RE: It would be funny if it wasn't sad
You are viewing a single comment's thread:
If you take a look at the sources, most of those are "leftist" sources, to whom Elon Musk is a villain. Meaning, the AI is really just a reflection of data created by humans. Now it would be interesting to know why those especially have been chosen but that's outside of this scope.
I'm pretty sure Elon knows about this issue and wants to fix it, but I doubt he'd put it offline, as that would go against everything he publicly stands for. Rather, he'll probably will try to get the AI training data as optimal, unopinitionated and truth-seeking as possible, which clearly is difficult in todays political landscape.
Here's a quick example of my viewpoint:
Given that you're building guitars, imaging building one with poor quality ingredients. Regardless how much effort you put into it, the output will be less good as if you're using the best resources available.
And I'd argue the same can be said about AI and data. If the data is poor, the output will be poor. If the data is optimal, the output will be optimal. This is especially true as AI is getting trained more and more on data created by AI.
So in the end, truth (aka good data) really matters to get an optimal AI.
hey brother, well... you do make a lot of sense, I hate to admit.
It has to be the case that the amount of people pointing out everything that is wrong about Elon, thus creating content on this subject must be magnitudes above the people speaking positively of the man.
Now that being said. How can this ever be corrected? Should it be corrected?
Would this not be an attack vector, an attack on any narrative. The more people become reliable on AI for verification of information the more this vulnerability "works".
I have a new thing to worry about now...
lol
Cheers my friend