Ha, finagled chatgpt to calculate fresh from non training data sources, essentially to take in information and create conclusions from that data without the heavy curation. The conclusions on financial collapse were pretty interesting. My guesstimate is by late 2027 if war with Iran kicks off this year, but almost certainly by 2030.
Timeline
Post
Remote status
Replies
7
@Heil_Honkler there is no calculation happening here, all the percentages appear out of thin air.
@WandererUber I didn't screenshot the 15 minutes of discussion and input/output before.
@Heil_Honkler the fact remains. these values appear ex nihilo for the reader of your post, even if there is the hypothetical possibility that chatGPT actually did a valid economics calculation to get to them. It probably didn't though. It probably made them the fuck up.
@WandererUber When asked initially about a possible financial collapse by 2030 the answer was around 20% even including a possible Iran war. This didn't seem quite right. I had to finagle it to start fresh using objective data collected on the various trends for employment, production, debt(both personal and governmental), population, % of native vs non-native, stock market analysis, every single possible major influence on the us economy that I could think of. Had it gather all of that data without referencing it's training data. Then I had it calculate the impact each had on the economy and consolidate it then have it come to some conclusions. As you can see in the conservative scenario where everything is flowers and butterflies, you're still talking around a 50/50 chance.
@Heil_Honkler this does nothing to answer my objection
@WandererUber Well I agree in that I wouldn't take anything by an llm, specially a public one as gospel, but being able to get it to admit that even as limited as it is in it's responses is interesting to me.
@Heil_Honkler >I wouldn't take anything by an llm, specially a public one as gospel,
That was not my point.
Don't get me wrong, I also enjoy finessing libtard AI to admit libtardism is wrong. My point was that this is not a calculation at all. And whatever it spit out before is also not a calculation.
If an LLM calculates something and it shows me the calculation, it doesn't matter that it's an LLM. I can see if it is correct. Just as I can with a human.
That was not my point.
Don't get me wrong, I also enjoy finessing libtard AI to admit libtardism is wrong. My point was that this is not a calculation at all. And whatever it spit out before is also not a calculation.
If an LLM calculates something and it shows me the calculation, it doesn't matter that it's an LLM. I can see if it is correct. Just as I can with a human.