Many users push back on claims that AI chatbots erode society's collective intelligence by citing faster learning and creativity gains, while others report personal cognitive decline from over-reliance.
No Digg Deeper questions have been answered for this story yet.
Most Activity
If you are comparing an LLM to a calculator to say “well, we didn’t stop thinking when we invented calculators”, I’m going to assume you are already beyond salvation.
What happens to society’s collective intelligence after two or three years of everyone outsourcing their thinking to a chatbot?
How many orders of magnitude dumber are we going to get?
Can we start talking about that yet, or nah?
@RoseSilicon @svpino Just like devices, chatbots are gonna be hardwired as human nature. And most people don't wanna fight that
There's no excuse to vibe code decompilations, yet people are still doing that
@svpino We can’t talk about it until it’s too late as per how society works
@svpino @grok Should we start talking about that yet?
@svpino This is a really old argument and goes at least all the way back to Socrates with writing. Its mostly fear of changed, mixed with some gatekeeping for flavor, masquerading as insight.
AI is literally just computers doing calculations.
Machines did a lot of thinking for us already, it's just the logic has been simpler if/then statements and loops.
Now it's if/then statements and loops doing matrix multiplication.
Google search just uses a keyword input string to score billions of links and serve the top 10 in a list.
LLMs just use a token input string to repeatedly score billions of tokens and then a pseudorandom number to print one of the tokens from the top_n scored, until the token printed would be an End-Of-Sequence token at which point the algorithm stops the scoring loop.
Technically LLMs would continue scoring and printing tokens after the End-Of-Sequence token if we wrote the code to do that, but the token predictions get pretty bad after an End-Of-Sequence token because the training data doesn't contain any strings with tokens after an End-Of-Sequence token.
@svpino I was doing some manual coding and damn I forgot a lot of things, I'm more hands on now
I’m not sure how this will evolve, but I don’t think we’ll get dumber overall. When calculators arrived, people probably thought we’d forget math. Instead, we spent less time on arithmetic and more time solving bigger problems. The same thing happened with computers.
AI feels like the next step in that pattern. Some skills will become less important, while others will become more valuable. The question isn’t whether we’ll think less. It’s whether we’ll use the time and mental energy we gain to think about better things.
Then again, maybe we’ll get dumber. Who knows?
@svpino Not sure how accurate this may end up being. I am finding that my knowledge is actually growing thanks to the "outsourcing" of my thinking to Claude or Codex. I feel so much better asking stupid questions to these models after I ask them to do things for me.
@svpino Gentle reminder that Socrates faulted writing for weakening the necessity and power of memory, and for allowing the pretense of understanding, rather than true understanding.
@svpino I'm getting more retarded by the day. I am aware of it.
Reading books Writing code by hand Doing leet code exercises
these things are helping me not lose my intellect
@svpino The people using AI to avoid thinking will get worse. The people using AI to think bigger will get better. The gap is what’s about to get scary.
@svpino is it actually making us dumber or just changing the skill set?
@svpino I’m not convinced the problem is using AI to help think.
We’ve always “borrowed” brains: books, teachers, mentors, search engines, smart friends, abacus, calculators.
The risk is accepting the output without doing the human part….
@svpino Humans have been outsourcing their thinking to the political class for 12,000 years. High automation allows for more direct contribution from the average person and far more conversation. What we need now are true guilds, basic income, and community driven organizations.
@svpino There are going to be safe cities that are controlled by the government with heavy Force and then the rest of the country will just turn into a dystopia where leaving your house maybe the last time you ever leave it.
@svpino just don't vibe code and you'll be fine
By then, the realization will be undeniable that the architecture was wrong to begin with. Cognitive friction is needed.
Our brain is bandwidth dependent. Higher bandwidth, but less time spent with it- we autopilot. Lower bandwidth, higher time spent with it, we retain better.
This is shown when we write things down vs typing on keyboard. We force our brain to use motor skills to manipulate the pencil, vs just tapping the keys. We retain written things better than the things we type. https://pubmed.ncbi.nlm.nih.gov/24760141/
In the future, we may have cognitive athletes, who exercises their facilities to maximize cognition, just because they're more visible at that point.
@svpino Read books, go the manual way sometimes, keep yourself sharp. You already know what they wanna archive with all this, be smarter
@svpino Idiocracy was a documentary