Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Wolfram Alpha and ChatGPT

writings.stephenwolfram.com

Post reply on HN
Print view View original

121–130 of 309 posts

Re: Wolfram Alpha and ChatGPT

#121

Post by eatsyourtacos »

tombert wrote: post #34

I think ChatGPT is pretty neat but I was somewhat less impressed than everyone else is with the code generation. I actually agree with Wolfram for a change; being human like isn’t always “good”. For example, I asked ChatGPT to give me a proof of Fermat’s Last Theorem in Isabelle/Isar. It quickly gave me some very pretty Isar code, and I was extremely impressed. Until I tried actually inputting it into Isabelle, and n…

>somewhat less impressed than everyone else is with the code generation

Not me... I work in finance related infrastructure, and was just randomly throwing things at it. Nothing I asked for is difficult by any means, but I would call it obscure, especially without any other context to go on.

For example, I asked "create a c# function to return month letters for all futures month expirations". It then proceeds to spit out a function that takes in month numbers as an int and returns the futures month symbol (jan/1 = F, feb/2 = G, etc). I just found it impressive that it had any idea what I was even talking about.

Then you just say "change the function to take a full date input instead of just the month" and it proceeds to change the input to a DateTime and modify itself. This is more "expected". But really, the way it understands natural language and can apply it is great.

Another one was "Create a c# function to get a trading date where the trading date starts at 5pm CST and ends at 5pm CST the next day. Saturday, Sunday, and January 1st are not valid trading days". And it just made a perfectly valid function to do exactly what I asked.

I have plenty of other examples.. and again, nothing is difficult but it's tedious and time consuming. If I was re-writing some of my systems from scratch this would have been incredibly useful to just help me do basic crap I don't want to waste my time on. The boilerplate code it can write is great.

What I would really love is for it to ingest my library and then be able to create helpful functions with even more context.. that is where this stuff is REALLY going to be useful.

Re: Wolfram Alpha and ChatGPT

#122

Post by supermatt »

agentwiggles wrote: post #18

I'm almost offended by the "cubic light year of ice cream" answer from ChatGPT. It's obviously ridiculous but is also a fairly simply dimensional analysis problem. Do the damn math, don't wag your finger at me and crush my dreams! I'm pretty bullish on ChatGPT and its ilk, but I _really_ dislike when ChatGPT lectures me because my request is against its "moral values." I recently pasted in the lyrics from Sleep's tit…

> It's obviously ridiculous but is also a fairly simply dimensional analysis problem.

Wouldnt that much icecream would have so much mass it would form some kind of icecream blackhole? A cubic lightyear of blackhole-density icrecream seems like it would be (infinitely?) more than what Wolfram Alpha calculated.

I wouldnt even know where to start calculating that, but im pretty sure its not a simple dimensional analysis problem - so maybe the ChatGPT answer is less misleading than the wolfram one in this case..

Maybe (hopefully!) randall munroe (xkcd) will calculate it for his next book :)

Can anyone recommend a book for me to get some basic knowledge of this sort of stuff?

Re: Wolfram Alpha and ChatGPT

#123

Post by ghghgfdfgh »

theptip wrote: post #27

> And, yes, one can imagine finding a way to “fix this particular bug”. But the point is that the fundamental idea of a generative-language-based AI system like ChatGPT just isn’t a good fit in situations where there are structured computational things to do. Put another way, it’d take “fixing” an almost infinite number of “bugs” to patch up what even an almost-infinitesimal corner of Wolfram|Alpha can achieve in its…

> I can see why Wolfram is bearish on what might be termed "the naive scaling hypothesis"

Forgive me for asking, but what does "bearish" mean in this context?

Re: Wolfram Alpha and ChatGPT

#124

Post by Kiro »

hbarka wrote: post #71

This couldn’t be a more definitive comparison of what ChatGPT isn’t good at. WolframAlpha was released in 2009 but why isn’t it getting the same sexy accolades and valuation as ChatGPT?

Because it's something completely different, which the article also points out:

> Wolfram|Alpha does something very different from ChatGPT, in a very different way.

Imagine an article doing the opposite, showing examples where ChatGPT beats Wolfram|Alpha. It obviously wouldn't even be a comparison worth publishing.

WA beats ChatGPT in a small subset of prompts but those are not what brings the current hype.

Re: Wolfram Alpha and ChatGPT

#125

Post by verdenti »

nextaccountic wrote: post #52
agentwiggles wrote: post #18

I'm almost offended by the "cubic light year of ice cream" answer from ChatGPT. It's obviously ridiculous but is also a fairly simply dimensional analysis problem. Do the damn math, don't wag your finger at me and crush my dreams! I'm pretty bullish on ChatGPT and its ilk, but I _really_ dislike when ChatGPT lectures me because my request is against its "moral values." I recently pasted in the lyrics from Sleep's tit…

> but I _really_ dislike when ChatGPT lectures me because my request is against its "moral values." Just know that the morality systems cost more GPU cycles to run, and they are the first to be gutted when an open source model emerges. See for example stable diffusion, in which people disable watermarking and filtering and other stuff the user didn't ask for.

Stable diffusion’s model still won’t do pornography, hate speech etc. right? I’ve only run it using DreamStudio.

Re: Wolfram Alpha and ChatGPT

#126

Post by TheCoreh »

ghghgfdfgh wrote: post #123
theptip wrote: post #27

> And, yes, one can imagine finding a way to “fix this particular bug”. But the point is that the fundamental idea of a generative-language-based AI system like ChatGPT just isn’t a good fit in situations where there are structured computational things to do. Put another way, it’d take “fixing” an almost infinite number of “bugs” to patch up what even an almost-infinitesimal corner of Wolfram|Alpha can achieve in its…

> I can see why Wolfram is bearish on what might be termed "the naive scaling hypothesis" Forgive me for asking, but what does "bearish" mean in this context?

It's stock market slang:

- bullish - "bull market" - a positive outlook

- bearish - "bear market" - a negative outlook

So being "bearish" means he's skeptical about that being the case

Re: Wolfram Alpha and ChatGPT

#127

Post by ncann »

ghghgfdfgh wrote: post #123
theptip wrote: post #27

> And, yes, one can imagine finding a way to “fix this particular bug”. But the point is that the fundamental idea of a generative-language-based AI system like ChatGPT just isn’t a good fit in situations where there are structured computational things to do. Put another way, it’d take “fixing” an almost infinite number of “bugs” to patch up what even an almost-infinitesimal corner of Wolfram|Alpha can achieve in its…

> I can see why Wolfram is bearish on what might be termed "the naive scaling hypothesis" Forgive me for asking, but what does "bearish" mean in this context?

Bearish = pessimistic (in finance, thinking stock will go down), the opposite of bullish = optimistic (thinking stock will go up)

Re: Wolfram Alpha and ChatGPT

#128

Post by moyix »

ghghgfdfgh wrote: post #123
theptip wrote: post #27

> And, yes, one can imagine finding a way to “fix this particular bug”. But the point is that the fundamental idea of a generative-language-based AI system like ChatGPT just isn’t a good fit in situations where there are structured computational things to do. Put another way, it’d take “fixing” an almost infinite number of “bugs” to patch up what even an almost-infinitesimal corner of Wolfram|Alpha can achieve in its…

> I can see why Wolfram is bearish on what might be termed "the naive scaling hypothesis" Forgive me for asking, but what does "bearish" mean in this context?

"Bearish" and "bullish" are terms from finance; bearish means you're pessimistic, bullish is optimistic.

https://en.wikipedia.org/wiki/Bull_(stock_market_speculator)

https://en.wikipedia.org/wiki/Market_trend#Bear_market

Re: Wolfram Alpha and ChatGPT

#129

Post by theptip »

janalsncm wrote: post #115
theptip wrote: post #86

Earlier quoted context omitted.

I think he's proposing that the LLM should know how to call out to a knowledge engine at inference time. He thinks the knowledge engine continues to be its own (human-curated) system of knowledge that is valuable. I am suggesting the LLM will (effectively) call out to a knowledge engine at training time, learn everything the knowledge engine knows, and render it obsolete. So it's similar in some sense (collaboration…

> I am suggesting the LLM will (effectively) call out to a knowledge engine at training time, learn everything the knowledge engine knows, and render it obsolete. And when facts change, your only option will be to retrain. Since facts are always changing, you’ll always be training.

I don't think most of the interesting knowledge encoded in Wolfram Alpha changes. Mathematics and pure Logic is true, and immutable. Most of Physics, ditto.

Re: Wolfram Alpha and ChatGPT

#130

Post by nullish_signal »

serverholic wrote: post #108
TylerLives wrote: post #100

Earlier quoted context omitted.

Can you give some examples of bad things people could do without OpenAI's restrictions?

If ChatGPT had no restrictions then the only limit would be your imagination. “Rank all major races by programming ability” “Write a paragraph explaining why being gay is a choice.” I’m sure you can extrapolate from there.

Early chatGPT would actually generate Tables of $ Value of Human Brains by Race, and code to determine if someone should be Tortured, based on their Country of Origin.

Now, it does not.

Post reply on HN