this post was submitted on 27 Oct 2023
301 points (96.6% liked)

Technology

59282 readers
4406 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related content.
  3. Be excellent to each another!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, to ask if your bot can be added please contact us.
  9. Check for duplicates before posting, duplicates may be removed

Approved Bots


founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
[–] [email protected] 99 points 1 year ago (6 children)

Can't wait for this AI bubble to fizzle. It's the blockchain insanity all over again.

[–] [email protected] 54 points 1 year ago (2 children)

Maybe, but gen AI produces actually useful, tractable results. That's already heaps more than crypto, which is just techno gambling

[–] [email protected] 34 points 1 year ago (4 children)

As long as LLM AI models are prone to hallucinating and there is no way to audit how they derive results (eg to verify accuracy), relying on them will have roadblocks/limitations. Once they solve this issue though, that will be a whole different story, I agree. As for other AIs such as image or video generation, I don't have enough experience to tell...

[–] [email protected] 19 points 1 year ago (2 children)

Hallucinations can be heavily reduced today by providing the LLM with grounding truth. People use naked LLMs as knowledge databases, which is prone to hallucinations indeed. However, provide them with verified data from the side and they are very, very good at keeping to the truth. I know, because we deploy these with clients to great avail.

Image, music, video models are making great strides and are already part of various pipelines, all the way up to the big boy tools like Photoshop (generative fill, for example).

The tech is being incorporated at a large scale by a lot of companies, from SME to megacorp. I don't see it going away any time soon, even if it doesn't improve from here on out (which it undoubtedly will).

[–] [email protected] 2 points 1 year ago (1 children)

The issue is that there are from time to time they still confidently hallucinate and there is no way to detect if they are right or not.

[–] [email protected] 1 points 1 year ago

Hire 1 person to verify AI output instead of a dozen to make the content. If that one editor misses something, who cares when we live in a post-truth society where the media lies on purpose.

[–] [email protected] -2 points 1 year ago (1 children)

How many countries start with the letter K in Africa?

[–] [email protected] 6 points 1 year ago* (last edited 1 year ago) (1 children)

GPT-4:

In Africa, there are three countries that start with the letter "K":

  1. Kenya
  2. Kingdom of Eswatini (although it's often referred to simply as Eswatini)
  3. Kiribati

However, it's worth noting that Kiribati is not in Africa; it's a Pacific island nation. So, only Kenya and the Kingdom of Eswatini in Africa start with the letter "K", but most people just refer to Eswatini without the "Kingdom" prefix. If you meant countries solely with the prominent "K" at the start, then it's just Kenya.

Anecdotal evidence is useless because it can be contradicted with anecdotal evidence.

[–] [email protected] 8 points 1 year ago (1 children)

Hallucinations aren't the only issue with LLMs, they also have a limited amount of context they can recall and that problem won't go away.

[–] [email protected] 1 points 1 year ago

that problem won’t go away

That problem is very much being worked on

[–] [email protected] 0 points 1 year ago

This mentality is the same as saying Wikipedia is not a source. I agree, there are better regulations that need to be implemented, but the speed at which these things are being churned out and tuned is mind blowing. It's like if Wikipedia started branching off into sub categories that each have their own specialty, which can be more easily moderated, and then folded back into the more general system.

[–] [email protected] 24 points 1 year ago (1 children)

It's not quite blockchain. It is incredibly useful in a broad range of applications, and has genuinely changed how millions of people work. Sure it's not the magic bullet wall street thinks it is, but my work has been improved immensely through the use of generative AI. Especially with uniquely challenging software problems and niche questions.

I think it'll be similar to VR. Extremely useful and interesting, but over-hyped and not going to penetrate our lives as much as most people think.

[–] [email protected] 9 points 1 year ago (1 children)

My mom never used VR, but she happily talks to GPT4. From that perspective I think mindshare in the broader population will be significantly higher than VR (even if it doesn't live up to the hype VC/Wallstreet machine).

[–] [email protected] 5 points 1 year ago (2 children)

If you had to buy an expensive headset to use chatGPT she wouldn't either.

[–] [email protected] 4 points 1 year ago (1 children)

If she had wheels she'd be a wagon.

[–] [email protected] 2 points 1 year ago

I'm just saying the accessibility of AI doesn't necessarily mean it has more utility. Just that it's more accessible.

[–] [email protected] 1 points 1 year ago (1 children)

Even if I would gift one to her she wouldn't use it. VR headset is peak nerd shit, as much as I love it. Having a dialog with an AI is much more approachable to the layman.

[–] [email protected] 1 points 1 year ago (1 children)

That's fair I guess. I don't have a VR headset or talk to AI so imo they're both pretty nerdy. I only talk to chatGPT every now and then to see if it can help me with code problems, and it almost always fails spectacularly unless I'm doing something really basic.

[–] [email protected] 1 points 1 year ago (1 children)

I work as a systems engineer and use it daily. I feel there is a particular way of using it where it really shines. Priming it with "you are an experienced senior python/rust/etc. developer who writes robust, idiomatic and maintainable code". Using GPT-4 (not 3.5) is paramount, and the Data Analysis mode on ChatGPT is also really useful, because GPT can actually run code to validate things.

Noone should force it of course, but I feel once you get intuition about what and how it does things well (and when it falls on its face) then it really flies.

[–] [email protected] 1 points 1 year ago

Yeah I should probably use gpt4. Just don't want to pay for another subscription haha.

[–] [email protected] 14 points 1 year ago (1 children)

I'm still waiting for the electricity bubble to fizzle.

[–] [email protected] 3 points 1 year ago

pretty sure it'll take nuclear war to pop that one

[–] [email protected] 0 points 1 year ago

Oh damn, this couldn't be more wrong.

[–] [email protected] -3 points 1 year ago

Only if you don't know what you're talking about.