Anthropic – potential rival to OpenAI

[spryfusion]

Blog post: https://claude.com/blog/1m-context-ga

[firestar464]

https://fortune.com/2026/03/26/anthropi ... abilities/

[firestar464]

How are there so many leaks?

[wjfox]

[raklian]

In a way it is a good thing - if it is that capable, it will be useful in areas where we have problems solving.

[weatheriscool]

[firestar464]

More on that by Reuters:

https://www.reuters.com/business/anthro ... 026-04-09/

[firestar464]

[weatheriscool]

[weatheriscool]

[weatheriscool]

[wjfox]

[raklian]

Did they ask Mythos what it would do if it fixed all of the vulnerabilities and still find a way around it?

[weatheriscool]

Did they ask Mythos what it would do if it fixed all of the vulnerabilities and still find a way around it?

We're going to have to start using models like shields in star trek. They'll block most hits and remodilate/or redesign their own defenses to the next attack. But we also have to consider making hacking a serious crime. If you walk into a store and steal stuff = you go to prison.(rob a bank and go to prison for life). We have to start throwing people into prison for doing it.

Passwords and other methods won't be full proof in the future and so isn't that lock on your bike in real life. That is the way I'll be. But if you're caught or found you pay a huge price.

Sadly this is part of the reason why they want to id you as want to know about you as they want to be able to punish you within a criminal way if you do this. They can't rely on passwords to protect themselves anymore so they're going to use real life methods to do so.

[firestar464]

Uhm, hacking is already punishable by prison time...

[wjfox]

[weatheriscool]

[weatheriscool]

[wjfox]

[weatheriscool]