The longer something has existed, the longer it is likely to continue existing. Going with Gmail or a custom domain is a really safe bet. There is no way I’d use an email service that’s less than 10 years old now.

> The longer something has existed, the longer it is likely to continue existing.

aka the (postulated) Lindy Effect

https://en.wikipedia.org/wiki/Lindy_effect

Just a different way of looking at survivorship bias (“all of these old buildings were built really well!”).

With the way Google and other providers have shutdown accounts with impunity, it’s hard to think of anything but a custom domain being a safe option.

You can move a custom domain easily enough. I am about to do that to move away for the GSuite’s fees. Worst case, I’ll invest into running a service on my own. Best case, I’ll just use Proton or FastMail, etc.

I guess registrars can go down too but it seems less likely.

> The longer something has existed, the longer it is likely to continue existing

horses for carriage seem to defy that kind of assumption

Gotta love these kind of sayings that have virtually no connection with reality.

Kind of a bad analogy because if Gmail is the horse, it can also become the carriage.

It’s a probability thing. Obviously everything old was once new. And everything dead was once old. But take any thing that exists right now and the most likely outcome is based on this rule.

If we take the C programming language, and one created last year, which one would you expect to still be used in 10 years? You can do this comparison for any two things where there is no inherent end date.

> If we take the C programming language, and one created last year, which one would you expect to still be used in 10 years?

That's a different thing altogether. Because it does not take in account market adoption. If C was a language on life-support with no major project using it, it could be as old as hell but I would not bet on it to be around and useful 10 years from now.

Also, that saying is stupid because it makes it look like you can predict the future of technologies by using a single factor, while reality a rich, multi-dimensional set of problems.

On the contrary, it is solidly rooted in math. On average, and without other context, a random sample will be close to the middle.

Given a random sample of any thing at an arbitrary point in time, on average it will be in the middle of its lifetime. Something that is X years old will on average, continue for X more years. Of course, additional information lets you refine that estimate considerably, but it is mathematically sound and surprisingly applicable to everyday life (e.g. using a single serial number on a product to estimate total production).

> On average, and without other context, a random sample will be close to the middle.

The thing is, we typically have a lot of context for every technology we discuss, so bringing back this saying that reduces everything to one-dimension is just silly.

You typically don't, you just think you do much like how stock pickers think they know but end up failing to beat the market average. There's too many unknowns and variables to even come close to an educated guess.

If you are truly an expert, you have a fighting chance. As a human, that means you are limited to a tiny number of specific subject matters.

Gmail isn't old enough. Go back another generation, AOL, Yahoo mail, etc. Best bet is probably some university email address. Now those are long-lived, if they let you keep it.

There is a sad irony behind that link…If I had to pay for bookmark management, it’d be Pinboard. What are the advantages?

I'm generally in favor of owning your data and domains so yes. There is so much good open source stuff now.

It's unlikely that Gmail will shut down though so there are a lot of reliable exceptions

I agree. I trust this one more than if Google would launch a free, customizable "start page" service like it. They're killing their free services like no tomorrow.