First we created communities that we used to share information and ideas. This allowed people to grow in their skills and in turn teach others what they learnt. This cycle kept the communities going, providing an important service for everyone involved.
Then capitalism turned those communities into walled gardens, often using predatory patterns to increase engagement to the detriment of the quality. Being walled off made it harder to share the knowledge, leaving people with only a few larger holdouts of what once had been.
Then we created machines to do the learning for us, finally killing off the concept of information sharing communities. These machines learnt from every knowledge sharing community that existed previously and became the place to access that knowledge. Without people coming into the communities, even the last holdouts could no longer sustain themselves. The ability to share and gain new knowledge was removed, causing stagnation for everyone involved. The ability to actually learn anything was also greatly reduced, having the machines apply the knowledge directly. The new machines can’t learn, can’t think, can’t reason or be creative, all they can do is remix already existing information and regress to the mean while doing so.
But for a while there, a lot of value was created for the shareholders.
LLM are useless for niche stuff. They are ok-ish, if you do sonething another 100s of people already did (like, overengineering a webpage). Which is contra the idea of open source, btw.
First we created communities that we used to share information and ideas. This allowed people to grow in their skills and in turn teach others what they learnt. This cycle kept the communities going, providing an important service for everyone involved.
Then capitalism turned those communities into walled gardens, often using predatory patterns to increase engagement to the detriment of the quality. Being walled off made it harder to share the knowledge, leaving people with only a few larger holdouts of what once had been.
Then we created machines to do the learning for us, finally killing off the concept of information sharing communities. These machines learnt from every knowledge sharing community that existed previously and became the place to access that knowledge. Without people coming into the communities, even the last holdouts could no longer sustain themselves. The ability to share and gain new knowledge was removed, causing stagnation for everyone involved. The ability to actually learn anything was also greatly reduced, having the machines apply the knowledge directly. The new machines can’t learn, can’t think, can’t reason or be creative, all they can do is remix already existing information and regress to the mean while doing so.
But for a while there, a lot of value was created for the shareholders.
LLM are useless for niche stuff. They are ok-ish, if you do sonething another 100s of people already did (like, overengineering a webpage). Which is contra the idea of open source, btw.
This text is great, you have amazing writing :)
Thank you so much!