Hacker News Reader

> and then their customers sent me unfriendly email and demanded that I provide support for the tool and solve their problems

Out of curiosity, what sort of tool were you offering? I’ve always heard about this but it’s never been my experience, so it really depends on the type of end user you have imo.

OP is currently trying to find a job and I think any attempt at open-sourcing his tooling (especially if it is useful) allows potential employers to review him in a better light. Abandoning it all seems wasteful when more and more companies try to find more reasons to filter out candidates.

Not the OP but, I can relate to this with a side-project I created many years ago that is an online tutorial for shell scripting. It's all open-source, requires minimal resources to run (small VM where I run a bunch of other stuff) and it is used most frequently by online schools to brush up on shell scripting. I get a lot of weird requests from random people which can be a bit annoying sometimes. This is for a site that gets around 100k visitors per month.

With that said I think there is some value in putting stuff out there without any marketing. Even if you don't have plan to monetize it can be rewarding in itself. The trick for me is to have a good way to minimize maintenance, which means use the same scripts and mgmt utils for several projects so you don't have any snowflakes that require special configuration.

The reality for the average person looking for a job is 1 out of 100 employers will ever look at an open source project you list on your resume.

Its better than nothing, better than others without projects, but it wont carry much weight unless its some widely used thing

Only exception is if your project matches the company's market or their products very closely. That may get you a meaningful advantage

a variety of backup scripts

For example, how to backup a running Linux server with rsync while using hard links to deduplicate files that didn't change.

I can relate, I used to have a small open source project (eventually transfered it to another maintainer). I often had quite rude and demanding Github issues from users.

What kind of apps do well these days if not niche ones? I am more of a web dev so curious if theres something inherent in app economics that makes it so.

App economics are just bad. Most of your competition will be free to download but then trick / bully people into a subscription. And maybe a few lootboxes?

CLV for a free game can easily be $5+ and that means when you market your niche app you compete for ad slots against people who are happy to pay $1-2 per install. But for a $1 app you can at most pay $0.7 per install after the store tax and then you're not making any revenue to pay for support, development, and the devices you need to buy for testing.

Consumer SaaS is booming on mobile, there's a lot of money to be made but it requires bigger teams and investments now than it did 5-10 years ago. OP is right in that it's harder to have a lifestyle business selling small apps.

> And my experience with releasing useful tools as open source was that companies started linking to them in tutorials, and then their customers sent me unfriendly email and demanded that I provide support for the tool and solve their problems (for free).

Which license did these tools have? AGPLv3 scares away most of the companies, especially stupid ones like those you described.

AGPL is the best license if you want to keep away stupid people. Our CTO spouted off that if we so much use anything with that license our valuation would plummet. People literally don’t understand the license and what it actually says. It’s hilarious.

You can forward support requests to an AI agent with infinite patience.

Not the parent, but can I ask why people are downvoting this?

On the one hand: yes, it's short and it's snarky

On the other hand: I feel like this is a legit use case for AI. ChatGPT is polite to the edge of obsequiousness and there has definitely been times when I wanted to say "No" but didn't want to lash out at the person - getting an AI to put together a first draft of a polite "No" is incredibly useful and a great starting point. If nothing it reduces the amount of time I spend agonizing over "Is this email ok? Is it at least reasonable?"

Probably because people here think that support requests are not always wrong. When I use a public open source project and I run into a bug, yes, I will request for a fix. I won't insult anyone for not doing it, though, I know I am not entitled to anything, when I did not pay for support, but I do not want my time to be wasted by talking to an AI. Just say no, if you don't want to fix bugs and maybe just make that plain on the project page. And those idiots who demand things can just be ignored/blacklisted (they often assume btw. that their company is paying for professional support, like they are for proprietary software).

You are also not entitled to a human response.

Only if that is made plain on the project/download page. The human social default and expectation is a human response. Also in OSS.

I do this for some code review comments where I want to say “what the actual fuck are you doing here. This makes absolutely no sense.” And I get a super nice output that needs very little tweaking.