AI can write a proper README. In fact, it's better than me at doing so and keeping it up to date. People writing README with AI are bothering to write it. In my experience AI won't automatically create README files for you when making projects with the exception of create project tools which create a default README, but in that case usually the AI ignores it and leaves it in the default state. People are just using a tool that lets them create without manually typing in each individual character.
Most manually written README's I come across are in a far worse state than an AI generated one. To the point that I will often ask an AI to summarise third-party projects for me because the README's are so abysmal.
Code is the expression of knowledge and can be protected by copyright.
A lot of the popular licenses on GitHub (like MIT) permits you to use a piece of code on the condition that you credit the original author. If an LLM outputs code from such a project (or remixes code from several such projects) then it needs to credit the original authors or be in violation.
If Disney's intellectual property can be stolen and needs to be protected for 95+ years by copyright then surely the bedroom programmers' labor deserves the same protections.
We're not talking about the expression of knowledge. What is used in AI models is the knowledge from that expression. That code is not copied as is, instead knowledge is extracted from it and used to produce similar code. Copyright does not apply, IMHO
So you can train AI on Disney Movies to generate and sell your own disney movies because "knowledge is extracted" from it ? Betcha that won't fly in the courts. Here is "Slim Cinderella" - trained and extracted from all Disney Cinderella movies!
Sure, you can do it illegally - by breaking the law and recognizing that you need to be a fugitive. You can give up civilization and live in the wilderness. People can do whatever they want on their 10 year old Dell as long as they don't sell/distribute products made from other people's true efforts.
But the license terms state under which conditions the code is released.
For example: MIT license states has this clause "The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software."
It stands to reason that if an LLM outputs something based on MIT-licensed code then that output should at least contain that copyright because it's what the original author wished.
And I saw a comment below arguing that knowledge cannot be copyrighted, but the code is an expression of that knowledge and that most certainly can be protected by copyright.
The blog just leads to https://research.swtch.com/qr/draw/, which is the demo page of the blogs [1] and [2] written by Russ Cox many years ago about putting pictures in QR codes by manipulating the error correction codes in them
Hey thanks a lot, that should be the actual link behind the submission. Very interesting technique that boils down to encode a url with a ton of carefully shaped random numbers hidden in the fragment to generate the actual art.
Messing with that demo page, it seems to manipulate the URL you put in there too adding some numbers after an anchor tag (#) at the end, rather than just error correction.
Interesting I was wondering about this, if you control the domain and web server technically you could even make a full path of arbitrary data that then redirects to the real target.
Or "relevant" in the sense that it's something I bought recently: I searched for vacuum cleaners, found one I liked an bought it. Now I will be seeing ads for vacuum cleaners for the next few months.
You write your markdown file, but add the code snippet at the bottom of yor document and save it with a .md.html extension. Then when you double-click it it opens and renders in your browser.
I save my notes in a Google Drive, and it's now replaced all the note taking apps I've tried over the years
"I couldn't be bothered to write a proper README, so I had the AI do it"
reply