I need to make a proper page for this, but it's a good start here: https://
x.com/alexjc/status/
1645771162897580032
… Where it's heading? I think there'll be licenses that make these conditions more clear. "The model was trained under Fair Use with this data, so you must respect those conditions."
ETHICS
-
Fair Use Licenses for AI Models: New Framework Emerging
By
–
-
Licenses Cannot Waive AI Model Training Regulatory Requirements
By
–
How does that work? There is no license that can waive regulatory requirements, so if model training relied on those they continue to be applicable. (Fair Use and TDM are granted conditionally based on the application, so the conditions must carry through by definition.)
-
Fair Use Training Data Liability in Commercial AI Applications
By
–
Remember, models that rely on "Fair Use" or "Text & Data-Mining" exceptions, those conditions carry through to the commercial application. Specifically, if you compete with the rightsholders whose data was used in training or harm their legitimate interests, you're still liable.
-
Parked Vehicles vs Robocars: Key Behavioral Differences
By
–
A personned vehicle stopped on Gough St is not 100% equivalent. They are subject to social shaming, they often respond to horns being blown, police can move them on. They intend to move soon. They are not waiting for rescue. The opposite of all these are true for robocars.
-
AI systems unlikely to experience genuine emotions or consciousness
By
–
I don't expect them to enjoy the paperclips. I don't expect them to have a worthwhile analogue of experienced happiness, or sadness, or wonder. I don't expect there to be many different minds experiencing different things and having empathy for each other.
-
Unaligned superintelligence would convert galaxies into worthless material
By
–
I expect that unaligned superintelligence will overwrite all nearby galaxies with stuff that is not very interesting or good from even the most embracing, cosmopolitan viewpoint on accepting diversity as valuable. I expect mostly lifeless paperclips, or stuff worth that little.
-
Reaction to gradient descent learning paper and concerns about implications
By
–
Jessicat via DM confirms that this is about the paper "learning to learn by gradient descent by gradient descent". I vaguely remember going "god damn it" about this paper, because it was a bad sign that what they'd tried *worked*. I do not recall preinventing, nor claiming to
-
Sovereign Agents: Exclusive Personalized AI Experiences
By
–
[ S O V E R E I G N A G E N T S ] Exclusively YOURS Sovereign Agents are designed to serve a single individual, exclusively. Bespoke Experiences | Formidable Privacy By: http://
MONTREAL.AI #AISovereignty #QuebecAI #SovereignAgent -
Copyright Law and Digital Products: Terminology Matters
By
–
It's not stealing because that's not a real product. There's no such thing as stealing in Copyright law. If you want to be better understood (comprehension begins with good writing), try using the correct terms and supporting your theories with references to legislation.