Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
Immigration or no immigration, our population will continue to grow. People live longer now and more babies are born than people die every year. Immigration doesn't tackle the core problem, only delays it just like the opponents of the minimum wage. It'll only be a matter of time before UBI takes place or something of that nature.
Re the bolded.....If we did away with Medicaid, etc...perhaps that would no longer be the case:
Probably too short. You'll need software engineers from maybe not writing say operating systems or whatever. But you'll still need software engineers to program embedded systems. Home automation, little tiny gadgets, cars, planes, medical devices, etc.. all contain processors and still has a need for engineers to write code because they are all different platforms, industries and have various regulations.
I don't understand the in and out of your field, but what he was saying was that you would ask a computer to basically create another computer that would have function x, and let it go to work.
The trend is more cores per chip due to Moore's law. The density of integrated circuits will keep increasing, probably for another decade. But then it'll hit a brick wall. As of now, density of ICs like GPUs or whatever has gotten so huge, you can put thousands of cores in a single chip. TPUs and DLUs are just subsystems designed to provide hardware support for a main general purpose processor (GPGPU). The new Nvidia Volta architecture contains something of a similar flavor as Google and Fujitsu and has up to 5120 cores.
Bottom line, processing power will always increase, right now the engineering challenge is, you have a massive budget of hardware to work with. How do you use all of it well + power it all.
With all this processing power, you can literally have cars or whatever contain multiple thousands of CPU/GPU cores to do parallel processing, hence AI. AI is very matrix-heavy computation. Thousands of cores can process matrices very quickly. Hence... those chips Google and Fujitsu are talking about, are typically integrated with the GPU core, not the CPU core like the desktop processors you'd find.
All this processing power availability is what is driving the automation change. Because it's so cheap to mass produce and a lot of computation can be done.
Me and Wee-Bey are talking about something else more fundamental to all these chip designs which I can't explain it all here.
The thing of it is ... and just a part of why I'm skeptical about the ultimate power of AI ... is that Moore's law is basically dead. Both Intel and AMD are at 7 nanometer gate length processes and that's about as far as you go. Below that you begin to see quantum tunneling effects. It's the process shrink more so than logic design innovation that has driven Moore's law. Make it smaller, pack those gates tighter and go faster. We're hitting a physical wall now. You're going to see uprocessor perf trailing off until some fundamentally different computing technology arises. Most of the performance headroom from here on out is in clever tweaks to the basic super-scalar design. More and smarter speculation both in terms of branch prediction and data prefetch, wider multi-way issue and retirement ... 4-way, 8-way SMT. It's going to be in the clever tricks, not the brute force "make it go faster with more cubic inches" philosophy of the past.
I just don't see staying on that perf doubling every 2 years trend line anymore.
I don't understand the in and out of your field, but what he was saying was that you would ask a computer to basically create another computer that would have function x, and let it go to work.
In concept it sounds easy, but in reality, it's more difficult than that. Software engineers are generally here to stay for a while.
I just don't see staying on that perf doubling every 2 years trend line anymore.
Doubtless that's true.
But there are other ideas such as carbon nanotube integration, optical computers, the IBM design using sheets of silicon surrounded by gate material, quantum computing, 3D integration / scaling, artificial synapses, etc.
The thing of it is ... and just a part of why I'm skeptical about the ultimate power of AI ... is that Moore's law is basically dead. Both Intel and AMD are at 7 nanometer gate length processes and that's about as far as you go. Below that you begin to see quantum tunneling effects. It's the process shrink more so than logic design innovation that has driven Moore's law. Make it smaller, pack those gates tighter and go faster. We're hitting a physical wall now. You're going to see uprocessor perf trailing off until some fundamentally different computing technology arises. Most of the performance headroom from here on out is in clever tweaks to the basic super-scalar design. More and smarter speculation both in terms of branch prediction and data prefetch, wider multi-way issue and retirement ... 4-way, 8-way SMT. It's going to be in the clever tricks, not the brute force "make it go faster with more cubic inches" philosophy of the past.
I just don't see staying on that perf doubling every 2 years trend line anymore.
Performance of a single thread won't be doubling, but the # of computations you can do per second will. It'll all be pushed towards multicore architectures to artificially extend Moore's law. The GPUs of tomorrow might probably have over 10000 cores and still be of a reasonable size in 7nm, or maybe 5nm if we can get there.
But it's interesting times in the computing field that's for sure. Since multicore architectures are here to stay, RISC is the way to go, not CISC. lol
I suspect corporations will pay most taxes in a world where automation has replaced much of the workforce. The big question is what will happen to our society? If half the population is unemployed, many being unemployable what happens. Our capitalist system will not work if half the population cannot work. If they are left in poverty they will revolt and destroy the nation and our society as a whole. Socialism? I suppose that is a likely result, especially since millenials want that now. Even if these folks are taken care of through socialist programs what will they do with thier time as robots do all the work? The uber smart and educated will likely be very wealthy, how will the other 50 percent really live? I'm not sure I want to see what a society that discards half its population in lieu of robots will look like. Taxes are just a small part of problem.
Status:
"81 Years, NOT 91 Felonies"
(set 26 days ago)
Location: Dallas, TX
5,790 posts, read 3,597,628 times
Reputation: 5696
Or, we can simply say that Homo sapiens itself as we know it is obsolete and deserves outright extinction. Not active steps to exterminate us, of course, but have the robots allow us to die of neglect - kind of like the Neanderthals (notwithstanding the 4% of Neanderthal in non-Africans) - OR if we manage to program empathy and compassion into the robots and AI, they'll set aside a "Homo sapiens preserve" for us. Maybe they'll gradually "provolve" (proactive evolution) Homo sapiens into cyborgs, and then, gradually replace our natural neurons for something electronic-based..so we can keep our consciousness living existence during the transition process. Eventually we can become essentially AI/robots ourselves, even as we might keep our memories of our personal living existence as sapiens
After all, if survivability and efficiently are the names of the game, then we must concede that AI and robots are stronger, faster thinking, and faster moving and overall simply more fit to survive than Homo sapiens. So the choices are to concede we deserve either:
(a)extinction, on account of our lesser ability to compete and survive
(b)placement into a "sapiens reserve" where we can't inconvenience the AI-robot population, or
(c)concede we have to be "upgraded" / "assimilated"* if we wish to continue existing as conscious beings.
OR we can admit that some things are indeed more important than mere survival and/or efficiency - like retaining the essence of what it is to be human (whatever that may be). (although that doesn't contradict point b).
Hey, were I a pure laissez-faire capitalist / Social Darwinist, this is what I would have to conclude - especially since humans are stupid enough to create an entity that can outcompete itself.
*Sci-Fi / Space Opera fans should immediately get this one.
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
