> It is "hard for anyone except larger companies to benefit from the underlying technology," OpenAI stated.
Ugh. This is the part of the recent AI advancement that annoys me the most. The web, at least in its infancy, was pretty accessible to individual hobbyists. You could run a web server on your own computer. AI technology seems highly centralized from the start.
If you expand your timeline a little bit you'll reach a spot where the computers, equipment and comms necessary to create a computer network were highly centralized and only run by large corporations and institutions. How many times do we need to see this pattern for it to be obvious?
> You could run a web server on your own computer. AI technology seems highly centralized from the start.
I compare the current AI hype to the mainframes of the past. Highly centralised systems that people logged into with a lightweight terminal to send commands.
Hopefully just like Personal Computers revolutionized computing, Personal AI models in the future will not run in the cloud...
How were PCs able to work on a smaller scale? I've been looking into the chips/model trends and I can't see local models that perform at the same level as GPT happening in the near future.
The web in its infancy only worked because of the massive infrastructure that allowed your server to talk to others. In the same way you can tinker with your own web server, you can tinker with AI tech. You just can't reasonably train it independently, though the cost in terms of compute resources is something like 1% of what it was many years ago.
Ugh. This is the part of the recent AI advancement that annoys me the most. The web, at least in its infancy, was pretty accessible to individual hobbyists. You could run a web server on your own computer. AI technology seems highly centralized from the start.