A 40 ish woman who works with economics said "you don' know anything about AI because you're only 16". I then proceeded to show her my shitty AI me and my friend made and explained how it worked. (It was really shit but would still consider it an AI)

She just stood there and was pretending to understand so when I was done explaining about it, I told her not to assume you're smarter than a 16 year old just because you're older and have read some article about AI on The Daily Mail.

AI? ffs, stop calling it that. ML is the correct term.

@magicMirror Ai is not ML, just like DL is not ML

The last line was Savage though

@-vim- could you please explain these abbreviations for me?

I agree with @magicMirror and @-vim- , stop calling machine learning (ml) an artificial intelligence (ai). People love grander names!

@irene no idea about dl, and can't even speculate!

It is artificial (duh), it is (somewhat) intelligent (due to some kind of algorithm), but it is not 'artificial intelligence'!

Edit: on second thought, algorithm running (rule check and flow), is not something I would call intelligent. Or better let me rephrase, the person who created the algorithm is intelligent but the machine which is running it, is not!

DL = Deep Learning (what google does)

@Wack what does it mean?

@irene I ddged weak ai and got something from there

"Weak artificial intelligence is a form of AIspecifically designed to be focused on a narrow task and to seem very intelligent at it. It contrasts with strong AI, in which an AI is capable of all and any cognitive functions that a human may have, and is in essence no different than a real human mind."

If I understand correctly, it is ai to do one thing and one thing only! Still doesn't make it (weak ai), if it is static, without learning new things (in scope of the task), and learning boundaries and weighing as function of machine learning!

@irene ok I concede to definition, imo I find it silly but yeah, weak ai!

@irene Nope. DL is not AI, not even a weak one.
any complex enough algorithem can be "weak AI" accourding to that definition. in order to qualify as AI, the algorithem needs to incorporate some sort of AL (active learning) without an oracle, and from what I saw, no such thing exists yet.

@magicMirror was trying to convey the exact point but lost, you did it!

@irene in that case, then you can say weak [random buzzword] to impress people in all cases. like weak blockchain, weak javascript, etc.
🤔🤔🤔

@irene it is the same as "paradim change". When someone uses this term, usualy wrongly, they try to give the impression that this is something new. It is not.

tl:dr Same as "clean coal", or "make america great again", "ms is great", "weak ai" is complete BS.

@irene so, essentialy, you agree with me. the only difference is semantics.

@irene I didn't intend in strong ai terms, I want it to do one thing and one thing only, just keep improving the way program works!
Also, even I agree weak ai is legit term today, I didn't know it before today and as magic says not too old.

Keeping the legitimacy of the term aside, this program is not even weak ai as it doesn't learn (unless you consider the machine learning aspect of it) and machine learning is not intelligence, it is pattern data analysis for the dimensions the ml algorithm is coded for (and not even a bit more).

@yendenikhil if you're still interested in an answer to your question: the best thing I can offer you is: I don't know. The second best answer is, you put layer after layer of a DNN and try to get a usefull output, without knowing if you do the right thing. Kind of.

@Wack the concepts are still big vague for me (as I don't understand neural networks intimately), but if machine learning is due to clever algo then deep learning learns due to neutral networking, but that's where my understanding ends!