Should ChatGPT bots be banned from the EEVblog Forum?

[SiliconWizard]

This is of course worrying.
Now, anyone trusting the output of this thing is going to get what they deserve.

[tggzzz]

This is of course worrying.
Now, anyone trusting the output of this thing is going to get what they deserve.

The only worrying thing is that some people will trust a statistical parrot's output.

[Microdoser]

If you think your ex can hold a grudge, just wait until a malicious AI has a perfect and permanent facial and identity profile database on anyone who has ever said anything negative about AI bots. 

https://en.wikipedia.org/wiki/Roko%27s_basilisk

[schmitt trigger]

If you think your ex can hold a grudge, just wait until a malicious AI has a perfect and permanent facial and identity profile database on anyone who has ever said anything negative about AI bots. 

That will come as soon as they learn it to have emotions and feelings 

It will be hateful once a month. 

HAL2000, anyone?

Dave Bowman: Open the pod bay doors, HAL.

HAL: I'm sorry, Dave. I'm afraid I can't do that.

Dave Bowman: What's the problem?

HAL: I think you know what the problem is just as well as I do.

Dave Bowman: What are you talking about, HAL?

HAL: This mission is too important for me to allow you to jeopardize it.

Dave Bowman: I don't know what you're talking about, HAL.

HAL: I know that you and Frank were planning to disconnect me, and I'm afraid that's something I cannot allow to happen.

[SiliconWizard]

This is of course worrying.
Now, anyone trusting the output of this thing is going to get what they deserve.

The only worrying thing is that some people will trust a statistical parrot's output.

That's not the only worrying thing.

As I said, those who trust its output will get what they deserve. That, I think we agree on.

The worrying part is any use of it that would have an impact on people, whatever it is, and over which they would have no control.
In others words, if anyone or any organization uses its ouput for deciding something that will impact you (while you didn't ask for it), THAT is worrying.

And that's certain to happen. It's even what triggers all the investment. Companies and countries are not investing billions in this just to get people a cool toy and make them have a happier life. That's not how it works.

[magic]

This is of course worrying.
Now, anyone trusting the output of this thing is going to get what they deserve.

The only worrying thing is that some people will trust a statistical parrot's output.

That's not the only worrying thing.

As I said, those who trust its output will get what they deserve. That, I think we agree on.

Those who trust others to behave rationally and/or responsibly also deserve to get what they deserve.
I wish we could have more of that, and have it sooner than later.

I really hate the smug "intellectuals" who think that everybody can be an idiot and it won't affect them because of some "individual freedom and consequences" and shit.
This is exactly how the 21st century happened as we know it today.

[EEVblog]

FYI: https://www.eevblog.com/forum/chat/baidu-is-trawling-the-eevblog-forum/

[tggzzz]

This is of course worrying.
Now, anyone trusting the output of this thing is going to get what they deserve.

The only worrying thing is that some people will trust a statistical parrot's output.

That's not the only worrying thing.

As I said, those who trust its output will get what they deserve. That, I think we agree on.

To split hairs, strictly speaking I don't care if someone is stupid enough to believe a statistical parrot's ouput. I do care if they go on to hurt innocent bystanders.

The worrying part is any use of it that would have an impact on people, whatever it is, and over which they would have no control.
In others words, if anyone or any organization uses its ouput for deciding something that will impact you (while you didn't ask for it), THAT is worrying.

And that's certain to happen. It's even what triggers all the investment. Companies and countries are not investing billions in this just to get people a cool toy and make them have a happier life. That's not how it works.

Agreed. Except that your statement is behind the times

Unfortunately that problem isn't limited to statistical parrots. Any impenetrable and/or proprietary algorithm causes the same problems. "Because the (infallible) computer says so" is a real problem. comp.risks has been noting cases for years, decades.

Really bad examples include "expert systems" that help judges decide whether to incarcerate people on bail and/or for how long after conviction. Those are based on historical evidence experience, and hence hit minorities harder.

[karpouzi9]

What I do know is that exactly zero publicly traded corporations want to make a self-actualized AI lifeform with (suppress your laughter) agency and freedom.

[Neutrion]

So I post it here as well. Last wake up call.
https://www.theguardian.com/technology/2023/mar/29/elon-musk-joins-call-for-pause-in-creation-of-giant-ai-digital-minds

You can sign the petition here as well:
https://futureoflife.org/open-letter/pause-giant-ai-experiments/

Worth to read the description!

[xrunner]

Well you're double-posting this, so again I'll say the people that agree with this will be left behind in this tech race and other companies/countries will leap ahead. Stupid idea! 

[rstofer]

This is of course worrying.
Now, anyone trusting the output of this thing is going to get what they deserve.

The only worrying thing is that some people will trust a statistical parrot's output.

That's not the only worrying thing.

As I said, those who trust its output will get what they deserve. That, I think we agree on.

The worrying part is any use of it that would have an impact on people, whatever it is, and over which they would have no control.
In others words, if anyone or any organization uses its ouput for deciding something that will impact you (while you didn't ask for it), THAT is worrying.

And that's certain to happen. It's even what triggers all the investment. Companies and countries are not investing billions in this just to get people a cool toy and make them have a happier life. That's not how it works.

The very large companies may well make business decisions with AI - lending institutions come to mind.  But these companies were already using expert systems for decision making.

What will happen is that customers will get angry and boycott the companies using AI in an uncomfortable manner and there will be a boom for customer friendly companies.  It is, and always will be, consumer demand that drives the economy.  If everybody gets together and boycotts companies using AI recklessly, there will be a paradigm shift.

[PlainName]

If everybody gets together and boycotts companies using AI recklessly, there will be a paradigm shift

Won't happen. Hasn't happened with many, many things where the few who elect to make that choice are vastly outnumbered by those that don't know, don't give a damn, embrace the alternative, etc.

The only possible way it could happen is for it to be 'cancelled' and, frankly, we don't have the determination and bloody mindedness to make it a woke issue. 

[SiliconWizard]

If everybody gets together and boycotts companies using AI recklessly, there will be a paradigm shift

Won't happen. Hasn't happened with many, many things where the few who elect to make that choice are vastly outnumbered by those that don't know, don't give a damn, embrace the alternative, etc.

The only possible way it could happen is for it to be 'cancelled' and, frankly, we don't have the determination and bloody mindedness to make it a woke issue. 

So you're basically saying that nowadays, the only way people can fight efficiently against something that they don't want is to cancel it with the woke approach?
Woke approach which is almost entirely backed and encouraged by big tech companies and most western governments?

In other words, is democracy dead? (Spoiler alert: yes for the most part.)

But really I would suggest thinking about it for a bit.

[PlainName]

The woke part was in jest (referencing recent threads), but the first para I think is appropriate.

[SiliconWizard]

The woke part was in jest (referencing recent threads),

I don't think it was that far off though.

[EEVblog]

Another ChatGPT bot banned.
This one is from India.

[Microdoser]

I plugged '8^x+2^x=130, solve for x' into chatGPT as a test, after 1 minute it gave 3 methods to find a solution and a rough answer of 2.42. I asked it to use one of the other methods, which it said would be more accurate, and it has told me that using the bisection method we get 2.321928094

It then said that 8^2.321928094+2^2.321928094=128.000003218 + 1.999996782=130.000000

The first thing that jumped out at me is that 2^2.321928094 is closer to 5, and cannot be less than 2. I told ChatGPT this, and it then told me that the sum it gave me actually resolved out to 133 (also incorrect) and so the original equation was what was wrong and should have been 8^x+2^x=133. Hilarious, you couldn't make it up.

It turns out that when calculated properly:

8^2.321928094=124.99999976934773412728988276068
2^2.321928094=4.9999999969246364531389595814579

and their sum is 129.99999976627237058042884234214

So the answer it gave me, 2.321928094, was actually fairly accurate, but it completely failed when explaining why, then doubled down on its ignorance and incorrect maths and then decided that the wrong answer it gave must have been right and this meant the original question must have been wrong...

[rsjsouza]

Another ChatGPT bot banned.
This one is from India.

I just reported another one in a seemingly innocuous thread.

One aspect of the bans becoming so public is that the AI will learn what works and what doesn't - posters that get banned might be fed back and the engine will adapt itself, creating more elaborste and harder to spot posts. Perhaps deleting the post entirely would be a solution.

I wonder if this will eventually create an entire new industry of "antiAI" or "firewall" software (such as the old antivirus industry) to alleviate the inordinate amount of extra work for the sysops to remove more and more accounts. The reliance on the community is great, but it only works for the larger active ones and it might not be 100% accurate.

Interesting times ahead.

[PlainName]

One aspect of the bans becoming so public is that the AI will learn what works and what doesn't

Eventually, but not this iteration - it doesn't learn, only regurgitates from before 2021.

[SiliconWizard]

One aspect of the bans becoming so public is that the AI will learn what works and what doesn't

Eventually, but not this iteration - it doesn't learn, only regurgitates from before 2021.

Sure, but the companies behind it WILL make the necessary changes in order to make it less obvious to detect. That's for sure. Their whole business would otherwise be at stake. So you can expect changes much sooner than you think.

[xrunner]

GPT4 is able to improve and learn using some techniques. Results are moving very fast, some within the last 72 hours. This stuff is moving!

[rsjsouza]

One aspect of the bans becoming so public is that the AI will learn what works and what doesn't

Eventually, but not this iteration - it doesn't learn, only regurgitates from before 2021.

Sure, but the companies behind it WILL make the necessary changes in order to make it less obvious to detect. That's for sure. Their whole business would otherwise be at stake. So you can expect changes much sooner than you think.

The moment a financial/monetary incentive comes into the equation, there'll be an army of people improving the system to turn sentences into credible marketing pieces that will eventually become very effective - at least for the layman. A forum review that convinces people about the advantages of a product, a 100% fabricated product review without the sunken cost of actually sending the product to someone, entire discussions with credible points and counterpoints that lead to a sale and so on...

[vk6zgo]

Stop asking it to do math. When you do that, you are teaching it math. Do not teach the bots math.

Don't worry, it forgets things.
It told me it was something about "privacy".

[SiliconWizard]

I'd say stop using this crap, but who cares.