this post was submitted on 19 May 2025
1510 points (98.2% liked)
Microblog Memes
7656 readers
2712 users here now
A place to share screenshots of Microblog posts, whether from Mastodon, tumblr, ~~Twitter~~ X, KBin, Threads or elsewhere.
Created as an evolution of White People Twitter and other tweet-capture subreddits.
Rules:
- Please put at least one word relevant to the post in the post title.
- Be nice.
- No advertising, brand promotion or guerilla marketing.
- Posters are encouraged to link to the toot or tweet etc in the description of posts.
Related communities:
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Except calculators are based on reality and have deterministic and reliable results lol
Edit: holy crap I would never have guessed this statement would make people wanna argue with me. I've never felt that my job is secure from the next generation more than I do now.
a transformer model is also deterministic, they just typically have noise added to appear "creative" (among other reasons) it is possible to use a fixed rng seed and get extremely deterministic results.
the results will still be frequently wrong but accuracy is a completely different discussion.
You're not wrong so you get an upvote but in the context of this conversation you know people are not using LLM tools with preseeded entropy. Also kind of a moot point because the idea of using some consistent source of entropy in a calculator is competly nonsensical and unnecessary.
Yeah but we heard the same arguments when they came out. Nobody will learn math people will just get dumber. Then we heard the same with the Internet. It's but trustworthy. Wikipedia is all lies. Turns out they were great tools for learning.
Your point is a false equivalence. Just because people said the same thing doesn't mean a calculator and an LLM are equivalent in their accuracy as a tool.
I'm not talking about accuracy. The Internet isn't accurate and they said the same things about it. Either AI isn't going away. Remain a troglodyte or learn to master it to enhance what you can do. That's how I dealt with it in the past.
Lmao I use LLM powered tools in my work daily, I understand their limitations and stay within them so say what you will. I still think your comparison is dumb.
You can make mistakes with a calculator. It’s more about looking at the results, verifying the data, not just blindly trusting it.
Your point has no bearing whatsoever on my statement. You could also misread a ruler but doesn't mean there's anything wrong with the ruler. Given I can reliably read a ruler, then I can 'blindly trust' it assuming it's a well manufactured ruler. If you can't that's definitively a you problem.
I mean it kinda does. If all you do is type numbers into calculator and copy results there’s a chance the result is wrong.
The same way some people use AI, which is wrong.
My point wasn't that people don't make mistakes they obviously do. My point is that calculators are deterministic machines; to clarify that means if they have the same input they will always have the same output. LLMs are not and do not. So no it's not the same thing.
I never said it was the same. I just said you have to be careful with tools you use. It applies to every tool.
You are implying that one must ensure the veracity of the output of a calculator in the same way that one must ensure the veracity of the output of an LLM and I'm saying no, that's strictly not true. If it were than the only way you could use an LLM incorrectly would be to type your query incorrectly. With a calculator that metaphor holds up. With an LLM you could make no mistakes and still get incorrect output.
I’m implying that you should be careful when you use tools, and not blindly trust the output.