The AI dooooooom thread

[DGuller]

All this talk about how AI stacks up against the best of humans IMO misses a much more subtle advantage:  accessibility.  This morning, while taking clean dishes out of my dishwasher, I noticed a part lying at the bottom.  I took a picture of it and asked ChatGPT what it was, without even revealing where I got it from.  It correctly identified the part, and the dishwasher it came out of, and explained to me what it was and how critical it was.

Did it do something that a human couldn't do?  No, of course not.  I could've gotten my answer in a few different ways, but none of them would've taken a minute and a single cell phone picture.  Just making intelligence more accessible, and not even better than a human intelligence, is already a huge win when multiplied over many such events.

[DGuller]

Once upon a time it made sense. Books were scarce. Remembering had value. Even without AI this hasn't been so for a while.
It would be nice to see a change in how things are done to encourage more application even at undergrad.

Life would be impossible without remembering.  We use basic math every day.  We use language every day.

I agree.  Maybe technology changed what we need to remember, but looking things up has just too much latency for some tasks.  For example, pilots have thick books with them on how to deal with emergencies, but some procedures are burned into memory because they need to be done immediately when the situation arises.  Also, if you're interested in research, you kind of have to keep many things in memory to connect the dots, knowing that you can look them up is not enough, although AI has changed that somewhat as well.

[garbon]

All this talk about how AI stacks up against the best of humans IMO misses a much more subtle advantage:  accessibility.  This morning, while taking clean dishes out of my dishwasher, I noticed a part lying at the bottom.  I took a picture of it and asked ChatGPT what it was, without even revealing where I got it from.  It correctly identified the part, and the dishwasher it came out of, and explained to me what it was and how critical it was.

Did it do something that a human couldn't do?  No, of course not.  I could've gotten my answer in a few different ways, but none of them would've taken a minute and a single cell phone picture.  Just making intelligence more accessible, and not even better than a human intelligence, is already a huge win when multiplied over many such events.

Yeah that sure is worth using up natural resources and increasing the cost of microchips  (and products using them) for marginally faster convenience.

[Admiral Yi]

I'm not so sure on academia being we're taught what we collectively know, but the contemporary take or interpretation of it. We are taught not the past but the present - which is built on those foundations.

The contemporary take on a matter is not a spontenous act of creation during the lecture; it is the result of *past* reflection and analysis and writing.  And thus searchable by LLMs.

[Sheilbh]

Sure - although I think there are spontaneous acts of creation/thought within seminars and tutorials (as a student, I imagine less so for the academic ). I agree in regards to lectures or set texts.

I don't disagree with that. I just disagree that academia is teaching "what we already collectively know".

[Josquius]

Once upon a time it made sense. Books were scarce. Remembering had value. Even without AI this hasn't been so for a while.
It would be nice to see a change in how things are done to encourage more application even at undergrad.

Life would be impossible without remembering.  We use basic math every day.  We use language every day.

Well obviously yes.
Do you really think the core concept of memory could be objected to?
What is not so necessary these days however is remembering the specifics of vast amounts of university level knowledge.
Learning how to apply the knowledge is far more key

[Admiral Yi]

Well obviously yes.
Do you really think the core concept of memory could be objected to?
What is not so necessary these days however is remembering the specifics of vast amounts of university level knowledge.
Learning how to apply the knowledge is far more key

We seem to be talking about different things.  I'm reacting to the video dude talking about AI "haunting" us. You seem to be defending the critical thinking component of education. 

[The Minsky Moment]

Sounds like all of academia.  We're taught  what we already collectively know, which was acquired in the past.  Libraries are full of books that were written in the past.

In academia, at least when it is working properly, books and articles are used as a touchpoint for discussion and further analysis.  The purpose is not to train people to just reproduce back what is written but to spark ideas, to build upon existing knowledge and push further.  That is very different from building a probability engine to predict the next set of characters based on a matrix of how those characters have been used in past publications.

[Admiral Yi]

In academia, at least when it is working properly, books and articles are used as a touchpoint for discussion and further analysis.  The purpose is not to train people to just reproduce back what is written but to spark ideas, to build upon existing knowledge and push further.  That is very different from building a probability engine to predict the next set of characters based on a matrix of how those characters have been used in past publications.

I think I see the source of the confusion.  When I said that's all of academia, I meant all of academia contains facts, wisdom, knowledge acquired  in the past.  Not that academia consists of nothing but facts.

[crazy canuck]

In academia, at least when it is working properly, books and articles are used as a touchpoint for discussion and further analysis.  The purpose is not to train people to just reproduce back what is written but to spark ideas, to build upon existing knowledge and push further.  That is very different from building a probability engine to predict the next set of characters based on a matrix of how those characters have been used in past publications.

I think I see the source of the confusion.  When I said that's all of academia, I meant all of academia contains facts, wisdom, knowledge acquired  in the past.  Not that academia consists of nothing but facts.

No, you don't see the problem

[Josquius]

Away from the higher discussions about AI for a moment....
Seen a lot of chat on social media lately on the Vichy hellscape's new fucked up development.
Increasingly basically any picture a woman posts of herself gets the grok command to put them in a micro bikini. Which is promptly generated.

Oh did I say woman?
No no. Children too of course.

Quite a cross-over with the general musk is a shithead thread but jeez that place really needs legal action.

[Zoupa]

"AI"

[HVC]

Away from the higher discussions about AI for a moment....
Seen a lot of chat on social media lately on the Vichy hellscape's new fucked up development.
Increasingly basically any picture a woman posts of herself gets the grok command to put them in a micro bikini. Which is promptly generated.

Oh did I say woman?
No no. Children too of course.

Quite a cross-over with the general musk is a shithead thread but jeez that place really needs legal action.

Looks like the found a way to correct this. People were using grok to generate images mocking Elon so X suspended the ability to generate images.

[crazy canuck]

Yet another person falls victim to relying on generative AI.  I believe this is the first reported case where someone relied on an AI tool to draft their materials in a law society disciplinary proceeding. Here are the material parts of the decision:

https://www.canlii.org/en/on/onlst/doc/2025/2025onlsth186/2025onlsth186.html

[9]          All the materials that the applicant filed were produced with the assistance of generative artificial intelligence. The tool the applicant used was "hallucinating".

[10]        Not a single proposition of law in any of the applicant's materials is supported by a reliable authority.

[11]        Once it became apparent to us that substantial parts of the applicant's materials made no sense, we directed the parties to appear at a case management conference. The direction included a 15-page chart of examples of the non-existent and misleading authorities in the applicant's materials. The parties were directed to address how they thought we should proceed in light of the applicant's materials.

[12]        The applicant wrote to the Tribunal on November 30, 2025, admitting that he had used generative artificial intelligence with the result that there were numerous errors in his motion materials. He admitted that he had failed to check the materials before submitting them. He apologized and undertook not to use generative artificial intelligence in preparing any further materials.

[Valmy]

I use AI just goofing around, mostly preparing RPG stuff for my games with my son, and it is rather shocking just how bad it is at keeping track of simple details and mixing things up that no human would ever mix up. It requires tons of handholding just in the silly things I use it for. I often think of that person wanting to generate an image of Jesus flipping over the tables of the money changers and for the AI to produce an image of gymnast Jesus doing backflips over tables. The money changers would have been impressed.

It can also be shockingly great at things. But fuck I would be super paranoid about using it in a professional or serious capacity.