They're telling you why: having a free tier means allocating a lot of resources to fraud and abuse, which are rampant in hosting, especially now that you can convert hosting dollars almost directly to cash. It's a big problem, and I have every reason to believe they're being forthright about why they've made this decision.

That may all be true but I don't see any other way to frame this other than them throwing in the towel for the whole platform and/or conceding to other platforms that are better at spotting fraud/abuse.

Require a credit card, require a deposit, allow free dynos for otherwise paying customers. Killing the whole program is a sign they don't care about growth, they don't care to be where a developer first launches a hobby that results in paying later, and they are fully embracing their Salesforce persona (not a good thing in my book).

I don't think you understand how much fraud and abuse is in the hosting world.

I've never said there wasn't any. Just that if you value a free tier (which you should if you want to attract developers) you find a way to deal with the fraud (or eat the cost). The fact that Salesforce either couldn't figure out (or didn't care to) the fraud and didn't want to pay for it spells bad news for the future of Heroku.

Again I don't think you understand the amount of resources that goes into anti-fraud teams in the hosting world. I used to work at a big hosting company and we had a team of 50+ people working around the clock to stop fraud ontop of tens of thousands of hours in engineering time to automate as much as possible. It STILL wasn't enough. I can only imagine how much worse it is when you have a free product offering.

They could've converted the entire free tier offering to a $2-5/month thing, which would still be very attractive for legitimate users. Fraud is a big problem, but it didn't mandate this specific solution.

They do require credit card for free dynos. Maybe it wasn't so in the beginning I can't remember.

You can still offer a "free tier for existing customers" and call it "development" or something. This lets people who are already known/paying to spin up test instances, without having to go through purchasing, which leads eventually to more paid usage.

I guess it depends on where you think the platform is going, and which segment of the market you think the revenue is going to come from. I'm just saying: it's a significant cost. It's maybe the most important thing to know about the hosting business.

OT: your HN profile says:

> All comments Copyright © 2010, 2011, 2012, 2013, 2015, 2018, 2023, 2031

I'm curious. If you were to continue that sequence would the continuation be:

A. 2044, 2065, 2099, 2154, 2242

B. 2044, 2065, 2099, 2154, 2243

C. Something else?

My HN test is an IQ test that I, myself, cannot pass.

The future ones (2023, 2031, ...) are primes in lunar arithmetic in base 4 written in base 4, of course.

https://oeis.org/A171122

But, in other terms, it means they no longer have enough tech talent in-house to solve hard problems.

This points to further decline down the line.

They want it to die. They don't want to keep supporting Heroku but they don't want the bad press of killing it outright. Or at least they're pivoting from independent hosting to an enterprise value add for SalesForce.

Or they're really short-sighted.

Salesforce acquired Heroku if you weren't aware. There's your "why" right there!

As someone in a number of non-corporate/dev Slacks on the free tier, I do wonder when/if Salesforce will tighten up on those as well. There are some I know were just for organization during a one-time event...

Um that was 11 years ago.

Which coincides with about the last time Heroku was relevant/competitive. This is just an example of the parent company tightening the screws some more (see also [0]).

[0] https://news.ycombinator.com/item?id=32174596

This is just wrong info. Search HN comments from ‘craigkerstiens, an early employee. Significant innovative product development continued for a few years after the acquisition.

That's always the case which is why I said "about the last time" instead of "the last time". There is always stuff in the pipeline, there are always still people who care for a while but they see the writing on the wall just like the rest of us, they leave, and it becomes a self-fulfilling prophecy. I'm not saying there aren't/weren't good developers/managers at Heroku, I'm saying the sale to Salesforce started the end for Heroku.

If the security incident a few months ago wasn't a clear "we don't care about Heroku" then I don't know what will wake some people up.

A quick Google of this dude's handle/name and the first link is his blog with a reflective article about his experience at and subsequent departure from Heroku:

https://www.craigkerstiens.com/2022/05/18/unfinished-busines...