OSDev.org

Brendan wrote:Intelligence is a word that was created before complex machines existed, that we use to refer to a set of familiar phenomena that makes humans different from inanimate objects, plants and machines.

"Humans are machines" is far more likely given the evidence than "humans aren't machines," and you haven't described what a not-machine would even be. Further arguments that humans aren't machines that lack any plausible, (hypothetically) testable alternative will be ignored.

Brendan wrote:You are only interested in diluting the meaning of "intelligence" to make it fit your unproven (and potentially false) "humans are machines" assumption; and because of this you are redefining the word to have a meaning that it could not have had when it was created (before complex machines existed, when humans believed they were not machines).

Using an existing word to describe a new thing that is explicitly designed to work and behave like the word's original referent is not diluting the definition, it's the entire point of human language. Further arguments about intelligence not referring to this behavior will be ignored.

Brendan wrote:By diluting the meaning of "intelligence" to make your flawed logic work, it loses any meaning at all - unintelligent (but complex) machines become intelligent, so slightly less complex machines are intelligent, so..... A dog turd is "intelligent" because when you step on it it "learns" your footprint.

Calling dog turds intelligent because you can get there by repeatedly removing complexity is a slippery slope fallacy. Talking about a complexity threshold when I never claimed complexity was a requirement for intelligence is a straw man fallacy. Further fallacies will be ignored.

Brendan wrote:Would you consider blindly trying all the possibilities intelligent (even though brute force approaches waste a huge amount of time on discarded work and are considered "least desirable" because of that)? I wouldn't - that's just an unintelligent machine creating an unintelligent machine.

This is a nice description of neural networks

No it's not.

Brendan wrote:VFS file format conversion stuff...

There is nothing "intelligent" about this in any way whatsoever.

Agreed. Good job writing another straw man fallacy though.

Hi,

Rusky wrote:

Brendan wrote:Intelligence is a word that was created before complex machines existed, that we use to refer to a set of familiar phenomena that makes humans different from inanimate objects, plants and machines.

"Humans are machines" is far more likely given the evidence than "humans aren't machines," and you haven't described what a not-machine would even be. Further arguments that humans aren't machines that lack any plausible, (hypothetically) testable alternative will be ignored.

It's not my fault if you're too stupid to understand that when the word intelligence was first given meaning "humans are machines" wasn't even close to being considered vaguely plausible; nor is it my fault if you're too stupid to understand that the word's meaning doesn't reflect the modern view of the human mind; nor is it my fault if you're too stupid to understand that your unproven "humans are machines" assumption is not the only possibility; nor is it my fault if you're too stupid to even consider the (less likely) "humans are not machines" possibility.

Rusky wrote:

Brendan wrote:By diluting the meaning of "intelligence" to make your flawed logic work, it loses any meaning at all - unintelligent (but complex) machines become intelligent, so slightly less complex machines are intelligent, so..... A dog turd is "intelligent" because when you step on it it "learns" your footprint.

Calling dog turds intelligent because you can get there by repeatedly removing complexity is a slippery slope fallacy. Talking about a complexity threshold when I never claimed complexity was a requirement for intelligence is a straw man fallacy. Further fallacies will be ignored.

Make up your mind. Either there is a threshold that distinguishes the difference between "large neural networks are intelligent" and "dog turds aren't intelligent", or it's a slippery slope.

Rusky wrote:

Brendan wrote:Would you consider blindly trying all the possibilities intelligent (even though brute force approaches waste a huge amount of time on discarded work and are considered "least desirable" because of that)? I wouldn't - that's just an unintelligent machine creating an unintelligent machine.

This is a nice description of neural networks

No it's not.

It's exactly how neural networks do work (potentially modified brute force for "training" that creates a system of weights).

Rusky wrote:

Brendan wrote:VFS file format conversion stuff...

There is nothing "intelligent" about this in any way whatsoever.

Agreed. Good job writing another straw man fallacy though.

What I described is GOAP. Are you saying GOAP isn't intelligent?


Cheers,

Brendan

Brendan wrote:Either there is a threshold that distinguishes the difference between "large neural networks are intelligent" and "dog turds aren't intelligent", or it's a slippery slope.

The threshold is there, but it's not in terms of complexity.

Brendan wrote:It's exactly how neural networks do work (potentially modified brute force for "training" that creates a system of weights).

No, it's brute force modified to the point that calling it brute force is silly.

Brendan wrote:What I described is GOAP. Are you saying GOAP isn't intelligent?

Yes, as I have from the start.

Brendan wrote:

Hellbender wrote:Anyway, intelligence is not be about the complexity of the machinery, but the complexity of the data. It does not matter if the mechanism is as simple as a look-up table. Intelligence is in building the table in the first place.

This is a nice description of neural networks - an unintelligent brute force approach to build an unintelligent table, with hype to scam fools into thinking it's "intelligent". This is the problem.
A processes submits a goal to the AI file engine, and (if necessary) the AI file engine uses intelligence to try to find a sequence of actions that achieve the goal. Of course new actions can be taught to the AI file engine at any time, to allow the AI file engine to learn to become better at achieving goals.
Now it's suddenly "intelligent". Yay!

You seen to consider intelligence as a binary value, whereas it should be considered a continuum (or maybe you just mix 'being intelligent' with 'having intelligence'). There is no intelligence in a rock, only a super tiny bit in a plant, a bit in an ant, some in a mouse, more in a dolphin, even more in a human, maybe a lots in some alien. Where on that continuum your system lands depends on the amount of data it can process and encode in a meaningful way (not just to store, but to generalize meaningful actions for situations not present in the data).

BUT: your system is not _intelligent_ JUST by being on that _intelligence_ continuum. It is considered intelligent ONLY if it can raise in the ranks of that continuum JUST by consuming more data, memory, and processing power (that is, it uses some 'learning'). We don't consider an ant intelligent although they can build extremely complex nests, cut leaves, cultivate mushrooms, etc. BECAUSE those ant keep repeating what they do even when it is not working. They don't learn from experience in a way we would expect an intelligent system to learn.

Let me rephrase that: AI system is something (that we believe) would be able to improve its decisions just by applying more number crunching, while non-AI systems (is known) to require a programmer to add better decisions. (That is what I meant by "Intelligence is in building the table in the first place" edit: but I should have said "being intelligent is in building a table that possesses intelligence", so I too intermix the terms, my bad..)

Brendan wrote: A processes submits a goal to the AI file engine, and (if necessary) the AI file engine uses intelligence to try to find a sequence of actions that achieve the goal. Of course new actions can be taught to the AI file engine at any time, to allow the AI file engine to learn to become better at achieving goals.

If your file system is able to add more file system descriptions by itself, without the aid of programmer, then it can be called AI and truly is different from non-AI solution. But if it has to be _taught_ to handle new filesystems by a programmer, then it is not an AI system.

That is also why a dog turd is not AI, while a neural network is.

Hi,

Hellbender wrote:

Brendan wrote:

Hellbender wrote:Anyway, intelligence is not be about the complexity of the machinery, but the complexity of the data. It does not matter if the mechanism is as simple as a look-up table. Intelligence is in building the table in the first place.

This is a nice description of neural networks - an unintelligent brute force approach to build an unintelligent table, with hype to scam fools into thinking it's "intelligent". This is the problem.
A processes submits a goal to the AI file engine, and (if necessary) the AI file engine uses intelligence to try to find a sequence of actions that achieve the goal. Of course new actions can be taught to the AI file engine at any time, to allow the AI file engine to learn to become better at achieving goals.
Now it's suddenly "intelligent". Yay!

You seen to consider intelligence as a binary value, whereas it should be considered a continuum.

Amount of intelligence is a continuum that ranges from zero to infinity. If something has a non-zero amount of intelligence, then it has intelligence (even if it's only a tiny amount), and if something has zero intelligence then it doesn't have intelligence. It's like "bool hasIntelligence(amount) { return (amount > 0); }".

Hellbender wrote:There is no intelligence in a rock, only a super tiny bit in a plant, a bit in an ant, some in a mouse, more in a dolphin, even more in a human, maybe a lots in some alien. Where on that continuum your system lands depends on the amount of data it can process and encode in a meaningful way (not just to store, but to generalize meaningful actions for situations not present in the data).

I wouldn't include plants as intelligent, for the same reason that I'd say parts of a human's brain that control involuntary actions don't really contribute towards human intelligence.

Hellbender wrote:BUT: your system is not _intelligent_ JUST by being on that _intelligence_ continuum. It is considered intelligent ONLY if it can raise in the ranks of that continuum JUST by consuming more data, memory, and processing power (that is, it uses some 'learning'). We don't consider an ant intelligent although they can build extremely complex nests, cut leaves, cultivate mushrooms, etc. BECAUSE those ant keep repeating what they do even when it is not working. They don't learn from experience in a way we would expect an intelligent system to learn.

For ants, nobody knows if they're intelligent or not.

I don't think storage and retrieval of data has anything to do with intelligence; even when it's the storage and retrieval of instructions (e.g. learning the recipe for a cheesecake by memorising it). I also don't think processing power (e.g. "floating point operations per second") has anything to do with intelligence.

As far as I'm concerned intelligence has much more to do with creating new ways to solve problems (without being told how to solve the problem).

Hellbender wrote:Let me rephrase that: AI system is something (that we believe) would be able to improve its decisions just by applying more number crunching, while non-AI systems (is known) to require a programmer to add better decisions. (That is what I meant by "Intelligence is in building the table in the first place".)

Perhaps, but AI currently doesn't do that, and I very much doubt AI will ever be able to do that.

What (some) "AI" currently does is solve a different problem. For example; rather than having a non-AI program where the programmer designs something to solve an "estimate the number of trees in a forest (from a description of the forest)" problem; they'll just have a non-AI program where the programmer designs something to solve a "generate a second non-AI program (from a description of the second program)" problem.

Hellbender wrote:

Brendan wrote:A processes submits a goal to the AI file engine, and (if necessary) the AI file engine uses intelligence to try to find a sequence of actions that achieve the goal. Of course new actions can be taught to the AI file engine at any time, to allow the AI file engine to learn to become better at achieving goals.

If your file system is able to add more file system descriptions by itself, without the aid of programmer, then it can be called AI and truly is different from non-AI solution. But if it has to be _taught_ to handle new filesystems by a programmer, then it is not an AI system.

Yes (assuming you meant "add/create more file format converters by itself"), but AI currently doesn't do that, and I very much doubt AI will ever be able to do that.


Cheers,

Brendan

Hi,

Note that your "humans are just complex machines" religion

A dog turd is "intelligent"

It's not my fault if you're too stupid to understand

Can you please stop wasting your time debating over whether something is intelligent or not? Also, can you stop flaming each other?

I think I have to agree with Combuster...


Regards,
glauxosdever

It may be a good idea to enumerate all messages and keep their text aside.
You can have short/devel messages as a comments in the source code and generate dictionary entries with designated source code scanner. Then you manually/semi-automatic translate these entries and compile into translation module.
During development it would require merging existing dictionary with newly generated list of entries.
It's up to you when you do it, so it should be rather convenient.
Additional advantage is that you don't need to be accurate with source-code messages, as they're going to be translated anyway.