Neal Stephenson Believes AI-Generated Creative Output Is 'Simply Not Interesting...
source link: https://slashdot.org/story/23/03/05/2223220/neal-stephenson-believes-ai-generated-creative-output-is-simply-not-interesting
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
Neal Stephenson Believes AI-Generated Creative Output Is 'Simply Not Interesting'
Follow Slashdot stories on Twitter
binspamdupenotthebestofftopicslownewsdaystalestupid freshfunnyinsightfulinterestingmaybe offtopicflamebaittrollredundantoverrated insightfulinterestinginformativefunnyunderrated descriptive typodupeerror
Sign up for the Slashdot newsletter! or check out the new Slashdot job board to browse remote jobs or jobs in your area.
Neal Stephenson "sees artificial intelligence in general, and ChatGPT in particular, as underwhelming," reports CoinDesk.
"I think it depends on how it's used," Stephenson told CoinDesk TV's "First Mover" on Friday. "What we've tended to see is that it's used in creative applications where I don't think it's at all interesting."
Stephenson said that with a painting or book, "what you're doing is having a kind of communion with the artist who made thousands of little micro decisions in the course of creating that work of art or writing that book." A decision that is generated by an algorithm, "that's simply not interesting," he said....
"Personally, I know a lot of writers who are putting a lot of effort into creating their own original works, and I'd rather support them and hear what they have to say than just look at the output of an algorithm," he said.
When asked if an AI could've written Snow Crash, Stephenson responded "Well, maybe one did."
But if that were the case, he added, a person would be reading only the output of an algorithm, "and if that's interesting to you, then fine."
-
›
Even a bad writer can figure out when they've written something completely awful and decide to change it or remove it.
AI isn't there yet.
However... what does Stephenson think he is other than an algorithm implemented by evolution in meat? There is nothing magical about being human, we're just (currently) far, far better than AI at intelligence. Which is to say, 'AI' has yet to have any of it... but there is no theoretical barrier based in the laws of nature to truly intelligent artificial minds, there is no reason to believe meat is inherently better than silicon... we simply haven't figured out how to do what evolution spent 4 billion years figuring out through massively parallel blind trial and error.
-
Even a bad writer can figure out when they've written something completely awful and decide to change it or remove it.
Are you sure? Have you read Seveneves by Neal Stephenson?
-
It was a dark and stormy night.
-
I unfolded my umbrella and spoke the magic words "Hey Siri, turn on flashlight."
-
As a follow up, I went to wordtune.com (no affiliation or connection other than I happened to see it mentioned) and put in my phrase. These are the suggestions I got back:
The night was dark and stormy.
A dark and stormy night greeted us.
Storms raged throughout the night.
There was a storm raging outside.
During the stormy night, it was dark and stormy.
Only the middle one is any better than the original, and certainly the last one needs work.
-
-
-
I died in that novel. If that is what it was. Even Enos Root couldn't save me.
-
I enjoyed it...
-
-
False.
-
you are not making a very convincing case...
-
-
While there is nothing magical, there is a lot be said in having actual shared subjective experiences. You and I both know what it feels like to stub your toe; to have the hairs on the back of your head rise up because something in the shadows moved; or to be gobsmacked by a sunset when you just happen to catch that golden moment. AI does not.
You could go to Manhattan and spend the day walking around and looking, and if you try hard enough are you'll observe something nobody, or at least very few people h
-
What do you think "AI" is besides massive pattern-recognition??
-
what do you think "HI" is besides massive pattern recognition?
-
-
Tell me you haven't watched the 2016 Ghostbusters reboot without telling me you haven't watched the 2016 Ghostbusters reboot.
-
Knowing something is awful is different from believing it won't be profitable.
-
-
Sadly, given the quantity of sheer dross that is uploaded daily to self publishing platforms, including Amazon Kindle, fan fiction sites etc, that is not self evident.
The vast majority of material released for consumption simply isnt of any value whatsoever, whether its written by an AI or a human author.
-
There's also no reason to believe that we'll ever be able to build silicon that matches the intellectual capacity of a human (or a mouse).
As for Stephenson, he's conflating the quality of the work with his empathy for the artist (which IS a valid part of appreciating a work of art). But I do agree the stories I've seen from ChatGTP are structurally sound and interesting from a technological standpoint, but they don't really work as good stories.
-
That's clearly nonsense, as the whole point of generative AI is to mimic human creativity.
-
Interesting FP, but I was looking for the response from ChatGPT. (Yes, someone created the account on Slashdot, but apparently hasn't yet said anything from that natural platform... It's an early 8-digit UID.)
I'm looking forward to your Funny responses, but not holding my breath. No jokes yet recognized for the story. Heck, the mods even rated that FP as insightful, which was a pretty generous stretch, if'n you ask me. (And the powers that bestow mod points never do.)
-
-
ChatGPT and its ilk are interesting and can yield useful answers. But it's also homogeneous to the extreme - given a certain of pattern of input, it shits out words to a certain pattern of output. It might not be obvious with the first response, but it certainly is with the second or third to a similar question.
-
But going with what Baron_Yam said above, are we really any different? We shit out some output in terms of action or words based on learned patterns. I think its amazingly like us. The main thing for me of value is getting non fiction answers for composite questions so quick. And for what Neal Stephenson is saying, the fiction can be interesting, just not in the same way as being transported to another persons imagination. The story patterns it shits out reveal archetypal patterns its learned across vast sw
-
It is an amalagator/conflator/collater/mosaic creator, a rearranger of things literally already imaged/written/said. It might be interesting, but we can't call any of it original.
-
I am already receiving essays from students using ChatGPT, and I grade them low. How do I know they are ChatGPT? First, the grammar is impeccable. Second, the style is turgid and completely predictable (hello pattern recognition!). Third, they say nothing beyond the thesis statement. Fourth, they never quote any authors I assign.
I don't bother telling my students I'm on to their tricks. I just tell them their essays suck.
-
-
-
It's too bad this one isn't titled "Neal Stephenson Believes AI-Generated Output Is 'Simply Not Interesting' and Everyone Else Believes NFTs about Snow Crash Are 'Simply A Scam".
Probably that's just too long.-
The only thing worse than NFTs and AI generated gobbledygook is AI generated NFTs.
-
-
You must be new here... no seriously, someone with your UID should know that Slashdot will often run consecutive stories with common threads. It's literally a normal practice across a wide range of topics and stories.
-
-
Wooo woo.. Someone, who shall remain anonymous but is a slashdot editor who has the words Editor and David consecutively as his name, has the hots for Neal Stephenson.
-
-
They do add to the discussion about AI tools and when/where/how they're actually useful or not.
ChatGPT is a useful tool, and it's not going away, or it might, but strong natural language processing tools aren't going anywhere.
Sooner we get all the people crying about jobs, or facts, or terminators to shut up and figure out what NLP actually does the better.
-
they do seem to add to the amount of clicks, which is the point i think. discussion is largely optional, and lately limited to max 2 levels of interaction, being 1 the most common. meaning: old men yelling at clouds all the way to click salvation.
anything goes, current trends: chatgpt, programming language popularity contests, chatgpt-fucking-up-ip-laws, crypto, china-basing, more crypto, russia-bashing, cyber-security-bullshit, more chatgpt, rust, and more russia-bashing because it's high season.
stephenson
-
-
Question: Neal Stephenson Says AI-Generated ChatGPT Is ‘Simply Not Interesting’. What do you think about that?
Answer: "As an AI language model myself, I don't have personal feelings or opinions. However, I can provide some context and perspective on Neal Stephenson's comment.
Neal Stephenson is a science fiction author and his opinion on AI-generated chatbots may be influenced by his experiences and expectations from science fiction. He might believe that AI language models like ChatGPT are not yet advanced enough to match the complexity and depth of human conversations or to replicate human emotions and creativity.
While AI language models like ChatGPT have made significant advancements in recent years, there are still limitations and challenges that need to be overcome. For instance, ChatGPT and other language models can sometimes generate irrelevant or nonsensical responses or perpetuate biases that exist in the training data.
However, AI language models like ChatGPT have shown tremendous potential in various applications, including natural language processing, chatbots, and content creation. These models can learn from large amounts of data and generate human-like responses, which can be useful in automating customer service or assisting with writing tasks.
In summary, Neal Stephenson's opinion about AI-generated chatbots being "simply not interesting" is subjective and might be influenced by his own expectations and experiences. While AI language models have their limitations, they have also shown remarkable potential in various applications and are continuously improving."
-
So many words to basically just say "Mr. Stephenson is entitled to his opinion." This is why AI still has a long way to go.
-
In more ways than that, a human response is "What do you mean not interesting?" The technology, the content it generates, the debate, what?
And then you have to be careful not to lead it with your questions like treeluvin did, "ai-generated chatgpt", because it ran with that and turned it into AI-generated chatbots, which is what exactly? So the whole response assumed Stephenson was talking about AI models and not generated content, (AI-generated-AI-model hallucination at that) but it sounded so right they
-
-
-
That's a good way of putting it. Who cares what some algorithm generated? Is it insightful? Is it prodigious? When I read a book, I trust that the author has a point they're trying to make, some kind of underlying thesis or argument. In the case of an algorithmically-generated word salad, that's completely missing. It could be useful for some purposes--such as the pulp trash you see on Kindle Unlimited--but you're ultimately just marveling at a particularly complex shape in the clouds. Nothing wrong with th
-
Can the AI write endings? Because that might still be an improvement.
-
Curently, they'll never pruduce anything relatively offfensive, so no killing, hurting, fucking, or even kissing passionately in whatever plot they can conjure up.
I'm not worried about AI rivalling what I write. Most people can't even match my twisted thoughts.
Recommend
About Joyk
Aggregate valuable and interesting links.
Joyk means Joy of geeK