I selfhost Stalwart for incoming mail with the free SMTP2GO tier for outgoing mail. This is for my personal email that I currently don’t use for anything. I use gmail for signing up for stuff because I’ve been using it for a very long time and haven’t picked an alternative setup yet for handling signing up to stuff with a different email. I also use Zoho mail for more important things like banking. Zoho makes its money from enterprise subscriptions and I frankly trust it more than Proton to not enshittify short or long term.

A fork is maintained here

Your 1234 example is stored in plaintext, which is bad practice but still happens sometimes. Usually 1234 will be hashed (irreversible and functionally unique string) and the user input gets hashed with the same algorithm to check against the server’s stored hash.

Passkeys are fundamentally different and work with public-private key pairs. It’s not possible for your public key to be “stored incorrectly” because it’s meant to be public. Your private key stays private and you never have to send it to the server to be checked against what they have stored. Instead, the server sends you a mathematical challenge that you solve with your private key. The resulting answer is also irreversible and functionally unique, but in such a way that the answer can be checked with the public key to make sure it was solved by the associated private key. It’s really clever math that’s worth watching or reading up on to understand.

I don’t understand the maths well enough to give you a pseudocode example, but I always found trying to simplify it made it difficult to understand anyway. Every analogy ended up falling short because it was an analogy and not what was actually happening under the hood.

I rent a free tier oracle vm that does nothing more than tunnel traffic using GOST. Ports 80/443/25 and a control port that my homelab can connect to to establish the tunnel. No ports open at my house, public IP is the cloud VM, and the raw encrypted tcp traffic is tunneled through whatever NAT shenanigans my ISP might have and straight into Caddy within a docker (podman) network. I’m pretty happy with it and it works well!

It’s a single binary and can parse a config file or command line parameters. If I ever switch public VM providers, it’s a single binary download and a systemd service file. Switch my DNS records to the new VM and I’m completely done. And since my homelab establishes the connection to the VM’s port, I don’t have to reconfigure anything if I move buildings, switch providers, get stuck behind NAT, or can’t open ports.

I love Symfonium so much it’s such a stellar app that’s actively developed and very customizable. It was so so so worth the $5 or $10 I paid for it.

Tldr the author of this article isn’t drawing a distinction between caffeine and coffee, even though she should be, and the data is less cut and dry than the headline suggests.

Most adults can safely consume up to 400 mg of caffeine each day—the amount found in up to five eight-ounce cups of coffee—and doing so may have cardiovascular benefits, according to a July 20 American Heart Association (AHA) statement published in Circulation.

Here’s the abstract for what’s being cited (emphasis mine):

Caffeine is one of the most commonly consumed drugs in the world. It is found in various naturally occurring substances and can be ingested in a synthetically derived pure form. The majority of human subject–based research is observational and has focused on beverages and foods that contain caffeine. The relationships between caffeine and cardiovascular risk factors and diseases are complex, exhibiting heterogeneity depending on the nature of the caffeine consumed and individual-level propensities. Acute versus chronic caffeine-associated cardiovascular effects are often different. Most studies suggest an inverse J-shaped relationship between consumption of naturally occurring caffeinated products and blood pressure. Data on relationships between caffeine and diabetes are not consistent, but habitual coffee consumption has been associated with a lower risk of incident type 2 diabetes. Whereas no clear relationship between caffeine and blood lipids is evident, unfiltered coffee raises low-density lipoprotein cholesterol. Caffeine, studied primarily in the context of coffee, has been shown either to have no relationship or to be associated with a lower risk of coronary artery disease and heart failure. Randomized controlled trial data among regular caffeinated coffee drinkers showed that caffeinated coffee decreases the risk of atrial fibrillation occurrence but increases the frequency of premature ventricular contractions. Data are fairly consistent that moderate caffeine consumption, again studied primarily in the setting of coffee consumption, was associated with a lower risk of stroke. Data on high doses of caffeine such as that found in energy drinks are limited and generally suggest cardiovascular harm.