I played with chat bots back in the AOL messenger days. There was this “Run-A-Bot” program I used to run at home, and when I was out I could chat with it and it was surprisingly interactive and realistic.
The responses were probably hard coded to expected inputs, done the hard way… as I feel it should be.
I thought wow if a company spent years add in more and more important data by hand this will be AMAZING in just a few years.
Fast forward 20 years and the AI companies did it the most lazy way possible. Just auto feeding data scraped from the web. Yikes, all the trolls and fake info included.
This new AI is just a sentence generator, it’s almost useless. They have some twisted algorithms to spit out junk but each one is leaning towards whatever twisted views the company that made it want it to spit out.
back then the buzz word was “expert system” and later “fuzzy logic”, all kinda sorta leading down the same path LLM 's have taken today, they just didn’t have the compute power back then.
I played with chat bots back in the AOL messenger days. There was this “Run-A-Bot” program I used to run at home, and when I was out I could chat with it and it was surprisingly interactive and realistic.
The responses were probably hard coded to expected inputs, done the hard way… as I feel it should be.
I thought wow if a company spent years add in more and more important data by hand this will be AMAZING in just a few years.
Fast forward 20 years and the AI companies did it the most lazy way possible. Just auto feeding data scraped from the web. Yikes, all the trolls and fake info included.
This new AI is just a sentence generator, it’s almost useless. They have some twisted algorithms to spit out junk but each one is leaning towards whatever twisted views the company that made it want it to spit out.
back then the buzz word was “expert system” and later “fuzzy logic”, all kinda sorta leading down the same path LLM 's have taken today, they just didn’t have the compute power back then.