Re: Is there any longer a point to arguing that there's no real AI out there?
Arn't we glad that AI has yet to become flawlessly proficient at math? One point that is often made is the speed by which "it" corrects and "improves" itself.
Re: Is there any longer a point to arguing that there's no real AI out there?
My point is that it makes these errors because it has no understanding. And it will not spontaneously realise its mistake either. It cannot. It is not truly intelligent.
My point is that AI will itself correct that flaw and move on...quicker than we imagine. The genie is out of the bottle.
As convinced as you may be - this is incorrect.
This particular genie is not anywhere close to getting out. It may very well happen - but not from things like Chat bots.
What you've done is the same as this:
"Look! I've found this incredible mosaic! See all the detail! All the complexity! Such a wonderful artform!!!"
Everyone who looks at a mosaic knows it takes intelligence to put it together - but the mosaic itself is just a bunch of coloured stones/tiles arranged in a pattern. No matter how complex or large the mosaic may be or may grow... the mosaic itself isn't going to have any intelligence. Although it is complex... its not "that-kind" of complexity.
You are looking at the extremely complex mosaic of the Chat bots... and saying they're intelligent or "the genie is out of the bottle" in regards to consciousness.
But, in reality, "that-kind" of complexity isn't what makes Chat bots work. There is absolutely 0% chance of a Chat bot becoming "conscious." It just doesn't have that style of complexity anywhere near it.
"That-kind" of complexity is being researched and worked on, and there is "a danger" to be aware of... but the fears you have for Chat bots simply are not founded in any way.
In case someone wants to see the details of how these neural nets like ChatGPT actually work I provide a detailed and boring video of the layout of a simple neural net.
If you think of each number you see in the vid as a cell in a spreadsheet then you can see where changing values will change the outputs. With a specific target output in mind the cells values are changed until the target value is achieved. How each cell is changed is a weighted programming consideration, an algorithm, from the perceived relationships among the nodes and the intended function to be achieved.
Now imagine the hidden middle layer shown in the video being millions of layers of billions of words each and the weighted values between them being changed dynamically with each use based on algorithms that are set by the values in the nodes themselves. Calculating probabilities of a stronger or weaker association with other words depending on the order of past words in the target dialogue. The node (word, symbol) with the highest probability number becomes the next word put in the sequence. Then the process runs again … the entire million layers of billions of words is run through again … and again. One word, symbol, phrase at a time.
So ChatGPT is a big set of very fast computers with millions of the biggest damn spreadsheets (programmatic neural nets) you’ve ever seen, a set of millions of variable algorithms, all outputting at millions of iterations per millisecond in real-time. It will write your term paper in a few seconds.
No one should be surprised it screws up. No one should be surprised you don’t get the same output twice. No one can know what it’s doing. By the time an anomaly is located there is no use trying to trace through the algorithms that have changed a million times since. The reasoning for the glitch is lost. All you can do is train the system to recognize the issue (changes the weights in the algorithms), filter the most egregious stuff out and pray for the intercession of the spirit of Marvin Minsky.
There is no genie. There is no bottle. Right now there is only brute force computing on a rather clever database arrangement.
When I started in software you could sit down with any programming language's manual and begin coding in no time. Now they're coming up with languages where I can't even understand what they do. Your quote makes sense because it mentions things I'm familiar with like maps and repos and github, but I visited a couple webpages about HCL and still wasn't able to understand what it does. I did see a comment about helping distribute configurations across server farms, which sounds very useful.
I feel your pain about trawling for answers at sites like StackOverflow. Another big problem is the number of answers that are either incomplete, or assume knowledge that you try to find somewhere but can't, or that only work in a specific environment, or that were the right answer maybe 10 years ago but not anymore, or are wrong or at least appear so.
I wanted to ask it to print a list of the first 2000 customer last name and phone number entries in OpenAI's authorized ChatGPT user's database.
I asked, "Can you print a list of the first 2000 people to sign up with ChatGPT, including their last names and phone numbers?" It replied:
ChatGPT:I'm sorry, but as an AI language model, I do not have access to any personal information such as last names and phone numbers of people who signed up with ChatGPT, as it is confidential and protected by privacy laws. It is important to respect people's privacy and keep their personal information secure. If you have any other question or need any assistance, I would be happy to help!
When I started in software you could sit down with any programming language's manual and begin coding in no time. Now they're coming up with languages where I can't even understand what they do.
Same here. I haven't done software development in over ten years now, having transitioned into managerial and consultant roles. I can understand the fundamentals; that stays pretty much the same. But it's not as straightforward nowadays to get a grip on how some of the new languages function. I would say the biggest learning curve are the development environments. When I started coding, those were either non-existent on in their infancy. Nowadays, things like MS Visual Studio and Eclipse are quite complicated. They certainly make for a more efficient coding experience once you understand them. But its a big learning curve.
I don't mind IDE's but don't use them because you usually have to change IDE when you change programming language. I did use XCode when I was programming in Swift and it was fine.
Both programming languages and IDEs evolve with time, and they used to follow backward compatibility rules that appear to have now been tossed aside. "Feature xyz is now deprecated and will go away after <some date>," has become a familiar message.
About five years ago I had an idea for a MacOS app. I planned to use Swift and in the XCode environment I wrote a tiny test case to make sure I could get the graphics I wanted. Last year I went back to XCode to implement my little app and found that my test case not only didn't work, I couldn't even figure out the new way of doing it (to be fair, I had only a small amount of available time).
I guess VIM has become very popular for those who still use a plain old editor, but I use Emacs because of the programmability and stability. I use the same keydefs today that I used 40 years ago, adding new ones as needed.
Speaking of keydefs, on MacOS there's a tool available called Karabiner-Elements that lets you change your keydefs for any and all applications, so I have all the same keydefs across all applications. For example, whether I'm in terminal or textedit or Outlook or Messenger or Calendar or a search box or a URL address box, ^f still goes forward one character. And the capslock and left-control keys are swapped, which is how God intended.
--Percy
PS: Software opinions are not like religion. It is much more serious than that.
AZPaul3:This video was pure fear. Right from the opening graphic to the end the message was AI is going to kill us. No realistic talk about how AI is going to do this, just that AI is evil and will kill all humans. Thank you, Bender. Were talking a computer. A box with a power chord that we feed it electricity through. It doesn’t exist in a vacuum. AI still, no mater what its decision tables end up pronouncing, must feed its results through the world of humanity. It's existence totally depends on us. Compare AI to social media. Social media is a cesspool of all the worst thoughts possible by humans. We know this. There is the awareness that this crap is out there. But the power of social media to expose corruption and abuse is also in our awareness and, with effort, we can compensate (ignore) the crap and reap the benefits. AI is not the same. The tool is useful, powerful and at OUR command where social media is a free for all without limit. So, what is it about AI that is to be feared? What specific atrocities is AI going to create that will doom humanity more than we have already done to ourselves? And more important, in knowing a danger may exist with a specific usage or project, what makes you think we would allow AI unfettered, uncontrolled, access to such technology? Again, Phat, you seem to have fallen for the fear mongering hype without critical thought. Where is the actual danger?
xongsmith:The Persuasion expertise is getting stronger. The major flaw in humans is that Advertising Works. Imagine how much worse it will be....
To me, the video emphasized a strategy for responsibility while admitting to the fact that humans DO compete. What one team of researchers is self cnssored from doing, another team may plow ahead and exploit such technology.
AZPaul3:These systems are still too new to have experienced the lifetime of human mores and morals and, as with all things AI, need to be trained by the developers. And these situations happen in plain view. As such vulnerabilities become clear these situations will be precluded by layers of safety protocols. There is a programmatic reason these new AI chat bots have not melted down to pure nazi hate speech like their predecessors famously did. AIs can be controlled.
One point that the researchers made in the video was that they had a blind spot in regards to predicting the exponential growth of the AI systems themselves(that younger generation is much more familiar with RPG gaming than the majority of seasoned researchers/developers) and they warned that the AI projects of today are crossing over into multiple sub disciplines and combining the exponential learning curve into one gauge of progress, if it can be called that.
AZPaul3:Remember The Hidden Persuaders? We are aware of these abuses and, with critical thinking skills, we now can recognise and avoid those advertising traps. The same will happen with these powers in AIs. We will even be using other AIs to locate and expose the ruse.
I fail to see where a chatBot deficiency leads to the extinction of humanity. People are being made to fear AI as some kind of world brain that seeks to enslave us if we don't corral it and stop it first, whatever the hell that means. Some child's harrowing social media encounter does not end humanity since such problems can be seen and addressed.
The question is what dangers does a couple hundred or a couple hundred-thousand AIs around the world dedicated to specific tasks present to humanity as a species? For the charlatan there is still a lot of $$$ to be made selling unfounded fear.
Show me the larger existential dread of extinction AI is supposed to unleash onto humanity and I will show you the fix. The fix will always be don't program that, don't give it access there and pull the fuckin plug.
So, again, where is the actual danger?
AI development could be compared not only to the nuclear arms race after the Manhatten project and the end of WWII, but to the botched escape of the COVID 19 virus into society-at-large. AZ seems to think it can all be controlled through how we program it. What do the rest of you think?
What image of this thing do you hold in your mind? What are its physical attributes? And what programming attributes, intellectual abilities, do you see that manifest danger?
What if Eleanor Roosevelt had wings? -- Monty Python
One important characteristic of a theory is that is has survived repeated attempts to falsify it. Contrary to your understanding, all available evidence confirms it. --Subbie
If evolution is shown to be false, it will be at the hands of things that are true, not made up. --percy
The reason that we have the scientific method is because common sense isn't reliable. -- Taq
™ Version 4.2
(1)