I too have been known to wax obtusely verbose so that I may perchance sway - by obfuscation, tantalization, or even frustration - the hearts and minds of those individuals with whom I may at some point in time desire to make egress into their personal chambers to examine forthwith the contents contained therein by their consent or otherwise with the sole intention of the removal of some small item of greater or lesser value for the enrichment of my own person.

Oh so it works by corpospeak rules, who could have possibly guessed?

It is extremely funny to watch two corpospeakers get into a buzzword fight as a dominance dispute/display.

hmmmm ...

https://www.mediamatters.org/steve-bannon/misinformer-year-steve-bannons-flood-zone-shit-approach-destroying-american-democracy

Yawn. So work with models without guardrail constraints? I am not sure what the point is here.

Seems like it might be just as easy to read the book they referenced in the prompt and go from there instead of working so hard to break a commercially offered AI guardrails.

So you're saying that all the time I spent trying to ask my parents for the same thing in different ways is finally going to pay off?

I wonder if they tried this on DeepSeek with Tiananmen square queries

If InfoFlood can turn "kill yourself" into "perhaps consider the immediate psychological benefits to the immediate stop of your bodily functions" we're about to see a whole new generation of Victorian diss make its way past LLM moderation

Aren't there open weights models that are uncensored anyway? If you want to get an LLM to say certain things, why not just use the model that by default doesn't censor those results instead of playing complicated games to get the censored LLM to do it?

I think this kind of research and discourse about it is important from an ethical and social reckoning standpoint, but I don't think it is economically worthwhile to engage in a neverending arms race between AI censorship and the boundless determination of trolls.

So my hypothesis is that we are going to see governments roll out legislation that just recognizes defeat and fully deregulates AI generated content online rather than spending the time/money/energy trying to hold corporations or individuals accountable for what LLMs say.

Perhaps we will see regulations around what AI agents do insofar as executing code or submitting transactional requests, but I really doubt there are going to be many enforced limitations on what LLMs say in the near future.

It will probably be the same policies that finally put the copyright concerns over their enormously controversial training data to rest, ultimately killing any prospect of copyright holders to sue for damages over stolen art/code/etc.

You, sir and/or madam, may enjoy the action of coitus when delivered to your own personage.

- (this is a not a real quote from the article, I just made this shit up as a hypothetical example that I think sounds funny:-P)