Matt LuckcuckDec 15, 2024
Luis Apiolaza
Some Profs at uni say that we should teach students how to use LLMs (like ChatGPT, Gemini, Claude, etc) ethically and responsibly. But how can we do that with tools that, at their core, are
1) based on stealing labor from creators,
2) are catastrophic from an environmental point of view due to its high energy consumption, and
3) are burning billions of dollars on speculative investment, to
4) produce little value? #AI
Dec 14, 2024 at 7:57amWeb
Show threadAdrian KosmaczewskiDec 14, 2024

@ojala I didn't address all issues you mention but IMHO the genie is out of the bottle by now

https://deprogrammaticaipsum.com/banning-adopting-reckoning/

Banning, Adopting, Reckoning

Last month, OpenAI, the company (in)famous for their ChatGPT product, released a course called "ChatGPT Foundations for K-12 Educators", an event that has raised more than a few eyebrows, and even some outrage. We must have a serious conversation about the value of a bullshit generator in the context of teaching programming skills to new generations.

De Programmatica Ipsum
Show threadsl1200Dec 14, 2024
@ojala It would be better to teach the #gemini smallweb protocol to students. Take a look:
gemini://sl1200.dystopic.world/links/linksgemini.gmi
Show threadDavid Delgado VendrellDec 14, 2024
@ojala Did you think about how unsustainable most likely is the construction process of the buildings where those teachers work from, too? Or the smartphone technology's carbon footprint they generate while claiming against the use of LLM in a shared forum?
Show threadJim GardnerDec 14, 2024
@daviddelven @ojala What’s the argument here? That LLMs will contribute less to climate change in total, than buildings and smartphones? If so, that doesn’t seem like a valid comparison to me. Especially buildings. You don’t have to condemn everything in order to criticise one thing…
Show threadDavid Delgado VendrellDec 14, 2024
@jimgar @ojala What I mean is that in order to condemn a new technology because the potential impact on environment, we first should look at the current conventional "technologies" that we the humans forgot they have, as the 40% of carbon footprint from buidling industry, and it doesn't seem a trending topic to talk about. Now the focus is on #AI... Ok.
Show threadJim GardnerDec 14, 2024

@daviddelven @ojala We can collectively talk about both, and individually choose to talk about either (or both, or even neither).

Choosing as an individual to criticise one independently from the other does not invalidate the arguments being made. If you’re saying ‘Sure, you’re concerned, but **what about** other bad things that already happen?’ You are right, that’s also bad. It’s just not what the OP was addressing here.

Show threadDavid Delgado VendrellDec 14, 2024
@jimgar @ojala You are right that LLM has an impact most likely. But I generally don't see the point of leveraging a claim against technology using another one that has even more that the one is targeted. That was all.
Show threadDave in 🏴󠁧󠁢󠁳󠁣󠁴󠁿Dec 14, 2024

@daviddelven @jimgar @ojala How about when you look at how (a) search result quality has declined since Google and Bing started incorporating LLMs into their search engines but (b) traditional search algorithms are much cheaper to run?

The other thing is that LLMs and the accompanying AI hype are being actively promoted by billionaires and not challenged nearly as much as, say, something like the oil industry. It's new territory in the fight against exploitation.

Show threadElizabeth SudduthDec 15, 2024

@daviddelven @jimgar @ojala Buildings serve useful purposes not easily replaced by other existing technology. Tents, for instance, are just not as good at doing all the things buildings do.

The LLMs that are being pushed everywhere are easily replaced by existing technologies that work better. For instance, rather than using an LLM to write a research paper that cites journal articles that don't exist, my students could write their own papers and cite real articles.

Show threadDavid Delgado VendrellDec 15, 2024
@hydropsyche @jimgar @ojala Maybe one day we won't find those students willing to "write" their own papers and cite articles... I just prefer not becoming a negationist on this topic.
Show threadElizabeth SudduthDec 16, 2024
@daviddelven @jimgar @ojala I definitely know that the spicy autocomplete stochastic parrot sure as fuck can't do science, so I'm not sure what happens then. At that point, I guess science is just dead and we can all pack up and go home?
Show threadProf. Catherine FlickDec 14, 2024
@ojala yeah it’s hard because the students are going to use it anyway.
Show threadchild of baphometDec 14, 2024
@ojala Professors are dumb.
Show threadWilfried KlaebeDec 14, 2024
@ojala »we should teach students how to use LLMs (like ChatGPT, Gemini, Claude, etc) ethically and responsibly« means »not at all« to me.
Show threadLars Marowsky-Brée 😷Dec 14, 2024
@ojala I mean, we do educate people to survive with capitalism, so ...
Show threadRyek DarkenerDec 14, 2024

@ojala

These points are also correct for many other products, not only in the IT area. You can’t rewind history, only try to do better in the future.

Show threadFuture SprogDec 14, 2024

Students wouldn’t have enough expertise to tell the difference between some AI bullshit and actual course material and would then waste their time trying to decipher the bullshit.

Uni was always hard and time consuming enough that just the course work took all my time and didn’t leave any leftover for trying to work out if AI slop was legit or not.

@ojala

Show threadMike DickisonDec 14, 2024
@ojala For all the “yes, but” people claiming smart phones, IT itself, and even university buildings (!) are also bad, this new, additional tech is an order of magnitude more wasteful than those. And also unnecessary. There can be some really lazy thinking from smart people when LLMs are discussed.
Show threadPriscilla Haring-KuipersDec 15, 2024
@ojala the only winning move is not to play 🤖