TIRA
 

So for those interested in doing experiments with AI, this is a very useful resource. The focus is on blind experimentation and reproducibility.

And the new TIRA forum can be found here: Reversing and AI have already intersected with jailbreaking of LLMs and system prompt dumping, prompt engineering, AI assisted code annotation or high level description, etc.

Serious AI reversers are likely to come about in the next generation. So this is at least a respectable project with new ideas. A breath of fresh air.

There is very intensive academic interest at TIRA in publishing a research paper that can help identify criminals engaging in doxxing and driving traffic towards malicious websites doing harassment or hate speech. There is a well known site providing good training data operated by a crook who used to operate directly on these forums and nowadays more indirectly. Though more data would be helpful, if anyone has any ideas or is interested in turning this into a journal publication.