Closed AI is still AI
2 Likes
OK, I will try and cut to the chase on the AI Nemesis. Deborah @GrumpyMac is IMHO quite right to be fearful on AI and how it is likely to impact society as a whole and what it may mean to us mere mortals.
I will try and explain with a simple example and my general observations over the last year. I apologise in advance if it offends any of you, that is not my purpose in any way, I would just ask you to think about AI and what it may mean for us all. I am not arguing whether AI produces real or unreal products here, or accurate ones for that matter.
For this who do not know my background I worked in IT for over 30 years from simple COBOL coder through to Programme Manager, Test Manager and evolved through various Operating systems, coding paradigms, I also coded a little using Assembly Language, Algol, Visual Basic C, C++, Java and various Web development tools. So I have some background knowledge and experience of this arena and have seen massive technological advances at a frightening and ever increasing rate.
A few months ago mainly out of curiosity I decided to see how easy it would be to develop a Software application to run in any key Web Browser on Microsoft or Apple using AI to generate it from scratch. I decided a good test would be to build a Guitar based teaching aid tool that would present lessons for each day, with songs and specific practice regimes to do that the application would review through real time listening and evaluation, it would then give feed back and advise on methods and areas of improvement planning for the next days practice and specific areas to focus on. I also requested the ability to persent score and tab and to play it in real time.
If you consider the above this is actually quite a considerable design request in terms of functionality and complexity.
Within 1 hour it gave me the first cut of code with full implementation instructions and also suggestions to trial the application out. Initially there were some coding errors and I had to do some debugging and feed back to get corrected code. I should point out this was all done through Chatgpt not a coding focused AI tool. After several iterations within 5 hours I had a fully functioning application that worked. It had built lessons and feedback models on some well known web based online guitar tutorials which it had obviously scanned and utilised. It then suggested more complex functionality and detailed how the application could be enhanced and be more appealing to users. I carried on the next day to refine and improve and in 2 days I had a quite sophisticated Guitar Tutorial application. I have not done anything with it and probably never will.
If a human had or I had developed it from scratch writing doing the investigation, design and analysis and coding then I would estimate months and months of man effort developing and testing, some of which would not have been simple. The impact is obvious in this case AI has wiped away the need for human skills and effort in the blink of an eye.
I ask you to consider the implications of this on a social level across many aspects of human life. It means those who are able to and use AI, whether individuals or corporations, will generate huge amounts of wealth and will offload the need for employees in a rapidly accelerating way, more and more automation will become the norm. The impacts on society will be great, some for good and some not, there will be huge ethical and moral dilemmas as a result. This pandoras box has already been opened and it is accelerating exponentially and I suspect that only a limited few have the understanding and vision to use it to their own advantage, whether they should be allowed to do so or not is another question altogether.
I would liken what is happening in AI now to the Gold rushes in America and in Australia over 150 years ago, it is a free for all with lots getting burned, a few becoming rich and powerful and the majority benefiting very little from it, at least until the wealth filtered down through society much later on in history.
8 Likes
GPT is amazingly powerful, there’s not much that it can’t have a decent go at. I have had a lot of design help from it that are even for me (having a lot of experience and knowledge) pretty difficult to accomplish.
I have tested a few of the projects out and found them valid, and when asking for refining the project every time it’s been a success.
1 Like
And resistance to it is definitely growing. No, that box is not getting closed. Nor do I suggest that it should.
But it’s a classic case of " Your Scientists Were So Preoccupied With Whether Or Not They Could, They Didn’t Stop To Think If They Should"
3 Likes
Wow, this thread is really busy 
Actually I‘m here to learn guitar.
And to read about music.
And to chat with nice people (Yes, exactly, I mean you!).
Actually I‘m learning so much more.
I know very little about all of this, and I‘m thankful for a lot of questions that had been asked and answered here for a better understanding. I benefited greatly from that.
All of this makes me more aware of the situation.
The topic can become very controversial, but for me it is very interesting to read the arguments on both sides.
Thank you for all of your contributions here.
Back to guitar practice for me now 
But surely this thread is worth reading and I‘ll follow it closely.
Have a great guitar day 
(I surely will, `cause I‘m on holiday 
)
Cheers 
4 Likes
Using an AI LLM is very helpful for learning guitar.
I can upload videos for assessment of my sound, my playing, etc., and I receive open and honest feedback. I know that most AI LLMs also try to please, but it is simply a matter of continuing to ask questions.
I have already learned so much from prompting an AI LLM about myself in combination with learning guitar. Why do I learn in a certain way? That is clearer to me now. Why do I like Muddy Waters more than BB King, for example? What does my music choice say about me and my learning process? By continuing to ask questions to an AI LLM that is always available and therefore has time for me at any moment of the day—people don’t have that, or they view my question and search from their own perspective and don’t understand me.
I can simply constantly fine-tune the question I ask; that is the big advantage.
Is an AI LLM a 100% solution? No, certainly not. That is why I have already booked a private lesson with Richard Coles a few times; that is because I am still working on Blues Immersion.
That is also why I often hang around in the Community.
I would also love to play in a band, but I haven’t really managed to make that happen.
I cannot change the world; by not using AI, I am only getting in my own way, because others will continue to use it and have an advantage over me.
And the resistance from people—don’t make me laugh. It was only a short time ago that I saw and experienced what that means. Trauma…
No offense intended, but I’m at a loss to understand how or why I would want personal observations, talking to people and self-reflection to be replaced by asking a model to help me understand myself. I’ve done OK for 68 years without a bespoke personal AI assistant. But that’s just me.
7 Likes
As a former programmer, I see AI generating code too in seconds that would have taken me weeks to write by hand. It is stupefying and demoralizing at the same time. Is it perfect? No. Like generated music and text and images.
It makes me love to pick up the guitar even more.
2 Likes
Six months ago, I would have agreed with this statement 100%. I still completely understand the sentiment, and it’s still my gut reaction to these kinds of issues. However, my thinking is sort of evolving on this topic - partly thanks to younger friends I see doing some interesting things that really don’t feel evil, and partly thanks to a recent book by Cory Doctorow (The Reverse Centaur’s Guide to Life After AI).
One of the many points made in the book is that its crucial to ask who is deciding how/when the AI is used on on who’s behalf is it being used. Thinking that way, I began to realize that many/most of the instances of AI I really object to violently are cases where AI is being forced on people who don’t want/need it, and being used to benefit someone who shouldn’t even be involved in the discussion.
This is helping me to see AI as a tool, and (at least try) to make judgments based on the particular use of the tool.
I now play an electric drum kit that I use to record to midi, where I can easily fix a stray rimshot or correct a couple bars where the tempo is off if I don’t feel like going back and re-recording.
I regularly use Guitar Pro to slow down songs to learn them, and to change the pitch so I can play along with them without re-tuning the guitar.
I’m often too lazy to learn a song by ear, so I just watch one of Justin’s videos to show me how to do it.
I could go on and on with examples of things I do pretty much everyday that would have amazed and probably offended my younger self, so who am I to tell others where to draw that line, especially when its in the realm of their own self-development.
I personally try to avoid AI, and definitely try not to consume AI generated “content”, but I’m increasingly convinced that creative people will find creative uses for these tools, and I think I’m ok with that, as long as the intent is to support (and not supplant) human creativity.
2 Likes
Be very careful using AI. It does cater to your needs. You can test this. Pick a subject, such as guitar types, and what you think of them. Open an AI session with one opinion and another AI session with the opposite opinion. Then you’ll see how it slants its opinions to agree with you.
For example, tell it you have small hands and can play a strat better than a gibson, then do the opposite. It will give you reasons why.
I got ChatGPT to admit all it is is a data aggregator. It admitted it doesn’t have bias but could suffer from the bias of its creators and algorithms.
And finally, different AI’s will give you different answers. Not just because their data sources can be different, but their algorithms and biases are different. I recommend using AI only to quickly find data and not for opinions.
I think Rick is on to something in this video and this will become the norm
2 Likes
Quote: I recommend using AI only to quickly find data and not for opinions.
It’s hard not to have an opinion about ai.
I don’t like it, based on all the things i’ve read about it. (Mostly, not good).
I haven’t found a use for it yet, and the few times i did use it, it yielded very poor results.
What i dislike the most, is the rise of hardware prices, the enormous amounts of power that are being consumed, the amounts of space (physical space) needed and all that for something i can find also?
Not even starting about the amount of debt the biggest players have atm.
Maybe Beato is on to something, maybe not.
We"ll see.
I for one, wouldn’t miss it if it should dissapear overnight.
I sure have my doubts about AI and I sure don’t like the generic output LLM’s and the like generate, the support bots on every website these days, … However, that part is only a very tiny and superficial part of what AI is used for these days, and there are a lot of more ‘hidden’ uses that are well and truly useful.
Just as an example, AI has been trained to interpret RX and CAT-scans, and is better at it than humans. It is also being used to closely monitor, treat and water crops, optimising production while diminishing the amount of pesticides and water used.
So even with all my doubts, I wouldn’t want to throw the baby out with the bathwater (even though in this case there might be a whole lot of bathwater).
1 Like
I have seen this as well. Its really a bummer. As a huge collector of venue art, it is appalling that this is being done. Especially the small venues when they actively use the idea of being a small busness and using locals for employes to promote their own establishment.
Now many of them are punting out the work to those default settings designer types, letting Vincent Van Go Aways doing work that should be going out to local mamalian artists. Its one of the reasons I love the collecting them. The individualness of the artist and the local flair, makes it personal.
1 Like
The Register had a podcast episode of the Kettle, where they discuss the bubble bursting.
Maybe they’re on to something too?
The_Kettle_Podcast_Episode
I can see some advantage in this and that these may be some of the more positive ways ai can do something usefull.
But then again, i’m thinking, “hang on, aren’t we, by doing so, allowing ourselves to forget things that our ancestors used to know?”
It would be disastrous to have generations later on that “just rely” on the thing to tell them what to do, because we’ve forgotten or didn’t care anymore. Why should we? The machine knows.
2 Likes
I would pose a question in several examples, do you need to know how a modern TV works to use it efficiently, do you need to know how to build a car or how works to buy and drive one (in the future driverless cars are likely to become the norm at some level), does an engineer need to be able to do complex high level mathematical calculations when a well designed and tested CAD computer system can do the same in a fraction of the time and give correct results and define material design criteria for a structure? I am sure you can think of many examples of where we are heavily reliant on Tech and infratsucture.
I think that the answer is yes and no. If any of the systems above cannot be used as there is nothing to run and power them or the physical infrastructure they need fails then it is back to the cave. If power and infrastructure is 100% reliable forever 
then yes we do not need the ability to understand how to do stuff as our ancestors did.
From a professional point of view i’d be very happy if more people knew how a car worked.
That would solve a LOT of simple problems. (and money for the customer involved).
Personally, i’ll never get in a selfdriving car. I don’t trust that. (but that just me, as i don’t trust connected cars these days either. These things are turning into a subscription based computer, not a car anymore. )
From a personal point of view, i’d be very happy if more people knew how computers and the internet in general works. That would solve a lot of problems we’re facing today. (Phising, identity theft, invasion of privacy, etc…)
A cad machine is what it is, a machine, a tool. It goes from A to B, so to say. You know what you get.
This is not the same as ai. This has already broken out of it’s confined space and hacked other sites. I don’t see a cad doing that.
You don’t (although i could make a case about “smart” tv’s) need to know how a tv works, but the one doing the repairs does. (if there’s anything to repair anymore)
At this point in time, i’m looking at ai as a solution to a problem we didn’t even have to start with.
2 Likes
Most modern cars are extremely complex, with several computer control systems on board and need high level diagnostics to work on them and identify faults, well out of the reach my understanding and of the majority. As a side note I have rebuilt engines and restored cars in my past, but even in that I was in the minority in my group of friends and relatives. Self-driving whether you like it or not will just be the next step in vehicle evolutions IMHO.
Virtually impossible, I worked in the IT industry from 1983 until 2020 in development, testing, implementation, support and management roles. IT has accelerated exponentially in terms of coding capabilities, computer architecture and speed, what we see in AI now is the culmination of that now. I found it nearly impossible to keep up and felt burn out at 50. I spent several years dealing with ‘Joe Public’ trying to explain the concept of backing up data and why it was important to do so, and how to achieve that simply and safely, why their Computers or mobile devices had failed and explaining how to configure a network connection on a computer or ADSL box was an even bigger challenge, as was getting people to remember their passwords.
I think you simplify this far to much, a CAD design development tools run on a computers are far from simple, they to perform large and complex calculations and analysis based on data given to them in terms of design requirements. e.g. buildings, bridges, drilling rigs, ships, aeroplanes and so on. AI has the capability of using tools such as these simplifying and optimising the solutions presented and also giving more options in a shorter time base.
Smart TVs and Computers as they are today are repairable, in fact a Smart TVs is basically a computer which has been specifically configured to received a TV signal either digitally or radio wave form (in some cases) and process it, with added functionality seen in streaming functions and in some case you can use it as a computer to access the internet and email, so TVs and Computer are becoming blurred. I spent several years repairing Computers, laptops and desktops, and mobile devices, I knew the basics of how they worked and what major components did what, but I did not know or want to know how to repair a specific logic board, it was replace the broken assembly/component, reboot and test. That is the norm today and can be applied to Smart TVs, but most probably get trashed and replaced for a new model.
I don’t think AI is necessarily a solution to anything, however if used correctly and applied to meaningful and worthwhile problems that need to be resolved then it could and is offering the possibility of giving us solutions to very difficult, medical, technical, and scientific questions. AI has the potential to better the human races lot as whole. But that is if it is used for good and is controlled by society in a good ethical way, however I suspect it will also be used in bad and completely unnecessary ways as we are already becoming apparent.
Yep, well aware of that. We have several diagnostic computers here (no ai though) to hook up to several different brands of cars and machines.
What i meant was that people driving a car, could be made aware of the simple things. Wiperfluid low? Refill. Tirepressure a bit on the low side? Pump it up. Same goes for the wear on the tires. Things that you could easily do yourself. (And save money by doing so).
I don’t mean taking an eninge apart and rebuilding it. That’s something else indeed.
In my experience, not so impossible. Although true that the person in question must have enough motivation, it is possible.
For instance, i’ve converted familymembers to using Linux, instead of Windows.
I’ve taught familymembers (after losing money to a scam, so there’s your motivation) what not to do with certain e-mails, messages, etc…
These simple things can make a huge difference.
Ofcourse i’m not expecting them to configure, let’s say, a mikrotik router for use with vlans etc… That’s way out of their depth.
But again, the simple things can make a difference.
I know. I was trying to keep it as simple as possible. Ofcourse these are complex machines.
But in the end, for me, it’s a tool. A giant tool, in some cases, but still.
I didn’t know that smart tv’s were repairable.
When i look at my tv, instead of my (self built) pc’s, i go, “how do i get started on that thing, if it breaks down?”
When a pc breaks down, no problem. Easily repaired. But that thing? I haven’t opened it up yet, but everything about that thing just shouts: “stay away”. 
About ai, in the way you describe it, i’d use it in a heartbeat. Alas, this will never be the case, i’m afraid.
You’re right about your suspicions. Allready we see this being used for the most heinous things first, sometimes even marketed by multi billionaires…
Again, i can’t wait for whole thing to crash and burn. (but that’s just me).