Read on Twitter
I keep seeing all kinds of crazy reports about people's experiences with GPT-3, so I figured that I'd collect a thread of them.
First, @gwern 's crazy collection of all kinds of prompts, with GPT-3 generating poetry, summarizing stories, rewriting things in different styles, and much much more. http://gwern.net/GPT-3
Automatic code generation from natural language descriptions. "Give me a page with a table showing the GDP of different nations, and a red button." https://twitter.com/sharifshameem/status/1282676454690451457
Taking a brief technical tweet about GPT-3 and expanding it to an essay which the author of the original tweet mostly endorses. https://twitter.com/Plinz/status/1283211048145711104
Acting as a more intense therapist than ELIZA ever was. https://twitter.com/nicklovescode/status/1283300424418619393
On the side of "it does still have weaknesses too", figuring out how to trick it at the Turing Test. http://lacker.io/ai/2020/07/06/giving-gpt-3-a-turing-test.html
Also, using the architecture behind it to generate images. https://twitter.com/xuenay/status/1273898995706998790
Redditor shares an "AI Dungeon" game played with the new GPT-3 -based "dragon model", involving a cohesive story generated in response to their actions, with only a little manual editing. https://www.reddit.com/r/AIDungeon/comments/hpkqij/using_the_dragon_module_i_just_generated_what_is/
The official announcement of the Dragon Model and examples. "The game invents a complex magic system and underlying theory behind why it works and describes the whole system as I read the book." https://medium.com/@aidungeon/ai-dungeon-dragon-model-upgrade-7e8ea579abfe
Another example of automatically generated code, this time giving GPT-3 a bit of React code defining a component called "ThreeButtonComponent" or "HeaderComponent", and letting it write the rest. https://twitter.com/hturan/status/1282261783147958272
From a brief description, GPT-3 correctly generates an explanation indicating that it's a case of asthma, mentions a drug that's used to treat asthma, the type of receptor the drug works on, and which multiple-choice quiz question this indicates. https://twitter.com/QasimMunye/status/1278750809094750211
GPT-3 tries to get a software job. https://twitter.com/lacker/status/1279136788326432771
Translating natural language descriptions into shell commands, and vice versa. https://twitter.com/harlandduman/status/1282132804034150400
Given a prompt with a few lines of dialogue, GPT-3 continues the story, incorporating details such as having a character make 1800s references after it was briefly mentioned that she's a nineteenth-century noblewoman. https://twitter.com/AndrewMayne/status/1271112770293862402
. @bioothod noted that not *all* of GPT-3's answers are that great. Fair. IMO, being able to produce good answers even *with* human selection is amazing by itself, and often it requires relatively little, but yeah it's not a general intelligence yet. https://aiweirdness.com/post/621186154843324416/all-your-questions-answered
So far I had only been sharing other people's experiences, which did occasionally make me wonder whether they're all real. Now I tried out GPT-3 myself, to generate some space opera. And whoa. Yeah I believe that they're real now. https://kajsotala.fi/2020/07/gpt-3-space-opera/
An earlier link pointed out that GPT-3 doesn't realize that nonsense questions are nonsense questions. But you can also just tell GPT-3 to treat nonsense as such. https://twitter.com/nicklovescode/status/1284050958977130497
Using GPT-3 to help you do gratitude journaling. https://twitter.com/nicklovescode/status/1283740861260369920
Source is an anonymous image board poster so could be fake, but: if you give an AI Dungeon character fake wolf ears and then ask her to explain formal logic to you, she may use the ears in her example. (source: https://yuki.la/vg/299570235 )
Even after seeing all the other results, I honestly have difficulties believing that this one is real. https://twitter.com/kleptid/status/1284098635689611264
Starting to see lots of "GPT-3 is overhyped and not that smart" articles now. Sure it's not actually intelligent, but the fact that a non-intelligent thing can do so many things is still significant and it will have lots of applications. https://twitter.com/anderssandberg/status/1285104499531698176
