=\ Mentifex has a defined crackpot value of 1.0, he is the benchmark, archetypal crackpot. He has earned this rating through 40 years of tireless effort. =|
Your (Matt) crackpot score is somewhere between 0.48 and 0.63. Since it's been several years and since you haven't proven that you are irredeemable, for the sake of being a decent person I'll give you another chance here. In all honesty, I'm not that hopeful. =\ Matt Mahoney via AGI wrote: > Like everyone else on this list, I do not have a working AGI. (crackpot score adjusted downward, congrats.) > It is easy to underestimate the enormity of the problem. The most obvious > application of AGI is to automate human labor. Globally this is a USD > $75 trillion per year problem. Cognitive error: conflation of issues. The applications and/or apocalyptic effects of unleashing AGI on the unsuspecting planet is not actually on-topic for the list. You are welcome to start a thread on that subject but recognize that it is strongly counter-productive to be thinking about the effects of AGI while working on solving AGI... (crackpot score adjusted upwards for obsession on this one single problem.) (crackpot score adjusted upwards for confused thinking.) > A working solution would have a ROI of > world GDP divided by market interest rates, about $1 quadrillion. This is not actually relevant. (crackpot score adjusted upwards for poor understanding of economics, displacing a quadrillion dolars of something doesn't give you a quadrillion dollars, it adjusts the cost of that thing downward, what your actual net revenue will be at the end of the venture must take into account the affect on the price structure of the market.) (crackpot score adjusted way upwards for assuming, without due consideration, the continued existence of money in a post-singularity world.) > When investors won't even put in $1 million into your project, they are > effectively setting odds of a billion to one against your success. > When you count the number of false promises and failures to solve AI > since the 1950's, that's not unreasonable. (crackpot score adjusted DOWNWARD for having a reasonable model of investor risk calculations...) > It's not that I haven't tried. Ten years ago I proposed an AGI > composed of billions of narrow AI specialists and a distributed index > to route your requests to the right agents. Building this would > require a global effort over decades and an economic infrastructure > that rewards providing useful services in a hostile and competitive > distributed computing environment. Agents would compete for attention > and reputation in a world where information has negative value. The > distributed index provides a message posting service, where all > messages are public, signed and dated and cannot be edited or deleted > once posted, and are routed to anyone who might care. You can find the > proposal at http://mattmahoney.net/agi2.html The Great and Powerful Goertzel's Singularity Net is basically precisely that, I am extremely skeptical but I'll leave your crackpot score unchanged on this. > Such computers exist but require over 1 megawatt of > electricity, compared to about 20 watts for the human brain. You would > need 7 billion of these computers to replace 7 billion workers, > requiring 7000 terawatts of power. Global energy production is > currently 15 TW. This kind of power reduction is not going to be > achieved by further shrinking transistors, which are already close to > the limits of physics. It will require computing by moving atoms and > molecules rather than electrons, something that biology has already > figured out but is still a long way off for us. (poor engineering acumen but acceptable back of the envelope computation, crackpot score increased but only slightly...) Hint: you would not replace workers 1:1, the ability to adjust the structure of an AGI agent would allow more efficient business structures, compounded by not needing entire departments such as HR and such which would be made redundant... > Perhaps there are more efficient solutions to AGI. Sure, for some > problems, like arithmetic. But deep neural networks are still the best > we have for vision and probably language. The problem is that > intelligence (measured by reward or prediction accuracy) over a wide > range of problems increases logarithmically with CPU time and memory > because the problems themselves have a power law distribution over > difficulty. This clearly shows as the economy grows linearly while > computing power grows exponentially. It shows in my own work in data > compression as a measure of natural language prediction accuracy. The > two graphs in http://mattmahoney.net/dc/text.html are 10 years old but > the relationship hasn't changed. (Crackpot score increased significantly because jumping to engineering an AGI rollout before doing the actual R&D...) > Ambitious, decades long AGI projects like OpenCog and NARS are > knowledge representation frameworks with empty knowledge bases and > inadequate hardware to solve any useful problems even if they were > populated. Why? Because software is easy and human knowledge > collection is hard. Gronk? Software is hard. =| > The design complexity of the human body cannot > exceed the compressed information content of our DNA, which I > estimated to be 300 million lines of code. (These numbers from my 2013 > paper "The Cost of AI" published in 2017 as a chapter in "Philosophy > of Mind: Contemporary Perspective" (Curado, Gouveia eds). [ugh] > http://mattmahoney.net/costofai.pdf ). This is doable for $30 billion, > a tiny, insignificant fraction of the total cost. What are we calling "This" again? > Knowledge collection is much harder. Little kiddies seem to do an amazingly good job at it though, and not for not much money either. > Human long term memory capacity is 10^9 bits, of which > 1% is unique to you and can only be extracted by speech or typing at 7 > bits per second costing $5 per hour at global average wage rates. This > is only practical (and still costing hundreds of trillions to collect > 10^17 bits) by giving up privacy and publishing your personal data. > Otherwise you are stuck with endless surveys and repeating the same > data over and over. How many times today did you have to type in just > basic info like your name or email? (crackpot score increased massively because of WTF.... An AGI that requires this much data collection to function does not actually qualify as an AGI.) While data collection is a huge subject in mainstream AI research, it is actually not interesting for AGI because the focus of research is on how to perceive and respond to the environment, not to somehow process massive quantities of data. (Crackpot score increased massively for failing to demonstrate any evolution in your thought process over many years...) > Hoping to stimulate some intelligent discussion. (crackpot score adjusted downward for acknowledging that other people might have something to say, unlike Mentifex who just wants everyone to bask in his brilliance..) -- Please report bounces from this address to a...@numentics.com Powers are not rights. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T731509cdd81e3f5f-Mf5dbaf0d818cfaed8b80c2ad Delivery options: https://agi.topicbox.com/groups
