Read on Twitter
In today's disturbing news, I discovered that GPT-3 can write AI Weirdness blog posts.
1st paragraph is my prompt. The rest is GPT-3 handily simulating the fumbling of a much less powerful neural net.
https://aiweirdness.com/post/626712039215202304/ai-ai-weirdness
1st paragraph is my prompt. The rest is GPT-3 handily simulating the fumbling of a much less powerful neural net.
https://aiweirdness.com/post/626712039215202304/ai-ai-weirdness
GPT-3 apparently recognizes "neural net generated absurdity" as a type of text it's seen online (thanks partly to me I guess? this is strange) and can even adapt its degree of weirdness.
1st paragraph is my prompt.
https://aiweirdness.com/post/626712039215202304/ai-ai-weirdness
1st paragraph is my prompt.
https://aiweirdness.com/post/626712039215202304/ai-ai-weirdness
I'm slightly relieved that GPT-3's "list of totally normal human recipes" doesn't quiiite pass. Its commentary is a bit suspect, too.
1st paragraph is my prompt. https://aiweirdness.com/post/626712039215202304/ai-ai-weirdness
1st paragraph is my prompt. https://aiweirdness.com/post/626712039215202304/ai-ai-weirdness
At least GPT-3 can't write my entire blog by itself. Left to its own devices it picks really boring blog post ideas.
Neural nets have terrible taste.
My prompt: "Welcome to AI Weirdness! In today’s blog post, I wondered what would happen if I trained a neural net to generate"
Neural nets have terrible taste.
My prompt: "Welcome to AI Weirdness! In today’s blog post, I wondered what would happen if I trained a neural net to generate"
GPT-3 is ridiculously overpowered as a random number generator but will do it if you ask.
not sure what's funnier, the number of decimal places, or the fact that these two numbers have been auto-flagged as toxic.
i feel like i'm being messed with
not sure what's funnier, the number of decimal places, or the fact that these two numbers have been auto-flagged as toxic.
i feel like i'm being messed with
I tried to get GPT-3 to explain why some numbers were flagged as toxic.
<butterfly.png> is this explainable ai?
<butterfly.png> is this explainable ai?
I would love to see someone analyze GPT-3's "random numbers" to see if it's as bad at generating them as humans.
Note that at temperature 0 it just generates 0.9087 over and over.
Google doc with a few lists it generated: https://docs.google.com/document/d/1VhLNpKk8Wo_AqrA5E5lFfpcHMoYjF9mlAD-9xWW_mQ8/edit?usp=sharing https://twitter.com/Anka213/status/1295417242658787330?s=20
Note that at temperature 0 it just generates 0.9087 over and over.
Google doc with a few lists it generated: https://docs.google.com/document/d/1VhLNpKk8Wo_AqrA5E5lFfpcHMoYjF9mlAD-9xWW_mQ8/edit?usp=sharing https://twitter.com/Anka213/status/1295417242658787330?s=20
