wtf happened to gpt models. the safety and morality updates are really annoying.
I had 5.6 going onto a GDPR consent side mission while building an unrelated feature
if they get any 3rd party service tos into their context they will follow it to the death. its insanity
i made an app that lets you send custom notifications via a webhook
☑︎ custom images
☑︎ custom links (on click)
☑︎ title and description
try it out now at https://t.co/fEacNfmw8g
@siddarthv66 i ernestly believe we have had and continue having a large amount of individuals at the level of those.
it’s just that their achievements aren’t as easily comprehensible as of their predecessors which are way more accessible due to the immaturity of the field at their time
@Monopotandpan@jonathan_gortz well if you want to be pedantic about it, then _mass_ toilet paper production began 8 years after gagarins first flight.
technically some of it existed in the country previous to that but limited to state elites
@iScienceLuvr because an experienced mathematician alongside fable worked through it
your average person will never be able to hit the required activations to find that counterexample
no i was referring to the fact that it might mirror swe
a good amount of swe’s don’t even read the code anymore yet nobody is dying cause the things they work on don’t really matter as much
but in fields where it does strong qc was already a thing pre LLMs and it will remain that way
we might get more carelesss but i doubt that we will for the stuff that matters
software can kill you with much more ease than mathematics.
even in your plane example software has an even higher probability of nose diving the plane downwards.
yet nothing happened despite vibe coding
my point was in areas were it truly matters oversight is not missed
regulations were already in place to prevent human miscalculations; they’ll stay there and be repurposed for LLMs