Shrikant Joshi
shrikant.bsky.social
Shrikant Joshi
@shrikant.bsky.social
Audiophile. Tech enthusiast. Lurker.

Website: https://42quirks.com

#audio #podcast #editing #storytelling #radio #science #tech #AI #Physics #Astronomy #FOSS #Python #Django #shitpost #meme
Fair, I accept your argument and somewhat agree with it.

This thread has been an eye-opener in many ways but I'm still not entirely convinced. I definitely do need to read up a lot more, that much is abundantly clear.

In any case, thank you for your time and patience. Really appreciate it. 🙏
October 5, 2025 at 4:16 PM
I guess my question right now boils down to something along the lines of: "Do we know what's the 'salicylic acid' part in AI?" But that's on me to find out, not on you.

Thank you for all your responses - truly appreciate it! 🙏

/end.
October 5, 2025 at 4:10 PM
Great explanation re: medicine and salicylic acid - thank you!

I'm still stuck on the knowing how generative AI works bit and I'll need some downtime to think and read up about it. Would you mind if I came back to you with some more questions later?

(contd.)
October 5, 2025 at 4:10 PM
Excellent examples, thanks.

Re: the medicine analogy, if we know it works but we don't know how it works, doesn't that still count as knowledge we have?

In the sense, we know the medicine works. Does it matter knowing how?

I'm asking because it's the same logic/explanation we use for LLMs, right?
October 5, 2025 at 1:11 PM
Thank you for indulging me with your response. And it makes perfect sense, what you have said: As long as someone has the knowledge, we can safely assume everyone knows it.

One more question, where should I go to read about the internal workings of LLMs? As in, how do they process their "thoughts"?
October 5, 2025 at 1:07 PM
...as I did learn today.

May I ask why "we as a society don't understand this" doesn't apply here?

Honestly, you don't have to answer it but I would really appreciate it if you did.
October 5, 2025 at 12:36 PM
Not an armchair diagnosis. I have suspected it and briefly discussed it with my psychiatrist.

I did not want to out myself like this but I am diagnosed with Adult ADHD and I have suspected it is AuDHD for a while now.
October 5, 2025 at 12:32 PM
Agree. Sufficiently advanced science is indistinguishable from magic, and all that.

But as tech advances, it does become difficult to keep up with because it covers a wider surface area and there's only so much you can spread yourself thin...

I'm trying and failing horribly, as I learned today.
October 5, 2025 at 12:29 PM
No, everything is 100% me.

I sound like that because I am also a people-pleaser and try to be careful with my words.

Plus, I might be an undiagnosed autistic who has trouble grasping abstract ideas, as someone pointed out in one of the many replies I received today...
October 5, 2025 at 12:26 PM
(contd.) Again, I am not trying to claw back a point but I'm genuinely curious about how AIs can maintain context in their "thinking".

Damn, I don't even have the right words to ask my questions. I need to read up much more. Ignore this post and the previous one, please.
October 5, 2025 at 12:24 PM
If you are referring to my declarative statement that we don't know how AI works, I still kinda-sorta stand by that statement.

Yes, plenty have pointed out that the mechanism of how we arrived at it is well-understood but from what little I understand the actual "thinking" mechanism is a black box?
October 5, 2025 at 12:24 PM
...everyone who had to spend their precious moments calling me a bot, LLM, etc

Sorry I sometimes find it difficult to grasp abstract concepts.

To those who pointed me to resources, thank you.

To those who were patient with explanations, thank you.

To the others, hope your day gets better. (2/2)
October 5, 2025 at 12:14 PM
I don't understand LLMs but I am trying. And this thread is helping me learn.

Well other than the derisive comments and sarcastic responses...
October 5, 2025 at 12:04 PM
Hi Ed, thank you for your kind words. I promise I wasn't "just asking questions" - I was genuinely trying to sort my own thoughts out.

I don't think I am an "AI booster" but I do have problems grasping abstract concepts some of it may have been a bit abrasive. I've apologised to those I needed to.
October 5, 2025 at 12:00 PM
Fair, I will learn art history and exercise my art appreciation muscle more often. Thanks for the clarification!
October 5, 2025 at 11:52 AM
Yes, I am probably an undiagnosed autistic. Sometimes I find it difficult to grasp abstract thought.
October 5, 2025 at 11:51 AM
Again, I am not claiming humans are stochastic or AIs are sentient. I am trying to understand the whole thing myself.

Questioning my humanity pits me against you and that helps neither of us. I'd much rather have the argument with you about the subject itself rather than about you & me.
October 5, 2025 at 11:50 AM
Apologies. I was trying to keep up with a lot of conversations and a lot of different things from different sides.

You did indeed post similar responses but I was trying to answer too many things and defending ad hominem attacks. Got them jumbled up, sorry.
October 5, 2025 at 11:46 AM
Thank you for the references. I will read them to inform myself better on the subject.

And you're right, I should have probably done that instead of... this, but I was taken aback by the brigade that swarmed by after Alex quote-posted & rushed to defend myself. Probably shouldn't have.
October 5, 2025 at 11:44 AM
Yeah, somebody should have...
October 5, 2025 at 11:40 AM
Okay, I don't know then. I have seen/experienced very little art and I have very subjective opinions about the art I have seen/experienced. Maybe I don't understand art at all, so I don't know what else to say.

I'll learn art appreciation and maybe come back to this thread later sometime...
October 5, 2025 at 11:39 AM
Sorry, what does having a soul have to do with orgasming? Do plants also have souls?
October 5, 2025 at 11:37 AM
This is incredible, thank you so much for this!!!

Loved this part, specifically: "GPT is intelligent (like a calculator), but it's not sentient. GPT (in theory) is not aware of itself, but has been programmed (in practice) to pretend that is. "
October 5, 2025 at 11:36 AM
My bad, I looked it up and I see it is Magritte. But Duchamp's Fountain evokes similar ideas for me, so I got my wires crossed.
October 5, 2025 at 11:34 AM