Which is when he can me potentially most dangerous to himself and us. Something we should prepare for. Suggestions?

Trump is a Real Estate guy whose primary success came from bullying people with lawyers is massively over his head

Great idea. Republicans will shut it down. Now if the Republicans came up with the idea....

That's cool. What a neat view.

Back to an earlier regarding power. The cool thing about LLM's is you can do a mass amount of training, and a power hungry system, but you can also add to it's training, or specialize it, on less powerful systems. So you don't have to start from zero. But quality of data sets is very important.

Finally, here is an example of what an AI can do, when properly trained in specifics, and how it can surpass human experts in a specific task. www.nature.com/articles/s41...

I made sure to use studies that didnt just favor AI, but showed where it is good AND where it is bad. As a tool, one of many in the toolbox, AI can be quite helpful. I found this on how Epic is using AI to help Doctors, these are the kind of things I like. www.healthcareitnews.com/news/how-epi...

To point out why training material is important, this study using Chat-GPT 3.5 and Chat GPT 4 (both of which are general knowledge LLM's) while doing well, was not enough to be used as a sole source. Because the LLM was not trained with the needed materials. sjtrem.biomedcentral.com/articles/10....

Next up, this one discusses how GPT-4V outperformed clinicians in closed book environments, but was beaten in open book. It does also point out issues with visual accuracies. (Also, I am definitely NOT pushing for AI to be making decisions, but as a helpful tool) www.nih.gov/news-events/...

Fair, I made the assertion after all. While showing definite need for improvement in some areas, it is discussed how specifically in skin cancer detection one of the AI's surpassed general practitioners and was comparable to specialists. www.nature.com/articles/s41...

It isn't a large step from training robots in a virtual factory to training drones in a virtual world for nefarious intent. As a society we need to recognize that, and do something. My thoughts harken back to Asimov and the Three Laws Of Robotics. I am both excited and fearful for the future.

While the opportunity for bad is immense, so is the opportunity for good. But that comes down to the humans, and that is where some of my fear comes from. I can only hope we figure it out, before it is too late.

Also, you hit the nail on the head with smart connected devices. "Adoption". AI however is being widely adopted. That's why I saw we aren't "unringing that bell". I mean even casual criminals are using AI. Which is not something I am happy to say, but it is truth.

We have to fight to things. Lack of messaging by the Dem's and the "everything is an emergency" messaging of main stream media. As long as that keeps happening with Republican's excelling at using fear to win elections, things will not improve.

LLM's get more accurate the larger the model's, which make them more useful. With chip companies already focusing on making more energy efficient AI specific processors, this is only going to improve. The question is, will society evolve, too? That I kind of have my doubts. lol

They definitely went overboard with Smart connected devices, and there was not a lot of adoption. AI however is not remotely comparable. Factories are adopting AI tech to train their robots in virtual factories that are copies of their real factories. Saving time and money, for example.