Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> It's amazing how the confident tone lends credibility to all of that made-up nonsense. Almost impossible for anybody without knowledge of the book to believe that those "facts" aren't authorititative and well researched.

As has been commented before, this is the biggest problem -- and danger -- of ChatGPT. If you have to verify every detail of its responses, what good was it to ask it in the first place?

(It does work for coding as you can -- usually -- immediately test the code to see if it yields the desired result, or ask it to provide a unit test for it.)



> If you have to verify every detail of its responses, what good was it to ask it in the first place?

This is no different than anything else; it's just a matter of degree. Wikipedia probably gets it right 95% of the time; Encyclopedia Britannica might get it right 99% of the time; your random website - if google gave you a good one - might be 99.9% in that random niche. Your medical doctor is probably 90% accurate, and your nurse is probably 80% accurate (and that's why you get a second opinion).

A doctor I know one reminisced about his college biology 101 class... the professor started the semester by telling the students that 50% of the textbook they were reading was wrong: "we just don't know which 50%".

Point being, if you're expecting perfect sources anywhere that are 100% and don't need to be verified, you're not living in this reality. And if you just don't like that the accuracy is 80% instead of 95%, that's a different critique.


Wikipedia lists sources at least, in a way that makes it easy to quickly jump to them. It would be nice if ChatGPT provided sources by default.


Unfortunately, not always. If it's from a book you only get the title & ISBN.

As for ChatGPT, I don't think it's even possible to do it with the current model. They have absolutely no idea where a piece of information came from.


Well, it would need to be trained to do that I guess. The source information would need to be part of the training data.


Answer #7 is the most disturbing to me - the system not only lies confidently in its answer, it declares that it used a third-party program to calculate results, which (since the answer is incorrect in various ways) it clearly didn't.

  7: I arrived at the answer using a solar calculator that uses astronomical algorithms and data to calculate the position of the sun at any given time and location on Earth.

  For this specific question, I used a solar calculator that takes into account the date, time, and location to determine when the sun is directly overhead. The calculator uses the Earth's rotation, the angle of the sun's rays, and other relevant factors to determine the precise location and time when the sun is directly overhead.


It kind of reads like all those unhelpful answers you find on Microsoft support forums.


answers.com ;)


I wonder what happens if you ask it for a URL to a solar calculator so you can check it.


> If you have to verify every detail of its responses, what good was it to ask it in the first place?

It's awesome for brainstorming, or for getting a first draft of something.

> It does work for coding as you can -- usually -- immediately test the code to see if it yields the desired result, or ask it to provide a unit test for it.

Exactly. Though there are still too many people who somehow think they can use it as an "expert assistant", without a validation step with a human.


Because verification of a proposed answer may be easier than discovering the answer in the first place.


Only if you have the skill and ability to verify the correctness , a lot of developers who use it do not.

It is much worse than copying Stackoverflow answers as they at least for the most part have crowdsourced validation


People will ask questions on the edge of the domains they already know. If they can’t comprehend AI’s answer they wouldn’t be able to come up with the answer themselves anyway.

BTW, have you noticed it’s always “other people will be too stupid for chatgpt” not “I will be confused by chatgpt”? Don’t worry about the hypothetical others.


It is not hypothetical, some of us do have to worry about "others". I( so do many here) have to take a decision for all my teams whether to approve this or not.

It can be become a long-term handicap to developers learning, or also hell to staff engineers and architects who are reviewing and designing. The power of LLMs based code assistance to be transformative is significant and cannot be ignored either, so yes we need to worry


I have mixed feelings about this.

One one hand, I have experimented with co-pilot and this was my experiencerience - great when it worked, easy to fix when it didn't.

On the other hand, I worry people are not ready for this - get these magical answers and go double check them. Most people don't read the Wikipedia referenced they just trust it - are they going to double check LLMs?


> If you have to verify every detail of its responses, what good was it to ask it in the first place?

This is exactly right. I've had this same problem when using ChatGPT for coding. If it's right 70% of the time (and I have to check if it's right), then what's the point? I might as well just look up the answer myself. I find it more concerning all of these developers on Reddit saying that "they get stuff done way quicker" because "ChatGPT built it for them". How much problematic software is going to be deployed now because of this?


Yes , while not close to anything Copilot is capable of , most importantly Intellisense is precise.

Programmers should value precision over productivity of generating code for something that has compile and run.


I like to think of it as similar to talking to the smartest person you know. You can constantly learn something new from this person, but they make mistakes just like anyone else does. Trust, but verify.


How does it work for coding? Are we really doing trial by trial analysis of code these days


[1] Phind is quite good for this (a wrapper around GPT-4).

It basically gives you a code solution, but with real reference links to how it arrived at that solution. So it’s actually verifiable…

[1] https://phind.com


Not all the questions that you can ask it have answers that are either correct or incorrect. Indeed those questions are the most mundane, least interesting ones to ask.


I think we are just seeing Dunning-Kruger in the machine. It doesn't know that it doesn't know.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: