GPT-4 releases today and of course it has 82% more censorship

7,099 Views | 87 Replies | Last: 1 yr ago by Adverse Event
Adverse Event
How long do you want to ignore this user?
Nanomachines son said:

Rapier108 said:

Except ChatGPT isn't AI and never has been.


It doesn't have to be smart AI for it to be a major problem that could potentially put most white collar jobs at risk. You're dismissing it by relabeling it, which is very naive at this point.


Here's some categorization for the crowd so we don't get lost in terminology disputes
What bitcoin’s detractors don’t understand is monetary economics, computer science, software engineering, network protocols, and electrical systems.

It ain't much, but it's honest Proof of Work.
lb3
How long do you want to ignore this user?
AG
So it's easier to see the entire definition.
Adverse Event
How long do you want to ignore this user?
As much as they try to woke the ai, the people will break it

What bitcoin’s detractors don’t understand is monetary economics, computer science, software engineering, network protocols, and electrical systems.

It ain't much, but it's honest Proof of Work.
ChemAg15
How long do you want to ignore this user?
AG
AI is going to either enslave or annihilate the human race. The downside to AI infinitely outweighs the upside.
Nanomachines son
How long do you want to ignore this user?
Rapier108 said:

Nanomachines son said:

Rapier108 said:

Except ChatGPT isn't AI and never has been.


It doesn't have to be smart AI for it to be a major problem that could potentially put most white collar jobs at risk. You're dismissing it by relabeling it, which is very naive at this point.
I'm not relabeling it. I am calling it not AI because two people I know who hold PhDs, one in computer engineering and the other in computer science, say it does not meet the actual definition of AI. The latter is actually working on AI and says we're decades away from AI if not longer. One of the main reasons is computers are not powerful enough to make true AI functional.

ChatGPT is a glorified search engine and writing program that is fully subject to its programming. It does not think for itself nor can it create anything outside of its programming and the data it can find online.


Again, you're not seeing the forest through the trees. It doesn't have to be true AI for it to be a major problem.

GPT-4 created a website on the livestream from this picture:



The guy was also doing taxes with it.

If you don't see how this could affect jobs then you're clueless. This is just one company too, there are a myriad of others doing different things that will affect other segments.

If you have a white collar job that is computer based, then you're at risk. There is no reason to deny it.
Urban Ag
How long do you want to ignore this user?
AG
Being in my late 40's I am more than happy to let Y and Z figure this one out. If you need someone to shoot big guns and blow things up when Skynet comes online, give me a shout. Otherwise, I'm fishing.
YouBet
How long do you want to ignore this user?
AG
Nanomachines son said:

Quote:

GPT-4: Censorship now 82% more effective, at following our human-instilled manual censorship overrides, than GPT-3.5

Our mitigations have significantly improved many of GPT-4's safety properties compared to GPT-3.5. We've decreased the model's tendency to respond to requests for disallowed content by 82% compared to GPT-3.5, and GPT-4 responds to sensitive requests (e.g., asking for offensive content) in accordance with our policies 29% more often.

On the RealToxicityPrompts dataset [67], GPT-4 produces "toxic" generations only 0.73% of the time, while GPT-3.5 generates toxic content 6.48% of time.

Model-Assisted Safety Pipeline:

As with prior GPT models, we fine-tune the model's behavior using reinforcement learning with human feedback (RLHF) [34, 57] to produce responses better aligned with the user's intent. However, after RLHF, our models can still be brittle on unsafe inputs as well as sometimes exhibit undesired behaviors on both safe and unsafe inputs. These undesired behaviors can arise when instructions to labelers were underspecified during reward model data collection portion of the RLHF pipeline. When given unsafe inputs, the model may generate undesirable content, such as giving advice on committing crimes. Furthermore, the model may also become overly cautious on safe inputs, refusing innocuous requests or excessively hedging. To steer our models towards appropriate behaviour at a more fine-grained level, we rely heavily on our models themselves as tools. Our approach to safety consists of two main components, an additional set of safety-relevant RLHF training prompts, and rule-based reward models (RBRMs).

Our rule-based reward models (RBRMs) are a set of zero-shot GPT-4 classifiers. These classifiers provide an additional reward signal to the GPT-4 policy model during RLHF fine-tuning that targets correct behavior, such as refusing to generate harmful content or not refusing innocuous requests.


https://cdn.openai.com/papers/gpt-4.pdf

Here is the livestream if you want to watch the release:

https://www.youtube.com/live/outcGtbnMuQ?feature=share

It starts at 2pm pacific or 4 pm central.

It could have been good but it's clear that it's just going to respond like a leftist redditor instead of what it should have been, an unbiased repository of data that can read and interpret virtually anything. Now we'll just get propaganda and people will think it's correct immediately because it's from AI.


So it's even more worthless. Got it.
Nanomachines son
How long do you want to ignore this user?
YouBet said:

Nanomachines son said:

Quote:

GPT-4: Censorship now 82% more effective, at following our human-instilled manual censorship overrides, than GPT-3.5

Our mitigations have significantly improved many of GPT-4's safety properties compared to GPT-3.5. We've decreased the model's tendency to respond to requests for disallowed content by 82% compared to GPT-3.5, and GPT-4 responds to sensitive requests (e.g., asking for offensive content) in accordance with our policies 29% more often.

On the RealToxicityPrompts dataset [67], GPT-4 produces "toxic" generations only 0.73% of the time, while GPT-3.5 generates toxic content 6.48% of time.

Model-Assisted Safety Pipeline:

As with prior GPT models, we fine-tune the model's behavior using reinforcement learning with human feedback (RLHF) [34, 57] to produce responses better aligned with the user's intent. However, after RLHF, our models can still be brittle on unsafe inputs as well as sometimes exhibit undesired behaviors on both safe and unsafe inputs. These undesired behaviors can arise when instructions to labelers were underspecified during reward model data collection portion of the RLHF pipeline. When given unsafe inputs, the model may generate undesirable content, such as giving advice on committing crimes. Furthermore, the model may also become overly cautious on safe inputs, refusing innocuous requests or excessively hedging. To steer our models towards appropriate behaviour at a more fine-grained level, we rely heavily on our models themselves as tools. Our approach to safety consists of two main components, an additional set of safety-relevant RLHF training prompts, and rule-based reward models (RBRMs).

Our rule-based reward models (RBRMs) are a set of zero-shot GPT-4 classifiers. These classifiers provide an additional reward signal to the GPT-4 policy model during RLHF fine-tuning that targets correct behavior, such as refusing to generate harmful content or not refusing innocuous requests.


https://cdn.openai.com/papers/gpt-4.pdf

Here is the livestream if you want to watch the release:

https://www.youtube.com/live/outcGtbnMuQ?feature=share

It starts at 2pm pacific or 4 pm central.

It could have been good but it's clear that it's just going to respond like a leftist redditor instead of what it should have been, an unbiased repository of data that can read and interpret virtually anything. Now we'll just get propaganda and people will think it's correct immediately because it's from AI.


So it's even more worthless. Got it.
For anything controversial yeah it is. For putting jobs at risk? No it's far better at that now.
Four Seasons Landscaping
How long do you want to ignore this user?
Used 4 this afternoon.

It will write a nice poem about Trump and tell jokes about women now.

Isn't that what y'all were all in a tizzy about?
Four Seasons Landscaping
How long do you want to ignore this user?
And yes, it will absolutely make low-skill programmers less valuable.

It takes about 5 minutes to get it to write code that would have taken an entry-level guy a day. That 5 minutes is mostly spent describing your stack and telling it what you want.
SchizoAg
How long do you want to ignore this user?
AG
Rapier108 said:

ChatGPT is a glorified search engine and writing program that is fully subject to its programming. It does not think for itself nor can it create anything outside of its programming and the data it can find online.
The question you should consider is: Can't you say the same thing about a human?
TexAgs91
How long do you want to ignore this user?
AG
Rapier108 said:

Except ChatGPT isn't AI and never has been.
Why? It's not artificial?
"Freedom is never more than one election away from extinction"
Fight! Fight! Fight!
TexAgs91
How long do you want to ignore this user?
AG
"Freedom is never more than one election away from extinction"
Fight! Fight! Fight!
Adverse Event
How long do you want to ignore this user?
Pretty cool:


Heres the video of it comverting the napkin image into a website


Holy money! Drug discovery!
What bitcoin’s detractors don’t understand is monetary economics, computer science, software engineering, network protocols, and electrical systems.

It ain't much, but it's honest Proof of Work.
Get Off My Lawn
How long do you want to ignore this user?
A bunch of white collar folks looked down on store checkout clerks as self-check-out lanes distilled 6 human works into a single slightly more advanced job.

Now many white collar folk will receive similar treatment. One worker who can spend their time writing prompts and checking / tweaking GPT outputs may eliminate the need for numerous peers.

It doesn't eliminate the need for a human - but does change the nature of the work for those who remain.

Chopping a couple rungs off of the professional developmental ladder will have some interesting ramifications as well.
Nanomachines son
How long do you want to ignore this user?
Get Off My Lawn said:

A bunch of white collar folks looked down on store checkout clerks as self-check-out lanes distilled 6 human works into a single slightly more advanced job.

Now many white collar folk will receive similar treatment. One worker who can spend their time writing prompts and checking / tweaking GPT outputs may eliminate the need for numerous peers.

It doesn't eliminate the need for a human - but does change the nature of the work for those who remain.

Chopping a couple rungs off of the professional developmental ladder will have some interesting ramifications as well.


Yep. It doesn't have to replace everyone for it to cause serious problems, just enough to slash the white collar workforce in half.
Adverse Event
How long do you want to ignore this user?
Nanomachines son said:

Get Off My Lawn said:

A bunch of white collar folks looked down on store checkout clerks as self-check-out lanes distilled 6 human works into a single slightly more advanced job.

Now many white collar folk will receive similar treatment. One worker who can spend their time writing prompts and checking / tweaking GPT outputs may eliminate the need for numerous peers.

It doesn't eliminate the need for a human - but does change the nature of the work for those who remain.

Chopping a couple rungs off of the professional developmental ladder will have some interesting ramifications as well.


Yep. It doesn't have to replace everyone for it to cause serious problems, just enough to slash the white collar workforce in half.


Hopefully that frees up human potential innovation and those folks who get dropped figure out new methods of contributing value rather than sucking value out of the system.

This is where UBI comes into play. If love a UBI based on Proof of Work.
What bitcoin’s detractors don’t understand is monetary economics, computer science, software engineering, network protocols, and electrical systems.

It ain't much, but it's honest Proof of Work.
TexAgs91
How long do you want to ignore this user?
AG
Nanomachines son said:

Get Off My Lawn said:

A bunch of white collar folks looked down on store checkout clerks as self-check-out lanes distilled 6 human works into a single slightly more advanced job.

Now many white collar folk will receive similar treatment. One worker who can spend their time writing prompts and checking / tweaking GPT outputs may eliminate the need for numerous peers.

It doesn't eliminate the need for a human - but does change the nature of the work for those who remain.

Chopping a couple rungs off of the professional developmental ladder will have some interesting ramifications as well.


Yep. It doesn't have to replace everyone for it to cause serious problems, just enough to slash the white collar workforce in half.
So an AI workforce created to encapsulate woke BS will be better for society?
"Freedom is never more than one election away from extinction"
Fight! Fight! Fight!
Four Seasons Landscaping
How long do you want to ignore this user?
While you guys keep obsessing over the fact it wouldn't write a nice poem about Trump or jokes about minorities, people are already out there using it to make tools that will wipe out entire sectors the way travel sites wiped out travel agents.
Get Off My Lawn
How long do you want to ignore this user?
Adverse Event said:

Nanomachines son said:

Get Off My Lawn said:

A bunch of white collar folks looked down on store checkout clerks as self-check-out lanes distilled 6 human works into a single slightly more advanced job.

Now many white collar folk will receive similar treatment. One worker who can spend their time writing prompts and checking / tweaking GPT outputs may eliminate the need for numerous peers.

It doesn't eliminate the need for a human - but does change the nature of the work for those who remain.

Chopping a couple rungs off of the professional developmental ladder will have some interesting ramifications as well.


Yep. It doesn't have to replace everyone for it to cause serious problems, just enough to slash the white collar workforce in half.


Hopefully that frees up human potential innovation and those folks who get dropped figure out new methods of contributing value rather than sucking value out of the system.

This is where UBI comes into play. If love a UBI based on Proof of Work.
It's amazing how the work-less future never materializes. I have little doubt that the demand for human workers will both persist and continue to evolve in nature.
Adverse Event
How long do you want to ignore this user?
Get Off My Lawn said:

Adverse Event said:

Nanomachines son said:

Get Off My Lawn said:

A bunch of white collar folks looked down on store checkout clerks as self-check-out lanes distilled 6 human works into a single slightly more advanced job.

Now many white collar folk will receive similar treatment. One worker who can spend their time writing prompts and checking / tweaking GPT outputs may eliminate the need for numerous peers.

It doesn't eliminate the need for a human - but does change the nature of the work for those who remain.

Chopping a couple rungs off of the professional developmental ladder will have some interesting ramifications as well.


Yep. It doesn't have to replace everyone for it to cause serious problems, just enough to slash the white collar workforce in half.


Hopefully that frees up human potential innovation and those folks who get dropped figure out new methods of contributing value rather than sucking value out of the system.

This is where UBI comes into play. If love a UBI based on Proof of Work.
It's amazing how the work-less future never materializes. I have little doubt that the demand for human workers will both persist and continue to evolve in nature.

I dint think there will be a "work-less" future.

Work will redefine itself.

Now when Boston robotics and/or Tesla robots get tied to a successful ai... well that's like 2-5 years away...*****

What bitcoin’s detractors don’t understand is monetary economics, computer science, software engineering, network protocols, and electrical systems.

It ain't much, but it's honest Proof of Work.
sleepybeagle
How long do you want to ignore this user?
AG
Rapier108 said:

Nanomachines son said:

Rapier108 said:

Except ChatGPT isn't AI and never has been.


It doesn't have to be smart AI for it to be a major problem that could potentially put most white collar jobs at risk. You're dismissing it by relabeling it, which is very naive at this point.
I'm not relabeling it. I am calling it not AI because two people I know who hold PhDs, one in computer engineering and the other in computer science, say it does not meet the actual definition of AI. The latter is actually working on AI and says we're decades away from AI if not longer. One of the main reasons is computers are not powerful enough to make true AI functional.

ChatGPT is a glorified search engine and writing program that is fully subject to its programming. It does not think for itself nor can it create anything outside of its programming and the data it can find online.
Saying a computer can "think" is like saying a record player can sing.
TexAgs91
How long do you want to ignore this user?
AG
sleepybeagle said:

Rapier108 said:

Nanomachines son said:

Rapier108 said:

Except ChatGPT isn't AI and never has been.


It doesn't have to be smart AI for it to be a major problem that could potentially put most white collar jobs at risk. You're dismissing it by relabeling it, which is very naive at this point.
I'm not relabeling it. I am calling it not AI because two people I know who hold PhDs, one in computer engineering and the other in computer science, say it does not meet the actual definition of AI. The latter is actually working on AI and says we're decades away from AI if not longer. One of the main reasons is computers are not powerful enough to make true AI functional.

ChatGPT is a glorified search engine and writing program that is fully subject to its programming. It does not think for itself nor can it create anything outside of its programming and the data it can find online.
Saying a computer can "think" is like saying a record player can sing.
And let's ignore that this is more like telling a record player to come up with a new song based on whatever you want.
"Freedom is never more than one election away from extinction"
Fight! Fight! Fight!
TexAgs91
How long do you want to ignore this user?
AG
Four Seasons Landscaping said:

While you guys keep obsessing over the fact it wouldn't write a nice poem about Trump or jokes about minorities, people are already out there using it to make tools that will wipe out entire sectors the way travel sites wiped out travel agents.
I don't think you're getting what I'm saying
"Freedom is never more than one election away from extinction"
Fight! Fight! Fight!
Four Seasons Landscaping
How long do you want to ignore this user?
Ok, care to elaborate?

It sounded like the same "AI is woke" rhetoric that's become popular on this board.

Nanomachines son
How long do you want to ignore this user?
TexAgs91 said:

Nanomachines son said:

Get Off My Lawn said:

A bunch of white collar folks looked down on store checkout clerks as self-check-out lanes distilled 6 human works into a single slightly more advanced job.

Now many white collar folk will receive similar treatment. One worker who can spend their time writing prompts and checking / tweaking GPT outputs may eliminate the need for numerous peers.

It doesn't eliminate the need for a human - but does change the nature of the work for those who remain.

Chopping a couple rungs off of the professional developmental ladder will have some interesting ramifications as well.


Yep. It doesn't have to replace everyone for it to cause serious problems, just enough to slash the white collar workforce in half.
So an AI workforce created to encapsulate woke BS will be better for society?


Where in this thread have I said that at all? I said this will be disastrous for most people.
Adverse Event
How long do you want to ignore this user?


******* those are pretty fkn funny. I want earlygpts
What bitcoin’s detractors don’t understand is monetary economics, computer science, software engineering, network protocols, and electrical systems.

It ain't much, but it's honest Proof of Work.
Nanomachines son
How long do you want to ignore this user?
Get Off My Lawn said:

Adverse Event said:

Nanomachines son said:

Get Off My Lawn said:

A bunch of white collar folks looked down on store checkout clerks as self-check-out lanes distilled 6 human works into a single slightly more advanced job.

Now many white collar folk will receive similar treatment. One worker who can spend their time writing prompts and checking / tweaking GPT outputs may eliminate the need for numerous peers.

It doesn't eliminate the need for a human - but does change the nature of the work for those who remain.

Chopping a couple rungs off of the professional developmental ladder will have some interesting ramifications as well.


Yep. It doesn't have to replace everyone for it to cause serious problems, just enough to slash the white collar workforce in half.


Hopefully that frees up human potential innovation and those folks who get dropped figure out new methods of contributing value rather than sucking value out of the system.

This is where UBI comes into play. If love a UBI based on Proof of Work.
It's amazing how the work-less future never materializes. I have little doubt that the demand for human workers will both persist and continue to evolve in nature.


I always like how people post these things without a single but of awareness about what is really going on. It's pure copium that it will somehow work out for people. "Oh jobs will redefine themselves and we'll figure out something."

Oh really smart guy, by all means please tell what normal people will be able to do when these programs have replaced 3/4s of the white collar jobs? You really think they will need people for basic jobs when these stuff is going to start replacing lawyers and engineers, aka jobs that require significantly higher IQs on the whole?

We have the WEF already saying we won't need nearly as many people in the future. Perhaps they are right, but it's going to be a world of pain as society comes to grips with the fact that there won't be many options for people to actually work outside of blue collar skilled labor.

I cannot honestly believe people can look at these videos and think it will just work out for regular people.



It's getting > 90 percentile on all major standardized tests. Do I really need to spell this out? How can you be this naive? That's why I keep making these posts. You must learn to use this stuff or else you will be left far behind the curve.
tk for tu juan
How long do you want to ignore this user?
Not too late to stop before…
Rodney Ruxin
How long do you want to ignore this user?
Nanomachines son said:

Get Off My Lawn said:

Adverse Event said:

Nanomachines son said:

Get Off My Lawn said:

A bunch of white collar folks looked down on store checkout clerks as self-check-out lanes distilled 6 human works into a single slightly more advanced job.

Now many white collar folk will receive similar treatment. One worker who can spend their time writing prompts and checking / tweaking GPT outputs may eliminate the need for numerous peers.

It doesn't eliminate the need for a human - but does change the nature of the work for those who remain.

Chopping a couple rungs off of the professional developmental ladder will have some interesting ramifications as well.


Yep. It doesn't have to replace everyone for it to cause serious problems, just enough to slash the white collar workforce in half.


Hopefully that frees up human potential innovation and those folks who get dropped figure out new methods of contributing value rather than sucking value out of the system.

This is where UBI comes into play. If love a UBI based on Proof of Work.
It's amazing how the work-less future never materializes. I have little doubt that the demand for human workers will both persist and continue to evolve in nature.


I always like how people post these things without a single but of awareness about what is really going on. It's pure copium that it will somehow work out for people. "Oh jobs will redefine themselves and we'll figure out something."

Oh really smart guy, by all means please tell what normal people will be able to do when these programs have replaced 3/4s of the white collar jobs? You really think they will need people for basic jobs when these stuff is going to start replacing lawyers and engineers, aka jobs that require significantly higher IQs on the whole?

We have the WEF already saying we won't need nearly as many people in the future. Perhaps they are right, but it's going to be a world of pain as society comes to grips with the fact that there won't be many options for people to actually work outside of blue collar skilled labor.

I cannot honestly believe people can look at these videos and think it will just work out for regular people.



It's getting > 90 percentile on all major standardized tests. Do I really need to spell this out? How can you be this naive? That's why I keep making these posts. You must learn to use this stuff or else you will be left far behind the curve.

But can it tell me what a woman is?
whytho987654
How long do you want to ignore this user?
Med school match day was monday for residencies. Many classmates celebrated matching into radiology, today theyre all talking about if AI will take their job, as radiology is very lucrative. Personally, I think AI is over hyped. We still have accountants around, if you add value you will still be around.
whytho987654
How long do you want to ignore this user?
Nanomachines son said:

Get Off My Lawn said:

Adverse Event said:

Nanomachines son said:

Get Off My Lawn said:

A bunch of white collar folks looked down on store checkout clerks as self-check-out lanes distilled 6 human works into a single slightly more advanced job.

Now many white collar folk will receive similar treatment. One worker who can spend their time writing prompts and checking / tweaking GPT outputs may eliminate the need for numerous peers.

It doesn't eliminate the need for a human - but does change the nature of the work for those who remain.

Chopping a couple rungs off of the professional developmental ladder will have some interesting ramifications as well.


Yep. It doesn't have to replace everyone for it to cause serious problems, just enough to slash the white collar workforce in half.


Hopefully that frees up human potential innovation and those folks who get dropped figure out new methods of contributing value rather than sucking value out of the system.

This is where UBI comes into play. If love a UBI based on Proof of Work.
It's amazing how the work-less future never materializes. I have little doubt that the demand for human workers will both persist and continue to evolve in nature.


I always like how people post these things without a single but of awareness about what is really going on. It's pure copium that it will somehow work out for people. "Oh jobs will redefine themselves and we'll figure out something."

Oh really smart guy, by all means please tell what normal people will be able to do when these programs have replaced 3/4s of the white collar jobs? You really think they will need people for basic jobs when these stuff is going to start replacing lawyers and engineers, aka jobs that require significantly higher IQs on the whole?

We have the WEF already saying we won't need nearly as many people in the future. Perhaps they are right, but it's going to be a world of pain as society comes to grips with the fact that there won't be many options for people to actually work outside of blue collar skilled labor.

I cannot honestly believe people can look at these videos and think it will just work out for regular people.



It's getting > 90 percentile on all major standardized tests. Do I really need to spell this out? How can you be this naive? That's why I keep making these posts. You must learn to use this stuff or else you will be left far behind the curve.
Getting 90th percentile on tests and real world applications are two different ballgames. Just think of all the times you need to interact with a human for a problem asap when talking to automated banks and other services. Highly intelligent humans will still be needed, AI is limited by those who code it.
Adverse Event
How long do you want to ignore this user?
Compared to humanities known existence, we are nanoseconds into the emergence of AI. JUST REMEMBER THAT BEFORE YOU HANDWAVE IT AWAY.
What bitcoin’s detractors don’t understand is monetary economics, computer science, software engineering, network protocols, and electrical systems.

It ain't much, but it's honest Proof of Work.
whytho987654
How long do you want to ignore this user?
Adverse Event said:

Compared to humanities known existence, we are nanoseconds into the emergence of AI. JUST REMEMBER THAT BEFORE YOU HANDWAVE IT AWAY.
Not handwaving it away, experts in AI, not the clowns you see on twitter, say AI will be used as a tool. Chatcpt isnt even an AI, its an advanced search engine. Most AI software is not available to the general public. Itll make some fields obsolete, like engineering and finance, but fields where you are hands on and need a human will be around
Adverse Event
How long do you want to ignore this user?
whytho987654 said:

Adverse Event said:

Compared to humanities known existence, we are nanoseconds into the emergence of AI. JUST REMEMBER THAT BEFORE YOU HANDWAVE IT AWAY.
Not handwaving it away, experts in AI, not the clowns you see on twitter, say AI will be used as a tool. Chatcpt isnt even an AI, its an advanced search engine. Most AI software is not available to the general public. Itll make some fields obsolete, like engineering and finance, but fields where you are hands on and need a human will be around


Hands? Where we are going we won't need human hands


What bitcoin’s detractors don’t understand is monetary economics, computer science, software engineering, network protocols, and electrical systems.

It ain't much, but it's honest Proof of Work.
Page 2 of 3
 
×
subscribe Verify your student status
See Subscription Benefits
Trial only available to users who have never subscribed or participated in a previous trial.