I’ve been on USENET back then (I wrote software for it), and it was nothing like the Internet of today. The amount of spam was neglectable, and perpetraters were quickly taken care of. Spam started to be a problem with Sanford Wallace (may he fry in the deepest hell!), and that was an email issue.
And the “expert systems” from back then were not even close to modern LLMs. They used rather sparse resources (even the big ones), they were “hand-made”, single problem oriented, and did not even try to look intelligent in any way. And nobody shoved them down your throat at every corner. A typical “expert system” of the time was an OCR system or a statistical analysis software for optimization purposes.
Spam certainly existed in the 90s. I’m baffled that you’re denying it – you were there!
AI hype cycles are nothing new. Today it’s LLMs. Past cycles have included deep learning, image recognition, expert systems, neural networks, etc. going all the way back to the 50s and 60s.
All of those existed during the 90s, unfortunately. Look up Usenet spam (1993) and for AI, the expert systems hype from the 80s.
I’ve been on USENET back then (I wrote software for it), and it was nothing like the Internet of today. The amount of spam was neglectable, and perpetraters were quickly taken care of. Spam started to be a problem with Sanford Wallace (may he fry in the deepest hell!), and that was an email issue.
And the “expert systems” from back then were not even close to modern LLMs. They used rather sparse resources (even the big ones), they were “hand-made”, single problem oriented, and did not even try to look intelligent in any way. And nobody shoved them down your throat at every corner. A typical “expert system” of the time was an OCR system or a statistical analysis software for optimization purposes.
I think you missed both my points.
You’re wearing nostalgia goggles.
I don’t deny those problems existed. But in comparison to today’s situation on the net, they were basically irrelevant.