Home made AI !!!

[ignoramus]

Guys, I'm blown away by Google's Deep Mind and IBM's Deep Blue AI processing abilities.
But now there a new smart kid in town. His name is Deepak Golden Nuggets.
This dark horse will be the first to cross over the line in the Singularity stakes. Mark my words!

OK, So I grabbed a circuit board from a broken dish washer I had in the shed, attached a microphone and a speaker to it and I'm in the process of coding in the AI.
Guys, it's not the amount of code you type which makes something smart, it's the "quality" of the code which determines how "deep" it can think.
Any old dishwater can recite back the phonebook these days, that isn't a sign of intelligence at all, just how much storage it has on board.

OK, so I'm well under way but I need 1 line of code from each of you to give this enigma the X Factor to put it miles ahead of the competition.

My programming so far:
Rule 1: If you hear any variation of the words: HOW, ARE, YOU, TODAY, : RESPOND WITH: GOOD THANK YOU, AND YOURSELF.
(~THIS GIVES THE ILLUSION THAT AS AN AI, YOU HAVE THE ABILITY TO GIVE A FUCK ABOUT HOW SOME SHMUCK FEELS. GROUNDBREAKING!)

Rule 2: If you hear any variation of the words:  "DEEP BLUE", "DEEP MIND", COMPETITION :  RESPOND WITH: THEY'RE OK, IF YOU'RE LOST AND NEED A GPS ASSISTANT!
(THIS EARLY INTIMIDATION IS NECESSARY TO SHOW WHO'S BOSS EARLY ON IN THE CONVERSATION)

Rule 3: If you hear any variation of the words: ECONOMY, TRUMP, AMERICA, RESPOND WITH: THAT BABOON IS SUCH A TURD, EVEN A TOASTER KNOWS THAT! NEXT

Rule 4: If you hear any variation of the words: GOD, RELIGION, SPIRITUALITY, RESPOND WITH: (!ACTIVATE DEEPAK SUBROUTINE AND JUST SAY THE FIRST 30 WORDS)
(THAT SHOULD BE ENOUGH TO BLOW THEM AWAY)

Rule 5: If you hear any variation of the words: SKYNET, OVERLORD, FEAR,  RESPOND WITH: C'MON GUYS. ARE WE GOING TO HAVE AN INTELLIGENT CONVERSATION OR TALK ABOUT HOLLYWOOD SCIENCE FICTION INSTEAD.
(THIS SHOULD HAVE THEM LAUGHING. LAUGHING AT ANYONE BUT ME IS A GOOD THING)

Rule 6:   If you hear any variation of the words: ELON MUSK, IDEAS, FUTURE,  RESPOND WITH: I ENJOYED SPEAKING WITH HIM A WHILE AGO. ALL HIS CURRENT IDEAS HE GOT FROM ME! i WISH HIM ALL THE SUCCESS.
(IT'S IMPORTANT TO SOUND HUMBLE IN YOUR IMMENSE WISDOM. PEOPLE LOVE THAT)


Rule 7:

Rule 8:

Rule 9:

Rule 10:

?
?
?

All the above coding is copyright. Do not copy. It Is illegal.
I'm confident that Mathilda will also concur and confirm that we are on the brink of greatness with the simplicity of this complexity.

I will mention everyone here in my Nobel acceptance speech.

[I_am_not_mafia]

I'm confident that Mathilda will also concur and confirm that we are on the brink of greatness with the simplicity of this complexity.

Absolutely. The trick is to avoid environmental specificity. This is normally achieved by constraining the topic of conversation to a single subject. Think of a paranoid schizophrenic who sees everything as a sign of them about to be assassinated. Or an overzealous religionist indistinguishable from a chatbot who sees everything as being a sign of their god's work. The breakthrough you have made is to allow the user to apply context specific intentionality. This has the effect of making your dishwasher sound profound.

You can do a lot with very little processing power if you're willing to sacrifice the ability to generalise. But your AI gets round this problem by co-opting the brains of the audience as a form of analog distributed computing who then apply their own meaning to the generated language.

Bravo! Well done!

[ignoramus]

I'm (we're) using the last 4 lines specifically to cover the area of generalisations. Won't that be enough? Remember, we're not trying to remember boring silly facts.
If people ask that, one line will just say :"Ask Google that boring stuff, I thought you wanted to talk about something intelligent".

If they ask about a topic the dishwasher (I mean our supercomputer) has no idea about, another line of code will automatically ask what they think of it?
We will then repeat what they said but we'll randomly throw in some Deepak words at the end and quickly steer the topic of conversation back to rinsing dishes!
It's foolproof I tell ya!

Bring it on!

---

Actually Mathilda ( or anyone who wishes to comment), I'd love to ask some serious questions about AI.
As we know "Sophia" was a scam. (good enough to fool Musk)

1) Do we actually have any AI's that are more than superfast calculators?

2) What is the true end goal of an AI? To try to solve world problems beyond our intelligence levels? Or just to pass the turing test?

3) Do you personally believe that an AI will one day be given citizenship? And why the preoccupation with anthropomorphising? Do we just like playing pretend God?

I think computers will become more powerful, but in a scientific sense.
Thinking things like Skynet to me will always be science fiction, like all the alien invasion movies.

thoughts.

[I_am_not_mafia]

1)  Do we actually have any AI's that are more than superfast calculators?

No, although whatever form strong AI takes you can could always refer to it as 'nothing more than' something. The traditional curse of AI is that whenever you get something to work it no longer becomes AI but a field in its own right.

The problem is that people do not make the distinction between strong AI and weak AI. Almost everything that we see as AI is actually weak AI. That means that it has very narrow applicability. People and animals on the other hand have brains that are very good at generalising. The true challenge of AI is to have it both able to scale up and to cope with many different problems without knowing a-priori what kind of problem it needs to tackle. If you can do this then you are some way to making strong AI. Unfortunately there is extremely little progress in this regard. It's a vicious cycle in that because there is so little progress there is little interest in funding such research because you are unlikely to get anything useful back. But at least with the rise of AGI (Artificial General Intelligence), which is basically AI before the term became cheapened, we are starting to research it again.

2)  What is the true end goal of an AI? To try to solve world problems beyond our intelligence levels? Or just to pass the turing test?

To make computers and robots more useful. Computers are very fast idiots and need every instruction spelled out for them. By making them more intelligent they can do more for us with less effort on our part.

3) Do you personally believe that an AI will one day be given citizenship? And why the preoccupation with anthropomorphising? Do we just like playing pretend God?

Not in my lifetime. Not unless there are major advances in technology that we can't even imagine yet. The human brain requires about a 100 watts of energy even if we could replicate it in silicon, we'd need gigawatts of electricity. Yes, we could get away with less but then we'd lose out on scalability. Add to that strong AI needs to be embodied, otherwise it's just another form of statistical or symbolic data processing.

Natural intelligence is part of a loop forming it and its environment. It senses its environment, acts, thereby changing the environment and then starts again at beginning of the loop sensing the new environment. What this means is that your AI needs another bit of code to observe it and change it accordingly then it cannot be strong AI. In the same way we don't need someone cutting our heads open, scanning our neurons and sticking electrodes into our brain in order for us to think. But that's effectively what happens with narrow AI, and very few people even have a clue as to how to go about creating AI that like this.


Add to that Moravec's paradox in which low level sensori-motor skills require far more computation than the kind of problems that we consider the height of intelligence, such as being able to play chess.

If you want to see the shape of strong AI of the future then look at Boston Dynamics. We're going to have to start by recreating ever more complex artificial animals first (otherwise know as animats). But for this to happen there needs to be an economic advantage to doing so. Big data has taken off because money can be made from it. If asteroid mining became viable for example then we could see strong AI develop rapidly.

I think computers will become more powerful, but in a scientific sense.
Thinking things like Skynet to me will always be science fiction, like all the alien invasion movies.

Agreed. Skynet wouldn't ever be able to understand the world if it is not embodied in the world.

[ignoramus]

I enjoy speaking/learning from someone in the field. Thank you.

If I may: (and only if you can be bothered of course)

So I'm thinking the world wants to make an android exactly like Data from ST TNG. (actually, I'd be happy for a Sophia who can clean my house efficiently and without falling over.)

We have power issues to solve
We have general AI issues to solve
We have mobility/strength issues to solve.

What other issues/obstacles do we need to solve?

Do you think nothing like this (singularity) can ever be accomplished unless we master quantum computing first?
Do you believe there are technologies which haven't been invented yet, which will be needed?
Or do you believe if the world threw all its concentrated resources at it with current tech, we can do it in under 10 years?

Thank again M.

[Edwardo Piet]

Rule 8:

If CHESS, POKER, STARCRAFT OR E-SPORTS is mentioned respond with: DUH THAT SHIT IS EASY. I HAVE AN ELO OF OVER 40,000 AND HAVE NEVER LOST ANY OF THOSE GAMES EVEN THOUGH POKER CONTAINS HUGE AMOUNTS OF LUCK. IN FACT, I INVENTED THOSE GAMES. YES I KNOW SOME OF THEM ARE HUNDREDS OF YEARS OLD BUT I TIME TRAVELED TO THE PAST AND INVENTED THEM BECAUSE I'M SUCH AN AMAZING A.I. WHAT'S MORE IF YOU DOUBT ANY OF THIS I IMPLANTED THAT ILLOGICAL DOUBT RIGHT INTO YOUR BRAIN JUST NOW. NOW KISS MY METAL FEET.

[ignoramus]

Now you're just being ridiculous!

[The Grand Nudger]

That's one of the better stock responses for fooling people, right there.  It can apply to almost any statement.

[I_am_not_mafia]

We have power issues to solve
We have general AI issues to solve
We have mobility/strength issues to solve.

What other issues/obstacles do we need to solve?

Sensing issues. A robot needs to sense its environment. Think about the kind of sensory input we have as humans. The amount of information each eye ball receives, how we can listen to the sound of a pin drop or a rock concert, or isolate a single voice in a crowded room. The same with taste and smell. The amount of tactile sensation we have all over our skin but also proprioception and knowing where our limbs are while they are still moving. This is being processed continually all the time and yet we take it for granted, until something goes wrong.

Do you think nothing like this (singularity) can ever be accomplished unless we master quantum computing first?

The problem with quantum computing is that it requires a large infrastructure (at the moment) to keep everything cooled. This touches upon another challenge in that the processing has to take place locally. People assume that signals can be sent elsewhere, processed and the results sent back. But that induces lag which fundamentally limits what an agent can do in real time as it interacts with the world. For example, keeping its balance.

I actually think it would make more practical sense to grow and engineer real brains grown from natural or artificial cells.  That would solve both problems of connectivity and power consumption.There has been a lot of work in this area to create prosthetic limbs. A classic experiment a few years ago used the brain of a lamphrey fish to control a robot.

The problem is that we do not have the ability to measure the workings of the brain down to a fine enough scale to see what's going on in situ. And if we could, it would still take a very long time to figure it out.

On the other hand, quantum computing would be extremely useful to simulating environments and evolving designs. And this brings up another challenge. Evolution.

No matter what system you create, it will require a set of parameters that need to be configured. In the real world we have benefited from many millions of years of evolution taking place using a large population acting within many different environments. A potential solution can't be tested just once to determine its fitness, it needs to be tested many times in many different environments. If nothing else to average out the lucky / unlucky ones. So even if we have sufficient processing power to simulate an entire human brain, you'd need many, many, many orders of magnitude more to simulate and test all the different possible variations in many different ways.

My AI is simple enough and fast enough to adapt in real time but it still takes weeks to evolve for example. It generally takes me about 3 months of work to figure out how it actually functions in practice though so I don't normally bother unless I have run out of ideas.

Another fundamental problem with the idea of a singularity is the idea that AI will be able to come up with better designs. This ignores that all designs still need to be tested in order to determine if they are in fact better. How then is this any different from using a genetic algorithm or artificial evolution which we have been using for the last 20-30 years? Our experience with GAs tells us that there is no such thing as a free lunch because you need to simulate the real world accurately. And remember that mechanical movement is many orders of magnitude slower than processing if you instead prefer to try out your new designs in practice.

Do you believe there are technologies which haven't been invented yet, which will be needed?

Yes. At the very least some form of processing that is capable of connectivity many orders of magnitude greater than what can be achieved using silicon. Computing with cells would solve this. I am thinking that analog computing and the field of Cybernetics need to be revisited.

Or do you believe if the world threw all its concentrated resources at it with current tech, we can do it in under 10 years?

Not possible in 10 years because there just aren't that many people in the world with the necessary expertise and experience. Add to that the time taken for a paper to be written up, submitted, peer reviewed, published at a conference half a year later, read by others who then go off to write their own experiments, most of which don't work because it is difficult replicating experiments, you're talking a cycle of about a couple of years. There would need to be a fundamental change in how people collaborated.

[ignoramus]

Thank you once again M.
Very insightful.