Stoic quotes real or generated?

Generates wise stoic quotes. Can you tell which stoic quotes are real and which ones are generated using GPT3 da-vinci? Are they more insightful than a 2000 year old stoic, I don’t know you tell me?

To be angry is to revenge the faults of others upon ourselves.

When we are alone, we ought to banish display, and put a seal upon our lips

These are pre-generated from playgrounds not refreshing using the API. This was more an experiment than an ongoing project.


philosophia - love of wisdom, gets a new meaning
The project was very interesting in its beginnings, unfortunately it became useless from the moment of commercialization,
Philosophy ends where business begins ; )
and similar is the case with the art
@m-a.schenk thanks for tagging me : )
I look forward to following your philosophical project further @simon


Is this giving more AI than real, I just clicked on one button repeatedly. (I’m so stoic, I don’t care either way lol.)

1 Like

Yeah, it wasn’t an even split 2/3’s are AI, this was slightly intentional as I thought if I had them at 50/50 it would be harder to measure the intrinsic 50/50 odds of picking real or ai, but I’m no statistician so I could be way off.

Interestingly I checked the analytics and the stats are interesting. People seemed to underestimate the real quotes (within a margin of error but I would have expected people to more certain of those)


@simon very interesting results, congratulations!
allow a question, whether the group of people
who participated in your experiment is statistically significant ?

I don’t know what size is statistically significant in this context but over 3000 users so I’d say it’s good enough.

1 Like

@simon 100 is the minimum sample population
Most statisticians concur that the minimum sample size to get any noticeable result is 100.
A reliable sample size for good quality data does not usually exceed 1000
A good sample size is typically around 10% of the population. The larger this number, the more reliable your data will be. In a population of 5000, 10% would be 500. In a population of 200,000, 10% would be 20,000.

Even when the number of respondents is only 200,000, a 1000-person random sample will normally be fairly accurate. @simon 3000 of your users is (very) statistically significant : )
I don’t think the number of real stoics today is more than 3000 ; )

1 Like

I think we don’t need to worry about infinite paper clips but infinite stoic bots, that’s the real threat to humanity

1 Like