Before we plunge into the fray with gusto and abandon (and aplomb, of course), I’d just like to say that my wife (Gina the Gorgeous) has long been asking me when we can get a humanoid robot to help her perform tasks around the house. What we are talking about here are tasks like loading and unloading the dishwasher, washing and ironing the clothes, dusting and vacuuming, feeding the cats and changing their litter, and perhaps a little light yard work to round out the day.

Gina is thinking along the lines of an EVE-type robot from Halodi Robotics (see Halodi’s Humanoid Robots are Already Amongst Us!). To be honest, I really don’t understand why we need one of these little beauties because my beloved has spent the past 20 years training me to do all these tasks.

As fate would have it, we went to see the American science fiction horror film M3GAN (pronounced “Megan”) a couple of days ago, and now Gina is no longer quite so keen on having a household robot, so that worked out alright LOL.

As I’m sure you are aware, artificial intelligence (AI) is popping up all over the place these days. Not-so-long ago, the apps-du-jour were AI art generators like NightCafe, DALL-E 2, and Deep Dream Generator. On the one hand I’m tremendously impressed with this technology. On the other hand, I have to ask, “Is this really art?”

Maybe I’m biased, but I have a visceral feeling that something generated by the human hand or conceived by the human brain has intrinsically more depth and worth than something generated by an AI. A few years ago, for example, I spent six months of weekends building a Welsh Dresser as a Christmas present for Gina. I’m convinced that my family and friends hold this in much higher esteem than if I’d simply purchased a machine-produced equivalent at a furniture store.

One of the latest “AI kids on the block” to impinge in the public consciousness is ChatGPT (Generative Pre-trained Transformer). This chatbot, which was launched by OpenAI in November 2022, quickly garnered attention for its detailed responses and articulate answers across many domains of knowledge. It also prompted loads of comments by teachers who fear that ChatGPT will make cheating easier than ever. Personally, I’m not convinced that this will be a problem because—having seen the quality of some young people’s writing attempts recently—it will be obvious that the output generated by ChatGPT is of higher quality than any handcrafted offerings from the students.

But we digress… Earlier today, I received an email from my friend Joe Farr in the UK. In this message, Joe spake as follows:

Hi Max, I’ve been playing with this new ChatGPT AI thing…

I tried the normal stuff to upset it… “What’s the meaning of life? “Are you religious?” Blah blah blah, and that was fun but a bit unproductive. We then had a long and fascinating conversation about the merits of having, or not as the case might be, an Earth connection in a mains electrical system. Was it actually safer to not have an earth? I couldn’t convince it, though it did acknowledge that in some instances it would be safer to not have an earth.

Then I moved onto programming.

“Can you write me a bubble sort in Visual Basic 6?” I asked.

Yes it can.. and it did… and it also provided a usage examples of how to call the routine.

Ha… “Can you write me a bubble sort in Z80 assembler?” I asked.

And, once again, yes it can… and it again gave me a complete source code listing and told me how to call and use the routine.

Then I asked, “Is it possible to write a program in C# that acts as a BASIC interpreter?” I knew the answer to this of course… and it concurred and gave me a reasonably detailed explanation on how you could do it.

So, then I asked, “Could YOU write a program in C# that acts as a BASIC interpreter?” And the darn thing responded:

Sure! Here is a simple example of a C# program that acts as a BASIC interpreter. This program can parse and execute a simple set of BASIC commands, including PRINT and INPUT.

And it proceeded to give me a complete C# program listing. Not a fragment or snippet, but a complete bloody program listing along with an example of it working. You can see the darn thing typing out the code in front of your eyes… in real-time. It’s really frightening to watch.

This was when I thought, “Time to give it a mental melt down,” so I asked, “Could you write a program in Visual Basic 6.0 that acts as a BASIC interpreter?” It had to think about this one for 10 seconds or so but, once again, off it went and did the task.

This is the point where I went for the jugular, asking, “Could you write a program in 6502 assembler that acts as a BASIC interpreter?”

Five seconds later it starts typing… gets around a screen full of assembler, and just stops… “I’ve killed it,” I thought. There is a button you can click that’s marked “Regenerate Response”, so I clicked it. This time it had a long think, and then spat out the following:

Writing a BASIC interpreter in 6502 assembler would be a complex and challenging task, as the 6502 instruction set is not well-suited to the tasks involved in interpreting and executing high-level language code.

In general, BASIC interpreters use a combination of string manipulation, parsing, and logic to analyze and execute BASIC code at runtime. These tasks can be difficult to implement in low-level languages like 6502 assembler, which do not have built-in support for higher-level programming concepts such as variables, loops, and conditional statements.

However, it is possible to write a BASIC interpreter in 6502 assembler with sufficient effort and expertise. To do so, you would need to design and implement custom algorithms and data structures to handle the tasks of parsing, analyzing, and executing the BASIC code. This would likely require a thorough understanding of the 6502 instruction set and a strong background in low-level programming.

I hope this helps! Let me know if you have any other questions.

It’s obviously learnt that it’s hard work and doesn’t want to do it again LOL. It’s also bang on that the 6502 is a terrible choice for this task. Indeed, many people have commented that the 6502 is useless for tasks like this.

Funnily enough, it’s now sulking and doesn’t want to complete any more of my programming challenges. I’ve either hurt its feelings or I’ve used my allotted CPU cycles.

For me, the good news is that the Visual Basic 6.0 code it generated for a BASIC interpreter doesn’t work. It used the keywords PRINT and INPUT as variables inside the code and these are of course reserved words, but this wouldn’t be difficult to fix.

Good Golly Miss Molly. I have to say that I hadn’t even thought about using ChatGPT in this way. I’ve been very impressed with GitHub’s Copilot, which looks at the comments and code you are writing and suggests code snippets and entire functions in real-time. What would happen if ChatGPT and Copilot got together and conceived an AI offspring?

I just got off the phone with my friend Ufuk Giriş in Turkey. I told Ufuk about Joe’s email, and he responded that many AI systems are set up to constantly learn and evolve, but that you can’t be sure as to the quality of the information they are receiving, and that one of the things that worried him was nefarious actors feeding an AI with bad information, which could result in a M3GAN-esque scenario.

Hmmm, on that note, I’d like to leave you with a couple of thoughts. The first is the world’s first AI Psychopath called Norman (by which I mean the world’s first AI Psychopath that happens to be called Norman, as opposed to the first to be called Norman out of all the world’s AI Psychopaths). This AI entity, which sees death and destruction in everything, was created by your* friends at the Massachusetts Institute of Technology (MIT) (*they aren’t friends of mine).

The second thought is as follows: do you think that ChatGPT or one of its cousins (like Salesforce’s Einstein GPT, for example) will ever evolve to the stage that it is capable of writing something like The Chronicles of Narnia by C.S. Lewis? And, if this should come to pass, would the fact that this was created by an AI make it intrinsically less worthy than if it had been created by a human intelligence? What say you?