Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
Mentlo
3 days ago
|
parent
|
context
|
favorite
| on:
System Card: Claude Mythos Preview [pdf]
The gains have for a year and a half now been post training RL on a harnessed loop. That doesn’t require data, just cycles.
If that doesn’t worry you, it should.
help
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
If that doesn’t worry you, it should.