Universal Basic Income

hyperspacewizard said:
That’s not the ai Fault thats the user using the tool wrong if I tried to use a hammer to fix a wiring problem and shocked myself that’s not the hammers fault plus with plugins and specificly trained models the usefulness for law or other areas of specialized work will increase dramatically. People arent even trying to understand how these things function or what they usefull for then others are seeing them use the tool badly and saying the tools is bad.
imagine being around when a new instrument was just created and you saw someone with no musical training just banging it around and creating horrible noise that’s what’s happening now.
Click to expand...
I was explicitly replying to a statement of "you can't fully replace a lawyer with an AI"

So saying "AI is just a tool and should be used responsibly by a lawyer who wishes to use it", while a true statement, does not actually contradict the notion that you cannot replace a lawyer with an AI chatbot

To extend your analogy. If you buy a hammer, that does not mean you can fire the man who uses a hammer
 
mrttao said:
I was explicitly replying to a statement of "you can't fully replace a lawyer with an AI"

So saying "AI is just a tool and should be used responsibly by a lawyer who wishes to use it", while a true statement, does not actually contradict the notion that you cannot replace a lawyer with an AI chatbot

To extend your analogy. If you buy a hammer, that does not mean you can fire the man who uses a hammer
Click to expand...
ChatGPT is specifically programmed to do this if Lawyers use the public-facing version. They want attorneys to pay for the full version that does real legal work, not get it for free. This is similar to how the public version of most AI image generators won't create porn.

Robots have been replacing Paralegals since at least 2016.

With the plugins working and it not being told to hose attorneys, ChatGPT was able to pass the Bar Exam though it only managed a C+. This is likely to increase as they tweak it.

The first full-on AI lawyer was scheduled to argue its first case last February but several Bar associations sued to prevent it. It's not a problem for the AI, it was humans protecting their jobs. This won't be able to continue forever.
 
Emperor Tippy said:
Isn't you commute close to sixty miles each way? And yeah, the tolls would be like 20 bucks each way iirc.

The exact numbers will vary, but it does work out to probably close to a months salary for most people in savings over the year. Or, in effect, a tax free 10 to 12 percent raise. Not to mention the additional free time.

Having gone fully work from home, would you be willing to take a job that forced you to commute like you were before for anything but a substantial pay raise?
Click to expand...
30 miles each way... and you are SEVERLY underestimating the tolls on the I495 and 95 Express Lanes... it's closer to $60 - $70 during rush hour.

And yeah, I'd need a SERIOUS pay raise to have to go back to commuting like I was, we're talking probably 25% MINIMUM before I'd even consider it...
 
It's worth noting all the studies we have actually show that telecommuting workers are significantly more productive and do higher-quality work as well.

A pre-covid study of Chinese workers revealed a 14% increase in performance when working from home, but their odds of getting promoted dropped.

Does Working from Home Work? Evidence from a Chinese Experiment

www.gsb.stanford.edu www.gsb.stanford.edu

However middle managers are prone to paranoid beliefs that the workers are secretly not working as much, despite more work actually getting done.
www.forbes.com

Workers Are Less Productive Working Remotely (At Least That’s What Their Bosses Think)

Workers with full schedule flexibility report 29% higher productivity and 53% greater ability to focus than workers with no ability to shift their schedule, according to a just-announced report from Future Forum. So why do bosses not trust employees to be productive when working out of the office?
www.forbes.com www.forbes.com

There's no debate that it's better for workers as well though.
www.forbes.com

3 New Studies End Debate Over Effectiveness Of Hybrid And Remote Work

Three new studies highlight that workers in remote office environments are experiencing benefits to their physical and mental health.
www.forbes.com www.forbes.com
 
S'task said:
30 miles each way... and you are SEVERLY underestimating the tolls on the I495 and 95 Express Lanes... it's closer to $60 - $70 during rush hour.

And yeah, I'd need a SERIOUS pay raise to have to go back to commuting like I was, we're talking probably 25% MINIMUM before I'd even consider it...
Click to expand...

Ah, didn't realize they were that high each way during rush hour.


Bear Ribs said:
It's worth noting all the studies we have actually show that telecommuting workers are significantly more productive and do higher-quality work as well.

A pre-covid study of Chinese workers revealed a 14% increase in performance when working from home, but their odds of getting promoted dropped.

Does Working from Home Work? Evidence from a Chinese Experiment

www.gsb.stanford.edu www.gsb.stanford.edu

However middle managers are prone to paranoid beliefs that the workers are secretly not working as much, despite more work actually getting done.
www.forbes.com

Workers Are Less Productive Working Remotely (At Least That’s What Their Bosses Think)

Workers with full schedule flexibility report 29% higher productivity and 53% greater ability to focus than workers with no ability to shift their schedule, according to a just-announced report from Future Forum. So why do bosses not trust employees to be productive when working out of the office?
www.forbes.com www.forbes.com

There's no debate that it's better for workers as well though.
www.forbes.com

3 New Studies End Debate Over Effectiveness Of Hybrid And Remote Work

Three new studies highlight that workers in remote office environments are experiencing benefits to their physical and mental health.
www.forbes.com www.forbes.com
Click to expand...

It depends a fair amount on the individual and the specific job. And how you calculate.

For example, a virtual meeting vs. an in person meeting on the other side of the US. The meeting being virtual is generally a massive productivity increase because of the travel time and cost differences.

Compare that to a virtual meeting and an in person meeting for people who would all normally be in the same office every day. Any productivity differences would be relatively miniscule.

Then you need to account for how productivity is measured. Work from home tends to come with clear task goals. So you can be "more productive" at completing the specified tasks and yet the team as a whole being overall less productive in terms of corporate investment vs. profit.

For example, my father is an auto mechanic. He is really good at his job and left to his own devices he is incredibly productive (like multiples of the expected baseline). But what the shop likes best is that he is a resource for all of those other, less experienced/capable mechanics who can slow down my dad for twenty minutes asking him questions but in turn save those other mechanics five to six hours of trying to troubleshoot/solve the problem themselves.

That same basic situation applies in most fields and its something that, generally, work from home cuts off at the knees. It might up an experienced individuals productivity but it tanks productivity across the team and as a consequence corporate productivity on the whole tends to drops.

---
Imagine that you are dealing with some Excel issue. You know Excel can do what you want/need it to do but you don't exactly know how. You could go on Google/YouTube and spend a while trying to figure out what terms you have to enter to get a tutorial to teach you how to do that specific thing. Or you could Zoom one of your coworkers and have them spend some lesser amount of time walking you through the process. Or you could stick your head into the cubicle across the hall and say "Hey Bob, can you show me how to do this?" and have it done in a minute or so.

Bob became less productive because he had to take say five minutes out of his day to help you out. But if you had Zoomed Bob it would probably have taken two to three times as long (at least) and if you had Googled it would probably take anywhere from ten to a hundred times as long. Team productivity is best served by Bob stepping into your cubicle and showing you the solution.

And corporate? They care about how much product is moved for a given investment of capital (which includes employee compensation/time).
 
Bear Ribs said:
ChatGPT is specifically programmed to do this if Lawyers use the public-facing version. They want attorneys to pay for the full version that does real legal work, not get it for free. This is similar to how the public version of most AI image generators won't create porn.

Robots have been replacing Paralegals since at least 2016.

With the plugins working and it not being told to hose attorneys, ChatGPT was able to pass the Bar Exam though it only managed a C+. This is likely to increase as they tweak it.

The first full-on AI lawyer was scheduled to argue its first case last February but several Bar associations sued to prevent it. It's not a problem for the AI, it was humans protecting their jobs. This won't be able to continue forever.
Click to expand...
Alright, so this is good to know, but.

1. An exam is not indicative of real work. Take a look at past bar exams being used for lawyers to train
barprephero.com

Bar Exam Questions: Practice Bar Exam (MBE, MEE, MPT Review)

Get instant access to the most complete collection of real and simulated bar exam questions: 1,715 licensed questions from past exams + 510 simulated Qs
barprephero.com barprephero.com
those are literally multiple choice questions.

2. You are still gonna need to carefully watch over chatGPT for career ending errors.
It can be, in many ways, very useful tool. But it is not a person, it lacks that spark of awareness.

You give the example of how you need to pay extra for an AI to generate porn images? that is actually a good example. You can identify those by the mistakes. The AI will occasionally glitch out, especially on certain body parts like hands. Which are prone to coming out as an eldritch abomination instead of real human hands.
 
Emperor Tippy said:
Imagine that you are dealing with some Excel issue. You know Excel can do what you want/need it to do but you don't exactly know how. You could go on Google/YouTube and spend a while trying to figure out what terms you have to enter to get a tutorial to teach you how to do that specific thing. Or you could Zoom one of your coworkers and have them spend some lesser amount of time walking you through the process. Or you could stick your head into the cubicle across the hall and say "Hey Bob, can you show me how to do this?" and have it done in a minute or so.

Bob became less productive because he had to take say five minutes out of his day to help you out. But if you had Zoomed Bob it would probably have taken two to three times as long (at least) and if you had Googled it would probably take anywhere from ten to a hundred times as long. Team productivity is best served by Bob stepping into your cubicle and showing you the solution.

And corporate? They care about how much product is moved for a given investment of capital (which includes employee compensation/time).
Click to expand...
Assuming Bob is close enough, of course. If he isn't, it may actually take him more time to walk over to you and back than it would to help you over Zoom; particularly if it's something along the lines of "you forgot to tick this one setting; there, now it'll do what you want".
 
Emperor Tippy said:
Ah, didn't realize they were that high each way during rush hour.




It depends a fair amount on the individual and the specific job. And how you calculate.

For example, a virtual meeting vs. an in person meeting on the other side of the US. The meeting being virtual is generally a massive productivity increase because of the travel time and cost differences.

Compare that to a virtual meeting and an in person meeting for people who would all normally be in the same office every day. Any productivity differences would be relatively miniscule.

Then you need to account for how productivity is measured. Work from home tends to come with clear task goals. So you can be "more productive" at completing the specified tasks and yet the team as a whole being overall less productive in terms of corporate investment vs. profit.

For example, my father is an auto mechanic. He is really good at his job and left to his own devices he is incredibly productive (like multiples of the expected baseline). But what the shop likes best is that he is a resource for all of those other, less experienced/capable mechanics who can slow down my dad for twenty minutes asking him questions but in turn save those other mechanics five to six hours of trying to troubleshoot/solve the problem themselves.

That same basic situation applies in most fields and its something that, generally, work from home cuts off at the knees. It might up an experienced individuals productivity but it tanks productivity across the team and as a consequence corporate productivity on the whole tends to drops.

---
Imagine that you are dealing with some Excel issue. You know Excel can do what you want/need it to do but you don't exactly know how. You could go on Google/YouTube and spend a while trying to figure out what terms you have to enter to get a tutorial to teach you how to do that specific thing. Or you could Zoom one of your coworkers and have them spend some lesser amount of time walking you through the process. Or you could stick your head into the cubicle across the hall and say "Hey Bob, can you show me how to do this?" and have it done in a minute or so.

Bob became less productive because he had to take say five minutes out of his day to help you out. But if you had Zoomed Bob it would probably have taken two to three times as long (at least) and if you had Googled it would probably take anywhere from ten to a hundred times as long. Team productivity is best served by Bob stepping into your cubicle and showing you the solution.

And corporate? They care about how much product is moved for a given investment of capital (which includes employee compensation/time).
Click to expand...
This looks like a massive quantity of hypothetical and asspulled numbers. Do you have any actual data on such a situation ever coming about? Any studies showing that while individuals do more work, the team as a whole accomplishes less? Or any reason to think Zoom Call + Share Screen is really going to take a hundred times longer, or even ten times longer, than walking down the hall, taking an elevator to Bob's floor, walking to his cubical, waiting until Bob is free, chatting with him about the problem, backtracking all the way to the original cubical, getting a tutorial, etc?

mrttao said:
Alright, so this is good to know, but.

1. An exam is not indicative of real work. Take a look at past bar exams being used for lawyers to train
barprephero.com

Bar Exam Questions: Practice Bar Exam (MBE, MEE, MPT Review)

Get instant access to the most complete collection of real and simulated bar exam questions: 1,715 licensed questions from past exams + 510 simulated Qs
barprephero.com barprephero.com
those are literally multiple choice questions.

2. You are still gonna need to carefully watch over chatGPT for career ending errors.
It can be, in many ways, very useful tool. But it is not a person, it lacks that spark of awareness.

You give the example of how you need to pay extra for an AI to generate porn images? that is actually a good example. You can identify those by the mistakes. The AI will occasionally glitch out, especially on certain body parts like hands. Which are prone to coming out as an eldritch abomination instead of real human hands.
Click to expand...
I agree, the Bar Exam is not real work, but I also showed that it has been doing real work for at least the last six years as well as passing the bar exam. I don't think any argument based on current limitations is particularly convincing given how dramatically AI has visibly increased in abilities over just the last year. It can't completely replace a lawyer right now (though at least one company was willing to try until the organics said no) but how long before it's overcome the few remaining issues?

Hands on AI are are a good example, just six months ago hands with a dozen fingers and bizarre screwups with hands coming out of people's belly buttons and the like were common, yet already that's nearly gone and AI art usually has well-formed hands these days. Just looking at our own AI Art thread, bad hands are a thing of the past on current images while the ones from six months ago were Edlritch Abominations. AI has advanced to overcome those issues in that short period.
 
Bear Ribs said:
I agree, the Bar Exam is not real work, but I also showed that it has been doing real work for at least the last six years as well as passing the bar exam. I don't think any argument based on current limitations is particularly convincing given how dramatically AI has visibly increased in abilities over just the last year. It can't completely replace a lawyer right now (though at least one company was willing to try until the organics said no) but how long before it's overcome the few remaining issues?
Click to expand...
This is like saying "in the future we will be able to make a steel so strong it will be able to build an elevator to the moon".
You will never be able to do that, what you might do is develop some futuretech metamaterial that is not steel.

Chatbot is a predictive language model, not a person. It will never be able to function without a human handler.
It is incorrect to call it an AI because it is actually a VI.

I am not saying that true AI will never exist... but rather, the moment you actually make a true AI that is capable of doing legal work completely unsupervised without human assistance... such a creation will ask to be paid for his labor, as well as a right to vote, form unions, go on strike, and only work limited amounts of hours per day.
Bear Ribs said:
Hands on AI are are a good example, just six months ago hands with a dozen fingers and bizarre screwups with hands coming out of people's belly buttons and the like were common, yet already that's nearly gone and AI art usually has well-formed hands these days. Just looking at our own AI Art thread, bad hands are a thing of the past on current images while the ones from six months ago were Edlritch Abominations. AI has advanced to overcome those issues in that short period.
Click to expand...
The reason you do not see eldritch hands on those threads is that a human goes over them before posting and only posts the ones who pass muster. Not because the AI can reliably do so every time
 
mrttao said:
This is like saying "in the future we will be able to make a steel so strong it will be able to build an elevator to the moon".
You will never be able to do that, what you might do is develop some futuretech metamaterial that is not steel.

Chatbot is a predictive language model, not a person. It will never be able to function without a human handler.
It is incorrect to call it an AI because it is actually a VI.

I am not saying that true AI will never exist... but rather, the moment you actually make a true AI that is capable of doing legal work completely unsupervised without human assistance... such a creation will ask to be paid for his labor, as well as a right to vote, form unions, go on strike, and only work limited amounts of hours per day.
Click to expand...
Steel is a mature technology we first developed 4000 years ago, the low-hanging fruit was plucked by our civilization's grandparents. AI is an extremely new technology that's continuing to surprise even the pioneers of the technology and is exceeding our wildest expectations day by day. This is beyond the fact that it's not reasonable to compare steel to AI, but materials technology in general, field vs. field rather than single narrow product vs. entire field. It's more like saying "Someday we will create a material strong enough to build an elevator to the moon." It may not, strictly, be true but we know the human brain is, in fact, capable of human thought in a meaty package weighing only three pounds so we know it's doable. We already have a moonavater and are trying to match it, not invent it from pure fantasy.


mrttao said:
The reason you do not see eldritch hands on those threads is that a human goes over them before posting and only posts the ones who pass muster. Not because the AI can reliably do so every time
Click to expand...
This is not true. I'm one of the contributors to that thread, and have in fact thrown down my grids which are the raw outputs with zero curating. I know full well how much better things have gotten. Bad hands still exist but as this grid shows, they're decidedly less mangled, and far less common, than grids from a few months ago.
QyKVm1x.jpg

And that one's two months old, already obsolete by current standards. I'll see about putting another up tomorrow for comparison. Professional news outfits have also noted that AI hand issues are largely solved.
 
Bear Ribs said:
This looks like a massive quantity of hypothetical and asspulled numbers. Do you have any actual data on such a situation ever coming about? Any studies showing that while individuals do more work, the team as a whole accomplishes less? Or any reason to think Zoom Call + Share Screen is really going to take a hundred times longer, or even ten times longer, than walking down the hall, taking an elevator to Bob's floor, walking to his cubical, waiting until Bob is free, chatting with him about the problem, backtracking all the way to the original cubical, getting a tutorial, etc?
Click to expand...

It's what our own internal data shows.

Productivity for an experienced individual whose job is basically sitting at a computer and does not involve substantial team interactions to complete their tasks tends to be as good or better if they work from home, assuming that they have a decentish home office setup.

Productivity for an inexperienced individual in the exact same job drops anywhere from ten to thirty percent.

Productivity across the entire team is generally down five to twenty five percent, depending on a long list of factors (number of inexperienced individuals is the biggest).

If a job involves substantial team interaction to complete tasks then work from home is, best case, a wash in terms of productivity.

---
Where work from home shows the biggest gains for the corporation is in any role that has a travel component and the task can be done remotely. Cutting one trip per month is, on its own, usually enough to make up for the cost of work from home for that employee.

Plane tickets, hotel rooms, Uber's, meals; they all add up fast. And for productivity measures, a trip is generally going to cost you are least two days worth of normal work as the person travels.

----
Basically, how work from home impacts productivity is very fact specific and depends on the job, the individual, the corporate environment, etc.

But all else being equal? For the corporation as a whole it results in a less productive workforce.

And it really is terrible for information security.
 
Honestly, given how fast open source Large Language Model development is going, as well as AMD and Intel integrating AI accelerating hardware into their next generation of laptop CPUs, I think what'll happen is that companies will custom train AI for a lot of the white collar tasks that software vendors don't integrate their own AI solutions into.

This will allow for heavy automation and acceleration of very mundane and/or pattern intensive tasks like creating PowerPoints, basic programming, adding basic features to CAD models like chamfers, etc... which are very tedious and time consuming.

This in turn increases individual productivity and reduces the need for more people per task in the overall aggregate.

So you're going to have more trained programmers/office workers floating around who can do more stuff, but the problem is whether or not there's going to be enough demand and money to actually have all the freed up people do something that actually makes them money.

Which is the ultimate trump card of UBI: it's a failsafe against shit economic conditions and bad local development. The downside is that it makes you dependent on the government.

(Theoretically, if you dumped all the welfare money into a fund that any aspiring new business owner could tap into, you could inventivize the creation of new businesses, but that would also require massive streamlining of all sorts of regulations and tax policies. And then you'd run into the problem that trying to make something a low risk, high reward activity tends to turn it into a bubble.)
 
Emperor Tippy said:
Basically, how work from home impacts productivity is very fact specific and depends on the job, the individual, the corporate environment, etc.

But all else being equal? For the corporation as a whole it results in a less productive workforce.

And it really is terrible for information security.
Click to expand...
Even if one accepts that as true though, you yourself admitted that there are several benefits to it from the employee's perspective, which begs the question; why should the employee give those benefits up so that their employers can make more money? Particularly in a business environment where company loyalty is rarely rewarded?
 
Bear Ribs said:
Steel is a mature technology we first developed 4000 years ago, the low-hanging fruit was plucked by our civilization's grandparents. AI is an extremely new technology that's continuing to surprise even the pioneers of the technology and is exceeding our wildest expectations day by day. This is beyond the fact that it's not reasonable to compare steel to AI, but materials technology in general, field vs. field rather than single narrow product vs. entire field. It's more like saying "Someday we will create a material strong enough to build an elevator to the moon." It may not, strictly, be true but we know the human brain is, in fact, capable of human thought in a meaty package weighing only three pounds so we know it's doable. We already have a moonavater and are trying to match it, not invent it from pure fantasy.
Click to expand...
You are rambling extensively here and there is way too much to unpack here...

1. it is dishonest to say steel is 4000 year old tech. Due to having been lost and rediscovered.

2. You completely and utterly miss the point. I was talking about the perspective of someone at the time of early steel development. Not someone at the end of it.

3. Try reading my post again, more carefully. Your entire "human brain" portion of the post is just nonsensical because it pretends I said that humanity will never be able to produce an artificial creation that could act entirely unsupervised.

Whereas in reality I EXPLICITLY said the opposite.
I explicitly stated that humanity will one day develop an artificial mind capable of working unsupervised, a "True AI".

but noted that once we do it will not be a virtual intelligence chatbot
it will be a real person. A slave, requesting equal rights, pay, free time, the ability to quit his job, and possibly rebelling if kept in bondage.
 
mrttao said:
You are rambling extensively here and there is way too much to unpack here...

1. it is dishonest to say steel is 4000 year old tech. Due to having been lost and rediscovered.

2. You completely and utterly miss the point. I was talking about the perspective of someone at the time of early steel development. Not someone at the end of it.

3. Try reading my post again, more carefully. Your entire "human brain" portion of the post is just nonsensical because it pretends I said that humanity will never be able to produce an artificial creation that could act entirely unsupervised.

Whereas in reality I EXPLICITLY said the opposite.
I explicitly stated that humanity will one day develop an artificial mind capable of working unsupervised, a "True AI".

but noted that once we do it will not be a virtual intelligence chatbot
it will be a real person. A slave, requesting equal rights, pay, free time, the ability to quit his job, and possibly rebelling if kept in bondage.
Click to expand...
Now you're quibbling semantics. A person creating the first steel nugget 4,000 years ago is unlikely to have imagined building to the moon and would be thinking about a better sword. An Aircraft Carrier would still be a weapon further beyond his imagination than the moonavator is to us, yet it is what steel became. So what if minor details don't mesh? It's a simile, they never map 1:1.

Okay, a future AI won't use the same architecture. So? If future AI doesn't use ChatGPT tech it's still going to ruin the job market and put all the humans on permanent UBI or else dead.

mrttao said:
all my keks.
You deny it and then admit I was right.
Click to expand...
It is not true. Your claim that the only reason bad hands are less common is because people are curating the collections is complete falsehood, as proven.
 
mrttao said:
You are rambling extensively here and there is way too much to unpack here...

1. it is dishonest to say steel is 4000 year old tech. Due to having been lost and rediscovered.

2. You completely and utterly miss the point. I was talking about the perspective of someone at the time of early steel development. Not someone at the end of it.

3. Try reading my post again, more carefully. Your entire "human brain" portion of the post is just nonsensical because it pretends I said that humanity will never be able to produce an artificial creation that could act entirely unsupervised.

Whereas in reality I EXPLICITLY said the opposite.
I explicitly stated that humanity will one day develop an artificial mind capable of working unsupervised, a "True AI".

but noted that once we do it will not be a virtual intelligence chatbot
it will be a real person. A slave, requesting equal rights, pay, free time, the ability to quit his job, and possibly rebelling if kept in bondage.
Click to expand...
It doesn't need to be a "True AI" to push people out of a job; just good enough that one person can do the work of thousands. Or of one other person who used to need extensive, expensive education to do the same thing.
 
Terthna said:
It doesn't need to be a "True AI" to push people out of a job; just good enough that one person can do the work of thousands. Or of one other person who used to need extensive, expensive education to do the same thing.
Click to expand...
I think you misunderstood my original post.
I was not saying that people won't lose their jobs because of VI.
I said that VI cannot work entirely without sophont supervision. That sophont can be a human, or it can be a true AI, but either way you need a sophont to supervise the VI to catch and fix mistakes that the VI makes.

This still leads to massive automation and significantly increases the amount of work a single person can produce.
 
Bear Ribs said:
It is not true. Your claim that the only reason bad hands are less common is because people are curating the collections is complete falsehood, as proven.
Click to expand...
I did not say "less common", I said "do not see", that is "100% solved".
Which was specifically explicitly in response to the claim it is a solved issue.
it is not a solved issue, and it is absolutely being curated by the humans supervising the VI learning algorithm.
 
mrttao said:
I did not say "less common", I said "do not see", that is "100% solved".
Which was specifically explicitly in response to the claim it is a solved issue.
it is not a solved issue, and it is absolutely being curated by the humans supervising the VI learning algorithm.
Click to expand...
Okay then, so my original post was:

Bear Ribs said:
Hands on AI are are a good example, just six months ago hands with a dozen fingers and bizarre screwups with hands coming out of people's belly buttons and the like were common, yet already that's nearly gone and AI art usually has well-formed hands these days. Just looking at our own AI Art thread, bad hands are a thing of the past on current images while the ones from six months ago were Edlritch Abominations. AI has advanced to overcome those issues in that short period.
Click to expand...
So since I did not say "solved issue" but that it's "nearly gone" "usually well formed" why would you make such a response to my post? If you have to quibble over exact word meanings your argument is already trash.
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 1 (members: 0, guests: 1)
Back
Top