I'm Conflicted About Future Of Open Source

Video Chapters

• 0:00 Is AI changing the fundamental nature of coding?
• 1:54 The project that exposed AI's role in open-source
• 3:01 Why almost every new project now uses AI
• 4:30 Valuing "Artisan" developers in an automated world
• 5:53 Major projects establish rules for AI contributions
• 8:10 What happens when developers only know AI tools?
• 9:30 The licensing crisis: Who owns AI code?
• 12:10 When maintainers use AI for code reviews
• 13:20 "Death by a Thousand Slops": The AI security report problem
• 16:10 The path forward for open-source and AI

Original Output

0:00 Is AI changing the fundamental nature of coding?
1:54 The project that exposed AI's role in open-source
3:01 Why almost every new project now uses AI
4:30 Valuing "Artisan" developers in an automated world
5:53 Major projects establish rules for AI contributions
8:10 What happens when developers only know AI tools?
9:30 The licensing crisis: Who owns AI code?
12:10 When maintainers use AI for code reviews
13:20 "Death by a Thousand Slops": The AI security report problem
16:10 The path forward for open-source and AI

Timestamps by StampBot 🤖
(542-i-m-conflicted-about-future-of-open-source)

Unprocessed Timestamp Content

0:00 The future of open-source, AI development, and coding in general.
0:35 Many developers, young and old, are using AI in their workflows.
0:54 The majority of people now use AI in some form of development.
1:10 Some rely heavily on AI, while others augment their development process.
1:54 This Reddit post about a Logitech Options clone for Linux inspired the video.
2:07 The Reddit post was temporarily removed, later edited to mention AI usage.
2:45 Claude, an AI, was one of the major contributors to the project.
3:01 Many new open-source apps now incorporate AI-generated code.
3:20 Basically, every project in the past two years has used AI.
3:40 The growing concern about new developers relying on AI tooling.
4:04 The speaker feels conflicted about promoting AI-assisted software.
4:30 Wants to highlight artisan developers who built software themselves.
4:50 AI tooling is becoming more and more accepted in open-source.
5:00 Not like the art space where AI art is heavily criticised and shunned.
5:20 In open-source, people use AI, some complain, but it's generally accepted.
5:40 What happens if AI causes problems in the long term for open-source?
5:53 The Linux kernel lays down rules for AI-generated code usage.
6:10 You take responsibility for AI-generated code if you use it.
6:34 NetBSD bans AI code due to concerns about tainted and GPL code.
6:55 KDE's Google Summer of Code recommends against AI for proposals.
7:20 Don't generate your proposal with AI; we want to talk to a person.
7:40 There's a consensus against AI code in open-source projects.
7:50 What happens when current young developers become the old guard?
8:10 A whole generation will grow up having access to AI tools.
8:30 What about projects that outright ban AI code in the future?
8:48 When AI code becomes indistinguishable from human-written code.
9:10 What if you say no AI, but can't tell the difference anymore?
9:30 How do software licenses apply to AI-generated code?
9:45 Licenses were built around people owning the code they produce.
10:05 AI-generated code is not copyrightable in the US.
10:15 Does a software license even apply to an AI-written codebase?
10:25 Is AI code now effectively in the public domain?
10:40 Licensing problems like accidentally bringing copyleft code into projects.
10:50 The core foundational problem: do code licenses even function with AI?
11:05 Different projects are making different assumptions about AI code.
11:25 Some project leaders actually don't like copyright, so AI breaking it might be good.
11:45 All of this is still a legal grey area until legally challenged.
12:10 The question of AI usage during code review process.
12:20 PotteRing uses Claude to review code in SystemD merge requests.
12:40 AI code review is a separate thing, for checking what humans missed.
12:58 Project Glasswing: securing critical software for the AI era.
13:10 Big companies like Google and Apple are collaborating to find vulnerabilities.
13:20 Daniel Stenberg's "Death by a Thousand Slops" post about AI security reports.
13:40 Curl bug bounty program was shut down due to AI slop reports.
14:00 Daniel acknowledges AI tools can be useful if used by skilled developers.
14:20 The important distinction is consent from project maintainers for AI use.
14:30 PotteRing explicitly consented to Claude reviewing SystemD code.
14:50 Project Glasswing participants are explicitly involved and control information.
15:10 Curl was inundated with random, untested AI-generated security reports.
15:30 Now every security report has to be treated as suspect.
15:50 It's hard to get clarification if a report is from an AI or a person.
16:10 The future of open-source and AI is an important, unresolved problem.
16:30 Even if you don't like AI, it's a problem that must be considered.
16:50 Do you use AI? Do you know people using AI for development?

Timestamps by StampBot 🤖
(542-i-m-conflicted-about-future-of-open-source)