• 0 Posts
  • 8 Comments
Joined 2 years ago
cake
Cake day: July 30th, 2023

help-circle
  • OpenAI noticed that Generative Pre-trained Transformers get better when you make them bigger. GPT-1 had 120 million parameters. GPT-2 bumped it up to 1.5 billion. GPT-3 grew to 175 billion. Now we have models with over 300 billion.

    To run, every generated word requires doing math with every parameter, which nowadays is a massive amount of work, running on the most power hungry top of the line chips.

    There are efforts to make smaller models that are still effective, but we are still in the range of 7-30 billion to get anything useful out of them.


  • Personally I think it’s fallen out of fashion. For my blog I’d either use a meme or other dump picture for each post. When generated images first came out I used a few for blog posts, it was new and interesting and said “I’m interested in technology and like playing around with new things”.

    Nowadays I’m back on the meme pics. I feel now it’s so much easier to generate images, it more says “I want to look professional but also spend no money and have no standards”.







  • This is why I find this stuff so bewildering, even Wikipedia says no one died in the square. It was hectic around Mudixi, with buses of soldiers being torched and the burnted bodies being strung up. People there were absolutely shot at and killed.

    It was an insane week with a lot happening, many different groups with their own motives, and so many details unclear. It’s weird that anti China rhetoric insists on something that didn’t happen, pushing a false narrative that’s so easy to dispell and distracts from the real violence and politics of the time.