I’ve been playing with ai, pushing the limits of it as a therapy tool (I’m being honest, but I’m using myself as a test case). I’m really trying to evaluate it in good faith, mostly to try to see what’s addicting folks to it these days.

I just cannot get it to pass my “Turing test”. It’s just so sycophantic. Every time I get close to feeling like there’s something there, it falls back into the “blow smoke up your ass mode.”

So far my hours of conversing has only convinced me that folks who follow chatbot advice for such important life decisions, probably deserve it. I’m happily married, but if a girl ever broke up with me because a LLM told her to, I’d probably login to thank it (joke).

So if I tell it that I'm literally a slave who is beaten and sexually violated daily, it's still going to tell me to stay in that situation? That seems helpful.

The US company said new ChatGPT behaviour for dealing with “high-stakes personal decisions” would be rolled out soon.

Fuck we've been in a black mirror episode all along

Does anyone else think that LLMs have been regressing in the last 2 years. At least in the area of deep questions. Initially you could push the limits and actually bounce crazy ideas off of them and get some creative results.

Now they basically they always give you a canned answer that sounds like it was pre-approved by a committee.

That AI bot must be saturated with break-up and "Delete Facebook, hit the gym!" advice..

Sometimes that’s what people need to hear.