News for hackers. Stuff that matters.

A comment on The 100k Whys of AI. Parent comment

Re:The 100k Whys of AI by Lerc on Sunday June 21, 2026 @06:44AM (#48617622)
I wonder how much variation there would be if you got a single model to produce a couple of gigabytes of tiny children's stories.

Might be an interedting research project.

[ Parent | on HN ]
Re:The 100k Whys of AI by pxagntuvzt on Sunday June 21, 2026 @07:11AM (#48617826)
There is one already: https://arxiv.org/abs/2305.07759 https://huggingface.co/datasets/roneneldan/TinyStories

6.5GB of tiny stories, as requested. ;)

[ Parent | on HN ]
Re:The 100k Whys of AI by sixtyj on Sunday June 21, 2026 @09:50AM (#48618999)
Texts in Gutenberg have 20GB, and full Wikipedia (English texts) have 80-110GB.

So to LLM-generate 6.5GB of tiny stories is quite a permutation in action :)

[ Parent | on HN ]
Re:The 100k Whys of AI by Lerc on Sunday June 21, 2026 @10:53AM (#48619475)
My comment was, in-fact, a subtle reference to this.

The best opening I got from my own TinyStories trained model was.

Once upon a time, in a small town, there was a large town.

Which I just love as an evocative idea.

[ Parent | on HN ]
Re:The 100k Whys of AI by aesthesia on Sunday June 21, 2026 @11:56PM (#48625472)
SimpleStories is a more diverse version: https://huggingface.co/datasets/SimpleStories/SimpleStories
[ Parent | on HN ]