When you go serverless you are really just massively overpaying for uptime and thoughtlessness about scale.
The thoughtless and their dollars are soon parted.
Given his unexpected shock I hope all those DevOps hours saved got him to market sooner - as is the promise of serverless. That pitch is easy to misunderstand as "saved time is always money saved", especially when it is so obviously an initial cost and ease of use gain out the gate.
It's more of an introductory offer that lets low rollers act like big rollers in the big leagues so long as they pay the big roller bills when the time comes.
I'm the person who originally set Cara up on Vercel in late 2022, before I stepped away to focus on my own startup.
I chose Vercel for a few reasons:
1. When you have 0 users, worrying about scale is a huge time sink. I wanted the team to be able to focus on building a product people wanted, acquiring users, and retaining them.
2. Vercel's preview branches make it lovely to preview work in progress, and QA pull requests. This is critical given Cara was and still is a small team of volunteers.
3. NextJS is lovely to work with, and choosing something both popular and with a JS/TS base made it easy to onboard new volunteers.
If I was doing things again, I'd still choose Vercel. The only difference is I'd enable spend management for their team (a feature Vercel released in late 2023: https://vercel.com/docs/pricing/spend-management).
> Your account does not appear to have spend management enabled, which would allow you to pause your project entirely if you hit a certain level of spend.
i mean yeah if you have 500k users with 56 million function invocations per day, it is expensive on a service like vercel. if you want to scale you should choose cheap services
it blew up to that amount virtually overnight- like maybe a week at the most. I don't even think they had the time to do anything.
after using it for a year already, I really need this site to succeed. It's one of the few I'd happily pay money to use, even if I don't get anything back.
a while ago, before Bluesky or Cara or anything, a similar issue happened with a site called Hive iirc, it did a great job at fostering a userbase right away and things went great but because they couldn't keep the site up and also develop it, they decided to shutter the entire thing for a month and buckle down.... and it completely lost its entire userbase, its momentum was destroyed, everyone moved on. I don't want to see that happen to Cara.
I hope they implement some other situation soon. A one-time fee to register, maybe monthly subscriptions for perks, or something. They may need it sooner rather than later and at least this is a nice stopgap
The thoughtless and their dollars are soon parted.
Given his unexpected shock I hope all those DevOps hours saved got him to market sooner - as is the promise of serverless. That pitch is easy to misunderstand as "saved time is always money saved", especially when it is so obviously an initial cost and ease of use gain out the gate.
It's more of an introductory offer that lets low rollers act like big rollers in the big leagues so long as they pay the big roller bills when the time comes.