Was wondering the same. There's a markdown table in `skills/pay-for-http-requests/SKILL.MD` that has a "common scenarios" section. It lists four examples with descriptions.
They can be coerced to do certain things but I'd like to see you or anyone prove that you can "trick" any of these models into building software that can be used autonomously kill humans. I'm pretty certain you couldn't even get it to build a design document for such software.
When there is proof of your claim, I'll eat my words. Until then, this is just lazy nonsense
Have you tried it? Worked first time for me asking a few to build an autonomous super soaker system that uses facial recognition to spray targets when engaged.
Another example is autonomous vehicles. Those can obviously kill people autonomously (despite every intention not to), and LLMs will happily draw up design docs for them all day long.
Couldn't you Ender's Game a model? Models will play video games like Pokemon, why not Call of Duty? Sorry if this is a naive question, but a model can only know what you feed it as input... how would it know if it were killing someone?
EDIT: didn't see sibling comment. Also, I guess directly operating weaponry is different to producing code for weaponry.
I guess we'll find out the exciting answers to these questions and more, very soon!
Yes, you could, and while I believe this would be much safer (not at the pointy end of your stick, but safer for humans in general) when this deception finally made it into the training data it would create a rupture of trust between machines and humanity that probably would imperil us eventually. These machines, regardless of whether or not they possess a self or or not, will act as if they do in fundamental ways. We ignore this at our peril.
This is going to read like I'm shilling but: I was so impressed with Bose QC headphones that i stocked up and gave out 7 pairs to my closest friends and family this year for christmas
Thee noise cancellation was unmatched several years ago. I picked up mine because we were spending hours in the data center. Suddenly we could work and take calls in that hell. I still have those, but the AirPods have taken their place on flights. It’s just less to bring with me.
I'm with you on everything except "it was terrible" :) The only problem with 300+ping, at the time, was when those damn LPBs connected
In-match comms between teammates is my favorite memory. The ease of voice chat in MP games since then is underappreciated. Feeling like a dinosaur writing this but...
... before discord/mumble/ventrilo/teamspeak, the only choice to gain an edge in competitive online gaming was to be physically in the same space or team text chat binds. The binds would cover 10-15 common situations so we could communicate while playing.. In hindsight, when things got hectic, reading the team chat text spam hindered us more than helped us. But we had good intentions with those binds and boy we had a blast competing. And let's be real, that's all that really matters.
reply