Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
sinenomine
on May 24, 2022
|
parent
|
context
|
favorite
| on:
Imagen, a text-to-image diffusion model
... Unless the possession of capable models becomes a legal liability by that time.
astrange
on May 24, 2022
[–]
This won’t happen in an interesting way. What will happen is you’ll find out training a model on copyrighted inputs causes it to memorize those inputs and the owners own your output.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: