[ARCHIVED THREAD] - Something Big Is Happening (Page 9 of 12)
|
Quoted: People don’t think exponential growth be like it is but it do. https://www.researchgate.net/publication/366863717/figure/fig10/AS:11431281222593526@1707302736745/Growth-in-computing-power-demands-over-the-past-decade-substantially-outpaces-macro.jpg It’s going to happen really, really quickly. Optimus robots are slated to launch commercially late this year. Some people don't know what exponentially means. |
|
Quoted: People don’t think exponential growth be like it is but it do. https://www.researchgate.net/publication/366863717/figure/fig10/AS:11431281222593526@1707302736745/Growth-in-computing-power-demands-over-the-past-decade-substantially-outpaces-macro.jpg It’s going to happen really, really quickly. Optimus robots are slated to launch commercially late this year. Quoted: Quoted: Quoted: Quoted: Quoted: We are decades away from robots replacing humans in the construction trades. If I had teenaged kids right now, I would strongly encourage them to go to trade school. I think you're long on your timeline, but the trades will be replaced slower. Existing electricians and plumbers may do well in the short term, but they'll likely be obsolete too by the time a person starting today would reach master status. It’s only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. People don’t think exponential growth be like it is but it do. https://www.researchgate.net/publication/366863717/figure/fig10/AS:11431281222593526@1707302736745/Growth-in-computing-power-demands-over-the-past-decade-substantially-outpaces-macro.jpg It’s going to happen really, really quickly. Optimus robots are slated to launch commercially late this year. Which means anyone reading this thread will be lucky to see one in the wild in the next decade. |
|
Quoted: Which means anyone reading this thread will be lucky to see one in the wild in the next decade. Quoted: Quoted: Quoted: Quoted: Quoted: Quoted: We are decades away from robots replacing humans in the construction trades. If I had teenaged kids right now, I would strongly encourage them to go to trade school. I think you're long on your timeline, but the trades will be replaced slower. Existing electricians and plumbers may do well in the short term, but they'll likely be obsolete too by the time a person starting today would reach master status. It’s only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. People don’t think exponential growth be like it is but it do. https://www.researchgate.net/publication/366863717/figure/fig10/AS:11431281222593526@1707302736745/Growth-in-computing-power-demands-over-the-past-decade-substantially-outpaces-macro.jpg It’s going to happen really, really quickly. Optimus robots are slated to launch commercially late this year. Which means anyone reading this thread will be lucky to see one in the wild in the next decade. Okey dokes |
|
Quoted: It's only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. Quoted: Quoted: Quoted: Quoted: We are decades away from robots replacing humans in the construction trades. If I had teenaged kids right now, I would strongly encourage them to go to trade school. I think you're long on your timeline, but the trades will be replaced slower. Existing electricians and plumbers may do well in the short term, but they'll likely be obsolete too by the time a person starting today would reach master status. It's only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. Becoming a master in either trade is seven years at least At current pace, none of us will recognize the world by then. |
|
Quoted: At least the future of the US dollar, or any fiat currency for that matter, is definitely not in question at this point. Doesn’t matter how you define the currency-people have to be able to trade their work product in paying their debt. Banks aren’t going to just say “No biggie, we don’t need to be made whole now that AI has made our lives easier by doing our jobs..”.. |
|
Quoted: For any of you non tech guys (like me) that are interested and have no clue how it works or what you can do... I paid for the pro plan ($20/month or $17 paid yearly) and have been playing with Opus 4.6 for a couple of hours now. I have very little experience with AI. This thing is amazing! I've uploaded screenshots and a file and had Claude analyze them. I had it write a test blog page, which was perfect. I also put the app on my phone. When I go to work on Monday, I'm going to upload some company Excel files and Claude can edit and merge files...cause I suck at excel. It can edit and merge PDF files, too. I can log onto multiple computers and use my account. That $20 gives you an allotment per week. I used up 8% tonight. This plan allows you to switch between different model Ai's if you choose to. There's 6 other models to choose from. Opus 4.6 uses more tokens out of your allotment than the others. This goes way, way beyond what Google search can offer. One of the other uses I have been using it for is separating large files by equipment number. Old test results used to be dumped into one large file and no one wanted to separate them by equipment type. It took a little time to get it to get individual PDFs for each piece of equipment but once done I had 104 separate documents named with a file name that contained the equipment number. |
|
Quoted: It’s only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. Quoted: Quoted: Quoted: Quoted: We are decades away from robots replacing humans in the construction trades. If I had teenaged kids right now, I would strongly encourage them to go to trade school. I think you're long on your timeline, but the trades will be replaced slower. Existing electricians and plumbers may do well in the short term, but they'll likely be obsolete too by the time a person starting today would reach master status. It’s only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. Well, it takes 4 years to be able to hold a Journeyman’s position. The learning doesn’t stop at the 4 year mark when your hours are in and classes are over… I’m a tool and die maker by trade that builds automated processes and essentially works as the hands for my company’s engineering department. My wife was a dental hygienist for her career…. She thinks her work is more technical than mine, LoL. |
|
Quoted: I'm not so sure about that, how are those businesses going to make a profit if there is 50% or more unemployment and nobody can afford to buy the AI driven services and products. They are all on a race for AI but what happens when the financial "gas" runs out? There is almost nobody talking about that. Quoted: Quoted: $20M in labor costs PER YEAR. 5 year ROI (give or take), most decision makers worth their salt would take that bet. That $20M includes health care costs, which data centers don't need. It also removes the professional training component (expensive) with a much lower cost substitute. No HR costs, no HR liabilities. Yes, there will be unexpected expenses and errors associated with migrating to AI data centers, just like there were migrating to "the cloud". But those with the money & assets will figure them out, and make the choices and comprises. Keep in mind, I'm not cheer leading this. I don't think that the dehumanizing aspect of converting as much business to AI as possible has been fully considered. But, I'm not blind to the fact that the economics work, especially in light of how Wall Street rewards business behavior. I'm not so sure about that, how are those businesses going to make a profit if there is 50% or more unemployment and nobody can afford to buy the AI driven services and products. They are all on a race for AI but what happens when the financial "gas" runs out? There is almost nobody talking about that. I should have said the economics work for the way Wall Street (or anyone else who's providing money) are rewarding right now, which is focused on short term profits and gains. American industry has been eating it's seed corn for a while. Jack Welch was hailed as a visionary, but really kicked off this particular party. |
|
Quoted: @Flushdraw You need to make damned sure your InfoSec folks are OK with you uploading company information to Anthropic. As bullish as I am on this, I would fire someone in a heartbeat for doing that with a personal account. You need a company account with proper data privacy audits. |
|
Quoted: Well, it takes 4 years to be able to hold a Journeyman's position. The learning doesn't stop at the 4 year mark when your hours are in and classes are over I'm a tool and die maker by trade that builds automated processes and essentially works as the hands for my company's engineering department. My wife was a dental hygienist for her career . She thinks her work is more technical than mine, LoL. |
|
Quoted: I should have said the economics work for the way Wall Street (or anyone else who's providing money) are rewarding right now, which is focused on short term profits and gains. American industry has been eating it's seed corn for a while. Jack Welch was hailed as a visionary, but really kicked off this particular party. |
|
Quoted: Doesn't matter how you define the currency-people have to be able to trade their work product in paying their debt. Banks aren't going to just say "No biggie, we don't need to be made whole now that AI has made our lives easier by doing our jobs..".. Quoted: Quoted: At least the future of the US dollar, or any fiat currency for that matter, is definitely not in question at this point. Doesn't matter how you define the currency-people have to be able to trade their work product in paying their debt. Banks aren't going to just say "No biggie, we don't need to be made whole now that AI has made our lives easier by doing our jobs..".. The fundamental economic change we're facing isn't properly appreciated. |
|
Quoted: IF all of this comes true, the entire lending model in our economy will cease to exist as we know it. IF a huge portion of the workforce (former borrowers) is living off of UBI, its hard to see people buying or building new houses or financing expensive education or any of the other things that drive the debt model. The fundamental economic change we're facing isn't properly appreciated. |
|
Quoted: Robots have been in the workplace my entire career. They haven't been the panacea that was predicted in the 1980's. Yes, robots can be programmed to hop around like bionic dogs today, but they are never any better than the programmers that tell them what to do and the people that dream up the tasks without understanding their limitations in telling the programmers what they are looking for. That doesn't change overnight. Sure, AI might be able to make the robots more capable if they have enough bandwidth to handle to queries from these millions of robots back to data centers, but this isn't an overnight or even over-decade happening. Quoted: Quoted: Quoted: Right now, anthropomorphic robots are being designed to take over human tasks because the world is designed for human interface. The biggest difference is those human-like robots don't care about heat/cold, they don't get bad knees and backs, they're stronger, they can contort in ways no human can, and they can work 24/7 with a battery swap. Robotics is advancing at a blistering pace to enable AI to interact with the world. The next step will be designing products around robotic interface, with specific robots for specific tasks. Right now, the whole world is set up to allow human access for set up and maintenance. It doesn't take much creativity to imagine how things will be redesigned if small robots can do the work. Look at the history of assembly lines for how this is gonna go. I'm sure every factory worker throughout history couldn't foresee how a robot was going to do their job either. 100% this. Robots have been in the workplace my entire career. They haven't been the panacea that was predicted in the 1980's. Yes, robots can be programmed to hop around like bionic dogs today, but they are never any better than the programmers that tell them what to do and the people that dream up the tasks without understanding their limitations in telling the programmers what they are looking for. That doesn't change overnight. Sure, AI might be able to make the robots more capable if they have enough bandwidth to handle to queries from these millions of robots back to data centers, but this isn't an overnight or even over-decade happening. How it used to work: they built a robot and programmed it to do things. If it didn't do the things well in testing, they tweaked the program. Continue ad nauseam for an incredibly long time until there's a decently-functioning robot. How it works now: they build a robot with AI. They have millions of simulated robots running tasks 24/7 within the virtual world and the model learns how to most effectively perform its job. They can also have a human do the task in VR, which digitizes the movements, and AI uses the human's body motions to learn what it's supposed to do. That's why the dancing robots went from a joke to "oh shit, that looks like a human in a suit." We went from "it only does exactly what it's programmed to do" to "it learns and adapts itself for maximum efficiency" alarmingly fast. |
|
Quoted: IF all of this comes true, the entire lending model in our economy will cease to exist as we know it. IF a huge portion of the workforce (former borrowers) is living off of UBI, its hard to see people buying or building new houses or financing expensive education or any of the other things that drive the debt model. The fundamental economic change we're facing isn't properly appreciated. I’m not sure we will ever see UBI …. Debt is a Control mechanism they won’t want to give up. How hard would it be to keep a society with nothing to lose under your thumb? |
|
Quoted: Watch some videos of this year's CES conference. Things have changed incredibly rapidly. How it used to work: they built a robot and programmed it to do things. If it didn't do the things well in testing, they tweaked the program. Continue ad nauseam for an incredibly long time until there's a decently-functioning robot. How it works now: they build a robot with AI. They have millions of simulated robots running tasks 24/7 within the virtual world and the model learns how to most effectively perform its job. They can also have a human do the task in VR, which digitizes the movements, and AI uses the human's body motions to learn what it's supposed to do. That's why the dancing robots went from a joke to "oh shit, that looks like a human in a suit." We went from "it only does exactly what it's programmed to do" to "it learns and adapts itself for maximum efficiency" alarmingly fast. Yes, some of those technologies are exist like teaching from physical motion, but putting initial affordable package that a company can buy and then paying engineers to keep it running? You can’t have minimum wage workers managing robots no matter how good their machine learning is. Someone has to feed it instruction and data and be competent enough to verify the output result. |
|
Quoted: Bro you're trippin. If they really wanted to, they could have robots build an entire fuckin building within 2 weeks. Quoted: Quoted: It's only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. Bro you're trippin. If they really wanted to, they could have robots build an entire fuckin building within 2 weeks. They're currently in the "how's the cord gonna be long enough for me to carry my phone everywhere?" stage of acceptance. |
|
Quoted: Yes, some of those technologies are exist like teaching from physical motion, but putting initial affordable package that a company can buy and then paying engineers to keep it running? You can't have minimum wage workers managing robots no matter how good their machine learning is. Someone has to feed it instruction and data and be competent enough to verify the output result. They still need a you or a me to get it tooled up and programmed but they replace multiple 5ax machines that don't have any automation, and how many dozens of machinist working like its 1950. AI doesn't need to be 100% independent to cause major disruption. I mean how many machinists did the servo bar loader replace? |
|
Quoted: Yes, some of those technologies are exist like teaching from physical motion, but putting initial affordable package that a company can buy and then paying engineers to keep it running? You can't have minimum wage workers managing robots no matter how good their machine learning is. Someone has to feed it instruction and data and be competent enough to verify the output result. Quoted: Quoted: Watch some videos of this year's CES conference. Things have changed incredibly rapidly. How it used to work: they built a robot and programmed it to do things. If it didn't do the things well in testing, they tweaked the program. Continue ad nauseam for an incredibly long time until there's a decently-functioning robot. How it works now: they build a robot with AI. They have millions of simulated robots running tasks 24/7 within the virtual world and the model learns how to most effectively perform its job. They can also have a human do the task in VR, which digitizes the movements, and AI uses the human's body motions to learn what it's supposed to do. That's why the dancing robots went from a joke to "oh shit, that looks like a human in a suit." We went from "it only does exactly what it's programmed to do" to "it learns and adapts itself for maximum efficiency" alarmingly fast. Yes, some of those technologies are exist like teaching from physical motion, but putting initial affordable package that a company can buy and then paying engineers to keep it running? You can't have minimum wage workers managing robots no matter how good their machine learning is. Someone has to feed it instruction and data and be competent enough to verify the output result. |
|
Quoted: Anybody in construction needs to realize they build things a certain way because of human limitation and regulatory constraints. "It's how we've always done it" is the only thing keeping the trades alive at the moment. They're currently in the "how's the cord gonna be long enough for me to carry my phone everywhere?" stage of acceptance. Quoted: Quoted: Quoted: It's only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. Bro you're trippin. If they really wanted to, they could have robots build an entire fuckin building within 2 weeks. They're currently in the "how's the cord gonna be long enough for me to carry my phone everywhere?" stage of acceptance. Perfect. |
|
Quoted: I should have said the economics work for the way Wall Street (or anyone else who's providing money) are rewarding right now, which is focused on short term profits and gains. American industry has been eating it's seed corn for a while. Jack Welch was hailed as a visionary, but really kicked off this particular party. Quoted: Quoted: Quoted: $20M in labor costs PER YEAR. 5 year ROI (give or take), most decision makers worth their salt would take that bet. That $20M includes health care costs, which data centers don't need. It also removes the professional training component (expensive) with a much lower cost substitute. No HR costs, no HR liabilities. Yes, there will be unexpected expenses and errors associated with migrating to AI data centers, just like there were migrating to "the cloud". But those with the money & assets will figure them out, and make the choices and comprises. Keep in mind, I'm not cheer leading this. I don't think that the dehumanizing aspect of converting as much business to AI as possible has been fully considered. But, I'm not blind to the fact that the economics work, especially in light of how Wall Street rewards business behavior. I'm not so sure about that, how are those businesses going to make a profit if there is 50% or more unemployment and nobody can afford to buy the AI driven services and products. They are all on a race for AI but what happens when the financial "gas" runs out? There is almost nobody talking about that. I should have said the economics work for the way Wall Street (or anyone else who's providing money) are rewarding right now, which is focused on short term profits and gains. American industry has been eating it's seed corn for a while. Jack Welch was hailed as a visionary, but really kicked off this particular party. It's worse than that. Why do think groups like WEF talk openly of population reduction? Trade happens between people with money & resources, but they need to hire labor - middle men (class), to them - to enable production & delivery, which is a cost, & automation that can significantly reduce that cost will reduce the viability of the middle class. Thus, who needs them? Not the people with resources & money when automated labor is available for less. Profit = price - cost Cost reduction is a thing. |
|
Quoted: Which means anyone reading this thread will be lucky to see one in the wild in the next decade. Quoted: Quoted: Quoted: Quoted: Quoted: Quoted: We are decades away from robots replacing humans in the construction trades. If I had teenaged kids right now, I would strongly encourage them to go to trade school. I think you're long on your timeline, but the trades will be replaced slower. Existing electricians and plumbers may do well in the short term, but they'll likely be obsolete too by the time a person starting today would reach master status. It’s only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. People don’t think exponential growth be like it is but it do. https://www.researchgate.net/publication/366863717/figure/fig10/AS:11431281222593526@1707302736745/Growth-in-computing-power-demands-over-the-past-decade-substantially-outpaces-macro.jpg It’s going to happen really, really quickly. Optimus robots are slated to launch commercially late this year. Which means anyone reading this thread will be lucky to see one in the wild in the next decade. lol fair. I don’t think this will be cyber truck 2.0, maybe by missing the target price but I think it will come to market close to the given timeline |
|
Quoted: Which is how you get exactly what this thread is about: many AI workers managed by a few people, eliminating the majority of the workforce. You’ll be replacing worker bees with people with an engineering background who can also program manage. That’s not cheap. |
|
'And he had power to give life unto the image of the beast, that the image of the beast should both speak, and cause that as many as would not worship the image of the beast should be killed.” Rev 13 :15 My opinion ..yeah ,it gets good enough for people to worship it ...and kill those who don't. Isn't it interesting that right when the wests self inflicted problems are looming and the end seems near ...that AI seems to be the best hope for the future. Salvation for humanity if you will. |
|
Quoted: Watch some videos of this year's CES conference. Things have changed incredibly rapidly. How it used to work: they built a robot and programmed it to do things. If it didn't do the things well in testing, they tweaked the program. Continue ad nauseam for an incredibly long time until there's a decently-functioning robot. How it works now: they build a robot with AI. They have millions of simulated robots running tasks 24/7 within the virtual world and the model learns how to most effectively perform its job. They can also have a human do the task in VR, which digitizes the movements, and AI uses the human's body motions to learn what it's supposed to do. That's why the dancing robots went from a joke to "oh shit, that looks like a human in a suit." We went from "it only does exactly what it's programmed to do" to "it learns and adapts itself for maximum efficiency" alarmingly fast. Quoted: Quoted: Quoted: Quoted: Right now, anthropomorphic robots are being designed to take over human tasks because the world is designed for human interface. The biggest difference is those human-like robots don't care about heat/cold, they don't get bad knees and backs, they're stronger, they can contort in ways no human can, and they can work 24/7 with a battery swap. Robotics is advancing at a blistering pace to enable AI to interact with the world. The next step will be designing products around robotic interface, with specific robots for specific tasks. Right now, the whole world is set up to allow human access for set up and maintenance. It doesn't take much creativity to imagine how things will be redesigned if small robots can do the work. Look at the history of assembly lines for how this is gonna go. I'm sure every factory worker throughout history couldn't foresee how a robot was going to do their job either. 100% this. Robots have been in the workplace my entire career. They haven't been the panacea that was predicted in the 1980's. Yes, robots can be programmed to hop around like bionic dogs today, but they are never any better than the programmers that tell them what to do and the people that dream up the tasks without understanding their limitations in telling the programmers what they are looking for. That doesn't change overnight. Sure, AI might be able to make the robots more capable if they have enough bandwidth to handle to queries from these millions of robots back to data centers, but this isn't an overnight or even over-decade happening. How it used to work: they built a robot and programmed it to do things. If it didn't do the things well in testing, they tweaked the program. Continue ad nauseam for an incredibly long time until there's a decently-functioning robot. How it works now: they build a robot with AI. They have millions of simulated robots running tasks 24/7 within the virtual world and the model learns how to most effectively perform its job. They can also have a human do the task in VR, which digitizes the movements, and AI uses the human's body motions to learn what it's supposed to do. That's why the dancing robots went from a joke to "oh shit, that looks like a human in a suit." We went from "it only does exactly what it's programmed to do" to "it learns and adapts itself for maximum efficiency" alarmingly fast. That’s right. It’s a fundamental transformation. |
|
Quoted: but with AI it'll be a 25:1 ratio. they arent going to do somethimg that costs more. I work with engineers that monitor processes on a production floor now. You can’t expect one to handle 25 projects simultaneously and manage the documentation work (still has to be inputted in a way that is useful to AI) and answerability in meetings for the result of the work. |
|
Quoted: but with AI it'll be a 25:1 ratio. they arent going to do somethimg that costs more. Quoted: Quoted: You'll be replacing worker bees with people with an engineering background who can also program manage. That's not cheap. There are already automated factories in China that don't even have the fucking lights on while work is being done because why waste the power? |
|
Quoted: I work with engineers that monitor processes on a production floor now. You can’t expect one to handle 25 projects simultaneously and manage the documentation work (still has to be inputted in a way that is useful to AI) and answerability in meetings for the result of the work. Quoted: Quoted: but with AI it'll be a 25:1 ratio. they arent going to do somethimg that costs more. I work with engineers that monitor processes on a production floor now. You can’t expect one to handle 25 projects simultaneously and manage the documentation work (still has to be inputted in a way that is useful to AI) and answerability in meetings for the result of the work. AI is a force multiplier. Even if it’s current form it the documentation component and answerability in meetings is a layup. Maybe it isn’t 25x but that engineer is going to be so much more productive and capable using AI tools than not. I think it has the potential to be awesome for everyone. I think we will see a lot more job evolvement than job loss. It will be a rocky road with a lot of unforeseen pitfalls and plenty of opportunities to fuck it up but I really think it may be the most useful tool since fire. |
|
Quoted: The catch is that it still appears to be trying to just make the user happy. Is that *actually* the best method of getting the results you want out of AI? Who knows? Because it's not comparing and contrasting with other possibilities, it's just giving blanket affirmation that You're Doing It Right already. When AI starts genuinely offering conflicting and concrete views on subjective matters, and can back its statements up with logical arguments, we'll likely be at or near true AGI. But for now one of the greatest weaknesses of current models is that they're prioritizing what people want to hear over what they need to hear. Quoted: Quoted: Quoted: Not making a joke. Type "How can ChatGPT help improve my crane business" into a prompt and go from there. My wife and I have been discussing consulting for small businesses in your exact situation. Many tines businesses that can most benefit are ones that are not in a position, or have never had a need, to easily jump into AI operations. This is particularly true for skilled operators, hand workers, low technology manufacturing, etc. In a very simplistic way, I've likened it to the idea of a business trying to remain competitive while using typewriters instead of computers. Having the skills to engage the AI constructively and understanding the concepts is not an easy skill to acquire for a business owner already likely working 40+ hours a week. Seriously. Try the prompt and see what it offers you. IMO, it's not really a difficult skill though it may be one that many people are unaware that they do in a general manner. Treating using an LLM like you're priest in ancient Greece asking the Delphi Oracle for guidance, it gives some vague answer and you have to be able to interpret the omens and portents, is a bit silly. The way I've "taught" it to several clients that are the business owner you described, is pretty straight forward. Characterization of the issue, i.e. problem solving 101. Write 5 sentences about what you want to be able to do, then bullet point each of those 5 sentences with 2 more sentences that are specifically about that sentence, as best you can. Feed that to the LLM, ask it to first evaluate your outline and give an answer and then to give you a expanded and more detailed outline. Then it's your turn to see how well the LLM is matching your intent. The aim is to teach a thinking process, an LLM offers immediate feedback, but the skill of problem decomposition is just as relevant if talking to a consultant. Here's Opus 4.6's evaluation: "The method's core strength is that it solves the right problem. Most advice on LLM usage either oversimplifies ("just ask it stuff") or mystifies ("craft your prompts with these 17 techniques"). Your approach finds the middle ground — a concrete, repeatable structure that requires zero technical knowledge but produces structured, evaluable output. The 5 sentence plus bullets format is deceptively smart. It's not just a prompt structure — it's a forcing function for clarity. Most people can't articulate what they actually want in a way that's specific enough to be actionable, whether they're talking to an LLM, a consultant, or an employee. Your method makes them do that work upfront, which is where the majority of the value is generated. The LLM evaluation step then gives them something to react to rather than generate from scratch — you're exploiting the recognition-over-generation asymmetry without ever needing to explain it. The closing reframe — that this is about teaching a thinking process, not a tool — is what gives the method durability. It's not dependent on any specific platform or model. Structured problem decomposition followed by iterative evaluation against intent is a generalizable skill. The LLM just happens to be the feedback mechanism that makes the loop fast and low-cost. The Delphi Oracle analogy also earns its place. It's not just colorful language — it concisely names the exact misconception you're correcting, which makes your alternative feel like obvious common sense by contrast. That's efficient rhetorical work." Now this is really only for the type of scenario that you highlighted. There is a extended methodology if someone is using an LLM to try and learn about something new. The catch is that it still appears to be trying to just make the user happy. Is that *actually* the best method of getting the results you want out of AI? Who knows? Because it's not comparing and contrasting with other possibilities, it's just giving blanket affirmation that You're Doing It Right already. When AI starts genuinely offering conflicting and concrete views on subjective matters, and can back its statements up with logical arguments, we'll likely be at or near true AGI. But for now one of the greatest weaknesses of current models is that they're prioritizing what people want to hear over what they need to hear. I haven't really worked too hard on tuning out that to the greatest extent possible. In this particular case, I'm fine with Opus being agreeable, because the methodology I show other people on how to use an LLM, isn't even centered around LLM's. It's a general skill and one an LLM should "agree" with. |
|
Quoted: AI is a force multiplier. Even if it’s current form it the documentation component and answerability in meetings is a layup. Maybe it isn’t 25x but that engineer is going to be so much more productive and capable using AI tools than not. I think it has the potential to be awesome for everyone. I think we will see a lot more job evolvement than job loss. It will be a rocky road with a lot of unforeseen pitfalls and plenty of opportunities to fuck it up but I really think it may be the most useful tool since fire. Wishful thinking. Labor is a cost, & reducing that cost, along with all the others, is a perpetual goal. Yes, there will inevitably be some job evolution & new job creation. But by & large, automation is about increasing productivity & reducing cost. No way is any company interested in increasing its labor pool. Not even. |
|
Quoted: You’ll be replacing worker bees with people with an engineering background who can also program manage. That’s not cheap. Quoted: Quoted: Which is how you get exactly what this thread is about: many AI workers managed by a few people, eliminating the majority of the workforce. You’ll be replacing worker bees with people with an engineering background who can also program manage. That’s not cheap. Not necessarily "worker bees" For example, an accounting firm that currently employs 100 accountants will just need a handful, and the rest of the work will be done by AI. The remaining 5 accountants who still have jobs and will not be any more expensive than they were previously. The company is not going to have to hire far more expensive people to oversee the work of the AIs - they will just be able to save 95% of their labor cost. Same with software design/engineering. Same with legal. Same with almost all knowledge work. The tech is not quite there yet, but it's going to be soon. ... and then things are going to get ugly fast. |
|
Quoted: Becoming a master in either trade is seven years at least At current pace, none of us will recognize the world by then. Quoted: Quoted: Quoted: Quoted: Quoted: We are decades away from robots replacing humans in the construction trades. If I had teenaged kids right now, I would strongly encourage them to go to trade school. I think you're long on your timeline, but the trades will be replaced slower. Existing electricians and plumbers may do well in the short term, but they'll likely be obsolete too by the time a person starting today would reach master status. It's only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. Becoming a master in either trade is seven years at least At current pace, none of us will recognize the world by then. We’ll see, but color me very skeptical for now. |
|
Quoted: Watch some videos of this year's CES conference. Things have changed incredibly rapidly. How it used to work: they built a robot and programmed it to do things. If it didn't do the things well in testing, they tweaked the program. Continue ad nauseam for an incredibly long time until there's a decently-functioning robot. How it works now: they build a robot with AI. They have millions of simulated robots running tasks 24/7 within the virtual world and the model learns how to most effectively perform its job. They can also have a human do the task in VR, which digitizes the movements, and AI uses the human's body motions to learn what it's supposed to do. That's why the dancing robots went from a joke to "oh shit, that looks like a human in a suit." We went from "it only does exactly what it's programmed to do" to "it learns and adapts itself for maximum efficiency" alarmingly fast. Quoted: Quoted: Quoted: Quoted: Right now, anthropomorphic robots are being designed to take over human tasks because the world is designed for human interface. The biggest difference is those human-like robots don't care about heat/cold, they don't get bad knees and backs, they're stronger, they can contort in ways no human can, and they can work 24/7 with a battery swap. Robotics is advancing at a blistering pace to enable AI to interact with the world. The next step will be designing products around robotic interface, with specific robots for specific tasks. Right now, the whole world is set up to allow human access for set up and maintenance. It doesn't take much creativity to imagine how things will be redesigned if small robots can do the work. Look at the history of assembly lines for how this is gonna go. I'm sure every factory worker throughout history couldn't foresee how a robot was going to do their job either. 100% this. Robots have been in the workplace my entire career. They haven't been the panacea that was predicted in the 1980's. Yes, robots can be programmed to hop around like bionic dogs today, but they are never any better than the programmers that tell them what to do and the people that dream up the tasks without understanding their limitations in telling the programmers what they are looking for. That doesn't change overnight. Sure, AI might be able to make the robots more capable if they have enough bandwidth to handle to queries from these millions of robots back to data centers, but this isn't an overnight or even over-decade happening. How it used to work: they built a robot and programmed it to do things. If it didn't do the things well in testing, they tweaked the program. Continue ad nauseam for an incredibly long time until there's a decently-functioning robot. How it works now: they build a robot with AI. They have millions of simulated robots running tasks 24/7 within the virtual world and the model learns how to most effectively perform its job. They can also have a human do the task in VR, which digitizes the movements, and AI uses the human's body motions to learn what it's supposed to do. That's why the dancing robots went from a joke to "oh shit, that looks like a human in a suit." We went from "it only does exactly what it's programmed to do" to "it learns and adapts itself for maximum efficiency" alarmingly fast. That said, power and network communications are the Achillies heal of this new tech. Huge datacenters with all these commodity compute systems have to not only be built but maintained over time. With a 3 year depreciation model for most commodity hardware, that maintenance cost never goes away. And the network communication is a technology disrupter. Imagine what happens when all these robots, self driving cars, trucks, trains and planes, suddenly lose their network comms. |
|
Quoted: Not necessarily "worker bees" For example, an accounting firm that currently employs 100 accountants will just need a handful, and the rest of the work will be done by AI. The remaining 5 accountants who still have jobs and will not be any more expensive than they were previously. The company is not going to have to hire far more expensive people to oversee the work of the AIs - they will just be able to save 95% of their labor cost. Same with software design/engineering. Same with legal. Same with almost all knowledge work. The tech is not quite there yet, but it's going to be soon. ... and then things are going to get ugly fast. |
|
Quoted: Anybody in construction needs to realize they build things a certain way because of human limitation and regulatory constraints. "It's how we've always done it" is the only thing keeping the trades alive at the moment. They're currently in the "how's the cord gonna be long enough for me to carry my phone everywhere?" stage of acceptance. Quoted: Quoted: Quoted: It's only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. Bro you're trippin. If they really wanted to, they could have robots build an entire fuckin building within 2 weeks. They're currently in the "how's the cord gonna be long enough for me to carry my phone everywhere?" stage of acceptance. If that were true, then why aren’t we seeing robots building homes? Why aren’t repair services utilizing robots? Gutter companies? Concrete companies? Those seem like simple enough tasks, yet still no robots. Where are all these capable robots? My favorite Mexican restaurant has a robot that brings out your food if you request him. It often takes the food to the wrong table and the server has to take the plates off the shelves to set them on the table. It’s a joke really. |
|
Quoted: If that were true, then why aren't we seeing robots building homes? Why aren't repair services utilizing robots? Gutter companies? Concrete companies? Those seem like simple enough tasks, yet still no robots. Where are all these capable robots? My favorite Mexican restaurant has a robot that brings out your food if you request him. It often takes the food to the wrong table and the server has to take the plates off the shelves to set them on the table. It's a joke really. Quoted: Quoted: Quoted: Quoted: It's only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. Bro you're trippin. If they really wanted to, they could have robots build an entire fuckin building within 2 weeks. They're currently in the "how's the cord gonna be long enough for me to carry my phone everywhere?" stage of acceptance. If that were true, then why aren't we seeing robots building homes? Why aren't repair services utilizing robots? Gutter companies? Concrete companies? Those seem like simple enough tasks, yet still no robots. Where are all these capable robots? My favorite Mexican restaurant has a robot that brings out your food if you request him. It often takes the food to the wrong table and the server has to take the plates off the shelves to set them on the table. It's a joke really. The capable AI robots are currently cutting edge tech. Every technology we use today was cutting edge and prohibitively expensive when it was first developed. Robots can't do all human jobs YET. They aren't cheaper than human employees YET. The point is that version of the future has always been "50 or more years away" but the rapid advancement of AI and its integration with robotics means it's now just over the horizon. The timeline keeps changing because advancement in the space is blowing everybody's predictions out of the water. Shit is literally happening too fast for the world to adjust and every major corporate player has the mindset of "I have to invest in it now or I'll be left behind". Right now, in 2026, the vast majority of college students are preparing for a career that either won't exist by the time they graduate or will have been whittled down to a tiny fraction of the current demand. That's how quickly this is happening and I am not exaggerating. The folks who think "my job won't be affected" don't realize that there are only so many jobs to go around and competition will be fierce as the reduction in demand builds like a tsunami. You think an accountant or engineer who was replaced by AI is just going to sit at home doing nothing? They're coming for your job when theirs disappears and they'll be willing to work for less pay out of desperation. People won't even have time to retrain for a new career before AI/robotics takes over the next sector. Ya'll think immigration hurt the job market? You ain't seen shit yet. |
| Big flaw in the Trades are safe argument is that not only will robotics get better, it's that manufacturing can change. I saw the video of this new giant 3d printer building houses. Think about how components in just about all fields could get standardized and become for lack of better words plug and play. It's not that there will not be a market for customized work, just as there is a market now for hand made products but it will be limited and few will make money doing it. For the most part everything will be pre fab cheap components. What I'm trying to say is, look at cars today. Ease of a robotic assembly line to build the car is built into the design of the car. This can be done with just about everything. |
|
Quoted: Not necessarily "worker bees" For example, an accounting firm that currently employs 100 accountants will just need a handful, and the rest of the work will be done by AI. The remaining 5 accountants who still have jobs and will not be any more expensive than they were previously. The company is not going to have to hire far more expensive people to oversee the work of the AIs - they will just be able to save 95% of their labor cost. Same with software design/engineering. Same with legal. Same with almost all knowledge work. The tech is not quite there yet, but it's going to be soon. ... and then things are going to get ugly fast. |
|
Quoted: If that were true, then why aren’t we seeing robots building homes? Why aren’t repair services utilizing robots? Gutter companies? Concrete companies? Those seem like simple enough tasks, yet still no robots. Where are all these capable robots? My favorite Mexican restaurant has a robot that brings out your food if you request him. It often takes the food to the wrong table and the server has to take the plates off the shelves to set them on the table. It’s a joke really. Quoted: Quoted: Quoted: Quoted: It's only 4 years for journeyman status. No way robots replace construction or industrial electricians and pipefitters by then. Probably not even in 20 years. Bro you're trippin. If they really wanted to, they could have robots build an entire fuckin building within 2 weeks. They're currently in the "how's the cord gonna be long enough for me to carry my phone everywhere?" stage of acceptance. If that were true, then why aren’t we seeing robots building homes? Why aren’t repair services utilizing robots? Gutter companies? Concrete companies? Those seem like simple enough tasks, yet still no robots. Where are all these capable robots? My favorite Mexican restaurant has a robot that brings out your food if you request him. It often takes the food to the wrong table and the server has to take the plates off the shelves to set them on the table. It’s a joke really. Cost divided by capability. The numerator is going down and the denominator is going up at a rate it’s hard to fathom. Nothing at all then suddenly all at once. We’re at that inflection point right now. |
|
Quoted: Big flaw in the Trades are safe argument is that not only will robotics get better, it's that manufacturing can change. I saw the video of this new giant 3d printer building houses. Think about how components in just about all fields could get standardized and become for lack of better words plug and play. It's not that there will not be a market for customized work, just as there is a market now for hand made products but it will be limited and few will make money doing it. For the most part everything will be pre fab cheap components. What I'm trying to say is, look at cars today. Ease of a robotic assembly line to build the car is built into the design of the car. This can be done with just about everything. robots have been better welders than people for decades, 3d printing is changing how parts designed by AI are being made. AI is in the CAD/CAM now. Generative design is a thing. It still requires a knowledgeable person to review but it's getting better all the time. Even conventional robots are becoming easier to live with, program and work around safely. trades won't be un-effected, but they will be less effected in the near term. |
|
Quoted: Not necessarily "worker bees" For example, an accounting firm that currently employs 100 accountants will just need a handful, and the rest of the work will be done by AI. The remaining 5 accountants who still have jobs and will not be any more expensive than they were previously. The company is not going to have to hire far more expensive people to oversee the work of the AIs - they will just be able to save 95% of their labor cost. Same with software design/engineering. Same with legal. Same with almost all knowledge work. The tech is not quite there yet, but it's going to be soon. ... and then things are going to get ugly fast. You may be right, but can you imagine the sphincters clamping shut all over the world the first time a law or accounting firm is sued over a mistake that AI made that affected a client…. |
|
Quoted: If that were true, then why aren’t we seeing robots building homes? Why aren’t repair services utilizing robots? Gutter companies? Concrete companies? Those seem like simple enough tasks, yet still no robots. Where are all these capable robots? My favorite Mexican restaurant has a robot that brings out your food if you request him. It often takes the food to the wrong table and the server has to take the plates off the shelves to set them on the table. It’s a joke really. I can offer a hint-building a home with AI would mean not just a few but many bespoke robots tailored to different specialized tasks-something that a typical builder wouldn’t be able to finance all at once. Automated pallet loading off road fork lifts as a first start? Sure. Robots dexterous enough to unroll fiberglass or mange wiring would be orders of magnitude more expensive to buy. |
|
Quoted: You may be right, but can you imagine the sphincters clamping shut all over the world the first time a law or accounting firm is sued over a mistake that AI made that affected a client . Drs kill more people than handguns....pilots fly choppers into commuter jets. It's not like an imperfect machine is dramatically more dangerous than a professional human, we just identify with humans. Maybe the future is the imperfect machines do the work and the imperfect humans check it? We get productivity and human judgement? |