SingularityNet (AGI)

And in the interim, that will created a marketplace for AI services, but in the longterm, the real goal is to create a conscious mind.

>the real goal is to create a conscious mind.
And before you say, "hey buddy, that's crazy, that's just nonsense", you're brain is just a collection of highly specialized subsystems that communicate with one another. One exception to that is memory, which has never been able to be pinned down to one specific location, but each area in your brain has a specialization and they all work together to create you. So SingularityNet is a pretty grand project to create something similar to a brain.

fuck off with your faggot nigger memes.

"Sophia the robot is dumb, but also not very different from past headlines about chatbots "passing" the Turing test. Theory of mind is fundamental to how we think, so it's very easy to fool us into lending human emotions & understanding to things that superficially mimic those."

- Francois Chollet (creator of Keras, the premier deep learning library)

AGI is a pipedream in 2018
Deep learning in general is deeply flawed and needs serious revisions before we can even roadmap a successful AGI attempt.

t. deep learning practitioner

>fuck off with your faggot nigger memes.
Well you have a good point there. This is all theoretical, basically it's just investing in research. So if you believe in that research going somewhere, invest in it, if not don't?

With all due respect to Dr. Goertzel and team, this coin is promising far too much far too early. Hoping for big things in 2020 and beyond.

Well, Mr deep learning researcher. I don't know if the human brain necessarily performs a gradient decent, but then again a sparrow's wings don't necessarily use the bernouli effect.

OP is retarded, but still, it's a cool project if neural nets are allowed to send money to each other for data.

imagine you train a neural net to retreive data on the network for learning a task mandated through natural language.

"hey sophia, make me a sandwich"
>sophia purchases sandwich-making algorithm
"hey sophia, perform sexual act 'x'"
>sophia doesn't know sexual act 'x', there is no neural net for sexual act 'x'
>sophia using NLP to determine what sexual act 'x' is relative to other concepts
>sophia purchases datasets that may be helpful in learning sexual act 'x'
>some of the datasets are useful enough for sophia to learn a passable form of sexual act 'x'

the future is bright.

Sophia is not dumb, she's a scam. Not AI at all. AGI is bound to be as much of a scam as that shit. I'm not touching this with a ten foot pole.

lol stay poor faggot

so buy now and hold longterm or just don't even bother until then? (yea I'm really new so what?)

Flush yourself down the toilet.

This is why I don't like the team whoring around Sophia. Oh well. I'm investing hoping they can use that money to make a better Sophia.

If we're using the metaphor of flight then what we have today is a glider made paper mache... it can perform limited tasks on a standard stream of input. But what we need is an airplane capable of dynamic updates to it's environment. And eventually AGI would be a factory which designs even better aircraft AND seacraft AND spacecraft. Does that make sense?

not even looking to buy, just interested in this coin because a friend who works in AI told me about this but whatever

Flush yourself down the toilet.

In a way. Better comparison would be desktop computers in the 90s to today's internet. So same difference but with machine learning algorithms.

it is interesting yes. thats how theyll get people to invest.
But the fact is that these shmucks wont invent AI, but maybe in the next 20 years someone like Elon Musk will make it.

There are startups selling pretrained neural nets, these are small enough to send over 2G networks and can do nearly everything a custom built net can do. There's also privacy laws that make distributed data pools very bad decisions legally. If you encrypt them, then you lose vital parameters for training your net. It's a lose-lose decision in the end...

But remember that 2018 neural nets will not look at all like 2020's neural nets (hopefully).

someone shilled this a week ago wish i bought then, the train is now gone.

Which is shittier this or dbc?

where would i invest in such a thing?

Deep learning is a lot like rugby. The ball moves up the field sideways, with a lot of little improvements here and there. What we need is a quarterback to throw a hail mary into the Next Big Thing. So until then we just wait ... (and push geniuses into machine learning classes)

>There are startups selling pretrained neural nets
Mr deep learning practitioner, you should know that 90% of the work is in getting the data formatted correctly. Shouldn't these start ups be selling data cleaning services not trained neural nets?

>(and push geniuses into machine learning classes)
lel I suppose that would be you?

DBC because my gainz lol

But these are just way too early. The promises are empty (even if they complete their roadmap). It's never a good idea to share training data for a neural net. That is the ONLY thing separating the winners and losers in AI. Google doesn't have better engineers, or better algorithms, they only have SO MUCH MORE data.

But meh ... hoping they prove me wrong.

Tidex or etherdelta if you can stomach it.

I'm just a cog in the machine

Fuck Silicon Valley

I made some gainz off DBC too. So where are you studying, Mr bigshot? Don't worry my school is shit so nothing to worry about there.

Bayesian stats, just google it.

It explains why things happen. And what will happen next

BTO is the chadder ai project of 2018. DYOR

Alright I guess it would be polite if I go first, I'm working on a masters in comp sci at Rutgers.

Ohhhhhhhh. Alright. I'll do that.

I actually already knew what that was, big shot. But thanks for the suggestion.

Well maybe these guys are sheisters. I'd like to think they're not. But if I can make AI research go a little faster. That's why I'm in this shit. for $50k. Maybe I'm a fool. But they'll probably make progress faster than I could. so. whatever.

It's all money I made trading crypto anyway.

Am I getting scammed out of my money? Maybe. I just have faith that a team of PhDs can do a better job than I can.

I mean look at this face. How could you not trust this face?

BAYESIAN STATS

SO YOU MUST BE ABLE TO TELL ME WHAT WILL HAPPEN NEXT

BAYESIAN FUCKING STATS

CAN ANY FIRST YEAR STATS STUDENT CHIME IN HERE

FUCK