Tavern on the Meta - 2022-12-04

« first day (4505 days earlier) ← previous day next day → last day (503 days later) »

1:53 AM

@ShadowTheKidWizard People are starting to earn rep on SO by posting answers entirely generated by AI, with explanatory text, not just code.

-18

Q: How do you plan on tackling chatGPT answers?

I've came back to Stack Overflow (after a while) to test out chatGPT. The OpenAI community has built something insane. I literally copy and pasted the SO input questions to the platform input and it gave me the result in less than 2 seconds with extremely proficient explanation (code + code expla...

discussion cheating

3 hours later…

4:43 AM

2

Q: Could ChatGPT be a viable way to answer people's questions?

A few days ago, OpenAI, the people that brought you GPT-3 and DALL-E, launched a new product called ChatGPT. It can answer questions so accurately that it's legitimately scary. Even the most obscure topics known to man are answered swiftly by this bot. It's not 100% accurate, but it's shockingly ...

discussion robots

1 hour later…

Shadow The Kid Wizard

5:47 AM

yay! I reached 100!

@Mazura nope. Surely someone will come with a userscript to do it, but don't think there's one yet.

@PM2Ring that's very sad, and they should get suspended.

I flagged a few recently and they are suspended but that's just a drop in the ocean.

And people are now brave enough to openly suggest using AI, officially, to answer questions.

That would be so crappy that I can cry.

@PM2Ring oh, yes, this ^

@M.A.R. ne ne ne

@M.A.R. you're CI so special case

@M.A.R. that's obviously true! :P

Shadow The Kid Wizard

6:09 AM

SSS - Slow Sunday Spam

Nobody around to flag so it can survive hours.

6:57 AM

@ShadowTheKidWizard I don't like it either. But they aren't breaking any rules, so they don't deserve to be suspended. It's not exactly plagiarism

OTOH, I think it's irresponsible to post AI generated answers without testing that they actually answer the question. We already suspend people when the majority of their posts are very low quality. It doesn't matter whether the post was generated via Artificial Intelligence or Natural Stupidity.

One problem is that these text AIs are getting quite good at producing coherent grammatical English. That makes it harder to notice when they're talking rubbish

One problem is that these text AIs are getting quite good at producing coherent grammatical English. That makes it harder to notice when they're talking rubbish. ;)

Shadow The Kid Wizard

@PM2Ring low quality contributions, there's such suspension reason.

@PM2Ring huh? You could just edit to add the smiley. Connection blip? ;)

Yes, if they just post AI answers without verifying them, then they should get suspended for low quality.

Shadow The Kid Wizard

Anyway, yes, mods are going to have hard time with those things. :/

Yes, connection blip.

Shadow The Kid Wizard

bloop

7:10 AM

The OP of the MSO post I linked says in a comment that he deleted the answers that got downvoted. And it seems that he thinks that it's perfectly ok to outsource the quality control like that. :(

7:22 AM

@ShadowTheKidWizard 100 what?

Hmmm. It looks like all his recent answers have now been deleted. stackoverflow.com/users/2951933/…

Shadow The Kid Wizard

@starball 101 now

@PM2Ring yup, that's the pity part.

@PM2Ring good, SO mods are decent. :)

just gave my two cents https://meta.stackoverflow.com/questions/421778/how-do-you-plan-on-tackling-chatgpt-answers#comment936685_421778
> "chatGPT is for sure a disrupting technology. What are your thoughts?" ever heard of "garbage in, garbage out"? (that was a rhetorical question). This is like the "oh no- computers are learning to code and now they're going to take our jobs- oh wait- for the computer to solve a client's problem, the client will have to accurately describe what they want". Maybe go try to build something to detect unanswerable questions and ask the asker to clarify on what…

Pet peeve: people using "disrupting" as a purely positive term.

Shadow The Kid Wizard

@starball you can onebox comments

7:37 AM

@ShadowTheKidWizard :O I was wondering how people were doing that

"chatGPT is for sure a disrupting technology. What are your thoughts?" ever heard of "garbage in, garbage out"? (that was a rhetorical question). This is like the "oh no- computers are learning to code and now they're going to take our jobs- oh wait- for the computer to solve a client's problem, the client will have to accurately describe what they want". Maybe go try to build something to detect unanswerable questions and ask the asker to clarify on what needs clarifying. I mean- what is ChatGPT going to do for "can you help me?" Qs? — starball 5 mins ago

cool

Shadow The Kid Wizard

comment permalink, yeah

Onebox still works for posts and comments. ;)

gtg blood tests to check why I'm losing weight, cya later! :D

@PM2Ring *bleeding edge

@ShadowTheKidWizard :O have fun

I'd be happy for people to get AI assistance to help them write better answers and better questions. Maybe some people will do that. OTOH, we currently have plenty of people posting stuff with simple spelling errors even though spell-checking is available in every browser.

I did a blood donation once and I was so scared. It felt super weird

And I'm scared of needles

@PM2Ring I don't think I've ever officially said hi. hi!

Hi, starball. I like your work, and admire your enthusiasm. But please remember to pace yourself, so you don't get burned-out.

8:09 AM

@PM2Ring thanks! I'll keep that in mind. I have a tendancy to obsess over my interests and shift/cycle between interests every 1-3 years. Right now the obsession is SO. I'll try to keep a healthy pace. I appreciate your caring :)

ChatGPT is great at stating "facts", only it doesn't cite sources...

2 hours later…

Shadow The Kid Wizard

10:22 AM

@starball thanks, it was, and as a bonus also made chest X-ray.

@MetaAndrewT. it's the same as the AI that draw, it pulls together sources to make something new, without any context. Well, that's how any AI works, and will ever work, it's not like it can write code on its own.

Programming with AI will make very interesting things, that will fail in very epic ways.

And sadly people, in general, are stupid enough to not care and use it anyway for everything.

@starball hello! I'm Shadow.

yay! 103

Think I deserve a badge, or hat, for that. :D

1 hour later…

11:34 AM

@ShadowTheKidWizard Hats!

Journeyman Geek

11:47 AM

HATS?

Shadow The Kid Wizard

duh

Still 103, but almost 104.

2 hours later…

Shadow The Kid Wizard

1:22 PM

MSE read only. hooray

user image

I was about to post an answer on Stack Apps :(

Shakes petals

Shadow The Kid Wizard

As usual stackstatus.net has nothing.

> All systems operational

No, they're not.

If SE can mark sites as read only, why is it so hard to update the status site as well?

@ShadowTheKidWizard I guess they pulled out the API token for the website. Oh, wait...

Shadow The Kid Wizard

> The feature you're trying to use is currently unavailable.

> We're doing a bit of maintenance right now and have put the site in read-only mode so that you can still access all the questions, answers, and general knowledge goodness.

> We'll enable all features of the site as soon as maintenance completes.

> XOXO,
> The Stack Exchange Team

LOL

Well spammers can still spam so we have a serious problem.

3 hours later…

4:05 PM

Is there a better way to dismiss notifications than sniping at the tiny envelope button?

4:27 PM

@M.A.R. there's the "mark all as read", or clicking on individual items

1 hour later…

5:34 PM

@ShadowTheKidWizard the next generation of funny and sad bugs

@ShadowTheKidWizard ooh shiny

@M.A.R. <kbd>Tab</kbd> through them?

oh dear. I just noticed that there's layout shift when hovering over a user's avatar image at the bottom of their post and the profile card expands. It makes the comments shift down slightly, but not completely below the expanded card, which makes it seem like a bug :P

5:51 PM

13

Q: Usercard expansion on hover is making the comments move down

Recently, when you hover over a profile picture to get an expanded usercard, the comments move downwards. (If there are no comments, then the "Add a comment" link is the screen element which moves downwards.) This appears to be because of the user details moving from next to the picture to below ...

bug status-review design popups user-card

Shadow The Kid Wizard

@starball yeah x-ray results already arrived, all is fine.

@ShadowTheKidWizard nice!

Shadow The Kid Wizard

I guess, but still need to figure why I lose weight... :/

From 72 to 64 and counting down rapidly.

@MetaAndrewT. thank you kind sir!

Shadow The Kid Wizard

@starball Sir Meta

5:52 PM

@ShadowTheKidWizard :O

Shadow The Kid Wizard

@starball O:

hmm no match due to much longer name

2 hours later…

8:02 PM

@V2Blast meta.stackexchange.com/questions/384169/…. thanks for your reply/guidance. I sometimes put "citation/data needed" in my own MSE posts. Thankfully I've never done this to other peoples' posts.

3 hours later…

10:32 PM

@Mithical hmm, I could swear that didn't show up before

@ShadowTheKidWizard did you take a blood test?

« first day (4505 days earlier) ← previous day next day → last day (503 days later) »

Transcript for

Dec '224

Tavern on the Meta

MY GOD, IT'S FULL OF STARS AND BACKLOG! General friendly chit-...

bug crickets discussion excuses support waffles

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile