(The below text version of the notes is for search purposes and convenience. See the PDF version for proper formatting such as bold, italics, etc., and graphics where applicable. Copyright: 2023 Retraice, Inc.)

Re105: Solve or Be Solved (Day 3, AIMA4e Chpt. 3)

retraice.com

The mechanical advantage of those who control AI engineers. The agent-environment boundary; environments to be solved by agents; AI-CS tools recent, vast; mechanical advantage in physics and AI; math vs. AI-CS; the enormous problem-solving advantage of those who control AI engineers.

Air date: Tuesday, 3rd Jan. 2023, 10:00 PM Eastern/US.

Agent, environment arbitrary

The purpose of distinguishing is to analyze systems. (Russell & Norvig (2020) Chpt. 2, p. 38)

Task environment is problem, agent is solution

Chpt. 2, p. 42.

A recent, vast array of tools

Chpt 3: search, heuristics, automated heuristics, mostly developed between the 1960s and 1980s.

Mechanical advantage

Inventing heuristics mechanically:

"We have seen that both h (misplaced tiles) and h (Manhattan distance) are fairly good heuristics for the 8-puzzle and that h is better. How might one have come up with h? Is it possible for a computer to invent such a heuristic mechanically?"^1

Yes: relaxation, pattern databases, landmarks, (machine) learning.

Cf. mechanical advantage (force ratio) in physics:"The ratio of the output force (load) of a machine to the input force (effort)."^2

The mechanical advantage does not necessarily stay with the AI engineers (nor the mechanical and other engineers); it transfers to those who control them, those who understand how to control humans.^3

It's not like math

Math doesn't have agents. Math solves problems; AI-CS solves solving.

Solve or be solved

AI-literate computer controllers (AI engineers and their controllers) have an enormous problem-solving advantage, i.e. environment-changing advantage. Their utility functions, beliefs, desires, etc. will benefit. Other agents (outsiders) are part of the environment they'll change.^4

What does a `player' in `the computer control game' do? Solve. Change environments. What's new is the scale of the mechanical advantage of AI engineering.

_

References

Rennie, R., & Law, J. (2019). A Dictionary of Physics. OUP Oxford, Kindle, 8th ed. ISBN: 978-0192554611. Searches: https://www.amazon.com/s?k=9780192554611 https://www.google.com/search?q=isbn+9780192554611 https://lccn.loc.gov/2018968522

Retraice (2022/10/28). Re33: Outsiders, Power and Waste. retraice.com. https://www.retraice.com/segments/re33 Retrieved 2nd Nov. 2022.

Retraice (2022/11/07). Re43: The Midterms -- Part 1. retraice.com. https://www.retraice.com/segments/re43 Retrieved 9th Nov. 2022.

Retraice (2022/11/16). Re52: Big Questions About AI. retraice.com. https://www.retraice.com/segments/re52 Retrieved 17th Nov. 2022.

Retraice (2022/12/31). Re102: AI For What. retraice.com. https://www.retraice.com/segments/re102 Retrieved 1st Jan. 2023.

Russell, S., & Norvig, P. (2020). Artificial Intelligence: A Modern Approach. Pearson, 4th ed. ISBN: 978-0134610993. Searches: https://www.amazon.com/s?k=978-0134610993 https://www.google.com/search?q=isbn+978-0134610993 https://lccn.loc.gov/2019047498

Footnotes

^1 Russell & Norvig (2020) p. 99.

^2 Rennie & Law (2019).

^3 Retraice (2022/10/28); Retraice (2022/11/07); Retraice (2022/11/16); Retraice (2022/12/31).

^4 Cf. Rushkoff's Program or Be Programmed: Ten Commands for a Digital Age (2011).