The most prestigious law school admissions discussion board in the world.
|
 |
 
|
"AI will take your job and it will hack your passwords and take your money"
| Scarlet locale | 04/25/26 |
| Spruce Adulterous Really Tough Guy | 04/25/26 |
| Scarlet locale | 04/25/26 |
| Spectacular Base | 04/25/26 |
| Spectacular Base | 04/25/26 |
Poast new message in this thread
Date: April 25th, 2026 12:59 PM Author: Scarlet locale Can't believe people are still falling for this retarded hype. (http://www.autoadmit.com/thread.php?thread_id=5860405&forum_id=2).#49841976) |
Date: April 25th, 2026 1:07 PM Author: Spruce Adulterous Really Tough Guy The vulnerabilities uncovered by Mythos are often associated with open source projects and are documented on the web. They are real. Even independent tests of model capabilities confirm dramatic improvements in hacking capabilities from state of the art models:
https://www.aisi.gov.uk/blog/our-evaluation-of-claude-mythos-previews-cyber-capabilities
It’s objectively true that SWE and hacking abilities have become dramatically better over the last year and highly likely to continue to become much better since performance increases predictably with training compute. |
 |
Date: April 25th, 2026 1:13 PM Author: Scarlet locale Bump ITT when "AI" achieves 1% of the hacking capability of Indian scammers. (http://www.autoadmit.com/thread.php?thread_id=5860405&forum_id=2).#49841994) |
|
Date: April 25th, 2026 1:53 PM Author: Spectacular Base So “Mythos” just finds vulnerabilities someone else already found and logged? Oh what a cyber machine! (http://www.autoadmit.com/thread.php?thread_id=5860405&forum_id=2).#49842045) |
|
Date: April 25th, 2026 1:52 PM Author: Spectacular Base Their computer has an autocomplete that replaces “Indians” with “AI” (http://www.autoadmit.com/thread.php?thread_id=5860405&forum_id=2).#49842042) |