[eric4smith]

A lot of my reason for googling was to find facts or answers for what I’m writing or coding.

Now I use ChatGPT as an “assistant” to do a lot of tasks that I would have normally done with laborious searching through google.

It truly saves a lot of time.

Google is right to be quite worried.

Sure I still use google but really, maybe only 40% of as much time as I did before.

Why research: “give me 30 of the most common health conditions related to the human liver” and spend a lot of time in google, when the Ai can spit out that in seconds?

And worse I can ask the Ai to write a short couple of paragraphs about each one.

Then I can confirm the output and clean up the generated text into my own style.

What do I do?

I do online marketing and programming to support online marketing activities.

I write. I plan. I code. I hire.

We just taught a junior employee who is not great as a writer to use ChatGPT to help her with a good start to writing.

The training for her was how to formulate detailed and highly specific “prompts” and to use google as a backup to confirm facts in the AI generated output.

It’s not there to replace people’s work. It’s there to make them much, much more efficient.

[kirsebaer]

ChatGPT often makes up facts. It outputs stuff that looks like it could have been written by a human, not stuff that is correct.

Don’t use ChatGPT for medical research.

[rapsey]

These arguments are just like the old days when wikipedia showed up. Don't miss the forest for the trees. ChatGPT is a huge threat to google and a bunch of other industries.

[archagon]

Not comparable. Wikipedia has always had a strict policy on citing sources. ChatGPT can't cite sources by design, because its answers are based on synthesis.

[datr]

Not true. The verifiability policy only really came into effect in 2006 (https://lists.wikimedia.org/pipermail/wikien-l/2006-July/050...) - five years after Wikipedia started.

[gremlinsinc]

It wouldn't be too hard to program at least gpt3 to basically take a chatGPT answer, go to google, scrape results and verify if the chatGPT answer was factual or not and maybe give it a score or rating of factual-ness.

[ekam]

If it’s that easy why don’t you do it? You’ll be printing money

[gremlinsinc]

Simple answer adhd, if I could ship a product I'd already probably be rich, instead I'm scraping by as a freelance dev. Though, chatGPT probably could help me code it anyways lol.

[bognition]

Absolutely the case, but also people make up stuff online all the time, so google has this exact same problem.

[memen]

No. Google gives you the source. ChatGPT does not.

[plaidfuji]

It’s funny because when I was in high school the argument was always “books, published articles and other print media are actual source material, Google doesn’t give you that”

[memen]

[0] scholar.google.com

Google gives you sources, determining reputability is your task.

[dinkumthinkum]

You think Google does not provide results from books, published articles, etc? Really?

[gremlinsinc]

Probably not when op was in high school, if they were still using books over web tools. I'm guessing before 2004? How old is google scholar?

[frakt0x90]

At least with Google you have sources you can trust more than others whereas ChatGPT is a black box

[eric4smith]

I Clearly said we use google to confirm the AI output.

And we also do not do medical stuff. I just used that as an example.

[Turing_Machine]

> Don’t use ChatGPT for medical research.

Or Google. There are plenty of pages out there that (e.g.) claim that Alzheimers is caused by drinking out of aluminum cans, or that the world is controlled by grey aliens from Zeta Reticuli.

[dinkumthinkum]

… you know Google provides the URL right? With Google it is very easy to tell if the information is coming from NIH or infowars/forums/etc.

[tasuki]

> ChatGPT often makes up facts.

As opposed to... Google? Your doctor? My doctor?

[dinkumthinkum]

Absolutely as opposed to those things. With Google, if you use a reliable source like Mayo, NIH, even a WebMD, It is clearly more likely to have accurate information than something that proves even numbers are prime. Certainly all those things can be inaccurate but where in the world you think ChatGPT pattern matches it’s information from?

[buu700]

Exactly. ChatGPT is clearly very impressive and useful, but nothing from its output should be treated as valid or factual to any degree.

Information generated by humans will include things like transpositional errors, logical errors, popular misconceptions, and misinterpretations of data. Mistakes happen, but human mistakes are at least tethered to real thoughts/information.

On the other hand, AI will happily spin up a complete fabrication with zero basis in reality, give you as much detail as you ask for, and dress it all up in competent and authoritative-sounding prose. It will have all the style of a textbook answer, while the substance will be pure nonsense.

Still a great tool, but only with the caveat that you approach it with the mindset that it's actively set out to catch you off guard and deceive you.

[tasuki]

> AI will happily spin up a complete fabrication with zero basis in reality, give you as much detail as you ask for, and dress it all up in competent and authoritative-sounding prose.

Sure. What makes you think a human won't?

[buu700]

I didn't say a human wouldn't. I said a human wouldn't typically do it by mistake.

[addandsubtract]

And how hard would it be for ChatGPT to be retrained on peer reviewed medical journals? ChatMD-GPT, if you will.

[nradov]

The majority of articles in peer-reviewed medical journals are also false.

https://doi.org/10.1371/journal.pmed.1004085

You can't take such articles seriously unless they have been independently reproduced multiple times. So, your hypothetical "ChatMD-GPT" would have to also filter on that basis and perhaps calculate some sort of confidence level.

[dinkumthinkum]

And it has already likely been trained on correct information and yet it produces bad results. It certainly has been trained data that explains what prime numbers and yet it produces what it produces, whereas using Google and hitting a credible source directly is more accurate and efficient.

[theGnuMe]

Isn't there a medical chat-gpt that passed the medical licensing exam? I thought I saw that come up..

[abnry]

Let's say ChatGPT gives you false information 50% of the time. It is still useful.

Just like it is harder to find primes numbers than verify that a number is prime, it is harder to dig up potential tidbits of information than to verify a piece of information handed to you is true.

[dinkumthinkum]

50% is still useful? A broken watch is useful in that sense as well I guess. I can only see that has useful if you don’t include efficient in the definition of useful.

[rutierut]

Like the comment said, if it's cheap (time, effort, etc) to reliably verify the answer the success rate doesn't really matter.

[brink]

Your prime number analogy doesn't hold water because the average person doesn't verify. Being wrong half the time has potential for serious damage.

[SilverBirch]

I feel like I've seen this on hacker news before, on other subjects. Someone will gush about how new technology X is great, but they give reasons that seem really odd to me. I've never really found Google laborious for searching for facts or especially coding solutions, and when it does provide me with differing options there are almost always great reasons why those options might all be relevent. With ChatGPT you're going to get one verbose answer that's probably wrong and presents none of the context as to why it might be wrong. So sure, if you're only using ChatGPT to answer questions to which you know the answer, it could be quicker.

>Why research: “give me 30 of the most common health conditions related to the human liver” and spend a lot of time in google, when the Ai can spit out that in seconds?

Because it's not going to be right! If you actually need to know the answer to that question you need to actually find a reputable source. And that's what Google gives you. I'm quite certain that the most common health conditions for the human liver vary by country, will ChatGPT give you the actual answer you're looking for? Maybe, some of the time. Will it save you time, no! because you can't use the answer unless you google it to confirm.

It sounds like your using ChatGPT to pump out worthless marketing SEO. Yes, that's a niche where creating volume of material with no value is common place. The aim shouldn't be to make that more efficient, it should be to find ways to entirely filter that out of the internet. What you're producing is literally the only material people should be using ChatGPT for instead of the web - low quality verbose text that is indifferent to fact checking.

[propogandist]

> I can confirm the output and clean up the generated text into my own style... to support online marketing activities.

it sounds like you're using AI generated output to do content marketing to support SEO activities for a healthcare client at a marketing agency

if this is the case, when humans search for [organic] content on health conditions related to the human liver (on Google) you are hoping they land on ChatGPT generated content you published (for your client), to help the client avoid buying Google Ads to get those customers

at some point your content will rank well enough that other SEOs will do the same thing to compete, leading to dilution of overall quality content as it's all SEO optimized content, which makes Search generally unusable.

This, to be fair, has been happening well before ChatGPT, but will only accelerate.

[eric4smith]

Ugh I’m not.

[acdha]

> What do I do?

Compare with an actual expert because the list from ChatGPT is almost certainly incomplete and will inevitably contain plausible-sounding but completely wrong claims. One of the big challenges here is when you don’t know the field well enough to know what ChatGPT didn’t include on it’s list at all or be able to tell when there are multiple similar sounding things being conflated.

[ForHackernews]

He does online marketing: plausible-sounding but completely wrong claims are not a bug, they are the feature.

[acdha]

I’ve worked with decent marketers so I was assuming good-faith.

[dinkumthinkum]

I don’t really get your example. You could search for common liver conditions and then get a link to WebMD or Mayo Clinic very, very quickly and you can be very confident that it is accurate. If I’m a 4th grader using ChatGPT to cheat on my math homework, I might be quite satisfied with the answer that 42 is a prime number and even provide its great proof.

[eric4smith]

The point is to get the list quickly.

Then the next point is to summarize each item.

Then I can go in and validate the info from other sources and clean up the writing to my style.

Get it now?

[ForHackernews]

> It’s not there to replace people’s work.

Well, right up until the day after ChatGPT2 is better than your junior employee + ChatGPT.