Chat GPT AI

  • Guest, it's time once again for the massively important and exciting FoH Asshat Tournament!



    Go here and give us your nominations!
    Who's been the biggest Asshat in the last year? Give us your worst ones!

Mist

REEEEeyore
<Gold Donor>
31,198
23,374
Project Managers for intangibles should be sweating.
A computer can neither be accountable nor hold people accountable, so not really.

A computer can't annoy me in the same way a PM can. A computer isn't going to call up my VP if a deliverable is late. A computer isn't going to manage client expectations or negotiate between multiple external vendors.

That said, bad PMs for purely internal projects should be scared, but not really because of AI, just because they've been useless for a long time and everyone is onto them.
 
  • 1Like
Reactions: 1 user

Seananigans

Honorary Shit-PhD
<Gold Donor>
13,699
34,209
A computer can neither be accountable nor hold people accountable, so not really.

A computer can't annoy me in the same way a PM can. A computer isn't going to call up my VP if a deliverable is late. A computer isn't going to manage client expectations or negotiate between multiple external vendors.

That said, bad PMs for purely internal projects should be scared, but not really because of AI, just because they've been useless for a long time and everyone is onto them.

From my experience, the most time consuming (by far) functions of PMs for intangible projects are things that can be automated by a LLM AI like GPT.

Backed up by the AI doing the time consuming shit, the other functions can be performed by a senior engineer(whatever) working on the project.
 

Mist

REEEEeyore
<Gold Donor>
31,198
23,374
From my experience, the most time consuming (by far) functions of PMs for intangible projects are things that can be automated by a LLM AI like GPT.

Backed up by the AI doing the time consuming shit, the other functions can be performed by a senior engineer(whatever) working on the project.
An engineer can't hold people accountable either.
 
  • 1OK, Sure
Reactions: 1 user

Pasteton

Blackwing Lair Raider
2,733
1,919

I just read (but did not understand) the paper on this. Really mind blowing, why is this in the chat gpt thread? Should put this under ai image generation if nothing else.

need a genius to tell me if this would be an big leap, but from what I saw and from the pace of advancement in ai lately, I don’t think we are far from an ai scanning your brain to make videos and create worlds just from your envisioning them. Like just picture old Sebilis in your mind, and the ai will just start generating the dungeon that you’re picturing in unreal 5. And of course you won’t need prompting for stuff like midhourney anymore , the ai can just take what you’re imagining and create an image from it.


Granted there is still the minor inconvenience of having to be enclosed inside a claustrophobic and noisy 1 million dollar machine in order for the ai to get the data it needs but I’m nevertheless looking forward to being alive for this.

think in the future, game dev studios are gonna just be a few creative guys and an mri machine that they spend most of their time in once they have a concept in their minds
 
Last edited:
  • 2Like
Reactions: 1 users

Chris

Potato del Grande
19,436
-10,733
So was the last South Park episode really written with help from ChatGPT as claimed?
 

Pasteton

Blackwing Lair Raider
2,733
1,919

this guy kinda breaks paper down and even more impressive can only assume he’s doing it between gas pumping sessions somewhere off a Jersey turnpike. I can see a role in the future of being a good ‘fmri’ boy where people that know *how* to think or picture things in their minds eye are going to be in super high demand. Imaging if you are a strong fMRI candidate because of your ability to visualize things and create a strong fMRI signal, you could be attached to an mr which is rigged to a stable diffusion model and a 3D printer, just pumping out perfect fleshlights all day
 
  • 1Like
Reactions: 1 user

Mist

REEEEeyore
<Gold Donor>
31,198
23,374
Lol says who?
The joke is that no sane company lets engineers have any authority over people. It's the whole reason they invented titles like Principal Engineer, so they can promote engineers past management salaries without actually making them managers.
 

Mist

REEEEeyore
<Gold Donor>
31,198
23,374

this guy kinda breaks paper down and even more impressive can only assume he’s doing it between gas pumping sessions somewhere off a Jersey turnpike. I can see a role in the future of being a good ‘fmri’ boy where people that know *how* to think or picture things in their minds eye are going to be in super high demand. Imaging if you are a strong fMRI candidate because of your ability to visualize things and create a strong fMRI signal, you could be attached to an mr which is rigged to a stable diffusion model and a 3D printer, just pumping out perfect fleshlights all day

I think you're missing how this works. It's still extremely impressive but it's merely correlating the brain activity to brain activity that was observed against other similar images were trained.

So it knows what a cathedral looks like in the brain scan, from looking at brain scans of people looking at cathedrals, and then you show someone else a similar cathedral image, and it's able to correlate that and then draw an image similar to the cathedral images it was trained on.

It is not drawing shit directly out of their brain.
 

pharmakos

soʞɐɯɹɐɥd
<Bronze Donator>
16,305
-2,234
A computer can neither be accountable nor hold people accountable, so not really.

A computer can't annoy me in the same way a PM can. A computer isn't going to call up my VP if a deliverable is late. A computer isn't going to manage client expectations or negotiate between multiple external vendors.

That said, bad PMs for purely internal projects should be scared, but not really because of AI, just because they've been useless for a long time and everyone is onto them.
Yet*
 

pharmakos

soʞɐɯɹɐɥd
<Bronze Donator>
16,305
-2,234
I just read (but did not understand) the paper on this. Really mind blowing, why is this in the chat gpt thread? Should put this under ai image generation if nothing else.

need a genius to tell me if this would be an big leap, but from what I saw and from the pace of advancement in ai lately, I don’t think we are far from an ai scanning your brain to make videos and create worlds just from your envisioning them. Like just picture old Sebilis in your mind, and the ai will just start generating the dungeon that you’re picturing in unreal 5. And of course you won’t need prompting for stuff like midhourney anymore , the ai can just take what you’re imagining and create an image from it.


Granted there is still the minor inconvenience of having to be enclosed inside a claustrophobic and noisy 1 million dollar machine in order for the ai to get the data it needs but I’m nevertheless looking forward to being alive for this.

think in the future, game dev studios are gonna just be a few creative guys and an mri machine that they spend most of their time in once they have a concept in their minds
10-20 years away from what you're imagining, at MOST
 

pharmakos

soʞɐɯɹɐɥd
<Bronze Donator>
16,305
-2,234

this guy kinda breaks paper down and even more impressive can only assume he’s doing it between gas pumping sessions somewhere off a Jersey turnpike. I can see a role in the future of being a good ‘fmri’ boy where people that know *how* to think or picture things in their minds eye are going to be in super high demand. Imaging if you are a strong fMRI candidate because of your ability to visualize things and create a strong fMRI signal, you could be attached to an mr which is rigged to a stable diffusion model and a 3D printer, just pumping out perfect fleshlights all day

Philip K Dick stuff right here
 

pharmakos

soʞɐɯɹɐɥd
<Bronze Donator>
16,305
-2,234
I think you're missing how this works. It's still extremely impressive but it's merely correlating the brain activity to brain activity that was observed against other similar images were trained.

So it knows what a cathedral looks like in the brain scan, from looking at brain scans of people looking at cathedrals, and then you show someone else a similar cathedral image, and it's able to correlate that and then draw an image similar to the cathedral images it was trained on.

It is not drawing shit directly out of their brain.
The fact that this was the very first result they were able to achieve shows us that the possibilities are pretty mind boggling. So the AI needs to be trained on the same images we are trying to mentally picture? Just improve the training data, make it more robust. Eventually you're gonna hit a point where the computer can pretty much, with.... Idk at least 75% accuracy predict what we are thinking. Sure, it will never be perfect. But if it's better than a coin toss...
 

Mist

REEEEeyore
<Gold Donor>
31,198
23,374
The fact that this was the very first result they were able to achieve shows us that the possibilities are pretty mind boggling. So the AI needs to be trained on the same images we are trying to mentally picture? Just improve the training data, make it more robust. Eventually you're gonna hit a point where the computer can pretty much, with.... Idk at least 75% accuracy predict what we are thinking. Sure, it will never be perfect. But if it's better than a coin toss...
Not the same ones, but similar ones.
 
  • 1Like
Reactions: 1 user

Mist

REEEEeyore
<Gold Donor>
31,198
23,374
It's definitely extremely cyberpunk. With the right tech, you could skip right past the holodeck if this gets iterated on. If you could get an fMRI down to the size of a VR headset, it could probably read various sensorimotor responses and then render and feed you back images based on your inferred movements, giving the sensation of being in the machine.
 

Control

Ahn'Qiraj Raider
2,993
7,898
So it knows what a cathedral looks like in the brain scan, from looking at brain scans of people looking at cathedrals
If you could get an fMRI down to the size of a VR headset
you're gonna hit a point where the computer can pretty much, with.... Idk at least 75% accuracy predict what we are thinking.
People thought we were gonna get a social credit score. Instead, we'll actually get literal, real-time, thought policing. They'll probably also throw in some magnetic field generators for some additional "thought-steering" and maybe even a cow-puncher for some extra proper-thought motivation.

What a time to be alive!
 
  • 1Like
Reactions: 1 user

ToeMissile

Pronouns: zie/zhem/zer
<Gold Donor>
3,169
2,058
The most powerful tools are also the most powerful weapons?

or something like that.
 
  • 1Like
Reactions: 1 user

pharmakos

soʞɐɯɹɐɥd
<Bronze Donator>
16,305
-2,234
The most powerful tools are also the most powerful weapons?

or something like that.
That would mean that the most powerful weapons are also the most powerful tools. And I'm not sure you can build anything with a nuclear warhead. 🤔
 

ToeMissile

Pronouns: zie/zhem/zer
<Gold Donor>
3,169
2,058
That would mean that the most powerful weapons are also the most powerful tools. And I'm not sure you can build anything with a nuclear warhead. 🤔
Well maybe not a warhead specifically, but harnessing nuclear energy.
 
  • 1Like
Reactions: 1 user