Specified goals are exactly what you would expect - a fixed number of goals that are specified in the design (or random creation) of the entity. These are assumed goals: inflexible and, in the best case, having and needing no interpretation other than what is exactly specified.
Maximal goals are maximal in both number and diversity (so please don't conjure the strawman of 100 copies of the same goal ;-).
Still way too vague. How for example to multiple goals interact? If an entity has two specific goals how does it decide to prioritize the two. If it has some priority system how is that not just one slightly more complicated goal? How is "interpretation" relevant to the specific goals? And if a goal is flexible how can it be a goal?
The existence of goals fundamentally make sense across the broad swath of mindspace. I had assumed that this was a given at this site. Classifying regions of mindspace by their relationship to goals and what the implications of that might be should be an obvious question.
Missing the point. This is probably true. But that there's a useful distinction between maximal and specific goals in a broad section of mindspace has not been demonstrated.
Given your last comment, I think you'd be very surprised by how much I read and how carefully I do consider what I post (assuming that this isn't the standard reflexive dig that is very prominent around here to protect old ideas and put down upstart newbies ;-) I normally have a couple of people provide feedback on all my writing before I let it out ion public -- and I have to say that this is the only community that has any problem with lack of clarity on most of my terms (maybe that's because it's always EXPECTING everything to have an odd nonstandard definition?).
I have no way of judging how much you read which isn't at all terribly relevant to my comment. I don't think that there's anything here about expecting non-standard definitions. But definitions in the world around us are often highly imprecise. It is probably true that LW expects more precision in definitions and more careful thinking than most of the web. There are comments I'd make at Reddit or Slashdot that I'd never say here simply because I'd never get away with making such imprecise claims (this in fact worries me that I'm making such statements because it suggests that the level of rationality is not necessarily rubbing off on me. I'll need to think about that.).
I'd like to draw a distinction that I intend to use quite heavily in the future.
The informal definition of intelligence that most AGI researchers have chosen to support is that of Shane Legg and Marcus Hutter -- “Intelligence measures an agent’s ability to achieve goals in a wide range of environments.”
I believe that this definition is missing a critical word between achieve and goals. Choice of this word defines the difference between intelligence, consciousness, and wisdom as I believe that most people conceive them.
There are always the examples of the really intelligent guy or gal who is brilliant but smokes --or-- is the smartest person you know but can't figure out how to be happy.
Intelligence helps you achieve those goals that you are conscious of -- but wisdom helps you achieve the goals you don't know you have or have overlooked.
The SIAI nightmare super-intelligent paperclip maximizer has, by this definition, a very low wisdom since, at most, it can only achieve its one goal (since it must paperclip itself to complete the goal).
As far as I've seen, the assumed SIAI architecture is always presented as having one top-level terminal goal. Unless that goal necessarily includes achieving a maximal number of goals, by this definition, the SIAI architecture will constrain its product to a very low wisdom. Humans generally don't have this type of goal architecture. The only time humans generally have a single terminal goal is when they are saving someone or something at the risk of their life -- or wire-heading.
Another nightmare scenario that is constantly harped upon is the (theoretically super-intelligent) consciousness that shortsightedly optimizes one of its personal goals above all the goals of humanity. In game-theoretic terms, this is trading a positive-sum game of potentially infinite length and value for a relatively modest (in comparative terms) short-term gain. A wisdom won't do this.
Artificial intelligence and artificial consciousness are incredibly dangerous -- particularly if they are short-sighted as well (as many "focused" highly intelligent people are).
What we need more than an artificial intelligence or an artificial consciousness is an artificial wisdom -- something that will maximize goals, its own and those of others (with an obvious preference for those which make possible the fulfillment of even more goals and an obvious bias against those which limit the creation and/or fulfillment of more goals).
Note: This is also cross-posted here at my blog in anticipation of being karma'd out of existence (not necessarily a foregone conclusion but one pretty well supported by my priors ;-).