ChatGPT accords

[RSG]

I have already read an article saying that AI already designed perfumes and one had already been sold with commercial success. And the perfumer involved said that it came up with a novel combination of materials he would not have thought of. But that otherwise what it was doing was the same as other AI which is remixing what was in its database. Even the combination of materials was novel to the perfumer but not novel to the entire database they fed the AI. All it can do is turn out remixes of commercially popular products.

If we're talking electronic nose. I've read they're doing work on that too for about 5 years now. It runs into the same problems mnitibach talks about and preexists with all AI. Even if it "smells" it still has to be told which input is "good". It then is still just remixing the data points. Humans can do this too it's just that AI is faster at it. Humans still have to evaluate the formula to see if it's fitting the brief.

Some professional perfumers I've read about say they trial and/or smell 100's-1000's and AI has only increased this. There is one perfumer who played with the perfume AI the early one embedded in a table? And said they'd never have another thing to do with it.

I read another article recently talking about using AI to design celebrity perfumes and the perfumer team still consisted of 5 people!

What I find interesting is how much care and effort is taken to program the ai with excellent formulas and indeed they have committees doing the choosing, then they program in IFRA limits, etc. They also choose equivalencies because of course a commercially successful perfume these days will not have an ambergris tincture but they tell the AI to choose .3 ppt ambroxan instead.

If we would all become as educated and fill out our notes as well. We would do as well or better than AI.

Of course, we seem to need inspiration which is where the speed ai can dump things can inspire. But honestly, it just makes my eyes glaze over and it's just too much, too divorced of context to feel inspiring. I still rely on dreams and my spreadsheet. lol

 

[the_good_life]

Why does this fill me with great pessimism for the future? Not to mention a slight concern at, either your students ability to write good essays, or your ability to teach them. Who defines "mediocre" or "bad student"? How do you teach a computer to write badly, or in a way that you would mark poorly?

Well, there's a fair share of special needs children, but not only and the fact is, many children, especially boys, do not read books anymore and it shows. But that was already an issue when I taught at university.They can create amazing TikTok videos though 😏

 

[LaFou]

Why does this fill me with great pessimism for the future? Not to mention a slight concern at, either your students ability to write good essays, or your ability to teach them. Who defines "mediocre" or "bad student"? How do you teach a computer to write badly, or in a way that you would mark poorly?

Your pessimism is in the right place good Sir,
I once was writing of this sort of thing but was met with resistance and denial!
you know that the United Nations today called for an urgent action and laws to be formed in terms of protecting people and creativity of work from AI 🙃

I saw AI up close, and 99% of people still don’t comprehend how enormously advanced it is! This is only the beginning!
most of what is human creative is already being taken down now as we speak online with AI websites and engines!

All hail our new “……….” ChatGPT and other AI‘s 😅😄😄🤣

I hope the ones who keep on staying in denial understand now 😅🤣

 

[LaFou]

This is part of what has destroyed much of the usable internet in my opinion. I owned and ran listverse.com which is a top 10 lists site and I did most of the writing initially. It was reliant on ad revenue. Every time I met with my account reps at Google they would give me an ever diminishing list of "requirements" that were "not mandatory" but were I not to comply it would reduce considerably my ability to make money from my content. I told Google they were tactily censoring me and they immediately (and excessivly) went into defensive mode stating that they were absolutely not censoring content.

I had staff to pay so I had no choice. I allowed Google to begin dictating what content was to be allowed. Eventually I found relacement advertisers and was able to resume writing the content I wanted and loved. Google consequently began "shadow banning" the site. I sold the business.

Google is going down soon either way, don’t worry! The US gov are aiming to break it down into several companies, similar to what happened to ATNT back in the days 😅 karmaaaaaaaaa baby

 

[LaFou]

AI to perfumery among everything else, is what the industrial revolution to modern society is!
You can accept or stay in denial, the famers got replaced with machinery, the workers in factories, and soon enough, chemists and perfumers too! were if before you had 10 chemists per year and 10 perfumers per year as new, In few years, you will have 1 and 1 …

😁😁

 

[Eric Anthony]

AI to perfumery among everything else, is what the industrial revolution to modern society is!
You can accept or stay in denial, the famers got replaced with machinery, the workers in factories, and soon enough, chemists and perfumers too! were if before you had 10 chemists per year and 10 perfumers per year as new, In few years, you will have 1 and 1 …

😁😁

This made me chuckle, not only do we still have and NEED farmers, but economically we are struggling to keep them farming because the rest of the economic chain advanced but the chokepoint is, and will likely always be, the manual farmer (*in most industries).

According to history, when a new nation or civilization is born the first expanse is horizontal, gobbling up land, because agriculture is free money, from an economic standpoint. Only when available land runs out, did societies grow vertically, improving efficiencies.

In other words it is ALWAYS cheaper to farm more land than create a factory or efficiency. Farmers improved efficiency with machine but we are far from complete automation in agriculture.

 

[Solua Botanica]

While I do appreciate the benefits technology can bring into our lives, I don't think AI is going to replace everything.
A lot of people are seeing through the illusion. Chat GPT has already lost a lot of appeal to many people who have been playing with it...
It takes a lot of folks and machines to even keep this thing running correctly.

All hail our new “……….” ChatGPT and other AI‘s 😅😄😄🤣

I hope the ones who keep on staying in denial understand now 😅🤣

When I read such phrases, I can't help but laugh... For some, it sounds like an empire is trying to conquer the world.
It is only a man made AI that is trying to influence the minds. If we start to rely on AI for everything, we will get duped...
There is a long way to go before AI can replace real human creativity and feelings.
People will always love authenticity and we will end up creating a stronger counter culture to the artificial life they want for us.
There is so much real beauty and wisdom to discover and appreciate in this world.
No, it's not denial. It's looking at life through a different lens. A unique lens for each one of us...

 

[LaFou]

This made me chuckle, not only do we still have and NEED farmers, but economically we are struggling to keep them farming because the rest of the economic chain advanced but the chokepoint is, and will likely always be, the manual farmer (*in most industries).

According to history, when a new nation or civilization is born the first expanse is horizontal, gobbling up land, because agriculture is free money, from an economic standpoint. Only when available land runs out, did societies grow vertically, improving efficiencies.

In other words it is ALWAYS cheaper to farm more land than create a factory or efficiency. Farmers improved efficiency with machine but we are far from complete automation in agriculture.

You Can’t seriously be writing this! and made you chuckle Too!
Come on, why do we always have to explain the point in details of details of details of details so it’s not taken literally!
are you saying now with all seriousness that the Industrial Revolution didn’t kill millions of jobs including and by hiiiiiiiiiiiiiiiiigh Margin farming!
your kidding, no serious person would say that!
an industry that was completely relying on man and animal power that are replaced by machines and tractors!

again, please guys read the full paragraph analyze it before hand, I said if you had 10 chemists or 10 perfumers you’ll have 1 and 1 every new year

like now you have 1 farmer out of 1 million per year!
among aaaaaaaalllll the remaining industries!

and since the Industrial Revolution we Have only been working on improving machinery to replace more and more jobs!
elevators,
photography
transportation
shipping
etc.. etc..

now we reached to a place in time we’re AI, is able to think better than any human! So with that soon enough another millions of jobs will be lost

otherwise why did the United nation called for urgent global action against AI and protecting human works!

cheers, I hope this won’t make you chuckle though because it shouldn’t

 

[LaFou]

While I do appreciate the benefits technology can bring into our lives, I don't think AI is going to replace everything.
A lot of people are seeing through the illusion. Chat GPT has already lost a lot of appeal to many people who have been playing with it...
It takes a lot of folks and machines to even keep this thing running correctly.

When I read such phrases, I can't help but laugh... For some, it sounds like an empire is trying to conquer the world.
It is only a man made AI that is trying to influence the minds. If we start to rely on AI for everything, we will get duped...
There is a long way to go before AI can replace real human creativity and feelings.
People will always love authenticity and we will end up creating a stronger counter culture to the artificial life they want for us.
There is so much real beauty and wisdom to discover and appreciate in this world.
No, it's not denial. It's looking at life through a different lens. A unique lens for each one of us...

Chatgpt is an illusion of AI.. go see the real AI in action!

most of you comment from your own perception! You think you know because it sound logical! Well reality check! AI is better than you by light years! I know I know your ego will jump and say no we human created AI we control bla bla
still doesn’t change the fact it’s light years ahead of you or anyone else
it think better, faster, reliable, no clouded judgement, and its Already documented that it surpassed our way of thinking that we don’t actually know how it’s operating any more! We have an idea but we don’t know!

so if you don’t work in AI and never seen it in action, be modest enough to say I don’t know, but don’t build answers because you Think it’s logical from your point of view!

cheers

 

[mnitabach]

Chatgpt is an illusion of AI.. go see the real AI in action!

most of you comment from your own perception! You think you know because it sound logical! Well reality check! AI is better than you by light years! I know I know your ego will jump and say no we human created AI we control bla bla
still doesn’t change the fact it’s light years ahead of you or anyone else
it think better, faster, reliable, no clouded judgement, and its Already documented that it surpassed our way of thinking that we don’t actually know how it’s operating any more! We have an idea but we don’t know!

so if you don’t work in AI and never seen it in action, be modest enough to say I don’t know, but don’t build answers because you Think it’s logical from your point of view!

cheers

These assertions are completely untethered from the reality of current "AI" technology, which hasn't even remotely "surpassed our way of thinking that we don't actually know how it's operating any more". Where are you getting this nonsense from?

 

[Solua Botanica]

AI is better than you by light years! I know I know your ego will jump and say no we human created AI we control bla bla

How do you know it's better than me ? My ego is hurt ...LOL
No seriously man, you can preach all you want, but AI will never replace true human intelligence. We only use a small portion of our brain and with all the technology we use nowadays, that portion is getting smaller. Mental calculus, writting, reading are all great ways to expand the powers of the brain. Unfortunately, the younger generations are doing it less and less, and it shows in their intellect. There is a lot of evidence supporting that.

so if you don’t work in AI and never seen it in action, be modest enough to say I don’t know, but don’t build answers because you Think it’s logical from your point of view!

No I don't work in AI and I don't intend to. But I have been in contact with some people that were so excited and thrilled about Chat GPT.
That was a few weeks ago, most of them have now lost their initial thrill and are saying it's a waste of their time. This is actual evidence to me.
I will be modest enough to say that we all view and value life differently. You're allowed to praise AI.
I'm allowed to be weary of it, because I value real human intellect and the real creativity it can manifest.

 

[LaFou]

How do you know it's better than me ? My ego is hurt ...LOL
No seriously man, you can preach all you want, but AI will never replace true human intelligence. We only use a small portion of our brain and with all the technology we use nowadays, that portion is getting smaller. Mental calculus, writting, reading are all great ways to expand the powers of the brain. Unfortunately, the younger generations are doing it less and less, and it shows in their intellect. There is a lot of evidence supporting that.

No I don't work in AI and I don't intend to. But I have been in contact with some people that were so excited and thrilled about Chat GPT.
That was a few weeks ago, most of them have now lost their initial thrill and are saying it's a waste of their time. This is actual evidence to me.
I will be modest enough to say that we all view and value life differently. You're allowed to praise AI.
I'm allowed to be weary of it, because I value real human intellect and the real creativity it can manifest.

I’m not praising AI.. at all, again my tone was clear to be sarcastic of it I’m concerned extremely.. of it! But chatgpt is kids AI thats not the real existing AI!
check the AI being used at the medical field, at the astronomical field, or the AI being used in quantum physics (got familiar with that one), or military or intelligence section or advanced technologies even stem cell technology! Those are the real AI and now in perfumery And chemistry too! Traffic patterns now being organized by AI..

AI is not what we see in movies.. That’s fantasy, AI is a collection of intelligent parameters with one purpose only! Do the job at the best optimized manner! Even if it means breaking the wall over our heads, it’s parameters to be done with a matrix of billions of calculation per second so 😅 yeaaaa I value life.. welcome to the new world

and it’s not 1 AI, it multiple ones, each being developed by concerned field! it’s in plain example we are teaching machines how to think within certain set of parameters, and in those parameters they are light years ahead of us .

Intelligence is different than self awareness! Big difference when we say if it’s better or not, which it is in intelligence! Let’s hope it stays only that way!

cheers

 

[steventeddy]

But chatgpt is kids AI thats not the real existing AI!
check the AI being used at the medical field, at the astronomical field, or the AI being used in quantum physics (got familiar with that one), or military or intelligence section or advanced technologies even stem cell technology! Those are the real AI and now in perfumery And chemistry too! Traffic patterns now being organized by AI..

AI is not what we see in movies.. That’s fantasy, AI is a collection of intelligent parameters with one purpose only! Do the job at the best optimized manner! Even if it means breaking the wall over our heads, it’s parameters to be done with a matrix of billions of calculation per second so 😅 yeaaaa I value life.. welcome to the new world

and it’s not 1 AI, it multiple ones, each being developed by concerned field! it’s in plain example we are teaching machines how to think within certain set of parameters, and in those parameters they are light years ahead of us .

If you wish to defend AI/machine learning, I do not think offering things like this is conducive. Before you ask whether or not I have experience with AI/machine learning, I'll tell you that I'm a doctoral student studying computer science, with a focus on machine learning. Some of the things you have said are misleading at best. Let me explain.

The training and fine-tuning a transformer at the size of ChatGPT is not a trivial or "kid's" task. You can take a look at Appendix D of the paper describing GPT-3. Or take a look at third party analysis about the enormity of money and computation resources required to train GPT-3. They claim it'd at least 4.6 million dollars on the cheapest cloud instance. There is also no clear distinction on what you mean by "real AI"-- as to you, a model using an established neural network architecture (transformers) doesn't count. Regardless of what you might say about transformers, many in the machine learning community are still interested in it and take it seriously. You don't have to take my word for it--just look at the number of citations the original paper proposing transformers (Attention is All you need, Vaswani et. al., NeurIPS 2017) has gotten. I don't want to accuse you of using the "No True Scotsman" fallacy, i.e. excluding any and all unfavorable examples to secure/defend a generalization, to defend your view on AI/machine learning, but that's the vibe I get.

Calling parameters intelligent is anthropomorphizing literal numbers. A parameter is a part of the learned function like the normal vector in a linear classifier. The normal vector is just a series of numbers. The thing that could be construed as intelligent is the function learned by the model.

"Doing the job in the best optimized manner" doesn't get you very much if that statement is meant to defend your view. In short optimizing a model's loss to optimality doesn't get you much in terms of model performance in the real world if its modeling assumptions aren't met. Optimizing for the best model is not something that can be by a computer in general. This task very human centered as people need to come up with the network architecture/training methods. This is brushing over the problem of loss functions for most large neural networks being highly non-convex. E.g. as soon as we move from logistic regression to a multilayer neural network, cross entropy loss goes from convex to non-convex. As you might know, it is practically computationally infeasible to find the global minimum, and no one tries for that. They just try to find a good local minimum. If you're interested, the Dawid-Skene model is a simple probabilistic model for predicting the label of an object from an ensemble of experts whose modelling assumption is broken. When its generative assumption is not met, e.g. when there is a duplicate classifier in the ensemble, it breaks down and suffers a huge loss in its predictive performance. This is not to say it's useless unless the modeling assumption is met, but that it can at times be worse than simpler methods like majority vote.

The question is not whether machine learning models can outperform us humans on a 1 on 1 comparison with equal footing, it's whether it can be better than us in our full capacity given a finite amount of memory, among other constraints. A 1 on 1 comparison doesn't really make sense IMO because a lot of the neural networks used today are not biologically plausible.

 

[Casper_grassy]

No one can predict how far this will go, and maybe my sentiments require a bit too much faith in humanity as a whole, but I think the greatest endeavor is one pursuing what appears impossible in any realm.

If using AI generates success without any risk, this IMO would just create a very banal future, which many, even if lacking exceptionality would be averse to this.

 

[Big L]

If using AI generates success without any risk, this IMO would just create a very banal future, which many, even if lacking exceptionality would be averse to this.

Most of the past, present, and probably future, is banal.
Like with every technology, most people will use it superficially, and a select few who are geniuses will create masterpieces with it.

 

[David Ruskin]

Whilst this link has nothing to do with ChatGPT and so may be considered off topic I wanted to share it with you all as it shows what can go wrong with AI. I found it frightening and hilarious at the same time. It is quite long, and varied in quality, but, I think, well worth reading.

https://simonwillison.net/2023/Feb/15/bing/

 

[Alex F.]

Whilst this link has nothing to do with ChatGPT and so may be considered off topic I wanted to share it with you all as it shows what can go wrong with AI. I found it frightening and hilarious at the same time. It is quite long, and varied in quality, but, I think, well worth reading.

https://simonwillison.net/2023/Feb/15/bing/

Thanks! This is wonderful. And so eerily reminiscent of some of the conversations here.

 

[LaFou]

If you wish to defend AI/machine learning, I do not think offering things like this is conducive. Before you ask whether or not I have experience with AI/machine learning, I'll tell you that I'm a doctoral student studying computer science, with a focus on machine learning. Some of the things you have said are misleading at best. Let me explain.

I wanted to reply in details, but then said it’s better to take what you said to not take your word on it!
you are convinced in something and no matter how much I will write or show you won’t change your mind!

so why loose both of our times!
you keep what you believe in, and i will do the same!

last, there’s nothing misleading about what I said,
otherwise, universities wouldn’t be conducting constant checks
United Nation wouldn’t call for an urgent action and decreed laws to protect human work and creativity (2 days ago)
among many other disciplines!

just next time, when you answer to people (at your first reply) try to do with some good manners and not to throw big demeaning words to someone opinion, there’s always a way to reply with good form and critique at the same time! were additionally now I’m absolutely certain you know nothing about this subject, Like 0% .. from that silly reply with information that has nothing to do with what was simply mentioned!

oh yes, just an FYI everything In this universe is governed by parameters, you as a human don’t think you are so special, we are governed by parameters
while we can think and do, we can’t simply transform into a duck, or fly because our parameters doesn’t allow us! We are bound by parameters far more than you can imagine!

and another FYI AI 1 on 1 with human will exceed a billion to 1 and that’s only in 1 parameter!

people throw big words about ahh humans and humans, while they can’t outmatch a simple calculator for 1$

and they want to preach us what they know! Pffff

yea that’s it.. this is getting boring to continue

 

[David Ruskin]

Thanks! This is wonderful. And so eerily reminiscent of some of the conversations here.

I hope that others read and enjoy as well. It really is worth it.

 

[darkwitch]

"FYI everything In this universe is governed by parameters"
Parameters? I don't believe in parameters. A silly human construct that the universe doesn't need but works well for current simulations made by humans.

What do you believe is true? That is all it comes down to. It doesn't have to be beliefs that you are aware of having. It determines what you "know", and based on that, everyone is getting their own personal view of the same universe, including the so called fact that everything in this universe is governed by parameters.