Social War Room Lounge Thread #325: PotWR Edition

Status
Not open for further replies.
Mr Holmes said:
I just finished listening to Huang on JRE. As he put it, AI will replace tasks, and the only people that need to be overly concerned are the ones where their job is a task e.g. data entry.
Click to expand...
I do think there are always concerns about how high the barrier is for an entry level worker. Like employers already don’t like having to invest the time in training due to lack of experience but technology also makes it harder because more and more grunt work gets automated so they have to be expected to do something higher level right off the bat. I don’t know how that will pan out but that’s a more believable scenario to me.
 
Some already said it but I do notice a recurring thing with these LLMs where some people test it out and see the mistakes it makes or getting to the completely wrong answer and see the current limitation. Then others praise it but don’t really have the use cases of it doing the world bending stuff they are excited about.

Like sometimes I think it’s both people wrote a prompt and pressed enter. One saw it generate all the text or file, gasped, stopped there and called it a day cause they thought the work was done. The other didn’t just see some output on the surface level but actually reviewed it and saw the errors, tried different ideas and ultimately saw the limitations and risks of what it currently does. It’s definitely a cool tool and it’s helped me in some daily tasks but they are very rigid / narrow currently.

I would like to see what the progress of operator shows in completely redundant work. Like there something in a spreadsheet called a macro where you record all the clicks, things you type, etc and then once compelted, you can do all those actions with one shortcut key. Again, it’s something that you need to be redundant for it to continually save you time. Anyways, I always wanted a macro for the operating system itself. Like why can’t my computer record, I went to x site, clicked y button, downloaded z output of page b and then just repeat that in the future for me. That seems like what operator and some of these other ones are getting to which would be a real time saver.

The other route is more connections between systems and the AI companies product. Like there’s so much mapping from company to company but it definitely seems like it will add value over time since a lot of that isn’t set up yet.

But overall automation of work, that hasn’t come close to being proven with this yet. And with LLMs using the internet, I think we may see a phase where it becomes a problem that it’s unknowingly reading AI content again to continue training it.

I definitely don’t shit on AI. I just don’t think we can arrive at the grand claims until they are actually here.
 
Limbo Pete said:
Disrupted? The word you used previously was "irrelevant". How exactly is it you think AI is going to make K12 education irrelevant? And that's even allowing for the weird "traditional" qualifier.
Click to expand...
Traditional Education(public k-12) will become irrelevant with the advent of ASI.

At the basic level traditional schooling is designed to do a few things. It serves to standardize knowledge at scale, train people for industrial/common place jobs/vocations, ensure that the pace of learning is centered around the average student and prepare people for a predictable labor market.

When we have ASI that entire model is irrelevant because all knowledge will be instantly and universally accessible. We’re essentially already here in this one. Almost all routine cognitive labor will be automated. The concept of an average student is completely outdated when you can teach at the level in each subject where the student is and labor markets will essentially cease to exist as we know them.

With AI you’ll have personalization, relevance to the student, better speed and outcomes and actual feedback.

Our model is study now and work later. When there isn’t work(as we know it) and credentials essentially will be meaningless since you’ll have the equivalent of 1000 PhDs at your disposal at any point, you won’t “need” school in the traditional sense.

It’s like the calculator. Long form math is essentially worthless now because we have a calculator. Education and credentials at large will be the same as long form math when you can ask your force of agents to solve anything better, more robustly, more accurately and faster than you ever could.

How you can’t see this inevitably is beyond me.

My 5 year old goes to a non-traditional AI centered school. In the school. There are 8 year olds who are at an 7th grade reading level and 5th grade math level. They learn at their own pace.


Here are examples of this school system. All the kids outpace publicly educated children.





But yeah, traditional education where nothing has changed in 50 years will definitely continue to be useful.
 
Rob Battisti said:
When we have ASI that entire model is irrelevant because all knowledge will be instantly and universally accessible.
Click to expand...
Couple of problems here, hoss. Well, more than a few lol. Not even dealing with the biggest, let's just stop for a second and consider that A) knowledge isn't static and B) it's not even remotely close to being captured by the internet lol
Aye carumba
 
Lead said:
Some already said it but I do notice a recurring thing with these LLMs where some people test it out and see the mistakes it makes or getting to the completely wrong answer and see the current limitation. Then others praise it but don’t really have the use cases of it doing the world bending stuff they are excited about.

Like sometimes I think it’s both people wrote a prompt and pressed enter. One saw it generate all the text or file, gasped, stopped there and called it a day cause they thought the work was done. The other didn’t just see some output on the surface level but actually reviewed it and saw the errors, tried different ideas and ultimately saw the limitations and risks of what it currently does. It’s definitely a cool tool and it’s helped me in some daily tasks but they are very rigid / narrow currently.

I would like to see what the progress of operator shows in completely redundant work. Like there something in a spreadsheet called a macro where you record all the clicks, things you type, etc and then once compelted, you can do all those actions with one shortcut key. Again, it’s something that you need to be redundant for it to continually save you time. Anyways, I always wanted a macro for the operating system itself. Like why can’t my computer record, I went to x site, clicked y button, downloaded z output of page b and then just repeat that in the future for me. That seems like what operator and some of these other ones are getting to which would be a real time saver.

The other route is more connections between systems and the AI companies product. Like there’s so much mapping from company to company but it definitely seems like it will add value over time since a lot of that isn’t set up yet.

But overall automation of work, that hasn’t come close to being proven with this yet. And with LLMs using the internet, I think we may see a phase where it becomes a problem that it’s unknowingly reading AI content again to continue training it.

I definitely don’t shit on AI. I just don’t think we can arrive at the grand claims until they are actually here.
Click to expand...
I don’t think anyone is saying we’re here. I’m saying we’re definitely going to get THERE.

Have you explored using Agents?
 
Limbo Pete said:
Couple of problems here, hoss. Well, more than a few lol. Not even dealing with the biggest, let's just stop for a second and consider that A) knowledge isn't static and B) it's not even remotely close to being captured by the internet lol
Aye carumba
Click to expand...
You’re really dense. Obviously it won’t be static. The pace in which ASI will be able to ingest knowledge and…
You know what, I’m wasting my time.

For whatever reason you feel compelled to ignore reality.

You’ve don’t nothing but be cynical and offered nothing by way of a retort in which you describe what life will be like with ASI.

By all means, have fun being an ignorant, arrogant cynic. Me? I’m preparing as best I can to get ready to live in this new world. Make as much money as I can now to escape from the race before the time comes and it’s too late.
 
Last edited:
Damn @Limbo Pete knows to get me rustled. For the sake of argument, what is something taught in traditional public k-12 schools which is not available to learn on the internet?
 
Rob Battisti said:
You’re really dense. Obviously it won’t be static. The pace in which ASI will be able to ingest knowledge and…
You know what, I’m wasting my time.

For whatever reason you feel compelled to ignore reality.

You’ve don’t nothing but be cynical and offered nothing by way of a retort in which you describe what life will be like with ASI.

By all means, have fun being an ignorant, arrogant cynic. Me? I’m preparing as best I can to get ready to live in this new world. Make as much money as I can now to escape from the race when the time comes and it’s too late.
Click to expand...
The reality in which your Grokzilla replaces human cognition? The reality where the Cylons conquer Earth? You're not only delusional, but really super ignorant. It's common with people that have that wonderful mix of intellectual arrogance, intellectual insecurity, and plain ol narcissism. Is this lingering from, like, not doing well in school or something? Question: How are you supposed to use AI to steal or plagiarize something that isn't digitally accessible?
 
Rob Battisti said:
I don’t think anyone is saying we’re here. I’m saying we’re definitely going to get THERE.

Have you explored using Agents?
Click to expand...
Yea, which is fine. I just don’t think there’s much weight in actually knowing or not knowing that’s the case, or at least how exaggerated the timeline might be.

My company has some limitations as to what types of AI we can leverage. I don’t have much downtime for personal use but used operator there.
 
Lead said:
Yea, which is fine. I just don’t think there’s much weight in actually knowing or not knowing that’s the case, or at least how exaggerated the timeline might be.

My company has some limitations as to what types of AI we can leverage. I don’t have much downtime for personal use but used operator there.
Click to expand...
I think that’s fair. It’s not like I’m saying it’s going to happen tomorrow.

I think you’d be surprised with some of the advances there.

I’m taking an AI Agentic course from MIT in January and hope to learn more.
 
Limbo Pete said:
The reality in which your Grokzilla replaces human cognition? The reality where the Cylons conquer Earth? You're not only delusional, but really super ignorant. It's common with people that have that wonderful mix of intellectual arrogance, intellectual insecurity, and plain ol narcissism. Is this lingering from, like, not doing well in school or something? Question: How are you supposed to use AI to steal or plagiarize something that isn't digitally accessible?
Click to expand...
Thanks for contributing nothing to this conversation.
 
Rob Battisti said:
Damn @Limbo Pete knows to get me rustled. For the sake of argument, what is something taught in traditional public k-12 schools which is not available to learn on the internet?
Click to expand...
The right questions to ask. What will you learn if you don't know what you don't know?
 
Status
Not open for further replies.

Latest posts

Forum statistics

Threads
1,280,886
Messages
58,325,569
Members
176,001
Latest member
Markooz86

Share this page

Back
Top