how come i keep seeing posts that "no they actually can code now" and also posts like "no they actually can't code now"
anyone wanna decipher this for me
0 Boosts
0 Likes
@majorlinux some people don't like the lying machine that convinces people to kill themselves and is a part of eugenicist cult project that's literally choking people while drinking their water and eating up every other resource...
@gildilinie I feel that wasn’t a part of the original post you made.
I was speaking more or less on the coding aspect of it.
But, since you’ve opened that can of worms:
A computer can’t lie to you because it doesn’t understand the concept of right or wrong.
It is unfortunate that people turn to computers to seek mental health, but we have a system that de-incentivizes seeking help and makes it out of reach for many people.
@gildilinie This overall tech predates the cult and said cult has all the money to shape the current iteration of this tech because unfortunately things only get done when the money is there to do it.
There are plenty of tools that do not do that and I’d argue that training models seems like a moot point now seeing where we are now and the capabilities of some of these models being able to run on things like smartphones.
The only way the harm goes away is with a radical change of society.
@majorlinux it wasn't a part of my original post but you've been begging for arguments!
And I'd like to know what LLMs don't do what I said? The local ones? The local ones still lie and convince people and do all kinds of nasty shit, and they still cause all the water, air, and resource consumption problems when they're being trained.
What does a good LLM look like?
These things, as long as they know natural language (English, French, Chinese, whatever) eat away your mind and turn it into cheese. Not your brain but your mind.
@gildilinie This isn’t me discounting that terrible things have happened. I’ve talked about them ad nauseam.
Let’s start with the LLM itself:
Had it been properly tested and vetted before being released, a lot of what we’re seeing could have been easily avoided. It’s like putting out bad code. If you know something like this happens, you make sure it doesn’t happen again.
@gildilinie As both of us pointed out before, these tools are nothing more than powerful autocomplete machines. There is no thinking behind them at all. It seems weird to me that people who acknowledge its software keep trying to attach personhood to it by saying it “lies” or “thinks”.
I’d argue that any “good” LLM would be something that is attached to a well-built harness with proper guardrails. Things that are built for specific tasks. Definitely not for fucking therapy and your BBQ flyer.
@gildilinie And, as I said for training, we don’t really need it anymore.
That should just stop.
No more data centers. Stop building them. Let the people have their RAM and storage back.
The only reason that they keep going is because everyone is trying to be the first to something that may or may not exist and to corner that market.
If the people controlled it, we’d have stopped a long time ago.
I mean, as you have mentioned, there are better ways and that could have been funded instead.
@gildilinie And I still don’t buy the “mind into cheese” argument.
I’ve been asking what that even looks like.
I’ve had more time to read, play video games, debate on the internet, do my son’s math homework (common core fucking sucks), be civically engaged, and still feel sharp.
So, what is the “mind into cheese” bit people keep referring to?
Offloading repetitive tasks?
Doing data analysis?
Answering a boss’s email?
People who look bad when the AI is taken away were never good to start.
@majorlinux they fucking combined the human need for social interaction, anthropomorphism, gambling, flattery, cognitive offloading and reinforcing people's beliefs into a single fucking thing. It's like a slot machine if a slot machine also kissed your ass and did your homework for you.
I think the last one you put "answering a boss's email" is really bad... now I can't even tell my boss to go fuck himself without going through Anthropic Eugenic Corp
@gildilinie Yes, that’s what the ruling class did.
The tool didn’t do it itself.
Most tech companies have doctors and experts on staff to keep people using the product more than they should.
These are people who shouldn’t be in control of shit.
Let Anthropic see it. I hope it cleans it up for me.
@gildilinie That’s been the argument people have been having about LLMs for quite some time.
I’d wait on people to make up their mind, but I have my thoughts on it and this one I’ll keep to myself before I get shouted down by angry nerds on the internet.