One of my bigger DeGoogle efforts so far has been getting off of Google Photos. The process took about a week, but I’m happy with the result and thought I’d share the process.

Motivation

I don’t want Google (or any one third party) knowing too much about me. With my devices synched to Google, I have to think twice about taking pics of sensitive docs, locations, and people thanks to their penchant for running AI/ML and OCR against all of the content I upload. They’ve collaborated with law enforcement in secret to perform unwarranted searches of user data, have complied with geofence warrants in the past, and they continue to self-deal and sell user behavioral data. I don’t care if it’s in aggregate or not. Put more simply, I simply do not trust them to care about me as a user, as they have no incentive to. Surely my own infra will still be hackable, but at least an attacker will have to work for it.

Read more...

There is a saying: “the Internet is forever” – but it isn’t automatically true. People work tirelessly behind the scenes to make it so.

A lot of content-intensive technology has “lookback” built in. Commercial sites do versioning, caching, backups, archives, and more. Sometimes they even do this in a way that lets ordinary users take advantage of it; one of the earliest examples in the web era was the ability to look at the “cached” version of a search engine result. Even if the site was down, the cached version might have given you what you were looking for.

Read more...

Note: since this was originally published, Mozilla released an updated Terms of Use that appears to grant them a license to slurp your data into an AI, and removed key provisions formerly committing them to not selling your data.

See here for the dicussion. For now, I’m striking it out of recommendations in favor of one of its forks until the implications of this change are either fully clarified or reversed.

Read more...

You can’t do much these days without connecting to the Internet. Unfortunately this means exposure to:

• getting hacked by randoms
• corporate malfeasance
• government malfeasance

Your network channels are probably the first thing you should thing about securing. There are lots of other things you should do as well, but most of them are worthless without at least some network security. There is no 100% foolproof approach, so I’m ordering this by bang-for-your-buck (in terms of time and/or money).

Read more...

Today’s Internet is in many ways worse than the one we had 30 years ago. It used to be easier and more fun to use computers and to be online in, say, the 1980s or 90s. This is not just nostalgia. While the hardware and software was certainly less capable, but it mostly got out of your way and didn’t take more from you than you were getting from it because for a brief time, personal computers were connected peer to peer or only intermittently on a young internet that commerce had not found or figured out how to coopt and monetize. This is no longer the case. And it couldn’t have happened at a worse time because computing has since pervaded all aspects of society. Ads and other behavioral nudges are in your pocket, buzzing your hip every few minutes.

Read more...

In the US our shared sense of reality is strained even more than usual.

I’ve been thinking a lot about the fact that we use the same part of the brain for both moral disgust and physical disgust. Neurobiologists’ explanation goes something like this: since the concept of moral wrong is new on evolutionary timescales (<50,000 years) selection has coopted a part of the brain that was chiefly evolved for somthing else, namely, detecting and rejecting food/water that might make us sick, to also help modulate our understanding moral and social limits. So we end up have trouble telling the difference between something that’s morally wrong and something that just disgusts us due to, say, unfamiliarity. In our confusion we intellectually conflate “different” or “unfamiliar” with “wrong” – after all, they feel really similar – and hang on to the possibly erroneous conclusion even upon minor reflection. Some of us even settle on the idea that “icky = bad” is a decent rule of thumb to live by. ¹

Read more...

When All You Have is Regex, Everything is a Pattern

I often see developers use regular expressions in the process of parsing languages such as DSL, data serialization formats, and config file formats. Regular expressions are powerful, concise, supported in lots of languages as a sort of sub-language, and they are already known to many programmers, so they can be a good solution here. Obviously some people are finding success using them this way.

Read more...

Some Basics

Keyboard Issues

Use polytonic keyboard to get the stress diacritic. Not needed with Duolingo but it’s helpful for reinforcing pronunciation.

In both Windows 10 and Gnome, the input toggle is [Super-Space].

The final-form sigma “ς” used at the end of letter-case words is normally where the US-EN “W” key is.

Intro Verbs

gr	en
είναι	to be
πίνω	to drink
τρώω	to eat

“Regular” Verb Conjugation

conjugation	ending	meaning
πίνω	ω	I drink
πίνεις	εις	you drink
πίνει	ει	he drinks
πίνουμε	ουμε	we drink
πίνετε	ετε	you drink
πίνουν(ε)	oυν(ε)	they drink

Pronouns

gr	en
εγω	I
εσύ	you
αυτός/αυτή/αυτό	he/she/it
εμείς	we
εσείς	you (pl)
αυτοί/αυτές	they (masc+neut)/they (fem)

Indefinite Article

These are the singular, nominative case versions:

Read more...

Hardware

A condenser mic is pretty key here, but you might as well try with whatever you have. FWIW my condenser runs through an M-Audio M-Track 2x2 (USB-C), which provides 48VDC phantom power to the mic, as well as driving the input used by ProTools.

It’s also essential to have headphones so the accompaniment does not leak back into the vocal track during recording. I use a cheap pair of earbuds, but over-the-ear headphones might be better for attenuating any potential leaks.

Read more...

Adding Math Support

Like all bad tech blogs, I’ve detoured into a customization yakshave. My problem: I want to support markdown-side input of TeX math with some level of accessibility on the reader side, i.e., I don’t want to render the math as images, I want to retain the semantics. So I went a googlin’. [^1]

MathML: Great, there’s a representational standard (1). Now the bad news: it’s not very concise – it is only ideal as an intermediate format – it is unsupported in Chrome, and it’s poorly supported in some other browsers. The upshot is you need to use js to do last-mile part of the render.

Read more...