aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authoralexpdp7 <alex@pdp7.net>2026-01-23 08:53:45 +0100
committeralexpdp7 <alex@corcoles.net>2026-01-23 07:54:09 +0000
commit349becf9bf42c6803019125de0881c2e3d0cb832 (patch)
tree0465ab7236eb74bd79a050f98ceedf78ef8bf663
parent2fd5fc939005a37ef4390e616c9d49f9dc2e7bc1 (diff)
Add note about publishing LLM output
-rw-r--r--programming/on-llms.md9
1 files changed, 9 insertions, 0 deletions
diff --git a/programming/on-llms.md b/programming/on-llms.md
index eaea6c6c..744c2a38 100644
--- a/programming/on-llms.md
+++ b/programming/on-llms.md
@@ -60,6 +60,15 @@ If you are using an LLM to create an illustration instead of commissioning an il
(As mentioned, I think very few authors would object to an LLM describing images to a blind person without making an obscene profit.)
+#### LLMs might need to be forced to publish all their output
+
+I read somewhere that LLM output should not be copyrightable.
+
+Making all LLM output public would solve a few issues:
+
+* Generated LLM content is now easy to identify by searching into the generated LLM output.
+* IP laundering would result in public material that can be easily replicated and would make commercial benefit difficult.
+
## LLMs are hype
### LLMs might not be effective for writing code