https://cloudposse.com logo
Join Slack
Powered by
Thanks <@UB2EH889X> for the opportunity to share m...
# random
h
Thanks @Erik Osterman (Cloud Posse) for the opportunity to share my application. One additional area I’ve been exploring is open-source licensing. I’ve noticed that more projects are adopting the AGPL, and I’ve come to the same conclusion that it offers stronger protections compared to licenses like MIT or Apache, even though it is not as widely adopted as those more permissive options.
sweetops 1
I used AGPL and commercial licenses for this application in order to eliminate security concerns.
👍🏽 1
p
@Hao Wang, you're right to use GPL, or AGPL for something that can be accessed as a service. Giving away work completely, as with an MIT license, breaks the "closure" of open source under modification. It might encourage adoption of existing open-source software but I think it discourages development of new open-source software. I use GPL for my work, plus the GNU Free Documentation License because I put a lot of effort into documentation and consider it a vital part of the product, as well as an adoption enabler. AWS's licensing patterns are fascinating. Of course, AWS holds back proprietary and profitable modifications to projects like PostgreSQL, and re-implements products like MongoDB whose licenses forbid closed-source modification. Non-revenue-generating public material like sample code gets an MIT-0 license, but stuff that's worthy of a press release gets a stricter license. In general, I worry that contributing to open source is only sustainable if one works for a large employer. In that case, the employer has a potential legal claim on the contributions, even if the employee doesn't realize. (A few states have laws about this, but they are vague. In California, the burden of proving that something isn't work-related rests with the employee, the weaker party!) The intellectual property assignment agreements that we all sign run the risk of poisoning open-source codebases. Intellectual property rights aside, I think the reality that large corporate employers indirectly fund most open-source contributions by paying the salaries of the contributors limits the kinds of problems that open-source software will address going forward, and the degree of innovation ( = risk ) that popular projects will tolerate.
💡 1
h
Amazing sharing @paulm, in last 10-15 years I thought GPL series will die but I do see the possibility it will live again and may be longer than others. 20 years ago when I knew license in open source software, I felt it is a good direction in a long run but it is a niche market only for people know software and laws. Now with AI, law and software are not big deal any more. Security I think may be a big concern. Can see from the development of Openclaw story 🙂
hope I didn’t sidetrack your points, or I may not fully understand your points, I’m open to any ideas
p
Not at all, I always enjoy a dialogue with you!
💯 1
💡 1
I think AI abets widespread plagiarism, further discouraging public software contributions. I always check the indemnification protections that AI/LLM products offer. There are so many exclusions, and here again, an open-source codebase can be poisoned. Do you remember the SCO issue, years ago, about teasing out what had been copied and what was original? Just as with employee intellectual property agreements, people should realize that indemnification is forward-looking. A claim won't arise until years after a piece of software is written, when it becomes popular and starts generating revenue for someone. It will be interesting to see what kinds of software development AI/LLM code generation replaces, and also, whether corporate adopters will return to human-written software if their generated code turns out not to be secure, or easily maintainable, or to miss the mark in terms of tapping the capabilities of the systems it calls. My open-source code uses very recent AWS features and specialized combinations of permissions, patterns that don't appear in training data. I get horrible results when, for fun, I ask AI/LLM products to accomplish the same things. But there's another sense in which software doesn't matter anymore. In the industry, adequate has become acceptable, and software buyers, chiefly large companies, don't place monetary value on quality. It's Don Norman's _The Design of Everyday Things_ except in software rather than physical products.
h
I remembered that SCO case and just checked again with AI 🙂
The case dragged on for *years (2003–2010+)*
last week I talked an ex-colleague, his father works for a telco here and the company doesn’t allow AI generated codes still
👍🏽 1
but it is hard for a company or a team not to use AI and compete with others… when AI appeared a couple of years ago, I told my wife that there is no way to go back to the life before if the companies continue training the models and then I heard Musk said the same thing Lol yeah we cannot go back
👍🏽 1
2 Views
Open in Slack
PreviousNext
Powered by