Current AI relies on search results and more broadly from what it has and is scrapping from the web. That is Ok for now, but AI is polluting the web with its own low grade output. And this trend is accelerating, so the low quality information is proliferating even faster now. As AI starts to scrape that low quality information, it will impact its own ability to create 'quality' output.
Put simply, the more garbage AI creates, the more of that garbage it eats which in turn degrades its output.
So, the current LLM model of 'AI' is OK, but unless it has a fundamental shift in how it works, it really won't be that long before it becomes useless.
I hope you have not invested too heavily in AI.
That's an Atom
When I spent one Sunday night working on a photograph in our basement laboratory last August, I was admittedly quite pleased with the results.
That spec in the middle, that's a single atom. Or, at least the light emitted from that single atom.
"Google employees demand the company pull out of Pentagon AI project"
I was trying to intelligently respond to this issue, but in reading the comments on the Verge article, I found a rebuttal that is much more concise and well reasoned than anything I could come up with. So here is commenter asarathy's response:
This attitude is so entitled and naive. Sure, the company should be careful in engaging in out right evil like say identifying which citizens to round up for the Purge, but to act as if there is no place for Google to help in defense of what is their home country is silly. Yes I get they are an international company but they are an American company and there are places where it makes sense to leverage it's talents for it's defense.
Photo by David Nadlinger
Painting by Susannah Bleasby
Photo by Reuben Wu
Photograph by Masha Ivashintsova
