Young People and Politics

[HVC]

But you weren't really sharing  your ideas, you were playing with your new favorite toy.

[grumbler]

But don't you see that you weren't actually sharing your ideas?  You were simply cutting and pasting something an application generated, which is now notorious for generating false information.

But we do that all the time, here: cutting and pasting from other sources to support (and often even replace) our own arguments or definitions.

DG was not claiming to have written that "definition."  He clearly was citing its source.  You can argue against what the source says, but arguing that we should not use/cite/quote external sources because then we are not "actually sharing [our] ideas" seems to be inconsistent with meaningful discussion.

[grumbler]

But you weren't really sharing  your ideas, you were playing with your new favorite toy.

I always enjoy a good strawman argument in the morning.  It smells like... Languish.

[HVC]

But you weren't really sharing  your ideas, you were playing with your new favorite toy.

I always enjoy a good strawman argument in the morning.  It smells like... Languish.

Glad to be of service

[crazy canuck]

But don't you see that you weren't actually sharing your ideas?  You were simply cutting and pasting something an application generated, which is now notorious for generating false information.

But we do that all the time, here: cutting and pasting from other sources to support (and often even replace) our own arguments or definitions.

DG was not claiming to have written that "definition."  He clearly was citing its source.  You can argue against what the source says, but arguing that we should not use/cite/quote external sources because then we are not "actually sharing [our] ideas" seems to be inconsistent with meaningful discussion.

He was using an application to produce an argument.  He wasn't making his own argument.

[DGuller]

But you weren't really sharing  your ideas, you were playing with your new favorite toy.

ChatGPT is not a toy for me, it's a very serious tool I use to increase my capacity to focus on what matters.  To be honest, I'm so normalized to it that I'm taken aback that a mere mention of it is apparently so triggering to so many people.  In my workplace, most (if not all) data scientists use it extensively, and I do too.  Coming up with new productive use cases for it is something that gets you respect, not scorn.

For example, I had a long conversation with it about an optimization problem I had.  I asked it whether I can set up an optimization problem I was working with as quadratic programming, and if so, how would I do that.  It got me 90% of the way there, and as a result I spent my time and energy making a more effective use of the optimization logic to solve a business problem, rather than figuring out the utlra-tedious details of setting up the matrices for it.

My math professor once said (paraphrased):  if a calculator can do it for you, then it's not a good use of your time for you to do it.  Use that time to figure out what you want to do with the calculator.  I think the same philosophy applies to LLMs, and I will continue thinking it.

[Grey Fox]

In my line of work, ChatGPT keeps making shit up.

[DGuller]

In my line of work, ChatGPT keeps making shit up.

That's why we have humans in the loop.  When I pasted ChatGPT's definition of "woke", I also said that I stand behind that definition.  I didn't stand behind it automatically, I stood behind it after I read over it and confirmed that we were in alignment.  If I saw that it didn't give me what I needed and made shit up, then I would've concluded that maybe my concept of woke is not as unoriginal as I thought, and that I would have to do the work myself defining it, but that didn't happen.

When I asked ChatGPT to give me a skeleton code for the quadratic programming, it got a detail or two wrong as well.  However, the stuff it got right got me started, and freed up my bandwidth to think about all the details.  If I had to do it all myself, I wouldn't be nearly as far along, and I'd probably get a few details wrong myself.  And I might not catch them, because then I wouldn't be checking someone else's details, I would be checking my own details.  It's a lot harder to be the doer and the checker than being just one or the other.

[grumbler]

He was using an application to produce an argument.  He wasn't making his own argument.

He was using an application to produce a definition.  His argument was that he agreed with that definition.  ChatGPT had no way to force him to agree with its definition.

It's no different than quoting a book or movie and saying that you agree with that sentiment.  One can argue whether the sentiment is plausible, but one cannot argue that the poster's agreement with the sentiment is untrue.

[garbon]

One can argue whether the sentiment is plausible, but one cannot argue that the poster's agreement with the sentiment is untrue.

What if you thought a poster was lying? (Not speaking of this moment but more generally)

[Tamas]

I guess where our mockery of DG's flippant use of "AI" was unfair is that we can be pretty sure that using search engines as we know it will be a thing of the past with "AIs" slowly taking their place.

We wouldn't get offended if somebody copy-pasted an article or definition they found using Google. But we are clearly not ready for "AI".


EDIT: and that we are being grumpy, tech-averse old people in a thread about young people and their politics is funny.

[garbon]

To be honest, I'm so normalized to it that I'm taken aback that a mere mention of it is apparently so triggering to so many people.  In my workplace, most (if not all) data scientists use it extensively, and I do too.  Coming up with new productive use cases for it is something that gets you respect, not scorn.

While I recognise why you wouldn't feel so in this moment, I think that is one of the strengths of Languish. Getting us out of the siloed thinking of our own immediate professional/social environments.

[garbon]

I guess where our mockery of DG's flippant use of "AI" was unfair is that we can be pretty sure that using search engines as we know it will be a thing of the past with "AIs" slowly taking their place.

We wouldn't get offended if somebody copy-pasted an article or definition they found using Google. But we are clearly not ready for "AI".


EDIT: and that we are being grumpy, tech-averse old people in a thread about young people and their politics is funny.

Of course, I'm also not sure many of us react positively to being quoted dictionary definitions either.

[Jacob]

That's true, I would not be interacting with ChatGPT. I would be interacting with ChatGPT generated content.

To elaborate, the interesting part of interacting with DGuller is stripped from the exchange by feeding it through ChatGPT.

The interesting part of interacting with me is when I write stuff that ChatGPT could've written? Ouch, that really hurts.  I thought the stuff I write that ChatGPT can't think of was the good shit.  You don't find that interesting, not even a little bit? 

Your reading here is like the exact opposite of the intended meaning.

You frequently have nuanced takes, and I very much appreciate your wit - which is very contextual, usually self-aware, and rather dry. That's what I appreciate about your posts.

Writing that originates from ChatGPT obviously lacks that. I does not express your personality, it lacks awareness of the larger context and of our interpersonal and forum-specific history, and is therefore pretty uninteresting to me.

My math professor once said (paraphrased):  if a calculator can do it for you, then it's not a good use of your time for you to do it.  Use that time to figure out what you want to do with the calculator.  I think the same philosophy applies to LLMs, and I will continue thinking it.

Which is fair enough, but...

... if you consider "... then it's not a good use of your time for you to do it" (and therefore I had ChatGPT do it) in relation to a conversation with someone, then you can get pretty close to suggesting "it's not a good use of my time to respond to this part of your argument" or even "it's not a good use of my time to have this conversation with you at all" (even if by accident).

It doesn't have to mean that, of course, but you're relying on your conversation partner(s) to share your (unspoken) assessment of which part of the conversation is a good use of time and which parts can be outsourced to ChatGPT.

In the specific case of your conversation with Valmy you relied on ChatGPT to provide a definition of "woke". This left Valmy with basically four options:

1) Agree that ChatGPT's definition is authoritative.

2) Agree that the definition of woke was not a critical part of the conversation, and therefore "not a good use of time" to dig into.

3) Take issue with ChatGPT's analysis and therefore start an argument with ChatGPT by proxy.

4) Get pissy about the use of ChatGPT.

None of those options are especially good, especially if one of Valmy's core issues is that the definition and application of "woke" in the current discourse is a bunch of bullshit (and that seems to me to be a part of his position).

[The Minsky Moment]

Donovan McNabb had such a weird career man. He took so much shit for no reason.

He played for the Eagles. No better reason to take shit than that.