AI Breakthrough

[Jonax]

Hi all. No one has missed the hype over the new AI tech.


There has mostly been buzz about the chatGPT from openAI and that Bard/Gemini which makes strange illustrations.


The true pioneer IBM, with DeepBlue and Watson, has on the other hand kept a low profile. Apparently they have been silently improving their tech without talking much about it. That approach has now paid off handsomely. Watson has made a quantum physics breakthrough which makes frequencey doubling of OLED monitors possible. I don't claim to understand the physics involved but the result is that the resolution of your computer monitor doubles. Pretty neat.
I for example have a 27" QHD monitor. After the doubling I can enjoy 4K resolution. Of course the monitor is still only 27" but the movies and images are much sharper.


Watson could make the upgrade on all internet-connected eligible computers on its own but is aware some people are skeptical about the new AI tech. So you got to initiate the process yourselves. It still doesn't work on Apple computers but for Linux and Windows the process is incredibly simple. Just press the ALT + F4 keys and the upgrade will begin.


Good luck and enjoy the enhanced monitors!

[SilverWarior]

This was a good one

[Jonax]

This was a good one

Thanks. Had to try something for the first of April

[davido]

It's very cool thank's for sharing.

snaptube vidmate

[Jonax]

I'm running low on creativity today. The best I can think of is to bump this thread.

[Ñuño Martínez]

I didn't see this thread.

I just add that ChatGPT and similar are the same than ELIZA (1960's). The only difference is the amount of data and the speed of the microprocessor.

[pitfiend]

Finally, some one with memory is aware of that gemme of code. The modern AI is no different of Eliza, better parser, humongous database, still a glorified IF!

[SilverWarior]

The modern AI is no different of Eliza, better parser, humongous database, still a glorified IF!

I can't agree with that.
ELIZA is generating its responses using a carefully chosen response patterns using keywords provided in a question. This greatly reduced the chance for ELIZA to provide false information. ELIZA has no ability for context tracking of entire conversation like modern AI's can.

Unlike ELIZA that used carefully chosen response patterns modern AI response patterns have actually been generated by analysing huge amounts of textual data using various machine learning algorithms. This does allow moder AI's to generate much more human-like responses.
But the main problem is the data that they are trained upon. Most modern AI's are trained on data found on the internet. And we all know that you can't trust everything you read on the internet as it contains a lot of BS (false data). Os if you train you AI on a bunch of false data you can't avoid that "false data" seeping into its answers.

While speaking of false data on the internet. In order to protect our children or elderly people we often talk about parental protections which in essence try to filter or prevent access of our children to such "false data".
Why don't we use similar approaches when choosing which data we use for teaching of the modern AI's?

[Jonax]

Super big subject. I have probably heard about ELIZA sometime but forgotten all and had to make internet search to get up to date. Impressive program. The newer AIs? Those poor enteties are indeed trained on a lot of crap content. I have never trained an AI but suggest we got a responsibility to make reasonable posts on the net. Though I for one will of course keep joking now and then and the AI got to deal with it. An example: the original post in this thread was obvoiusly fake. But fake because first of April. A sufficiently advanced AI should probably be able to figure that out, eventually.

Anyway the AIs won't disappear and they seem to become more and more advanced. The future? Who knows? The sky is the limit, perhaps.

[SilverWarior]

An example: the original post in this thread was obvoiusly fake. But fake because first of April

It is interesting how last years April first fools joke is turning into a real conversation about AI. I wonder why?