Singularity (The A.I. Thread)

[MR2 Fan]

While we're on that topic...one of the best episodes of ST:TNG

[Crazed_Insanity]

Hard to know if intelligence and emotions are closely tied.
We don’t have a lot of examples to compare. Yes even if you include other mammals, they have brains similar to ours, shared evolution. There are traits that we see in these animals that we might describe as being emotions. Happy, sad, jealousy, fear - these do seem to show up. Obviously our smarter brains, more complex relationships give us a bigger range of emotions (and our ability to discuss, compare emotions)
Maybe the best place to look is in some other smart animals, birds and octopus. Do they have emotions at all similar to ours?

Are these emotions the result of surviving on our environment? AI will have such a different environment, it doesn’t have the same basic requirements - food, shelter, company, reproduction.
To reach the singularity it requires evolution, but who decides the environment and the goals, risks and rewards.
That could have a massive impact on what possible emergent side effects come out - maybe emotions are just useful side effects of intelligence as we know it.

Yeah, to me, that's the fascinating part about AI research... to be able to figure out whether emotions are just merely side-effects... or perhaps actually a critical ingredient that you can't do without. Currently, this is something we just can't know for sure.

My guess is that emotions are the key ingredients... because at least nothing in observable nature suggests that such emotionless creatures exist... in the not very observable realm, even God has emotions. It'd be truly amazing if we humans can create a conscious self aware being without emotions such as commander Data.

If emotion is really key, that'd be truly dangerous... for we all know how emotional beings can turn out. If Commander Data is truly possible, then we have less to worry about with AI.

[MR2 Fan]

I was searching, I KNEW we had an A.I. thread somewhere.

Since the last post here I feel like there's been some major shit happening.

I've just been learning about the OpenAI developers and "Dall-E 2" and also their coding and writing algorithm stuff with GPT-3

Both are basically ways of interpreting our verbal suggestions into, in this case, pictures, with insane fidelity. (Video below). This to me feels like having something like Google in super-overdrive where it scours millions (?) of possible reference materials and not only does a kind of smash together of existing images, but completely interprets something brand new.

The most basic issue with this is that it could put a lot of freelance artists out of work, or severely limiting the work they can do. This is still a program only limited to a small number of people, but the information is out there...it won't be long until this becomes ubiquitous. Also, the coding stuff, scientists are predicting that by 2040 (I predict even sooner) all code for programs will be written by A.I.. That doesn't surprise me at all, the fundamentals are already there.

It's difficult to wrap my head around these developments, honestly.

[G'day Mate]

Wow, that's impressive

[Tom Servo]

Alternately, what it came up with when prompted with the Cats movie that came out a couple of years ago.

[Crazed_Insanity]

Wow! It’s like Star Trek replicator of images…

However, I’m a bit skeptical of it being real. Pixar can take a long time to create digital images, but this AI software is actually able to create multiple versions based on your verbal instructions? The amazingly fast multiple creations is making me think that it’s probably just cut and paste jobs based on what the AI could google online.

If it’s truly original creations, Why are they still beta testing it? Video examples shown are already super impressive! It’s not like autonomous cars having safety issues to worry about. If Siri and Alexa are ready for public, why not this?

[MR2 Fan]

Wow! It’s like Star Trek replicator of images…

However, I’m a bit skeptical of it being real. Pixar can take a long time to create digital images, but this AI software is actually able to create multiple versions based on your verbal instructions? The amazingly fast multiple creations is making me think that it’s probably just cut and paste jobs based on what the AI could google online.

If it’s truly original creations, Why are they still beta testing it? Video examples shown are already super impressive! It’s not like autonomous cars having safety issues to worry about. If Siri and Alexa are ready for public, why not this?

There's safety concerns with what it could create....illegal or disturbing or misleading images, like Queen Elizabeth stabbing Big Bird to death or something (to note something silly)

[Crazed_Insanity]

I thought they have coded such safety restrictions in place? Such as no famous people doing anything that’s not rated G?

Anyway, I know we’ve made lots of advancements in field of AI, but besides not seeing driverless Tesla Ubers, I really don’t think Siri and Alexa have gotten smarter after how many years now?

Until they can really fool me with realistic conversation, I kinda doubt AI will be able to take over creative efforts just yet… my feeling is that AI today are just very good at mimicking, but not really ‘conscious’ like us yet.

Creation of these images, if genuinely original, could be getting closer to ‘consciousness’ I guess. The only way to test it is to cut itself off from Google and see if it can really continue to create original images…

[G'day Mate]

They are original (as opposed to cut/paste) - didn't you see how it started with the dots?

[Crazed_Insanity]

I know that’s what they claimed, but I’m skeptical about that… considering how this AI could create so many different variants of ‘original’ works so quickly… as if it’s a google search.

Pixar probably couldn’t render a frame of their movie so quickly even if we ignore all the human involvements.

After companies like Theranos, claiming that They’d be able to do blood tests with a single drop of blood… until they official release and sell their products for actual use, it’s hard to tell whether if they’re for real or not. Sometimes too good to be true is probably too good to be true. But of course other times a truly disruptive product really revolutionize everything… you just can’t really tell before it’s released.