Updoots for The 4400. Might be my favorite show of all time. The theme song is still playing in my head.

Do /not/ make the mistake of watching the cw’s 4400 (without a The). I would rather watch Tommy Wiseau’s The Neighbors over it.

Add Tucker and Dale vs Evil to that category.

Hi! Although your post is full of reasonable advice on maintaining privacy online I want to challenge you on the statement that the threat model matters. The contrapositive of the statement “I don’t need privacy if I have nothing to hide” is “I have something to hide, if I need privacy”. This puts those marginalized groups you mentioned in a position where simply by using a privacy tool or technique, they draw suspicion to themselves. It might immediately raise subconscious alarms in internet communities like facebook, where the expectation is that you use your real name.

The only way privacy measures work for anyone, is if they’re implemented for everyone.

Further, I’d like to challenge the concept that a cis white tech bro has nothing to hide. There’s a big invisible “for now” at the end of that statement. The internet, mostly, never forgets. We’ve had waves of comedians get “cancelled” over tweets they made years ago. Times change, people grow, laws regress. Posting statements about abortions is as of this year, suddenly unsafe. Maybe posting about neurodivergence comes next. Who knows with the way the world is going, maybe 5 years from now you’ll regret having posts on /c/atheism associated with you.

I think a good way to be considerate of privacy is to think in terms of identities, what those identities are for, and what links those identities. Does your identity on github need make comments about your political leanings? Should your resume have a link to your github? Does your identity on etsy need to have a link to your onlyfans? Does your dating profile need a link to your reddit account? Your “2nd” reddit account? Not all of these are clear yes or no answers, they’re just things to consider and make decisions about. Also, consider what class identities you only have one of, and what class of identities are for the most part unchangeable, e.g. attaching your phone number to two separate identities functionally links them.

… but cd is a built-in

John Oliver take on it https://youtu.be/gYwqpx6lp_s

I really think that’s the secret end game behind all the AI stuff in both Windows and MacOS. MS account required to use it. (anyone know if you need to be signed in to apple ID for apple ai?) “on device” inference that sometimes will reach out to the cloud. when it feels like it. maybe sometimes the cloud will reach out to you and ask your cpu to help out with training.

that, and better local content analysis. “no we aren’t sending everything the microphone picks up to our servers, of course not. just the transcript that your local stt model made of it, you won’t even notice the bandwidth!)”

Are you using PersistentVolumes? If your storage class supports it, looks like there’s a volume snapshot concept you can use, have you looked into that?

Not sure what you’re doing, but if we’re talking about a bog standard service backed by a db, I don’t think having automated reverts of that data is the best idea. you might lose something! That said, triggering a snapshot of your db as a step before deployment is a pretty reasonable idea.

Reverting a service back to a previous version should be straightforward enough, and any dedicated ci/cd tool should have an API to get you information from the last successful deploy, whether that is the actual artifact you’re deploying, or a reference to a registry.

As you’re probably entirely unsurprised by, there are a ton of ways to skin this cat. you might consider investing in preventative measures, testing your data migration in a lower environment, splitting out db change commits from service logic commits, doing some sort of blue/green or canary deployment.

I get fairly nerd-sniped when it comes to build pipelines so happy to talk more concretely if you’d like to provide some more details!

The two factors at an ATM are possession of your bank card + knowledge of your pin. (it also takes your photo, for good measure)

GitHub will happily accept a smart card or whatever, if an extra plastic rectangle jives with you more than an OTP generator.

Your two factors shift to possession of your password vault + knowledge of the password to it. You’re okay IMO.

You also still get the anti-replay benefits of the OTPs, though that might be a bit moot with TLS everywhere.

Not a security scientist, but in my interpretation, it’s the “categories” of the factors that matter. Ideally, you use some two of three of:

• something (only) you know - generally represented by passwords
• something (only) you have - most commonly represented by some device. you prove that you have the device by providing a token only that device can generate.
• something (only) you are - generally represented by biometrics

the goal then is maintaining the "only"s.

if you tell someone your password, or they see you type it in, or they beat it out of you with a wrench, it’s no longer something “only” you know, and it is compromised.

if you use the same password on two websites, and one website is compromised, the password is compromised.

OTPs from a key fob or yubikey or something are similarly compromised if the device that provides them is left out in public/lost/stolen/beaten out of you with a wrench.

biometrics are again, are compromised if it’s not “only” you with access to them - someone scans you face while you’re asleep, or smashes your finger off with their wrench.

having multiple factors in the same category, like having two passwords, or two otp tokens, or two finger prints, doesn’t significantly improve security. if you give up one thing you remember, it’s likely you’ll give up more. if one fob from your keychain is stolen, the second fob on that keychain is of no additional help.

you can start shifting what categories these things represent though.

if you write down your password in a notebook or a spreadsheet, they become thing you have.

OTPs can become something you know if you remember the secret used to generate them.

knowing many different things is hard, so you can put them in a password vault. the password vault is then something you have, which can be protected by something you know. so although your OTPs and passwords are in one place, you still require two factors to get access to them.

you still need to protect your "only"s though. and don’t put yourself in situations where people with wrenches want your secrets.

I use passwordstore.org/ as my password manager, including for my otp codes. It’s backed by a git repo. I get a backup of it on every device it is cloned to.

Further, in terms of safety, having a large display built into your dash showing you navigation is much better than a small device you jerryrig onto a vent or something. It’s easier to see via your peripheral vision, and won’t put you in a situation where you need to go find it off of the floor when it falls off.

just to give you the term to search for, these types of applications are called snippet managers. for example, https://snibox.github.io/

there’s a ton of them around. I don’t have a particular one that I recommend, since it’s not something I use in my workflow.

The password to my password manager: a few randomly chosen words that will definitely just sound like nonsense dementia-talk probably.

Geocaching is free and usually lots of fun in cities. It’s like a big database of dead drops - people hide small containers with pieces of paper to sign, and post their GPS coordinates online. Frequently they’re hidden near points if interest, as well so you might find some cool shops or bars as a side effect.

yep. they’re still here. they got smaller, and we call them “tracking pixels” now.

it’s just an image, which, server side, you can count the number of times it got loaded. easy to embed and no js required.

That’s interesting, okay. Is svn doing compression of those binaries for you?

Not to say “you’re holding it wrong”, but I’m curious about your workflow here. You clone these binaries every time you come back to a project?

I don’t get it, who in their right mind hosts development stuff on a Windows clunker?

Same question, but Subversion. Switch to git. Import your repos with git-svn.

Hey, thanks for that, I appreciate you sharing your list.

One option you can consider with fairmail and gmail is to use an “app password” to authenticate to IMAP, instead of oauth. That might work better when backing up with neo backup?