rez0__

4 ms·

1.
Using Genius AI Models for Important Things: Cheating at NetHack(josephthacker.com)
4 points by rez0__ 1mo ago|0 comments
2.
Bug in AI Toy Console leaked 50k kid's conversation(wired.com)
27 points by rez0__ 8mo ago|3 comments
3.
by rez0__ 1y ago
what product does he sell
4.
Yes, vibecoders, you need row level security. UUID-obfuscation isn't enough(josephthacker.com)
4 points by rez0__ 1y ago|0 comments
5.
How to Hack AI Agents and Applications [Guide](josephthacker.com)
7 points by rez0__ 2y ago|0 comments
6.
by rez0__ 2y ago
> Now that we are an administrator of Air Transport International... LOL > Unfortunately, our test user was now approved to use both KCM and CASS smh...
7.
Comprehensive Prompt Injection Primer(github.com)
3 points by rez0__ 3y ago|0 comments
8.
by rez0__ 3y ago
"On May 2nd, 2023, we identified that the Flask session secret for the points[.]com global administration website used to manage all airline tenant and customer accounts was the word 'secret'." And so many insane vulnera
9.
Leaked Secrets and Unlimited Miles: Hacking the Largest Rewards Vendor(samcurry.net)
5 points by rez0__ 3y ago|1 comments
10.
ChatGPT plugins Account Takeover via Prompt injection(rez0.blog)
42 points by rez0__ 3y ago|2 comments
11.
by rez0__ 3y ago
The new code interpreter model allows real RCE unlike the previous RCE hallucinations people submitted. But there's very little impact. I'm mostly shocked openai can support a container-per-user architecture.
12.
Actual RCE on ChatGPT(twitter.com)
2 points by rez0__ 3y ago|1 comments
13.
Import Alignment: A Library-Based Approach to AI Alignment(danielmiessler.com)
3 points by rez0__ 4y ago|3 comments
14.
by rez0__ 4y ago
Daniel Miessler suggests a way to help with the alignment problem.