You think that the farther in the future you go, the less likely it is that any existing population will know how to fix what breaks? That strikes me as oddly pessimistic, and frankly unlikely.
I wouldn't be so sure, it's wildly common. Reverse engineering products to figure out how to make them again is its own sub-industry in every manufacturing field, especially anything tool & die.
There's an enormous chasm between when companies started mass producing things and the digitization of blueprints. More often than not it isn't even the original company, it's one of their customer's customers 50+ years after they went out of business.
Im more convinced we are heading that way every year. Certainly not anytime soon, but 100, 200, 1000 years from now? Seems plausible to me. Most people don't even know how their house electricity system is setup or works, or how their plumbing system works despite being incredibly simple, not to mention more advanced technologies. We are surrounded by internal combustion engines and yet less and less people actually understand them each day and it becomes more and more specialized knowledge. How many people know how a refrigerator works despite being incredibly simple technology and the basics necessary to understand it covered by atleast freshman highschool science class if not earlier?
We have built a disposable society so people most people never have to learn how anything works. They press a switch or button and it doesn't work? Throw it away and buy a new mass-produced model. And as time goes on we only need less and less people to understand a technology to mass produce it and sell it.
What a strange attitude. It certainly is not anyone else's job to maintain the home I own and live in. That is my job. If something breaks, I don't get to tell someone else that they have not done their job. I am responsible for fixing it. Yes, I can delegate a specific fix to a professional electrician or plumber, but that is a project-by-project choice.
If people want to approach life by hiring out every little thing, they can certainly do so, but those choices do not make my home maintenance someone else's job.
Yeah, but he is extrapolating something going from common knowledge to specialist knowledge as a step towards it being something that nobody knows, and that's just not what it is at all.
It's not that I think that future generations are dumb or anything like that, it's just that the longer an assumption is held, the less likely people are to be prepared for it being invalidated.
Many of the people who designed our electrical infrastructure are still alive. Far fewer are needed to keep it running. If we must rethink our priors, I'd rather have both groups in the room while we do it.
With the state of how the ai agents stuff already is, if you don't think that most tech jobs will be automated away in our lifetime, I don't what to tell you.
Within 10-15 years, AI research automates itself and no one ever has a software engineering job ever again.
I think sometimes of how crazy it would be if we could send GPT 4.5 in the past somehow, what effect that could have, just a magical almost all knowing being from the future
Computing power requirements and the fact that my subjective appraisal of the performance of these LLMs doesn't match the insane scaling curves that AI companies put out showing that their capabilities double every 6 months. Even if a 100x increase in computing power could equal a software engineer (and that's far from certain), that would be more expensive than a software engineer.
But really the burden of proof is on you here since you are making extraordinary claims of AI superintelligence replacing all jobs in 10-15 years. You are making the trillion pound baby argument, so you need to back it up.
It's not, this isn't a lawsuit, it's a casual discussion, and I don't really care whether I'm convincing you or not, I'm taking about this because I enjoy it. I can tell you though that DeepSeek R1 & RL approaches do have shown the power to scale coding and reasoning much further without increasing model size or data requirements much, & that new improvements come non stop in the field from the billions being ingested and all the minds being focused on this all day long as it's so obvious to everyone that it's powerful
I feel like you're coming from a place where GPTs value lies only in coding. I use it for planning all the time, saving me hours per week.
For example, the security team gave me a list of 100s of policies that need to be implemented. I was able to dump that list in and get a rollout plan over the next two months in a matter of minutes. This would easily have taken me half a day before GPT.
Isn't that a bit pessimistic? Assuming machine natural language understanding and general reasoning improves dramatically (which seems possible, based on recent history), it is likely that (given that we still have the data) at some point in the future anyone will have the ability to acquire these skills, or that machine agents will be skilled enough to guide human or other types of agents to do things.
The post is about solar events, good luck getting machine agents working in case this happens...
It is pessimistic yes, but what in recent history did happen to not have a pessimistic outlook on things?
As far as I am aware, the vast majority of electrical components will still function during an extreme solar particle event. The major risk is to electrical distribution, but, as you are surely aware, there are other ways to access electricity off grid.
There's an enormous chasm between when companies started mass producing things and the digitization of blueprints. More often than not it isn't even the original company, it's one of their customer's customers 50+ years after they went out of business.