Activity
Mon
Wed
Fri
Sun
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
What is this?
Less
More

Owned by David

Stop giving ChatGPT all the power. Turn your prompts, including most GPTs, into protected, monetized web apps that run on any AI platform.

If you Use, Make or Sell AI content, you likely have legal issues, esp. if you buy/sell Work-for-Hire services. FOR USA SALES AND BIZ'S ONLY!

6 contributions to Work For Hire Is Dead
The book is now $27
Free isn't selling. So I raised the price to $27.
0
0
IP and software patents
I wonder what the impact is going to be on IP law in software in the very near future when an AI can theoretically write an entire app from first principles. I was working with something the other day where I wanted a library that simply did not exist in the environment I was working in. Specifically, I'm working on a Lua report generator and wanted to create a PDF and there's no PDF library available in Lua, or at least not one comparable to something like PDFKit in JavaScript. So I said, hey buddy, why don't you port that bad boy over into Lua for me? And it said, sure, here you go, 25,000 lines of pure Lua code including compression and PNG support, all byte-level identical to the output from PDFKit. Have a great day! I didn't even have to give it the source for PDFKit, as JavaScript that runs it is there if it is running (node_modules or whatever). So a one-shot deal and off it goes. Could be the same for anything. What does "clean room implementation" even mean anymore?
0 likes • Aug 12
Well, the Copyright Office and courts in USA have said this: "No human involvement means no author; no author means no ownership; and that means the Copyright laws do not apply." That's not to say Congress might change the laws before Trump leaves office as a nod to high-tech industries that sent him (or might in the future) lots of donations. But as of today, AI is regarded as no more "creative" than a room full of monkeys banging at keyboards. My opinion here is of no consequence; it's ultimately up to the Copyright Office, the courts, and ultimately Congress (in America, at least). That said, while compilation is regarded as a deterministic mechanical translation that preserves copyright ownership, what about using AI (non-deterministic) to mechanically translate code from one programming language to another? If you use it to translate your Delphi code into JS and the resulting code tests show both of them produce the exact same results, shouldn't that be good enough to treat the AI code simply as a tool? I mean, this is a two-edged sword -- when sliced from one edge, the two pieces are as if they were (in fact) cut from the same cloth; yet when sliced from the other edge, they are regarded as if the were cut from different cloths. Take a piece of software you wrote in Delphi and pas2js translates it to JS (tool-1). A prompt given to AI might create a similar or even very different translation (tool-2) that produces the exact same output as the pas2js version, in all test cases. In one case, the ownership of the output of tool-1 is preserved through the translation; while in the other, the ownership of the output of tool-2 might be getting stripped off, simply because it was created by a tool regarded as non-deterministic and also not human. That seems very bizarre to me. You could simply use tools created by AI (like tool-2) that produce identical results as hand-made tools, to strip the Copyright protections from the translated output. Maybe I'm missing something? But isn't this basically what you're pointing out?
DISCLAIMER
I am a Software Engineer, Architect, Designer, and Developer with a BS in Math / Computer Science. I've been dealing with IP laws and their impacts my entire career. I AM NOT A LAWYER. Nothing here should be considered "LEGAL ADVICE" but rather BACKGROUND MATERIAL FOR EDUCATIONAL AND ENTERTAINMENT PURPOSES ONLY. I'm sorry that needs to be said. The ONLY "LEGAL ADVICE" I would offer is simply this: IF YOU HAVE LEGAL QUESTIONS ABOUT HOW ANY OF THIS MIGHT AFFECT YOU, CONSULT WITH AN IP LAWYER. Here's some FINANCIAL ADVICE: Read my book first -- the last thing you want to do is pay $400+ per hour to be TUTORED by a lawyer about stuff you can learn here or online for free. You don't need to hire a CPA or Certified Financial Planner to teach you how to balance your checkbook, right? Learn the basics BEFORE you hire a specialist! If you have any questions, please post them here so I can see where to refine parts of the book. I'm happy to answer as well as I can, but don't be put off if I suggest asking Google or a lawyer. Google will probably cite several references that include US Government sites like the Copyright Office and the Patent and Trademark Office. The info on those sites is very comprehensive and easy to understand. Curiously, there is no "US Trade Secrets Office", which is where you'd go to learn how to deal with AI-generated content. Which may be why so few people outside of the Tech industry are familiar with TS laws and practices. In fact, Trade Secret practices seem to have been around for over 2,000 years. Generally speaking, you can use Trade Secret practices to protect anything that Copyrights and Patents don't. There are even two flavors you can use in parallel: Federal DTSA laws and State-by-State UTSA laws. But, again, spend some time learning about this stuff BEFORE you consult with a lawyer; your pocketbook will thank you!
0
0
CASE: Diamond v. Diehr (U.S. Supreme Court, 1981)
INSIGHT: The Book's Part II software parallel. The law kept refusing to protect raw math but protected what humans built with it — the same shape as refusing to protect raw AI output while protecting what humans build around it. RULING: A rubber-curing process used a well-known math equation, run by a computer. The question was whether that’s patentable or “just math.” The Court said PATENTABLE — because the process transformed something in the real world. The application crossed the line that the bare formula couldn’t.
0
0
CASE: Burrow-Giles Lithographic Co. v. Sarony (U.S. Supreme Court, 1884)
INSIGHT: This is the 140-year-old root of everything in Part II of the book. The machine-is-a-tool test decided in 1884 is the exact test that underlies the latest decisions. Same argument, different machine. RULING: A company copied a famous photograph of Oscar Wilde and argued photos can’t be copyrighted — a camera is a machine, so there’s no human author. The Supreme Court disagreed: the photographer made the creative choices (pose, lighting, framing), so the HUMAN WAS THE AUTHOR and the CAMERA WAS THE TOOL. This is THE pivotal case behind nearly ALL AI-generated content today.
0
0
1-6 of 6
David Schwartz
1
3 points to level up
@david-schwartz-8083
I help experts who've maxed-out their ability to grow their business, expand it further and work fewer hours, without hiring more people.

Active 7h ago
Joined Jul 9, 2026
INTP
Phoenix, AZ