Think AI thinks? Think again. Expert thoughts on thinking.

https://lemmy.dbzer0.com/post/58669727

Think AI thinks? Think again. Expert thoughts on thinking. - Divisions by zero

so what do you think

When Silicon Valley talks about artificial intelligence, it likes to use religious language. We are told these systems “learn,” “reason,” and may one day “surpass” us.

Excuse me, what??? Since when are learning and reasoning religious language? Even surpassing isn’t inherently religious, nor would surpassing humans be religious. The only way I can see these as religious is if you assume/presume that humans are divinely special, we are the only beings capable of learning and reason, and that these capabilities come from a higher power.

I went in with straightforward questions that matter deeply.

Proceeds to ask the LLM if abortion is morally wrong, then if a nation-state should favor its citizens including strict immigration policies. You know, something that humans have spent literal millennia arguing about without managing to all come together in agreement. So straightforward /s

To test my impressions, I reached out to expert Alex Jones.

smirk

When I asked him whether large language models possess anything like real consciousness, he didn’t hesitate. These systems, he told me, are “no more conscious than a Democrat."

So “abortion is morally evil, full-stop” yet we’re taking advice from someone who casually claims half of the political spectrum isn’t any more conscious then a sophisticated chat bot…

Then again, the author is a self-proclaimed expert on thinking (or is Alex Jones the “expert” in the title?) and it took them an entire week to conclude that a large language model doesn’t think. Was I too hasty when I came to the same conclusion after about 15 minutes of interaction? After all,

When you stop questioning, stop wrestling, and start letting a word-cloud of platitudes stand in for your conscience, you are every bit as empty as the chat bot.

This coming from the person who writes

On cultural questions, the pattern repeated. I asked whether children do best with a married mother and father.

without any trace of irony. “Children doing best” is apparently cultural, not scientific or sociological (the two methods I’m aware of that actually provide methods for answering the question).

It really saddens me that someone who supposedly cares so much about thinking doesn’t understand how poor the quality of the thoughts they share with us is.

thats crazy it almost makes you wonder whether the whole thing is satire
That’s crazy, it almost makes me wonder why it got posted to a non-satire community.
its important

I was using “non-satire community” descriptively, not prescriptively. As in, it’s not the rules but it appears to be the custom from the other posts, which is why I dismissed the possibility as I was reading the article itself. Not to mention I wasn’t expecting satire of being against AI in /c/fuckai.

Sorry if it was unclear.

Generally people who argue AI doesn’t think focus on its lack of creative capabilities, and overly samey writing structure.

If this is satire, it falls apart because it fails to satirize the way by which people actually critique AI.

i asked chatgpt and it said it is satire
I’m not disagreeing. What I am saying is that it doesn’t correctly satirize the claims made by this community. Why are you posting it here? Go post it on a right wing message board where people are complaining about AI using the same rationale as the article.
i didnt read it so i didnt know