Re: [sl4] How much do languages matter for AGI programming?

From: J. Andrew Rogers (
Date: Fri Nov 28 2008 - 12:13:30 MST

On Nov 28, 2008, at 6:36 AM, Edward Miller wrote:
> MapReduce does sound like a good idea. It is open source, and this
> has allowed people to port it to Java.

MapReduce is designed to solve a very specific kind of problem, and
using it for anything outside of that scope is going to be a waste of
time relative to alternatives. I have never seen significant evidence
that AGI might be a MapReduce-optimal problem, and so using MapReduce
would probably be a bad idea since it generalizes poorly.

> We need to find the best language.

Even if I accepted that there was a quasi-objective "best" language,
which I do not, you have not defined the problem the language would be
applied to well enough to demonstrate a language's "best"-ness. You
can't simply assert "best"-ness, so what are the specific metrics that
prove optimality of a language for AGI purposes that we can use to
evaluate the claim?

> Would you want an AGI that is prone to buffer overflows? These sorts
> of problems can make the AGI very vulnerable to hacking,
> malfunction, or complete failure. I am assuming it would be best to
> implement the most rock-solid language(s) possible.

This is a programmer problem, not a language problem.

What language do you think these "rock-solid language(s)", libraries,
and operating systems were implemented in by mortal programmers?
Throwing problems over walls do not make problems go away. Not taking
responsibility for a problem is not the same as solving the problem.


J. Andrew Rogers

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:03 MDT