ChatGPT

[Uncle Rico]

It's so frequently wrong, though.

That just means it could fly through Marquette's dental school

[Scoop Snoop]

It's so frequently wrong, though.

So? Considering the decisions that we have had to deal with for the past few months, I doubt that AI could possibly do worse.

Forgetful forgot that the AI developed college courses with AI gradable assessments could be easily handled by students using AI for the exams.

[jesmu84]

Resurrecting this from the archives of Scoop, instead of creating a new thread.

OpenAI released new models this week that function at a PhD level. In all honesty, AI could reliably replace most jobs right now.

Curious as to what people on here think will be the first jobs to go, and which ones are the hardest to replace.

Personally, in medicine, at this point I'd rather be able to work with a chatbot regarding my symptoms, which could then prioritize and order tests (i.e. blood tests, imaging, appointment with experts), then find the most convenient nearby location to provide those services, than I would to work with a primary care physician.

Most business related tasks, could be automated with a much better digital paper trail, and audit record, than by humans.

And in education, AI with proper guidance, can build entire courses, including AI-gradeable assessments.

What do others on here think? Doom?

So, what do the citizens of this AI-driven economy do? How do they get paid?

[Jockey]

That is not a concern for these people. Too much $$$ to be made.

Same as it ever was. Same as it ever WAS.

[tower912]

That is not a concern for these people. Too much $$$ to be made.

Same as it ever was. Same as it ever WAS.

Jb's favorite line is ...
Look where my hand was....

[NCMUFan]

If AI has access to super intellectual sites like MUSCoop, I have to think no job is safe.

[Scoop Snoop]

If AI has access to super intellectual sites like MUSCoop, I have to think no job is safe.

Rocky is toast!

[mu_hilltopper]

ChatGPT was released to the public 29 *months* ago. 

While I've seen firsthand AI errors .. it's getting better, literally every minute.  12-18-24 months from now it'll be incredibly accurate.  AI is literally improving AI, relentlessly getting a tiny fraction better 24 hours a day, 7 days a week.

It's inevitable that AI will make workers enormously more productive and efficient, and that will come at a large and increasing loss of human employment.

In the short-term you won't lose your job to AI, but you will lose your job to someone who knows how to use AI better than you.

In the long-term, from capital's perspective, the absolute worst thing is employing humans.  They're hard to train, costly, make mistakes, whine a lot, and quit on you. Add in robotics, and it's only a matter of time before each McDonald's is run by one person. 

Anyone who doesn't see a massive amount of doom coming our way isn't paying attention. 

Capitalism, optimization, and innovation are relentless. 

[forgetful]

It's so frequently wrong, though.

Depends on the AI algorithm, these newer iterations are quite accurate. More accurate than your average employee.

[forgetful]

So? Considering the decisions that we have had to deal with for the past few months, I doubt that AI could possibly do worse.

Forgetful forgot that the AI developed college courses with AI gradable assessments could be easily handled by students using AI for the exams.

Actually I didn't. Some of the articles I was reading right before making my post talked about this specific conundrum.

There are two other aspects of this problem though:

1. Many states are going to standard curriculum regardless of school. It is more cost effective and equal quality of materials to use AI.

2. Students can already use AI to cheat their way through most college courses. An AI course would be no better or worse in that regards.

[Uncle Rico]

ChatGPT was released to the public 29 *months* ago. 

While I've seen firsthand AI errors .. it's getting better, literally every minute.  12-18-24 months from now it'll be incredibly accurate.  AI is literally improving AI, relentlessly getting a tiny fraction better 24 hours a day, 7 days a week.

It's inevitable that AI will make workers enormously more productive and efficient, and that will come at a large and increasing loss of human employment.

In the short-term you won't lose your job to AI, but you will lose your job to someone who knows how to use AI better than you.

In the long-term, from capital's perspective, the absolute worst thing is employing humans.  They're hard to train, costly, make mistakes, whine a lot, and quit on you. Add in robotics, and it's only a matter of time before each McDonald's is run by one person. 

Anyone who doesn't see a massive amount of doom coming our way isn't paying attention. 

Capitalism, optimization, and innovation are relentless. 

Wrong.  With the onslaught of onshoring heading our way, there shouldn't be any reason to worry.  At least in America

[forgetful]

ChatGPT was released to the public 29 *months* ago. 

While I've seen firsthand AI errors .. it's getting better, literally every minute.  12-18-24 months from now it'll be incredibly accurate.  AI is literally improving AI, relentlessly getting a tiny fraction better 24 hours a day, 7 days a week.

It's inevitable that AI will make workers enormously more productive and efficient, and that will come at a large and increasing loss of human employment.

In the short-term you won't lose your job to AI, but you will lose your job to someone who knows how to use AI better than you.

In the long-term, from capital's perspective, the absolute worst thing is employing humans.  They're hard to train, costly, make mistakes, whine a lot, and quit on you. Add in robotics, and it's only a matter of time before each McDonald's is run by one person. 

Anyone who doesn't see a massive amount of doom coming our way isn't paying attention. 

Capitalism, optimization, and innovation are relentless. 

These align with my thoughts. And the incredibly rapid progression and accuracy of models is mind boggling.

Open AI's 3o model (released recently) is quite powerful in terms of accuracy. And if you know which models are most efficient at different tasks, you can be extremely efficient and accurate, well surpassing your average person.

There's even more coming down the pipeline soon. Doom, is a possibility.

And as Jesmu84 commented, what do you do about the loss of jobs? I know many say that every time there was a technological innovation, lost jobs were replaced by new ones we were unaware of.

It really feels like this evolution of technology is different. You can use AI to write code, test and refine that code, and implement it once it reaches a threshold right now, that is equal to or exceeds most coders (you can train it on the best examples). The same is possible for developing new AI algorithms or implementing algorithms for specific tasks. So when are the individuals who know how to use AI even partially made obsolete?

[mu_hilltopper]

To add to the doom ..

One human "skill" that some have suggested is important is the ability to prompt AI the right way to give you good answers.

Except now?  You can ask AI "I need to do X.  Ask me a few questions about what I'm trying to do, then give me an amazing prompt to use, so AI gives me a great answer."

There goes one more human competitive advantage.

[jesmu84]

To add to the doom ..

One human "skill" that some have suggested is important is the ability to prompt AI the right way to give you good answers.

Except now?  You can ask AI "I need to do X.  Ask me a few questions about what I'm trying to do, then give me an amazing prompt to use, so AI gives me a great answer."

There goes one more human competitive advantage.

Too bad the white house economic/tarrif team didn't know about this shortcut