--- On Sun, 9/21/08, David Hart <[EMAIL PROTECTED]> wrote: >On Mon, Sep 22, 2008 at 10:08 AM, Matt Mahoney <[EMAIL PROTECTED]> wrote:
>> Training will be the overwhelming cost of AGI. Any language model >> improvement will help reduce this cost. > How do you figure that training will cost more than designing, building and > operating AGIs? Unlike a training a human, training an AGI for a specific > task need occur only once, and that training can be copied 'for free' from > AGI-mind to AGI-mind. If anything, training AGIs will cost ludicrously *less* > than training humans. Training the first few generations of AGI individuals > (and their proto-AGI precursors) may be more expensive than training human > individuals, but the training cost curve (assuming training for only the same > things that humans can do, not for extra-human skills) will eventually > approach zero as this acquired knowledge is freely shared, FOSS-style, among > the community of AGIs (of course, this view assumes a soft takeoff). An organization is most efficient when its members specialize. It is true we don't need to build millions of schools to train child AI's. But every person has some knowledge that is unique to their job. For example, they know their customers, vendors, and co-workers, who to go to for information. It costs a company a couple year's salary to replace a white collar employee, including the hidden costs of the new employee repeating all the mistakes made by the previous employee in learning the job. I estimate that about 5% of the knowledge useful to your job cannot be copied from someone else. This fraction will increase as the economy becomes more efficient and there is more specialization of job functions. -- Matt Mahoney, [EMAIL PROTECTED] ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244&id_secret=114414975-3c8e69 Powered by Listbox: http://www.listbox.com
