Jump to content

Welcome to FutureTimeline.forum
Register now to gain access to all of our features. Once registered and logged in, you will be able to create topics, post replies to existing threads, give reputation to your fellow members, get your own private messenger, post status updates, manage your profile and so much more. If you already have an account, login here - otherwise create an account for free today!
Photo

What was your "Holy Shit! We're living in the future" moment?

future shock we live in the future The Future™

  • Please log in to reply
112 replies to this topic

#101
Yuli Ban

Yuli Ban

    Born Again Singularitarian

  • Moderators
  • PipPipPipPipPipPipPipPipPipPipPip
  • 22,091 posts
  • LocationNew Orleans, LA

If you are an old'un, Caltrek, then I have no doubt that you've experienced futureshock before and can attest to how different the times are today from where they used to be.


And remember my friend, future events such as these will affect you in the future.


#102
starspawn0

starspawn0

    Member

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,960 posts

I think seeing what GPT-3 can do is a "we're living in the future moment".  It's better than just about all the previous Deep Learning progress I've seen over the past several years.  Perhaps the reason is because of the jump from GPT-2.  I was also impressed with GPT-2, but nowhere near to the same degree as GPT-3.  GPT-3 can literally pass a Turing Test.  Maybe not a very involved one, and not one where someone probes too deeply; but it can answer commonsense-type problems most of the time, and keep track of context.  It blows all other chatbots out of the water (when used as a chatbot -- a language model isn't a chatbot); better than the recent best by Google and Facebook, combined -- it's not even close.  And it can do tons of other things, like write code.

 

I think I will probably be less impressed with the jump from GPT-3 to whatever comes next, say GPT-4.  The next thing to impress me is probably a major advance in robotics or in applying BCIs to improve AI -- or even whatever Elon is going to announce about Neuralink in another month.  

 

This decade is going to be a wild ride!  We have some scary-impressive AI already, 2020 isn't even over yet!  Just think about what 10 more years of advancement will mean!  



#103
caltrek

caltrek

    Member

  • Members
  • PipPipPipPipPipPipPipPipPipPipPip
  • 11,365 posts

If you are an old'un, Caltrek, then I have no doubt that you've experienced futureshock before and can attest to how different the times are today from where they used to be.

 

 

Well, I am an old'un. Still what strikes me is not so much the future shock moments, as I have discussed in this thread, but the overall continuity of things.  Sure, if I compare computers of the early 1980's to something like GPT-3, it makes for a dramatic contrast.  Still, I experienced that as an evolutionary process stretched out over forty years, so it does not seem all that Future Shockish to me.

 

I also had the benefit fairly early on of reading a book entitled All That Is Solid Melts Into Air.  A theme of that book, as I remember it, is that the process of development itself is always developing.  The way things developed in the past will soon become obsolete as a method of development.  I am here discussing community development, as in construction of infrastructure.

 

Reading that book very much helped me prepare as a professional for a long career in community infrastructure development. I realized very quickly how fast my skills could grow obsolete. That helped a lot.   

 

From the dystopian angle, Covid-19 is very much a new experience for me.  Reading history, you realize such pandemics were not that unusual.  With the partial exception of the AIDs epidemic, I never experience in my life something like the current health crisis.  

 

The AIDs epidemic was different in that it was primarily a sexually transmitted disease (except for drug users utilizing dirty needles). So it was easy to avoid the virus by being very careful about selecting sexual partners and staying away from injecting illicit drugs.  

 

So Covid-19 just does not jive with anything in my living memory.


The principles of justice define an appropriate path between dogmatism and intolerance on the one side, and a reductionism which regards religion and morality as mere preferences on the other.   - John Rawls


#104
starspawn0

starspawn0

    Member

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,960 posts

Well, I'm also an old'un.  Not as old as you, but still old enough to be the father of most of the people on this forum.  I am older than Will by a several years (less than 10, more than 5).

 

... And I have to say that GPT-3 is a quantum leap beyond what came before about 2014, say.  There is no comparison.  I remember how shitty AI was in the 1980s.  And I learned about neural nets in the early 90s.  Back then, they couldn't do very much.  That remained true through the early 2000s, up to 2011.  After that, progress was  very rapid; but the jump in NLP AI over just the past 3 years has been greater than in the 20 years before then (before 2017), in terms of generality and ability to crack untouchably hard problems.  Remember IBM's Watson?  Its NLP capability was nowhere near as powerful as GPT-3 -- it's so weak in comparison, it doesn't even register.



#105
Yuli Ban

Yuli Ban

    Born Again Singularitarian

  • Moderators
  • PipPipPipPipPipPipPipPipPipPipPip
  • 22,091 posts
  • LocationNew Orleans, LA

Being a young'un, I feel a strange emotion when looking back to my "futurist break" back in the spring of 2014 (the first time I believe you joined this forum as star0)— indeed, one big reason why has to be because of your enthusiasm for deep learning which, at the time, was apparently "about to eat the world." I was also desperate for anything new to come in the world of AI, and Watson being shrunk to the size of a pizza box was the most exciting news of the decade, it seemed.

Yet an interesting phenomenon kicked in around 2016 or so, where I started seeing various AI and robotics developments from before 2015 as being unusually poor. I don't know why it's 2015 in particular; I felt this way even before I learned of GANs and their rapid growth from 2014 onwards.  Even now, "2014 generative AI" sounds much more like it's going to be an algorithmic magic trick than even "2016 generative AI". Of course, at the same time, I also feel that anything older than a year is outdated. GPT-3 has already vastly surpassed GPT-2 so completely that the thrill I used to feel over seeing GPT-2 generated texts and MuseNet has been dampened. 

 

I'm fairly sure that wasn't something you were going to feel in the 80s— something from 1981 probably wasn't outdated circa 1989 just because the ungodly weak computers of the era meant everything was an algorithmic magic trick.

 

Of course, by "algorithmic magic trick," I mean that it was one third clever programming to get around limitations, one third cherry picking the absolute best examples, one third hyping people to expect the best results consistently, and one third actual success.

 

It's only in the past two years that I've felt that AI began evolving beyond being just magic tricks, or at least starting seeing such successes without needing them to be cherry picked all the time. Neural networks started actually working as promised. And to that end, it was discovering GPT-2 and ThisPersonDoesNotExist, both on Valentine's Day 2019, that shook me completely. Generative AI actually feels like there's something tangible happening. And GPT-3 is a world apart even there.


And remember my friend, future events such as these will affect you in the future.


#106
starspawn0

starspawn0

    Member

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,960 posts

In the 80s, it was hard to really know what was real, regarding AI, as there was so much scifi entertainment around the concept -- and it was always supposed to be far away, in a computer in a government, military, or university lab somewhere; so there weren't really expectations of it being a consumer product.  Was KITT from Knight Rider a real possibility?  What about WHOPR from War Games?  Or the robots in Asimov's novels?  Or Westword?  Or HAL from 2001?   We just didn't know; and it was easy to believe in the fantasy, because few people understood the complexity of the problem.  But you could definitely see that everything claimed to be AI you could actually interact with, like an Eliza program, was just a mere trick; and a very bad one at that.

 

So what to believe?  That real AI actually exists in a secret facility somewhere?  That's probably what people thought.  And so, when real, useful AI started to actually creep into existence in the early-to-mid-2000s, they were underwhelmed.  "But I thought this existed 30 years ago.  Why don't you just do what existed back in 80s?"  But it didn't exist.  It was a fantasy.  A mirage.  

 

And now it's getting really useful, and widespread.  It's still not up to the level of the fantasy stories; but it's getting there at a rapid clip.  

 

Anyways, all those stories from the 80s leave one with the impression that we've not really advanced that much, all things considered; or that it happened at a slow, evolutionary pace.  The reality is, though, that progress has been like a hockey-stick graph -- flat for a long time, with some gradual upward turn, and then a sudden spike.



#107
Yuli Ban

Yuli Ban

    Born Again Singularitarian

  • Moderators
  • PipPipPipPipPipPipPipPipPipPipPip
  • 22,091 posts
  • LocationNew Orleans, LA

I wish I had more technical knowledge and STEM capabilities to be so totally amazed by GPT-3; I'm still learning of its abilities as you tell us of them. 

If anything, it was Kernel's BCI breakthrough that really gave me future shock, but that was more just the thought of what it could potentially do. 

 

Edit: I just realized that both Kernel's announcement and GPT-3 happened in a single month. May 2020 is what May 2014 was trying to be! (If anyone remembers that period of my life)


And remember my friend, future events such as these will affect you in the future.


#108
Metalane

Metalane

    Member

  • Members
  • PipPip
  • 34 posts

I think seeing what GPT-3 can do is a "we're living in the future moment".  It's better than just about all the previous Deep Learning progress I've seen over the past several years.  Perhaps the reason is because of the jump from GPT-2.  I was also impressed with GPT-2, but nowhere near to the same degree as GPT-3.  GPT-3 can literally pass a Turing Test.  Maybe not a very involved one, and not one where someone probes too deeply; but it can answer commonsense-type problems most of the time, and keep track of context.  It blows all other chatbots out of the water (when used as a chatbot -- a language model isn't a chatbot); better than the recent best by Google and Facebook, combined -- it's not even close.  And it can do tons of other things, like write code.

 

I think I will probably be less impressed with the jump from GPT-3 to whatever comes next, say GPT-4.  The next thing to impress me is probably a major advance in robotics or in applying BCIs to improve AI -- or even whatever Elon is going to announce about Neuralink in another month.  

 

This decade is going to be a wild ride!  We have some scary-impressive AI already, 2020 isn't even over yet!  Just think about what 10 more years of advancement will mean!  

It's likely that we'll experience about a 100 years of progress in this decade (from a 2000's standpoint I assume). 



#109
caltrek

caltrek

    Member

  • Members
  • PipPipPipPipPipPipPipPipPipPipPip
  • 11,365 posts

 

So what to believe?  That real AI actually exists in a secret facility somewhere?  That's probably what people thought.  And so, when real, useful AI started to actually creep into existence in the early-to-mid-2000s, they were underwhelmed.  "But I thought this existed 30 years ago.  Why don't you just do what existed back in 80s?"  But it didn't exist.  It was a fantasy.  A mirage.  

 

And now it's getting really useful, and widespread.  It's still not up to the level of the fantasy stories; but it's getting there at a rapid clip.  

 

I tend to agree with you except to offer one relatively minor tweak.  It was not so much a suspicion of it existing in a secret facility somewhere. It was more like a a "well what took you so long?"  kind of response.  

 

Television, and even movies, have really fostered a false sense of how fast things develop.  Overwhelming problems are solved in the space of an hour. Sure, the viewer may be given some cues that more time has elapsed than the length of the TV show, and that it should be taken as a compression of events.  But even this doesn't really overcome that "everything must be resolved immediately" kind of orientation. Collectively, we end up suffering from a very short attention span. Anything that spills beyond that attention span seems like a very long time.  So you end up with something that is quite the opposite of future shock.  Of course, if events do things like overtake your ability to stay competitive in the work force, then we are back to a discussion of future shock.

 

Science fiction shows that offer us a view into future events thus change our orientation from being surprised by new technologies to being disappointed that they took so darn long to develop. Not always, but enough of the time.


The principles of justice define an appropriate path between dogmatism and intolerance on the one side, and a reductionism which regards religion and morality as mere preferences on the other.   - John Rawls


#110
wjfox

wjfox

    Administrator

  • Administrators
  • PipPipPipPipPipPipPipPipPipPipPip
  • 12,396 posts
  • LocationLondon

EyOB8pI.png



#111
wjfox

wjfox

    Administrator

  • Administrators
  • PipPipPipPipPipPipPipPipPipPipPip
  • 12,396 posts
  • LocationLondon

^ 2001 vs. 2020.



#112
wjfox

wjfox

    Administrator

  • Administrators
  • PipPipPipPipPipPipPipPipPipPipPip
  • 12,396 posts
  • LocationLondon

2LdEjFV.jpg



#113
wjfox

wjfox

    Administrator

  • Administrators
  • PipPipPipPipPipPipPipPipPipPipPip
  • 12,396 posts
  • LocationLondon

lyLzOsV.jpeg







Also tagged with one or more of these keywords: future shock, we live in the future, The Future™

0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users