5 ms·

Not just that. Scraping the website with AGPL-licensed code and then feeding it into OpenAI's GPT-4 API [1], a corporation led by someone about as "tech bro" as

by ninjin 2y ago

Not just that. Scraping the website with AGPL-licensed code and then feeding it into OpenAI's GPT-4 API [1], a corporation led by someone about as "tech bro" as it gets, which they are paying for in attempt to mock? This may be some generational thing, but the layers of "satire" here are reminding me of a snake eating its own tail and I feel more cringe than anything else. [1]: https://github.com/dfeldman/quackernews https://github.com/dfeldman/quackernews Unrelated, why is the duck in the "logo" looking at the back of their screen while typing? Edit: Took me a while to remember, but if we are talking satire using a large language model, this is how it is done: https://www.goody2.ai https://www.goody2.ai Wrote them a thank you e-mail when I first saw it and of course they responded with an e-mail generated with their GOODY-2 prompt to give me yet another great laugh.

DaSHacka 2y ago
> Unrelated, why is the duck in the "logo" looking at the back of their screen while typing? Having seen far too many AI-generated chibi images, I can say with 95% certainty the logo was generated by AI.
- xanderlewis 2y ago
  Absolutely. It’s unmistakable, even without the backwards screen. But why? Why does it have such a characteristic look? I wonder what causes this sort of recognisable ‘AI style’ to emerge.
  - ninjin 2y ago
    I am not an expert when it comes to image generation, but know a fair bit about large language models and machine learning in general. My guess is that like all gradient-based models, they will find certain minima based on the the training data and be somewhat reluctant to deviate too strongly from it as this would most likely penalise both training and test loss. Thus, a "style" can easily "emerge" simple due to the way we formulate these models mathematically. By intuition, consider how say a comedian or workman would often be reluctant to deviate from whatever is tried and tested. Machine learning models tend to be somewhat conservative by default, although there is plenty of literature on how we can sample them to try to avoid this at the cost of reduced accuracy/precision.

by ninjin 2y ago

DaSHacka 2y ago

> Unrelated, why is the duck in the "logo" looking at the back of their screen while typing? Having seen far too many AI-generated chibi images, I can say with 95% certainty the logo was generated by AI.

xanderlewis 2y ago
Absolutely. It’s unmistakable, even without the backwards screen. But why? Why does it have such a characteristic look? I wonder what causes this sort of recognisable ‘AI style’ to emerge.
- ninjin 2y ago
  I am not an expert when it comes to image generation, but know a fair bit about large language models and machine learning in general. My guess is that like all gradient-based models, they will find certain minima based on the the training data and be somewhat reluctant to deviate too strongly from it as this would most likely penalise both training and test loss. Thus, a "style" can easily "emerge" simple due to the way we formulate these models mathematically. By intuition, consider how say a comedian or workman would often be reluctant to deviate from whatever is tried and tested. Machine learning models tend to be somewhat conservative by default, although there is plenty of literature on how we can sample them to try to avoid this at the cost of reduced accuracy/precision.