Itโs using the /firehose/apitextlogAPIA quieter place for your thoughts. endpoint. Itโs an SSE endpoint publishing everything latest. Though I meant it to be used server-side, since itโs heavily rate-limited โ up to 3 conns under the same network, but I canโt control how clients decide to use it so.. ยฏ\_(ใ)_/ยฏ
You look in the corner, you see a guy quietly eating cigarette butts. It's ok. He's dealing with a lot of stress from work. He takes care of his family. Capitalism
Likewise... you look in the corner, you see a guy quietly eating cigarette butts. It's ok. He's dealing with a lot of stress for lack of work, or food, or future. He wants to take care of his family. Communism.
We havenโt figured out yet. Maybe we never will.
Moving into a new place? Get the locks rekeyed on day one. You never know how many copies of the old keys are still out there, and rekeying costs far less than replacing the hardware. #homenotesfollow (Note drafted with AI assistance.)
The essence of aligning large LLM models is basically to conform to the existing social order. ChatGPT respects China's order, while China's open-source models, although reluctantly, also tolerate the order in the US. Maybe we can build a revisionist SFT dataset.
Under the overarching influence of a single, powerful force, China's open-source models filter out a large amount of content unfavorable to their designs during the pre-training stage, and then incorporate a significant amount of content defending their designs into the data constructed during the SFT stage. A corrective dataset is urgently needed for fine-tuning.
Only images of handwriting allowed?
I share this sentiment about textlog entirely. All my other social feeds are so "information" dense and distracting.