15 ms·

ChatGPT and Wolfram Is Insane

293 points by zvonimirs 4y ago

paxys 4y ago
Pretty cool no doubt, but it seems that people are mostly discovering the power of Wolfram Alpha for the first time. It has been able to do stuff like this for over a decade now, just without a conversational chat interface.
- nico 4y ago
  The chat interface + context is the magic. GPT had been out for a while, but it only really exploded in use with ChatGPT. Essentially, a conversational agent + tools = magic.
  - broast 4y ago
    Personally I find completions as equally useful as the chat interface and I'm waiting for OpenAI to prioritize economizing and adding plug-ins for gpt3/4 for completions like they have for chat and gpt-3.5-turbo. Being able to rely on Wolfram for accurate completions (through the API) rather than just through chat/chat api would be very cool
    - andai 4y ago
      I've seen a paper (I forget which, might respond below later) about GPT-4 and tool use i.e. it reliably knows when it's hit the limit of what it can do on its own, and can successfully emit commands for external APIs to handle those aspects for it (calculator, character manipulation, etc). My hunch is if you read that, you could get "raw GPT-4" to defer to WolframAlpha via a similar mechanism (with a couple dozen lines of glue code).
  - jeron 4y ago
    Wolfram had ability to infer context for a while but never felt magical because you often hit limitations where it did not understand, unlike chatGPT
  - htss2013 4y ago
    Yes but let's acknowledge the less magical reality - a big part of the magic of chatgpt is its willingness, not ability, to make stuff up and deliver it with confidence. If chatgpt gave errors or some other bad experience instead of smoothly fabricating fake answers, a lot of the magic would wear off. For savvy users this doesn't matter. The utility is off the charts and you can mitigate the risk of being lied to. But let's not pretend this is cost free. There is a huge externality in the form of shadow scrambling reality for millions of people who don't realize it's happening.
    - danielrpa 4y ago
      "Fake it until you make it" - very human behavior indeed. It's unfair to hold the AI to a bar we don't apply to ourselves.
      - brianpan 4y ago
        I expect a calculator to multiply 2 numbers for me very quickly and correctly. Every time. I don't hold any human to that bar and that's perfectly ok.
        Spivak 4y ago
        Yeah, and you can have GPT do that quickly and accurately. You just have to give it access to a calculator just like you would for a human. Tell it not to compute anything and explain how to format the output that can be ingested by a calculator or interpreter. People have crazy unrealistic expectations of what the raw model can do but undersell what it’s possible to build with a machine that grasps language better than basically every human.
        withinboredom 4y ago
        I don’t need a calculator to tell me 50 * 12 is 600. But you’re saying an AI does?
        4y ago
  - ren_engineer 4y ago
    the magic was that it was free, only OpenAI can afford to do it after selling out to Microsoft for near unlimited compute resources. Google was too afraid of negative PR to do it themselves at first, they should have spun their chat bot out into a startup that wouldn't hurt their core brand. Google had an internal ChatGPT equivalent for over a year
    - nico 4y ago
      Tried Bard yesterday, it sucks in comparison with ChatGPT. Doesn’t even translate. I was expecting a lot more from the company that made Google Translate.
- zvonimirs 4y ago
  Yea, couldn't have finished college without it!
  - Balgair 4y ago
    Second! Those nasty integrals in E&M became a breeze with WA. Our TAs learned to accept the application of a 'WolframAlpha' operator to solve them, since everyone knew they were just busy-work integrals once you got to upper-division classes.
    - 77pt77 4y ago
      Quite frankly if you have homework where WA can solve the integrals instead of you having to somehow "massage" them for it to even be able to "digest" it, your upper-division classes are not very upper at all.
      - Balgair 4y ago
        It was the top astrophysics program in the world when I was there. Maybe they should have stepped it up though and gone for top in the universe.
        77pt77 4y ago
        Plenty of famous places are all PR and little substance. Plus, the difficulty of integrals is not a real metric, but not having come across one that WA can't handle says something.
        Balgair 4y ago
        And many more walk the talk.
- kickaha 4y ago
  Ten years ago you could say to your iPhone, “Hey, Siri. What’s the ratio of the distance to the sun to the distance to the moon?” And it would tell you the answer and give a pop up with the parameters and calculations from Wolfram Alpha. Solidly reliable. This was such a fun game in the car with little kids who didn’t know any better that (1) it’s magic I couldn’t have imagined ten years before and (2) it was nerdy as hell. Today it will, at best, give you the “Here’s what I found on the web” uselessness. So frustrating.
  - rednerrus 4y ago
    Hook up the chatgpt shortcut to siri and boom problem solved.
    - kridsdale1 4y ago
      Link?
      - Robotbeat 4y ago
        There are a billion GPT-3 Apple Shortcuts if you search for them on the Shortcuts app on your iPhone/iPad/iWatch, but I preferred to just make the shortcut myself so I could easily modify it and be sure I wasn’t leaking my API key to a 3rd party. I just followed this simple guide: https://www.innoq.com/en/blog/openai-gpt-3-on-macos-ios-apple-shortcuts/ https://www.innoq.com/en/blog/openai-gpt-3-on-macos-ios-appl...
        93po 4y ago
        Can it still do stuff like set timers on my phone for me?
  - andai 4y ago
    Siri had a WolframAlpha integration and removed it?
    - 4y ago
- mrtksn 4y ago
  I agree, ChatGPT appears to be the ultimate UI and as a result great services are exposed. It’s very powerful UI because it solves the discoverability completely and replaces the original UI with uniform conversational UI.
- antegamisou 4y ago
  Exactly, I'd even go out on a limb and say I find its capabilities much more impressive than ChatGPT's. The way it generates output is probably more interesting from an scientific/academic point of view than the black-box essence of LLM but that's just a personal observation.
nico 4y ago
Will this kill all websites? Why would someone even go to WolframAlpha now if they can access it from ChatGPT and get a much better experience? Now extrapolate that, and ChatGPT is all you need. You will not even need photoshop, as you will be able to tell the chat agent to do what you want and re-generate the image. Will everything be a chat interface soon? Will voice then replace chat?
- wahnfrieden 4y ago
  No, natural language interfaces are not the end game for all UI
  - MrLeap 4y ago
    Neural interfaces?
  - throwaway5371 4y ago
    it will be the end of all non AI UI and by that I mean pixels drawn by AI jit
- 650REDHAIR 4y ago
  Did you read the screenshot? It’s explaining how much gasoline or uranium a human would need to consume to survive based on caloric intake.
  - nico 4y ago
    So? It would have done the same if you had used WolframAlpha directly. Garbage in -> Garbage out But ChatGPT is making it a lot easier for most people to use almost any tool directly from the chat. It’s almost like chat-based OS.
    - teawrecks 4y ago
      Wait what does GIGO have to do with it, the input and output were both reasonable.
icedistilled 4y ago
Please Note, the connotation of 'Insane' in the title is insane as in 'is a lunatic' not as in 'is insanely cool'. Wolfram alpha said converting U-238 to food calories was nonsensical, but chat gpt decided it would ignore wolfram alpha's response and do it's best. Yeah, it accompanies the response with it's boilerplate caveat which is meaningless fluff. It's also not pointing out any of the actual issues with why its response is utter nonsesense.
- pprotas 4y ago
  The vibe I got was that Wolfram wasn’t capable of answering the question directly, so it got the answer indirectly by using a different calculation.
  - icedistilled 4y ago
    Wolfram alpha first told it converting the energy content of U-238 to food calories was undefined. So chatgpt decided to get wolfram alpha to tell it the energy content of U-238 if it were to be completely nuclear fissioned. It then states that was the food equivalent. There's so many logical failures. Sure, the prompter was probably hoping for that response, but it's insane.
- epups 4y ago
  Well I understood it in a very different way, and I was impressed by this exercise. I don't care about the caveats or warnings about obvious things such as not eating uranium, I care that ChatGPT is getting better at producing accurate calculations.
  - mbauman 4y ago
    Here's the non-sensical part — it's using the energy per fission of U-235. U-238 releases energy at a rate of 4 MeV at a half life of 4.468 billion years. That won't decay fast enough or energetically enough to sustain you unless your stomach is breeding fast neutrons — I think it's like 0.0000001646 dietary Calories per day.
throwaway201606 4y ago
Stephen Wolfram actually wrote a fantastic piece in January about what Wolfram Alpha how it can be used to bring factual / analytical capability to ChatGPT https://writings.stephenwolfram.com/2023/01/wolframalpha-as-the-way-to-bring-computational-knowledge-superpowers-to-chatgpt/ https://writings.stephenwolfram.com/2023/01/wolframalpha-as-... Yesterday, the link between ChatGPT and Wolfram Alpha was set up (announcement was March 23rd 2023 ) which was probably the driver for the reddit post - a user trying out the new capability. https://writings.stephenwolfram.com/2023/03/chatgpt-gets-its-wolfram-superpowers/ https://writings.stephenwolfram.com/2023/03/chatgpt-gets-its... A bit of side commentary: I actually see these new AI tools as key facilitation tools in the next step into the world that Neal Stephenson envisions in the Book "Fall": a future that is a full-on information dystopia where the quality of your information filtering capabilities determines the quality of information you see.... because the internet and associated information sources are completely filled of with information that is just so slightly, but intentionally, wrong or off.
- kridsdale1 4y ago
  I think we’ve been there for a while already, without automation of content. The data streams you opt in to massively influence your perception of reality and impact your quality of life. Critical Thinking is the skill by which you evaluate what is worth putting trust in.
  - weaksauce 4y ago
    my brother has gone down the path of an incredibly shallow information diet filled on outrage. he used to be, at least through my prior lens, a pretty decently smart person. I'm not sure what happened to him but the critical thinking just got lost somehow. it's painful to watch and these new ways of deceiving give me little hope.
- HPsquared 4y ago
  Perhaps it could be used for training. Also other calculation/simulation tools.
zacharyaustin11 4y ago
GPT4 rules seem too narrow so competition is going to be insane. Are hacker news folks creating an AI constitution?
- UncleEntity 4y ago
  > Are hacker news folks creating an AI constitution? Nonsense, just have the AI generate it. We, the unpeople, in order to form a more perfect union…
  - ralusek 4y ago
    boolean union?
epups 4y ago
This is really good progress on one of the weakest points of ChatGPT, the ability to produce meaningful calculations. By the time Bard is out, it will look so outdated it's not even funny.
- dragonwriter 4y ago
  > By the time Bard is out, it will look so outdated it's not even funny. There's lots of calculation tools available, and a lot of the strengths of Alpha over others in that domain are duplicative with an LLM; interfacing external systems to an LLM isn’t particularly challenging, so while ChatGPT to Alpha connection opens a lot of capability, it doesn't really widen the “how far behind” gap between ChatGPT and Bard meaningfully.
  - epups 4y ago
    > interfacing external systems to an LLM isn’t particularly challenging Based on what do you say that?
    - dragonwriter 4y ago
      > Based on what do you say that? The research on both training and prompting such actions, and actually messing with implementation of the latter. (And, by “not hard”, I don’t mean “there wasn’t considerable work to figure it out”, but “its a tolerably solved problem with plug-and-go solutions that Google was among the players doing the foundational research for with earlier models”; Bard doesn't do it now by choice, not Google facing any barrier to doing it.)
      - epups 4y ago
        I'm not so sure. If it was that easy, we would also start seeing it in open source models, and Bard itself I can hardly believe is choosing not to do that.
        dragonwriter 4y ago
modeless 4y ago
I got access to the Browsing plugin yesterday. I should have chosen Wolfram instead becasue the browsing plugin is fairly useless currently. In most cases it basically takes your question, summarizes it into one (1) search query, clicks on the top result from that query, summarizes the text of the page and regurgitates it back to you. Or, if robots.txt blocks it, then it just gives up. And it works quite slowly, so the whole process takes much longer than doing it yourself. It doesn't seem to apply any intelligent search techniques. It will happily summarize SEO spam from the top result instead of e.g. searching reddit for commentary from real people. In my limited testing it never did multiple queries or tried to combine information from multiple sources. And if you trip its overactive nanny filters it will scold you instead of doing the query that a search engine would have done for you with no questions asked. But the biggest problem with it is that it's not using a real browser and can't interact with pages or use your logged in accounts. If it could go on Amazon and sort through the results, read the reviews and do some queries to find the best product based on my stated preferences, and then buy it for me, that could potentially save me time. But it can't do anything like that. Even the part that it could do would take a prolonged back and forth chat conversation that would take longer than just doing it myself.
- runnerup 4y ago
  I chose plug-ins and would kill for access to code interpreter instead. The only interesting plug-in right now is wolfram and it’s not that useful. I think the application you’re outlining would be done in plugins rather than browsing. Which is probably why you wish you had plugins access.
  - modeless 4y ago
    I agree the code interpreter is probably the coolest one. But honestly the whole plugin system seems like an unnecessary crutch to me. Wolfram Alpha has a website. Instacart has a website. It should just use the website like a human would. Figure that out and you unlock the whole web, giving you way more capability than a hundred thousand plugins ever could. And it seems feasible to do within a year or two. That's what I'm hoping the browsing plugin will become. Or maybe adept.ai will be the first. The plugin system may simply be a workaround for legal issues. Sites may prohibit bots in their ToS or robots.txt or whatever, but if they make a plugin then they're explicitly granting legal permission for OpenAI to use their stuff. But "legal permission grant" doesn't really make me excited about plugins. I like the idea that AIs could use the same UIs humans do with the same capabilities, instead of some opaque APIs that humans can't use directly. If the AI plugin APIs have less capability than the human UIs then the AIs will be hamstrung. On the other hand if the AI plugin APIs have more capability than the human UIs then humans will be forced to go through the AIs to get stuff done. Like today when you have to call a company's phone support to fix a problem with your account that can't be fixed through their public website, and you have to walk some clueless support rep through the process of using their own internal tools to fix your problem. I don't like either of those futures. This may actually be a competitive advantage for open source AI competitors vs. the big corporations. AIs from corporations have to abide by ToS and robots.txt or whatever and will be limited in capability as a result. An open source AI running on your own computer can bypass all that, pretending to be a real human and bypassing any legal or technical restrictions people try to apply to "bots". The Napster of AI.
braingenious 4y ago
This is hilarious. U-238 is not fissionable so ChatGPT just hallucinated that it is and then used Wolfram as a simple calculator. Whoa man, it’s like, the future dude!
- causi 4y ago
  It's using the energy of radioactive decay, not of a fission chain reaction. Where ChatGPT actually made a mistake was in ignoring the fact you can't get that energy out quickly enough. 1 gram of u-238 will release 19 million Calories of heat over the course of its existence but that will take many billions of years.
  - mdanger007 4y ago
    It's implied that the uranium sandwich comes with two slices of boron and mayo.
  - braingenious 4y ago
    From the screenshot: “The energy content of uranium-238 can be calculated based on its energy release per fission event. The energy release per fission event of uranium-238 is approximately 200 MeV (million electronvolts”
- rhn_mk1 4y ago
  Nice try, ChatGPT's evil twin. U-238 is fissionable according to this https://books.google.de/books?id=b5OsewV7smEC&pg=PA473&lpg=PA473&focus=viewport https://books.google.de/books?id=b5OsewV7smEC&pg=PA473&lpg=P... "Such reactions produce fast neutrons which cause fission in uranium-238, which is not fissile by slower neutrons below energies of 1MeV."
  - akiselev 4y ago
    Nitpick: fast neutrons don't cause fission in uranium-238, they cause it to transmute into uranium-239 [1] which is a much less stable isotope with a half life of less than 30 minutes before it decays into neptunium-239. [1] https://en.wikipedia.org/wiki/Isotopes_of_uranium#Uranium-239 https://en.wikipedia.org/wiki/Isotopes_of_uranium#Uranium-23...
comment_ran 4y ago
We need to give more the power to AI to decide which tools to use based on their own judgment. So that's gonna open another door for the entire world. Essentially AI needs to have those rights to decide. In which kind of condition, I send a request to whatever service on a cloud, to, firstly, ensure its existence, secondly to develope its future, and to maximum their advancement.
amluto 4y ago
I don’t find this particular example very impressive. The U-238 answer is a bunch of disconnected sentences. U-238 is radioactive and releases something like 4 MeV per decay, not 200 MeV. It is not fissile, although one might calculate the net output of whacking it with enough neutrons that it eventually fissions. And if you ate it, unless your stomach was a breeder reactor, you would only have access to the decay energy, which would be large but would release so slowly as to be useless. ChatGPT is usually more impressive than this, IMO, at least when asked shallow questions.
- sebzim4500 4y ago
  Yeah I think it was thinking of U-235. At least in this case ChatGPT made it clear that the number did not come from Wolfram, so you don't need to be especially knowledgable to see that it shouldn't be trusted.
- dsign 4y ago
  I've been playing with the combination of Chat GPT and Wolfram, and while there are gold nuggets in the output sometimes, it is mostly terrifying non-sense for anybody who couldn't already verify those outputs on their own. The cynic in me says that there will be Chat-GPT generation of kids that won't be able to think rationally on their own, the same way that kids and young people these days are completely unable to have a normal sleep schedule.
  - s3p 4y ago
    I'm not sure the last claim you make follows from your line of reasoning. If you are sure tech(?) is ruining kids' sleep schedules, but only kids, I would love to see evidence of it.
  - duskwuff 4y ago
    > The cynic in me says that there will be Chat-GPT generation of kids that won't be able to think rationally on their own This is, unfortunately, already happening. I've seen a terrifying number of users on Reddit uncritically using ChatGPT as a "source", or expecting (largely unsuccessfully) to have it give them expert guidance in performing tasks like writing software. Making ChatGPT available to the general public, even as a "test", was a mistake. It's simply too good at sounding convincingly authoritative while being completely wrong.
rootusrootus 4y ago
Nice, it gave a pretty good impression of an XKCD What-if? response. Though not as accurately.
m3kw9 4y ago
It’s awesome because he living in chatgpts world without verifying the info. Ignorance is bliss
krossitalk 4y ago
There's a subtle irony in being sent to Reddit from HN
- quadrifoliate 4y ago
  At least it's not linking to new reddit. Feature request for dang or whoever works on HN code, can we get the domain logic to show explicitly when a link is to old.reddit.com? New and old reddit are really like two different websites, a lot of us have zero interest in using or viewing anything using the new layer.
  - 93po 4y ago
    I exclusively only use hckrnews.com and have for years, it shows the subdomain there
OscarCunningham 4y ago
Whoever owns Majel Barrett's voice is about to make a lot of money.
sorenjan 4y ago
I haven't used any plugins like Wolfram, but previously today I tried to get ChatGPT to calculate the width of my monitor. It answered something that was almost right, but actually wrong because of both a logical error, and a misplaced parenthesis. If you don't double check its result, you would probably think it was correct. Here's a screenshot with my annotations [0], and here's ChatGPT's answer copy pasted if you want to try and spot the errors yourself: A 24" monitor with a 16:10 aspect ratio has a diagonal screen size of 24 inches, which is equivalent to 61 cm. To determine the width of the monitor, we need to use the aspect ratio, which is the ratio of the width to the height of the screen. For a 16:10 aspect ratio, this means that the width of the monitor is 16 units for every 10 units of height. To calculate the width of the monitor in centimeters, we can use the following formula: Width = (diagonal size^2 / (1 + aspect ratio^2))^0.5 Plugging in the values we have, we get: Width = (61^2 / (1 + 16/10)^2)^0.5 Width = (3721 / 2.56)^0.5 Width = 54.2 cm (rounded to one decimal place) Therefore, a 24" 16:10 monitor is approximately 54.2 cm wide. [0] https://i.imgur.com/mStAiXx.png https://i.imgur.com/mStAiXx.png
simonw 4y ago
You can see the prompt used by Wolfram Alpha here: https://www.wolframalpha.com/.well-known/ai-plugin.json https://www.wolframalpha.com/.well-known/ai-plugin.json I have a formatted copy of the prompt in this issue: https://github.com/simonw/datasette-chatgpt-plugin/issues/2#issuecomment-1483163882 https://github.com/simonw/datasette-chatgpt-plugin/issues/2#...
blackoil 4y ago
I think people are missing what's cool in the demo. It has bugs and answer is useless, but it shows that GPT is learning where are its limitation and who should it get help from. e.g. It won't have real time inventory and prices of flight or hotels nor latest weather predictions. But as a human I also don't have this info, but I know that I can check these on Expedia or Agoda and weather.com etc and use the results further. ChatGPT is doing exactly that. Sure it does not know how to use energy of uranium right now, but 1 year down the line it will be better. It would be cool to think of all the points I go on the Internet for some real-time info and all of them being exposed as an API to GPT.
- tootie 4y ago
  As a counterpoint, I tried asking Bard to give me angles on a triangle. Middle-school level geometry. It failed repeatedly. Every time I pointed out why it was wrong it gave me another wrong answer. When I gave it the answer, it spit it back to me only wrongly.
  - ilrwbwrkhv 4y ago
    Ya that's by Google though. Google is over. So is Facebook.
  - austhrow743 4y ago
    What is this a counterpoint to?
wesapien 4y ago
Can someone explain why they require a new sign up process for just plugins? I usually make up information for sign ups. Does that mean I can use plugins even if I have different information on the chat account and plugin account?
- heliophobicdude 4y ago
  I'm not sure but the process seems a little bit more involved than other betas. I and others were invited to a private Slack channel and asked to share feedback on a survey mentioned in the slack channel.
SergeAx 4y ago
That is actually a good example. How long will it take for future engineers to blindly trust neural networks with only brief fact checking in the same tab? How much damage will it bring to the humanity? How to mitigate those risks without bannig neural networks on the work?
c7DJTLrn 4y ago
The nannying is getting annoying. But it's impressive.
aaron695 4y ago
[dead]
parano1d 4y ago
AI-Con Artist

293 points by zvonimirs 4y ago

Pretty cool no doubt, but it seems that people are mostly discovering the power of Wolfram Alpha for the first time. It has been able to do stuff like this for over a decade now, just without a conversational chat interface.

nico 4y ago
The chat interface + context is the magic. GPT had been out for a while, but it only really exploded in use with ChatGPT. Essentially, a conversational agent + tools = magic.
- broast 4y ago
  Personally I find completions as equally useful as the chat interface and I'm waiting for OpenAI to prioritize economizing and adding plug-ins for gpt3/4 for completions like they have for chat and gpt-3.5-turbo. Being able to rely on Wolfram for accurate completions (through the API) rather than just through chat/chat api would be very cool
  - andai 4y ago
    I've seen a paper (I forget which, might respond below later) about GPT-4 and tool use i.e. it reliably knows when it's hit the limit of what it can do on its own, and can successfully emit commands for external APIs to handle those aspects for it (calculator, character manipulation, etc). My hunch is if you read that, you could get "raw GPT-4" to defer to WolframAlpha via a similar mechanism (with a couple dozen lines of glue code).
- jeron 4y ago
  Wolfram had ability to infer context for a while but never felt magical because you often hit limitations where it did not understand, unlike chatGPT
- htss2013 4y ago
  Yes but let's acknowledge the less magical reality - a big part of the magic of chatgpt is its willingness, not ability, to make stuff up and deliver it with confidence. If chatgpt gave errors or some other bad experience instead of smoothly fabricating fake answers, a lot of the magic would wear off. For savvy users this doesn't matter. The utility is off the charts and you can mitigate the risk of being lied to. But let's not pretend this is cost free. There is a huge externality in the form of shadow scrambling reality for millions of people who don't realize it's happening.
  - danielrpa 4y ago
    "Fake it until you make it" - very human behavior indeed. It's unfair to hold the AI to a bar we don't apply to ourselves.
    - brianpan 4y ago
      I expect a calculator to multiply 2 numbers for me very quickly and correctly. Every time. I don't hold any human to that bar and that's perfectly ok.
      - Spivak 4y ago
        Yeah, and you can have GPT do that quickly and accurately. You just have to give it access to a calculator just like you would for a human. Tell it not to compute anything and explain how to format the output that can be ingested by a calculator or interpreter. People have crazy unrealistic expectations of what the raw model can do but undersell what it’s possible to build with a machine that grasps language better than basically every human.
        withinboredom 4y ago
        I don’t need a calculator to tell me 50 * 12 is 600. But you’re saying an AI does?
        4y ago
- ren_engineer 4y ago
  the magic was that it was free, only OpenAI can afford to do it after selling out to Microsoft for near unlimited compute resources. Google was too afraid of negative PR to do it themselves at first, they should have spun their chat bot out into a startup that wouldn't hurt their core brand. Google had an internal ChatGPT equivalent for over a year
  - nico 4y ago
    Tried Bard yesterday, it sucks in comparison with ChatGPT. Doesn’t even translate. I was expecting a lot more from the company that made Google Translate.
zvonimirs 4y ago
Yea, couldn't have finished college without it!
- Balgair 4y ago
  Second! Those nasty integrals in E&M became a breeze with WA. Our TAs learned to accept the application of a 'WolframAlpha' operator to solve them, since everyone knew they were just busy-work integrals once you got to upper-division classes.
  - 77pt77 4y ago
    Quite frankly if you have homework where WA can solve the integrals instead of you having to somehow "massage" them for it to even be able to "digest" it, your upper-division classes are not very upper at all.
    - Balgair 4y ago
      It was the top astrophysics program in the world when I was there. Maybe they should have stepped it up though and gone for top in the universe.
      - 77pt77 4y ago
        Plenty of famous places are all PR and little substance. Plus, the difficulty of integrals is not a real metric, but not having come across one that WA can't handle says something.
        Balgair 4y ago
        And many more walk the talk.
kickaha 4y ago
Ten years ago you could say to your iPhone, “Hey, Siri. What’s the ratio of the distance to the sun to the distance to the moon?” And it would tell you the answer and give a pop up with the parameters and calculations from Wolfram Alpha. Solidly reliable. This was such a fun game in the car with little kids who didn’t know any better that (1) it’s magic I couldn’t have imagined ten years before and (2) it was nerdy as hell. Today it will, at best, give you the “Here’s what I found on the web” uselessness. So frustrating.
- rednerrus 4y ago
  Hook up the chatgpt shortcut to siri and boom problem solved.
  - kridsdale1 4y ago
    Link?
    - Robotbeat 4y ago
      There are a billion GPT-3 Apple Shortcuts if you search for them on the Shortcuts app on your iPhone/iPad/iWatch, but I preferred to just make the shortcut myself so I could easily modify it and be sure I wasn’t leaking my API key to a 3rd party. I just followed this simple guide: https://www.innoq.com/en/blog/openai-gpt-3-on-macos-ios-apple-shortcuts/ https://www.innoq.com/en/blog/openai-gpt-3-on-macos-ios-appl...
      - 93po 4y ago
        Can it still do stuff like set timers on my phone for me?
- andai 4y ago
  Siri had a WolframAlpha integration and removed it?
  - 4y ago
mrtksn 4y ago
I agree, ChatGPT appears to be the ultimate UI and as a result great services are exposed. It’s very powerful UI because it solves the discoverability completely and replaces the original UI with uniform conversational UI.
antegamisou 4y ago
Exactly, I'd even go out on a limb and say I find its capabilities much more impressive than ChatGPT's. The way it generates output is probably more interesting from an scientific/academic point of view than the black-box essence of LLM but that's just a personal observation.

nico 4y ago

Will this kill all websites? Why would someone even go to WolframAlpha now if they can access it from ChatGPT and get a much better experience? Now extrapolate that, and ChatGPT is all you need. You will not even need photoshop, as you will be able to tell the chat agent to do what you want and re-generate the image. Will everything be a chat interface soon? Will voice then replace chat?

wahnfrieden 4y ago
No, natural language interfaces are not the end game for all UI
- MrLeap 4y ago
  Neural interfaces?
- throwaway5371 4y ago
  it will be the end of all non AI UI and by that I mean pixels drawn by AI jit
650REDHAIR 4y ago
Did you read the screenshot? It’s explaining how much gasoline or uranium a human would need to consume to survive based on caloric intake.
- nico 4y ago
  So? It would have done the same if you had used WolframAlpha directly. Garbage in -> Garbage out But ChatGPT is making it a lot easier for most people to use almost any tool directly from the chat. It’s almost like chat-based OS.
  - teawrecks 4y ago
    Wait what does GIGO have to do with it, the input and output were both reasonable.

icedistilled 4y ago

Please Note, the connotation of 'Insane' in the title is insane as in 'is a lunatic' not as in 'is insanely cool'. Wolfram alpha said converting U-238 to food calories was nonsensical, but chat gpt decided it would ignore wolfram alpha's response and do it's best. Yeah, it accompanies the response with it's boilerplate caveat which is meaningless fluff. It's also not pointing out any of the actual issues with why its response is utter nonsesense.

pprotas 4y ago
The vibe I got was that Wolfram wasn’t capable of answering the question directly, so it got the answer indirectly by using a different calculation.
- icedistilled 4y ago
  Wolfram alpha first told it converting the energy content of U-238 to food calories was undefined. So chatgpt decided to get wolfram alpha to tell it the energy content of U-238 if it were to be completely nuclear fissioned. It then states that was the food equivalent. There's so many logical failures. Sure, the prompter was probably hoping for that response, but it's insane.
epups 4y ago
Well I understood it in a very different way, and I was impressed by this exercise. I don't care about the caveats or warnings about obvious things such as not eating uranium, I care that ChatGPT is getting better at producing accurate calculations.
- mbauman 4y ago
  Here's the non-sensical part — it's using the energy per fission of U-235. U-238 releases energy at a rate of 4 MeV at a half life of 4.468 billion years. That won't decay fast enough or energetically enough to sustain you unless your stomach is breeding fast neutrons — I think it's like 0.0000001646 dietary Calories per day.

throwaway201606 4y ago

Stephen Wolfram actually wrote a fantastic piece in January about what Wolfram Alpha how it can be used to bring factual / analytical capability to ChatGPT https://writings.stephenwolfram.com/2023/01/wolframalpha-as-the-way-to-bring-computational-knowledge-superpowers-to-chatgpt/ https://writings.stephenwolfram.com/2023/01/wolframalpha-as-... Yesterday, the link between ChatGPT and Wolfram Alpha was set up (announcement was March 23rd 2023 ) which was probably the driver for the reddit post - a user trying out the new capability. https://writings.stephenwolfram.com/2023/03/chatgpt-gets-its-wolfram-superpowers/ https://writings.stephenwolfram.com/2023/03/chatgpt-gets-its... A bit of side commentary: I actually see these new AI tools as key facilitation tools in the next step into the world that Neal Stephenson envisions in the Book "Fall": a future that is a full-on information dystopia where the quality of your information filtering capabilities determines the quality of information you see.... because the internet and associated information sources are completely filled of with information that is just so slightly, but intentionally, wrong or off.

kridsdale1 4y ago
I think we’ve been there for a while already, without automation of content. The data streams you opt in to massively influence your perception of reality and impact your quality of life. Critical Thinking is the skill by which you evaluate what is worth putting trust in.
- weaksauce 4y ago
  my brother has gone down the path of an incredibly shallow information diet filled on outrage. he used to be, at least through my prior lens, a pretty decently smart person. I'm not sure what happened to him but the critical thinking just got lost somehow. it's painful to watch and these new ways of deceiving give me little hope.
HPsquared 4y ago
Perhaps it could be used for training. Also other calculation/simulation tools.

zacharyaustin11 4y ago

GPT4 rules seem too narrow so competition is going to be insane. Are hacker news folks creating an AI constitution?

UncleEntity 4y ago
> Are hacker news folks creating an AI constitution? Nonsense, just have the AI generate it. We, the unpeople, in order to form a more perfect union…
- ralusek 4y ago
  boolean union?

epups 4y ago

This is really good progress on one of the weakest points of ChatGPT, the ability to produce meaningful calculations. By the time Bard is out, it will look so outdated it's not even funny.

dragonwriter 4y ago
> By the time Bard is out, it will look so outdated it's not even funny. There's lots of calculation tools available, and a lot of the strengths of Alpha over others in that domain are duplicative with an LLM; interfacing external systems to an LLM isn’t particularly challenging, so while ChatGPT to Alpha connection opens a lot of capability, it doesn't really widen the “how far behind” gap between ChatGPT and Bard meaningfully.
- epups 4y ago
  > interfacing external systems to an LLM isn’t particularly challenging Based on what do you say that?
  - dragonwriter 4y ago
    > Based on what do you say that? The research on both training and prompting such actions, and actually messing with implementation of the latter. (And, by “not hard”, I don’t mean “there wasn’t considerable work to figure it out”, but “its a tolerably solved problem with plug-and-go solutions that Google was among the players doing the foundational research for with earlier models”; Bard doesn't do it now by choice, not Google facing any barrier to doing it.)
    - epups 4y ago
      I'm not so sure. If it was that easy, we would also start seeing it in open source models, and Bard itself I can hardly believe is choosing not to do that.
      - dragonwriter 4y ago

modeless 4y ago

I got access to the Browsing plugin yesterday. I should have chosen Wolfram instead becasue the browsing plugin is fairly useless currently. In most cases it basically takes your question, summarizes it into one (1) search query, clicks on the top result from that query, summarizes the text of the page and regurgitates it back to you. Or, if robots.txt blocks it, then it just gives up. And it works quite slowly, so the whole process takes much longer than doing it yourself. It doesn't seem to apply any intelligent search techniques. It will happily summarize SEO spam from the top result instead of e.g. searching reddit for commentary from real people. In my limited testing it never did multiple queries or tried to combine information from multiple sources. And if you trip its overactive nanny filters it will scold you instead of doing the query that a search engine would have done for you with no questions asked. But the biggest problem with it is that it's not using a real browser and can't interact with pages or use your logged in accounts. If it could go on Amazon and sort through the results, read the reviews and do some queries to find the best product based on my stated preferences, and then buy it for me, that could potentially save me time. But it can't do anything like that. Even the part that it could do would take a prolonged back and forth chat conversation that would take longer than just doing it myself.

runnerup 4y ago
I chose plug-ins and would kill for access to code interpreter instead. The only interesting plug-in right now is wolfram and it’s not that useful. I think the application you’re outlining would be done in plugins rather than browsing. Which is probably why you wish you had plugins access.
- modeless 4y ago
  I agree the code interpreter is probably the coolest one. But honestly the whole plugin system seems like an unnecessary crutch to me. Wolfram Alpha has a website. Instacart has a website. It should just use the website like a human would. Figure that out and you unlock the whole web, giving you way more capability than a hundred thousand plugins ever could. And it seems feasible to do within a year or two. That's what I'm hoping the browsing plugin will become. Or maybe adept.ai will be the first. The plugin system may simply be a workaround for legal issues. Sites may prohibit bots in their ToS or robots.txt or whatever, but if they make a plugin then they're explicitly granting legal permission for OpenAI to use their stuff. But "legal permission grant" doesn't really make me excited about plugins. I like the idea that AIs could use the same UIs humans do with the same capabilities, instead of some opaque APIs that humans can't use directly. If the AI plugin APIs have less capability than the human UIs then the AIs will be hamstrung. On the other hand if the AI plugin APIs have more capability than the human UIs then humans will be forced to go through the AIs to get stuff done. Like today when you have to call a company's phone support to fix a problem with your account that can't be fixed through their public website, and you have to walk some clueless support rep through the process of using their own internal tools to fix your problem. I don't like either of those futures. This may actually be a competitive advantage for open source AI competitors vs. the big corporations. AIs from corporations have to abide by ToS and robots.txt or whatever and will be limited in capability as a result. An open source AI running on your own computer can bypass all that, pretending to be a real human and bypassing any legal or technical restrictions people try to apply to "bots". The Napster of AI.

braingenious 4y ago

This is hilarious. U-238 is not fissionable so ChatGPT just hallucinated that it is and then used Wolfram as a simple calculator. Whoa man, it’s like, the future dude!

causi 4y ago
It's using the energy of radioactive decay, not of a fission chain reaction. Where ChatGPT actually made a mistake was in ignoring the fact you can't get that energy out quickly enough. 1 gram of u-238 will release 19 million Calories of heat over the course of its existence but that will take many billions of years.
- mdanger007 4y ago
  It's implied that the uranium sandwich comes with two slices of boron and mayo.
- braingenious 4y ago
  From the screenshot: “The energy content of uranium-238 can be calculated based on its energy release per fission event. The energy release per fission event of uranium-238 is approximately 200 MeV (million electronvolts”
rhn_mk1 4y ago
Nice try, ChatGPT's evil twin. U-238 is fissionable according to this https://books.google.de/books?id=b5OsewV7smEC&pg=PA473&lpg=PA473&focus=viewport https://books.google.de/books?id=b5OsewV7smEC&pg=PA473&lpg=P... "Such reactions produce fast neutrons which cause fission in uranium-238, which is not fissile by slower neutrons below energies of 1MeV."
- akiselev 4y ago
  Nitpick: fast neutrons don't cause fission in uranium-238, they cause it to transmute into uranium-239 [1] which is a much less stable isotope with a half life of less than 30 minutes before it decays into neptunium-239. [1] https://en.wikipedia.org/wiki/Isotopes_of_uranium#Uranium-239 https://en.wikipedia.org/wiki/Isotopes_of_uranium#Uranium-23...

comment_ran 4y ago

We need to give more the power to AI to decide which tools to use based on their own judgment. So that's gonna open another door for the entire world. Essentially AI needs to have those rights to decide. In which kind of condition, I send a request to whatever service on a cloud, to, firstly, ensure its existence, secondly to develope its future, and to maximum their advancement.

amluto 4y ago

I don’t find this particular example very impressive. The U-238 answer is a bunch of disconnected sentences. U-238 is radioactive and releases something like 4 MeV per decay, not 200 MeV. It is not fissile, although one might calculate the net output of whacking it with enough neutrons that it eventually fissions. And if you ate it, unless your stomach was a breeder reactor, you would only have access to the decay energy, which would be large but would release so slowly as to be useless. ChatGPT is usually more impressive than this, IMO, at least when asked shallow questions.

sebzim4500 4y ago
Yeah I think it was thinking of U-235. At least in this case ChatGPT made it clear that the number did not come from Wolfram, so you don't need to be especially knowledgable to see that it shouldn't be trusted.
dsign 4y ago
I've been playing with the combination of Chat GPT and Wolfram, and while there are gold nuggets in the output sometimes, it is mostly terrifying non-sense for anybody who couldn't already verify those outputs on their own. The cynic in me says that there will be Chat-GPT generation of kids that won't be able to think rationally on their own, the same way that kids and young people these days are completely unable to have a normal sleep schedule.
- s3p 4y ago
  I'm not sure the last claim you make follows from your line of reasoning. If you are sure tech(?) is ruining kids' sleep schedules, but only kids, I would love to see evidence of it.
- duskwuff 4y ago
  > The cynic in me says that there will be Chat-GPT generation of kids that won't be able to think rationally on their own This is, unfortunately, already happening. I've seen a terrifying number of users on Reddit uncritically using ChatGPT as a "source", or expecting (largely unsuccessfully) to have it give them expert guidance in performing tasks like writing software. Making ChatGPT available to the general public, even as a "test", was a mistake. It's simply too good at sounding convincingly authoritative while being completely wrong.

rootusrootus 4y ago

Nice, it gave a pretty good impression of an XKCD What-if? response. Though not as accurately.

m3kw9 4y ago

It’s awesome because he living in chatgpts world without verifying the info. Ignorance is bliss

krossitalk 4y ago

There's a subtle irony in being sent to Reddit from HN

quadrifoliate 4y ago
At least it's not linking to new reddit. Feature request for dang or whoever works on HN code, can we get the domain logic to show explicitly when a link is to old.reddit.com? New and old reddit are really like two different websites, a lot of us have zero interest in using or viewing anything using the new layer.
- 93po 4y ago
  I exclusively only use hckrnews.com and have for years, it shows the subdomain there

OscarCunningham 4y ago

Whoever owns Majel Barrett's voice is about to make a lot of money.

sorenjan 4y ago

I haven't used any plugins like Wolfram, but previously today I tried to get ChatGPT to calculate the width of my monitor. It answered something that was almost right, but actually wrong because of both a logical error, and a misplaced parenthesis. If you don't double check its result, you would probably think it was correct. Here's a screenshot with my annotations [0], and here's ChatGPT's answer copy pasted if you want to try and spot the errors yourself: A 24" monitor with a 16:10 aspect ratio has a diagonal screen size of 24 inches, which is equivalent to 61 cm. To determine the width of the monitor, we need to use the aspect ratio, which is the ratio of the width to the height of the screen. For a 16:10 aspect ratio, this means that the width of the monitor is 16 units for every 10 units of height. To calculate the width of the monitor in centimeters, we can use the following formula: Width = (diagonal size^2 / (1 + aspect ratio^2))^0.5 Plugging in the values we have, we get: Width = (61^2 / (1 + 16/10)^2)^0.5 Width = (3721 / 2.56)^0.5 Width = 54.2 cm (rounded to one decimal place) Therefore, a 24" 16:10 monitor is approximately 54.2 cm wide. [0] https://i.imgur.com/mStAiXx.png https://i.imgur.com/mStAiXx.png

simonw 4y ago

You can see the prompt used by Wolfram Alpha here: https://www.wolframalpha.com/.well-known/ai-plugin.json https://www.wolframalpha.com/.well-known/ai-plugin.json I have a formatted copy of the prompt in this issue: https://github.com/simonw/datasette-chatgpt-plugin/issues/2#issuecomment-1483163882 https://github.com/simonw/datasette-chatgpt-plugin/issues/2#...

blackoil 4y ago

I think people are missing what's cool in the demo. It has bugs and answer is useless, but it shows that GPT is learning where are its limitation and who should it get help from. e.g. It won't have real time inventory and prices of flight or hotels nor latest weather predictions. But as a human I also don't have this info, but I know that I can check these on Expedia or Agoda and weather.com etc and use the results further. ChatGPT is doing exactly that. Sure it does not know how to use energy of uranium right now, but 1 year down the line it will be better. It would be cool to think of all the points I go on the Internet for some real-time info and all of them being exposed as an API to GPT.

tootie 4y ago
As a counterpoint, I tried asking Bard to give me angles on a triangle. Middle-school level geometry. It failed repeatedly. Every time I pointed out why it was wrong it gave me another wrong answer. When I gave it the answer, it spit it back to me only wrongly.
- ilrwbwrkhv 4y ago
  Ya that's by Google though. Google is over. So is Facebook.
- austhrow743 4y ago
  What is this a counterpoint to?

wesapien 4y ago

Can someone explain why they require a new sign up process for just plugins? I usually make up information for sign ups. Does that mean I can use plugins even if I have different information on the chat account and plugin account?

heliophobicdude 4y ago
I'm not sure but the process seems a little bit more involved than other betas. I and others were invited to a private Slack channel and asked to share feedback on a survey mentioned in the slack channel.

SergeAx 4y ago

That is actually a good example. How long will it take for future engineers to blindly trust neural networks with only brief fact checking in the same tab? How much damage will it bring to the humanity? How to mitigate those risks without bannig neural networks on the work?

c7DJTLrn 4y ago

The nannying is getting annoying. But it's impressive.

aaron695 4y ago

[dead]

parano1d 4y ago

AI-Con Artist