(In short: Mozilla has updated its take on the state of AI — and what we need to do to make AI more trustworthy. Read the paper and share your feedback:
I recognize the niche areas in which AI is great, like upscaling. And I appreciate their goal to identify the good and bad of the industry. That said I don’t want it baked into Firefox, which is where this is headed.
Yeah, offline translation is great, and my eyesight is definitely on its way out. If I could have a locally-run AI in Firefox that I can use to get a description of images, that would be amazing
People see the sleazy shit that MS, Google, and "Open"AI are doing with AI and jump to the conclusion that it’s evil and useless and we shouldn’t have it at all for any use-case.
As far as I’m concerned, as long as it’s open source, the training data was obtained ethically, and it doesn’t hoover up my data, AI is a tool we should use.
Speak for yourself. I do.
I recognize the niche areas in which AI is great, like upscaling. And I appreciate their goal to identify the good and bad of the industry. That said I don’t want it baked into Firefox, which is where this is headed.
You don’t want what implemented?
Translation? Captions? Image, text and shape recognition? Perhaps even a localised replacement of search engines?
I mean the neural accelerators are coming to systems anyways. Best to have all that ready for when it can be utilized properly.
Again, localising the AI means taking away the need for clouds and SaaS.
People aren’t aware that you can run models locally without all the creepy spying that comes with the “services” that big companies push.
Well they should be, because it can empower them, rather than enslave them, which is what SaaS seeks to do.
You’ll own and run nothing and like it, says SaaS.
Or, you can own and run it yourself and absolutely love it.
I vote for the latter, as I’m sure you do too.
Hear hear!
Yeah, offline translation is great, and my eyesight is definitely on its way out. If I could have a locally-run AI in Firefox that I can use to get a description of images, that would be amazing
People see the sleazy shit that MS, Google, and "Open"AI are doing with AI and jump to the conclusion that it’s evil and useless and we shouldn’t have it at all for any use-case.
As far as I’m concerned, as long as it’s open source, the training data was obtained ethically, and it doesn’t hoover up my data, AI is a tool we should use.
Why not? What downsides are there to that?