lnxist @lnxist@qoto.org

@freemo A PWA being a "Progressive Web Application" https://developer.mozilla.org/en-US/docs/Web/Progressive_web_apps

How many of you actually use PWA applications?

@fribbledom "Each package fulfils a single purpose" isn't this the same line of thinking that npm in the situation where packages like "leftpad" overwhelmed it's package ecosystem?

@fribbledom I solve that by using Ethernet over Powerline (or what some people call "dirty networking"). It's a very underrated solution to the problem that's cheap and easy to implement.

@freemo I'll be trying Nuxt.js next for this to see how it compares to my experience with react-static

Really looking forward to Paper Mario or Bug Fables being released to the Nintendo Switch.

@khird True! Which is why I'm still going to verbally say "kilobyte" and "megabyte". Only difference I'll probably make is using KB vs KiB where appropriate now in writing.

Apparently the distinction is for the conflicts with the SI prefix of kilo and mega being defined as powers of ten (not in relation to 1024).

I just learned that kilobyte and megabyte refer to the 1000 scale units while "kibibyte" and "mebibyte" are apparently the names for the 1024 scale units. Everything I know is a lie!

Related XKCD: https://m.xkcd.com/979/

Don't you hate when you try to Google search an error message and you only get 4 results. One of them is always in German and the others are usually someone asking the same question as you back in 2012 with either no answer or the poster saying "figured it out on my own".

@freemo @InvaderXan That same line of thinking is also what brought about the concept of derivatives in calculus.

So far, the architecture is 728x128x32x10, learning rate of 0.001 learning rate, sigmoid activation for all neurons, and using stochastic gradient descent. It's definitely lowering the error over time, but it's at such a slow rate that it'd take ~2 months to actually start guessing inputs correctly (which is definitely not ideal).

Tricky bit is getting the architecture and settings done in such a way that it can train against the MNIST dataset within a sane amount of time on my VPS.

Decided to finally implement a deep neural network engine from scratch.

I'm looking forward to the future where all the "basic" animals (Go Gopher, Rust Crab, etc) have already been assigned as mascots to existing programming languages and the new ones will be left with the "bottom of the barrel" options like the sea cucumber and dung beetle.

@freemo Are you doing it truly from scratch, or are you piggybacking off of Ethereum like a lot of "from scratch" cryptocurrencies?

@design_RG Right now I have my Intel NUC, MacBook Air, monitor, Ethernet switch, and USB power supply for my Raspberry Pi collection running off if it. I'd put my gaming rig on it, but I'd need a beefier UPS to throw it on there alongside everything else, so it's just using the surge protector.

This weekend I finally put all of my home workstations / servers on a UPS!