for a battery-powered IoT application. You'd need a pretty giant battery pack to make it last a long time.
Nordic's nrf52 is an order of magnitude better for a typical "sleep-burst transmit-sleep" cycle, and can be suspended to <5uA current. Pico W is $6, Seeed has a $10 nrf52 MCU, or you could get a "just hook up USB and power" bare module for $5-6.
Load up the OP's page with Chrome dev tools network tab open.
Connection start: 90ms (60ms of which is SSL handshake) Request/Response: 30ms request / 30ms response
So the whole post is about yak shaving not splitting the 30ms response portion of a request that already takes 5x that (150+ms).
Sure it's a bit faster, but your users will not notice the difference between a 14kb page and a 15kb page over https (which you hopefully have on).
Takes way less time to park and you dont need to lug multiple 5lb locks (unfortunately thats how heavy secure locks actually are).