3 ms·
> What's so bad with this? It's lying. It's an open source project and a business model built on deceit. I guess I care about clear rules, clear intentions, an
by ezekg 1y ago
> What's so bad with this?
It's lying. It's an open source project and a business model built on deceit. I guess I care about clear rules, clear intentions, and I care about integrity above all. The AGPL is ambiguity; unclear rules, veiled intentions. And these same people will relicense without a thought, too. I think we should care about these things, otherwise we repeat history over and over again.
- jraph 1y ago> It's lying. I don't see how it can do this. The AGPL is a license text. It states what people can do with the licensed code. What is lying in this text? > I guess I care about clear rules, clear intentions, and I care about integrity Me too > The AGPL is ambiguity; unclear rules, veiled intentions That's where I don't agree: The rules are written in the license text and I see no ambiguity there. Where is the ambiguity in the AGPL text? What is not clear about it? What granted rights are we not sure about? > veiled intentions The original intent of the AGPL authors (the FSF) was clear and simple to me: ensure the end user's freedom. It was GPL, but address the SaaS situation where someone can modify networked GPL software, make users use it from the network, without having to redistribute the modification since the program runs remotely and not on the user's machine. And that intent is perfectly align with what I want for my code. Sure, people with bad intents will use the AGPL, and so what? People kill with knives, but I'll definitely keep cooking with mine. The AGPL is a tool. It doesn't, by itself, has intents, especially veiled intents. I'm not going to stop using the AGPL because someone wants to use it to trap users. This is all abstract, I'd appreciate concrete examples where: - people have done that, without a CLA (because yeah, I'm convinced AGPL+CLA can be bad). - the AGPL doesn't work well for someone using it with the original intent in which the AGPL was written.
- ezekg 1y agoHere's a thread from a couple hours ago on X: https://x.com/_m27e/status/1962563736142565882 https://x.com/_m27e/status/1962563736142565882. Read the images and follow the quoted posts. The example comes straight from the mouth of a COSS founder.
- jraph 1y agoSorry, but this yes-answered AI slop subtly makes no sense, it just reads comfortable to you. You should not delegate writing your opinion to the llm and even less use it to shape your opinion, this is my main takeaway from your screenshot. The AI answered your biased question in a way that pleases you, it's well known that they do this. This stuff is actually scary, and that clever people like you rely on this and don't notice the glaring issues is even more. The agpl is ambiguous because users may choose not to fully use the freedom they were given? Sorry but this is bullshit. I'm glad I haven't started using this stuff yet (for other reasons). I'm sure I wouldn't notice such issues in topics I'm not at ease with and I can see how easy one can be seduced by this stuff. Now, I'm convinced AGPL can be misused. What's more, I'm certainly quite happy that a side effect of the AGPL is that Google won't touch it, to mirror the comment you point to (whose author is wrong by the way, the intent behind the AGPL was not to exclude big tech, but to promote/protect user software freedom). All the fud around the AGPL actually comes from companies like Google, so respectfully fuck them all. Nobody could possibly have weaponised the AGPL against them had they not started spreading all the fud in the first place. But even considering that picking the AGPL to scare big tech away is bad and weaponizes it (which I can hear, and let's assume), I believe you are wrong that nobody chooses the AGPL for the user freedom genuinely. There are a lot of examples of software under AGPL in good faith seemingly to me. Examples: Nextcloud, Joplin, CryptPad, Overleaf, Passbolt, Univention... I doubt any of these commercial projects from friendly (?) companies choose the AGPL to fuck the world. I don't know the numbers, maybe it's a minority. You may not have numbers as well. Are you against the AGPL when used in good faith? If so, what to you suggest as an alternative? I'm with you with the wish people were altruistic. But the system we live in doesn't exactly help being altruistic. Not being altruistic is certainly not a trait of people using AGPL, it's virtually everyone in a commercial setting (although some of us try to do their best to be good humans and virtuous). If anything, the AGPL was born from a ideal and that was certainly driven by something like altruism. All this blame towards the AGPL because people are out there to make money really feels weird to me. Anyway, I don't think we'll reach an agreement here and that's OK. Thanks for the discussion, despite the strong disagreement it is/was stimulating.
- ezekg 1y ago
- NoGravitas 1y ago> veiled intentions The intentions behind the AGPL were never veiled. The intent was to close the loophole in the GPL3 that is opened by providing software as a network service. If you were to release software that provides a network service under the GPL3, someone else could use and modify it without sharing their changes, but while also providing it as a service. AGPL3 closes this loophole by ensuring that everyone who can use the software must have access to the source including any changes. I don't see any deceit, and I don't even see any anti-competitiveness.
- ezekg 1y agoIf you read the conversation higher up, instead of taking what I said out of context, you'd see that I was talking about the author veiling their intentions, i.e. using the AGPL as an effective non-compete while masquerading as "open source." They (the author aka the startup) wouldn't have chosen the AGPL if not for enterprises/bigtech/competition being scared of it. They didn't choose AGPL to increase freedom 0 (AGPL's purpose), they chose it to decrease it. Thus, veiled intentions.