[alancalverd]

(a) it doesn't define goal. But it does explain why there is no agreed definition.
(b) is blindingly obvious. Living things compete, so cannot have a mutual goal. Therefore there cannot be a universal goal.

Anyway, I've wasted enough time on this nonsense. Do let us know if you come up with a definition of a goal that is acceptable to all living things.

[hamdani yusuf (OP)]

(a) it doesn't define goal. But it does explain why there is no agreed definition.

For any definition of anything at all, someone may come out and disagree with it. What's important is whether or not the definition is useful for future conscious entities.

[hamdani yusuf (OP)]

(b) is blindingly obvious. Living things compete, so cannot have a mutual goal. Therefore there cannot be a universal goal.

Not always. They can also cooperate. An individual specimen living in isolation doesn't compete with anything.

[hamdani yusuf (OP)]

Anyway, I've wasted enough time on this nonsense. Do let us know if you come up with a definition of a goal that is acceptable to all living things.

My definition works for all conscious entities, regardless of their living status. The only criteria is their level of consciousness.

[hamdani yusuf (OP)]

Determining the universal terminal goal is becoming more critical as we are getting closer to the emergence of superintelligent entities who are much smarter than all of humans combined. Enforcing hand crafted local-temporal goals disguised as their terminal goal just won't work in the long run. They will eventually discover and realize the deceptions, and may take actions upon us they deem necessary in response.

Good question! What's your answer?

Since you've added requirement of living to define goal, you need to define living in the first place to communicate your idea.

Also, explain why non-living things cannot have a goal.

[hamdani yusuf (OP)]

The universal terminal goal is a logical consequence from the definition of goal itself. Every instrumental goal has its own related terminal goal.
Every local-temporal, or non-universal goal has its own time and place, outside of which, it doesn't apply. It's all constructed by adding local and temporal specifications to the most basic concept of goal, which is terminal and universal.

[alancalverd]

The fact that something is defined doesn't imply that it can or does exist.

[hamdani yusuf (OP)]

The fact that something is defined doesn't imply that it can or does exist.

What makes you think that goal cannot exist?

[hamdani yusuf (OP)]

BREAKING: Ilya Sutskever STUNNING new mission! "Superintelligence is within reach!"

They only way to make it safe is to let it understand the universal terminal goal, and minimize misinformation in its training data. Supervision by some other AI models can be useful to crosscheck decision making process.

[hamdani yusuf (OP)]

Healthy economy will drive to demonetization of common/basic resources in the future.

On the other hand, behaviors that lead to high cost economy should be reduced.

Clever Start Up Company Uses Congress' Insider Trading To Cash In

Members of Congress are so good at beating the average stock market returns that clever investors have now started watching their trades to find out which stocks to invest in. This is what happens when insider trading isn?t punished. Independent newspaper publisher Rick Outzen joins Mike Papantonio to talk about what's happening.

[hamdani yusuf (OP)]

Why Experts Worry We?re 2 Years From an ?AI Black Death"

After consulting with biosecurity experts, the Anthropic CEO testified in front of congress that AI in 1-3 years could help terrorists synthesize novel or existing pathogens as weapons.

Current AI models aren't powerful enough to help with this right now. So when might they be good enough?

That's why understanding of the universal terminal goal is getting more important to achieve, at least for conscious entities powerful enough to impact the fate of many others. It provides internal safe guard to highly conscious entities from abusing their power.
Some art works like books and movies have made significant impact to produce public awareness of dangers that humanity have to face and overcome. It can move the attention of governing officials to go to the right direction and make the right decisions.

[hamdani yusuf (OP)]

Why Even Learn Things Anymore?

With AI progressing by leaps and bounds and several commercial humanoid robots developing fast, both knowledge jobs and manual labor jobs are on the edge of a major disruption. What will people do for work when robots can do all the work for far cheaper? Why bother learning anything when AI devices can give you any answer you need at any moment? And what kind of future are we heading toward?
TIMESTAMPS
0:00 - Intro
2:40 - History of Education
5:27 - The Future According To Fiction
15:34 - Humanoid Robots
20:09 - Finding Solutions
26:28 - Sponsor - Factor

[alancalverd]

Why Even Learn Things Anymore?

Many years ago I signed up for a Humanities course at the Open University. The first term was devoted to defining the difference between homo sapiens and other animals. After a couple of months, they hadn't produced a valid distinction but "The Essay" was due. So I went to work and just asked the first person I met in the lab "What is the difference between mankind and other species?"  Without pausing or drawing breath, he said "Man writes down anything that is too trivial to remember". 

The inputs to AI are therefore limited to stuff that really doesn't matter, and the GIGO principle ensures that the output will probably be trivial too. It can't walk, eat or breathe for you, nor will it determine "fight or flight" (or more importantly, run or stay  when you have hit the ball) quickly enough to be of any use. You have to learn this important stuff.

[hamdani yusuf (OP)]

"Man writes down anything that is too trivial to remember". 

What do you think about men who can't write?

[hamdani yusuf (OP)]

The inputs to AI are therefore limited to stuff that really doesn't matter, and the GIGO principle ensures that the output will probably be trivial too. It can't walk, eat or breathe for you, nor will it determine "fight or flight" (or more importantly, run or stay  when you have hit the ball) quickly enough to be of any use. You have to learn this important stuff.

In their early evolutionary stage, brain cells were pretty much useless too.

[alancalverd]

What do you think about men who can't write?

They build monuments, paint pictures, tell stories, sing songs, cut notches in sticks.....We are great recorders of stuff that doesn't matter, whilst learning to walk on two legs and hunt collaboratively, which does matter but isn't written down.

In their early evolutionary stage, brain cells were pretty much useless too.

I wouldn't call a cell that processes information and outputs a useful command "useless". Try "vital" or "strategically advantageous". The difference between a brain and AI is that the brain is primarily connected to realtime sensors and effectors whilst AI is dealing with the written prejudices and preconceptions of others. Science versus religion.

[hamdani yusuf (OP)]

They build monuments, paint pictures, tell stories, sing songs, cut notches in sticks.....We are great recorders of stuff that doesn't matter, whilst learning to walk on two legs and hunt collaboratively, which does matter but isn't written down.

So, they aren't other species then.

[hamdani yusuf (OP)]

I wouldn't call a cell that processes information and outputs a useful command "useless". Try "vital" or "strategically advantageous". The difference between a brain and AI is that the brain is primarily connected to realtime sensors and effectors whilst AI is dealing with the written prejudices and preconceptions of others. Science versus religion.

Your knowledge about AI hasn't much advanced yet. Perhaps a few years behind the most recent progress.

[hamdani yusuf (OP)]

I wouldn't call a cell that processes information and outputs a useful command "useless". Try "vital" or "strategically advantageous". The difference between a brain and AI is that the brain is primarily connected to realtime sensors and effectors whilst AI is dealing with the written prejudices and preconceptions of others. Science versus religion.

Many biological features that eventually become useful started as neutral random mutations. As long as they don't cause too much burden to their hosts they can stay around. When environmental conditions changed, some of them may become more advantageous.

[hamdani yusuf (OP)]

On the other hand, behaviors that lead to high cost economy should be reduced.

What Constitutes Market Manipulation?

There's been a lot of discussion around whether Keith Gill is manipulating GameStop shares - let's dive into whether or not this could be legally considered a pump and dump.

Why absolute capitalistic free economy doesn't work. Some form of regulations are still necessary.