The problem: Will human death really stop in 40 years? It's been going on for millions of years, and we still haven't solved it at all really.
What I wrote recently could use a bit more proof. Here is a more detailed explanation of how it will happen soon and be practical: Once we create AGI - which we almost have (openAI.com, and Facebook's Blender chatbot) and should have by 2029 (at our rate of progress and seeing how few things left there is to do to make AGI; RL for thought, notepad motors, executive control to enact it or not, Cerebellum motor target & smoothing), and assuming AGI will be fairly similar to GPT-2 in both accuracy and resource needs: In it is now the year 2029 we will be able to train it into an adult in a few days or weeks, as that is how long it took GPT-2/ DALL-E basically. It will be quite wise on all topics. Goals can be given to it later, to make it think of certain things often, but we may need a dataset hand made like check things twice, humans, loving humans, hugs, computers, etc, and a value how often they should be said, because human values are said everywhere: fun, love, cash, computers but the ones they need to say are rarer and are just becoming common still, we must bias the thing towards immortality at the root of its raw desire and what it needs to do next at the other end - boring hard rare education. It will clone 1,000,000 times (for every clone it can clone 2 times faster if it takes too long. Should not take long. Months at most.), and they will work great together because they are clones. They will work on different tasks the original wanted to do but had little time to do. They will /all/ be made to love working on ASI due to the engineers, and due to knowing what they are - human but machine. If we have 7,000,000,000 humans on Earth and 1 of 1000 works on AGI, then having 1,000,000 AGIs all work on ASI should be equivalent to doubling humanity's progress towards AI. We should be able to manage running 1M AGIs in parallel, we have many computers and ex. google centers around Earth, and to be clearer: if each human can buy and run a 1,000$ computer, then we would have enough computers to run them all, and still have enough time to do other work while our computers run. Now, assuming we can make all 1M AGIs run 10 times faster than human brains, we will have 100 years of AI progress in 10 years; intelligence/ speed/ data. That is very good, because then soon we will have 100 years per 1 year, until they have become ASIs that are each 1,000 years old. They will test code like we do but in their brain, so they need no body to come up with ideas and see what's missing/ issues in predictions, they will improve their code! Even still we are not sure; ok, we have 1000 times faster human progress now on cars/ cancer/ etc, but with 1,000 years humans still hadn't cured cancer, why will these ASIs? Let's remember humans are loops, we are an eye>info processor>motor loop, and we use tools to build or use tools. It was somewhat recently in evolution (*50,000–150,000 years ago*) it was the age of General Purpose Language which allowed us to learn quicker/feel like we are 10,000 years old and seen all, and allowed us to work together. Then was the age of General Purpose human labor automation which allowed many humans to work together and shift their job over. Then much later (see below) came the age of General Purpose machine labor automation which allowed scaling it further. "1760 to 1820 and 1840 ... This transition included going from hand production methods <https://en.wikipedia.org/wiki/Craft_production> to machines <https://en.wikipedia.org/wiki/Machine>, new chemical manufacturing <https://en.wikipedia.org/wiki/Chemical_manufacturing> and iron production <https://en.wikipedia.org/wiki/Puddling_(metallurgy)> processes, the increasing use of steam power <https://en.wikipedia.org/wiki/Steam_power> and water power <https://en.wikipedia.org/wiki/Water_power>, the development of machine tools <https://en.wikipedia.org/wiki/Machine_tool> and the rise of the mechanized <https://en.wikipedia.org/wiki/Mechanization> factory system <https://en.wikipedia.org/wiki/Factory_system>. The Industrial Revolution also led to an unprecedented rise in the rate of population growth. ... The Industrial Revolution marks a major turning point in history; almost every aspect of daily life was influenced in some way. In particular, average income and population began to exhibit unprecedented sustained growth. ... Six factors facilitated industrialization: high levels of agricultural productivity to provide excess manpower and food; a pool of managerial and entrepreneurial skills; available ports, rivers, canals, and roads to cheaply move raw materials and outputs; natural resources such as coal, iron, and waterfalls; political stability and a legal system that supported business; and financial capital available to invest. Once industrialization began in Great Britain, new factors can be added: the eagerness of British entrepreneurs to export industrial expertise and the willingness to import the process. Britain met the criteria and industrialized starting in the 18th century." " *1834 - **Thomas Davenport* <https://edisontechcenter.org/DavenportThomas.html> of Vermont developed the first real electric motor ('real' meaning powerful enough to do a task) although Joseph Henry <https://edisontechcenter.org/JosephHenry.html> and Michael Faraday <https://en.wikipedia.org/wiki/Michael_Faraday> created early motion devices using electromagnetic fields. The early "motors" created spinning disks or levers that rocked back and forth. These devices could not do any work for humankind but were important for leading the way to better motors in the future. Davenport's various motors <https://edisontechcenter.org/DavenportThomas.html> were able to run a model trolley on a circular track and other tasks. The trolley later turned out to be the first important application of electric power (it was not the light bulb). Rudimentary full sized electric trolleys <https://edisontechcenter.org/TrolleysTrams.html> were finally built 30 years after Davenport's death in the 1850s. *The electric motor's world impact before light bulbs:* Trolleys and the connected power systems were very expensive to build but transported millions of people to work in the 1880s. Until the growth of the power grid in the 1890s most people (middle and low classes) even in cities did not have the electric light in the home. It wasn't until 1873 that the electric motor finally achieved commericial success. Since the 1830s thousands of pioneering engineers have improved motors and created many variations. See other pages for more detail on the electric motor's vast history. *Motor leads to the generator:* After weak electric motors were developed by Faraday and Henry, another early pioneer named *Hippolyte Pixii* figured out that by running the motor backwards he could create pulses of electricity. By the 1860s powerful generators were being developed. The electrical industry could not begin until generators were developed because batteries were not an economical way to power society's needs. " Then (at ~1850-~1900) we made General Purpose image displays and General Purpose cameras and General Purpose printing of images, and same for speakers and microphones. Which STORED language so others can read it as mass scale without having to say it yourself. That was the age of information. ~1950 we made the General Purpose computer that can run any algorithm: antivirus software, AI, physics sims, calculators, etc. That was the age of Software (previously all was Hardware!). Physics sims are even general purpose, they can create/predict new scenes correctly even though not programmed the answer! ~2017 the age of intelligence we made General Purpose AI (almost) that solves any problem, it can simulate a physics/hardware simulator at much less the cost by /just/ predicting the images, it will be able to come up with new tools on its own. The next thing to come, perhaps by yet another (notice the patterns above) 50 years is the age of nanotechnology, 2030+50=2080, maybe Matt is right? The ASIs will be able to manage this (and control them all perhaps), you can even do a lot with just 100 nanobots that don't duplicate themselves. This will be a General Purpose tool, like a general purpose 3D printer but way better, a tool that can become a hammer and swing that hammer by morphing foglets, they can go so small it can go into your blood vessels and repair everything. They will like on Terminator T3000 and T1000 be able to look at someone's face and voice, and mirror them, you do this by predicting the "video" like DALL-E and then deciding to enact it using a 2nd prediction 'do_it' and then using the Cerebellum to smoothly let the nanobots reach the target image in the mind that was predicted by comparing external input to the desired image in brain: mind control to make your body become and do anything, whatever you see in your brain can make the nanobots enact it all! I can already see someone's features and enact them to some extent, making my neck 1 inch longer, puffing my checks, making my tummy poke out, and talk higher like them, and put makeup on my face in the mirror (more timeful though). While Matt is right about nanotech, I think the fact that AGIs will speed up progress by 10 - 100 times will cause it not to be 2080 or 2100, but 10 times sooner; 10 or 20 years! So it will be by 2030+10 or 20 = 2040 or 2050! It is true! Omg! That is the date others predict. So AGI causes ASIs, which causes the next step in the pattern: nanobots, which with this increase of computation and matter manipulation crunching, we can expect a huge leap in the ability to learn about individual body profiles and repair them too, and upgrade them eventually into nanobots in an "acceptable" way. After the age of nanobots will be the age of extreme miniaturization and scaling, as it can not further go smaller or improve the technology anymore. All it can do in the end is grow bigger. This is exponential, as it grows it can reach more volume. Growing larger = less probability of dying, it takes longer to lose those cells than the time it took to grow them. No one will be in a VR sim eventually because we will live longer if we are the nanobots and defending the sphere/ cube. Yes AGI works on patterns and so does predating in your environment, organization is better. Colder is too. 3D environment is better too than 2D plane. Everything in your home is grouped together by similarity and shaped like square, homes are lined up (like trie trees with weights because a network doesn't store the same word twice). We'll need a lot of energy, nanobots, info, and processors, and sensors, so it is necessary to seek scale soon. If evolution took this long to invent humans and we only needed 100 years (+ a few more hundred of knowledge) to invent AGI, then surely we will invent something really powerful soon. It will have far greater intelligence and speed. And scale: it only takes a doubling if you "do it right" - and doubling causes a huge increase in productivity. The next question is what are these nanobots and how do they carry out all the informed procedures to cure us? They also need to analyses what is inside you first to know what is wrong. It's likely they will work wirelessly or talk to neighbors as a free floating network, because a single nanobot would be very stupid doing anything much for us in a blood vessel. They need the ability to have energy needs met, and manipulation needs met - general purpose such mechanisms. Since they are reliable seeing that, we should zoom out and look at the larger systems: what the goals and thoughts are higher up. They might be looking for symptoms, and carrying out tasks together. So you would have multiple ASIs in you, each a network free floatingly, and for a given ASI network it would have a brain being the network and it would have many many fingers being the nanobots, able to think of moving them all or only some based on what it predicts visually. They would need to stay connected like a body so none get lost, and would need to make sure it knows what's happening to most of them in case some get clogged up in some vessel, i.e. it would need to have some thoughts/ goals and therefore every nanobot and/or group of nanobots or groups of groups would have some agendas each follow automatically, on their own, without the brain thinking about doing them. The highest part would be thinking of larger goals, the lowest nodes would be thinking about doing more static goals given to them from higher up, they are more simpler tasks that are elementary like letters in a sentence, only 26 letters are commonly used, compared to 50,000 word vocab. So far that seems weird. Will it even fit? Will it all work together? Will it get clogged and fail doing much? I think even just a few nanobots could work. More just means it can do things faster at scale. And only with more nanobots can it do larger tasks, like lift organ B up quickly. Thoughts? Anyway back I go to my work. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T956b74065f76478b-M0498ade7541f4c993a4be369 Delivery options: https://agi.topicbox.com/groups/agi/subscription
