New 8008 Program

[Slob]

There aren’t all that many 8008 programs, so here is a new one. You put in a date >= 01-JAN-1800 <= 31-DEC-9999 and it tells you the day of the week for that date. Not all that interesting, but I was fascinated by the poem “Monday’s Child” when I was a kid. Naturally, I was born on a Saturday… doomed at birth to be dull but hardworking. The program fits in 2K and has bitbanged serial I/O at 2400 baud at the nominal 8008 clock speed, 500 kHz. The bitbanged parts are from Jim Loos’ code, but this can easily be modified for parallel or UART type I/O. Be careful not to use -any other registers- than are shown in those bitbanged procedures.

No effort whatsoever was made in optimization or cleanup. The math parts (16 bit unsigned addition, 16 bit unsigned subtraction, 16 bit unsigned division) were done with the help of the free version of Chat GPT a few months ago, which is or at least was not very good at 8008 coding. But I eventually beat it into submission and got it to work, possibly with as much effort that it would have taken to do it myself. The division routine certainly could have been looped to save a lot of code space, but it might be a good example of “brute force” coding time/space tradeoffs because looping would undoubtedly involve storing and loading more temporary data and division would be less RAM/ROM, but slower. How much? IDK.

I am planning to modify this program into a universal calendar type program, where it will print out the calendar for the years 1800-9999 on demand.

I realize that my use of ChatGPT for assistance may be controversial. I should have done those by myself, but I wanted to try out ChatGPT. Anyway, it does work, and if you want to improve it, please do!

I don’t think ChatGPT (free version) of several months ago is even close to something like Claude Code of today. Where the version that I used fell down was in:

1. Old vs. new mnemonics, and on top of that it would occasionally use 8080 features
2. Inability to handle running out of registers
3. Didn’t seem to learn from mistakes very well (we would go three steps forward, then one back).

 

[rico]

Nice @Slob - I'll give this a go on my FPGA implementation! I found myself using the old mnemonics on the first go around, but settled on the cpu 8008new for the AS Macro Assembler. Dig around here if you wanna see any test/bring up programs https://github.com/robertrico/intel-8008-vhdl/tree/main/test_programs

Let me know if you build anything else! I got ways to run it and play with it. How are you running these programs? I even have the softcore running on Apple IIe so I can use real vintage tech to write and assemble the programs.

 

[johnx993]

Well if using AI to write code is a learning experience, it should be good.
But in the overall scope - this is a hobby and we should enjoy doing it.
I think using AI all the time for this is akin to a sports car hobbyist who buys a Ferrari, then hires a chauffeur for it.
But if it's used for learning, it's more like letting a professional race driver use the car to give you a few pointers!
-J

 

[rico]

Well if using AI to write code is a learning experience, it should be good.
But in the overall scope - this is a hobby and we should enjoy doing it.
I think using AI all the time for this is akin to a sports car hobbyist who buys a Ferrari, then hires a chauffeur for it.
But if it's used for learning, it's more like letting a professional race driver use the car to give you a few pointers!
-J

I think some of us are in the hobby strictly and purely for learning. I am too young to have nostalgia and am not a collector hoarder. The bits, bytes and 8 bit architecture are legacy tooling that is hard to find useful learnings for. AI provides an incredible tool for compiling resources, aid in debugging and most importantly offering wrong suggestions that you, the human, must reason about and fix.

Using AI to generate something for the end product is different than someone using AI to figure out that the 8008 has weird address based stack, its nuances around limited registers, address registers and clobbering registers.

But if people find joy in having a robot build a piece of software that they don’t understand, who am I to judge that? I don’t find sports enjoyable but I don’t find myself telling others that its not something they should be doing because I don’t enjoy it.

 

[Istarian]

@rico It's really not that hard to find good information if you know what you are looking for and a great deal of information is available simply by reading the relevant datasheets (where available).

And while finding a good resource to directly teach you how to program an 8008 may be quite difficult, using AI to try and find it may not be a good strategy.

Expecting the AI to do your reasoning for you will ultimately makes you extremely dependent on the AI, which is a very bad long-term plan.

That is especially true when you consider that the objective of the folks developing it is, as usual, to part you from as much of your money, attention, time, etc as they can.

P.S.

A big part of "X is hard to find pn the internet" has to do with the reality that it has become more and more cluttered over time with low-quality information and useless nonsense like advertising. This is almost entirely due to corporations getting involved in using it as an e-commerce platform and product/information distribution system.

 

[rico]

@rico It's really not that hard to find good information if you know what you are looking for and a great deal of information is available simply by reading the relevant datasheets (where available).

And while finding a good resource to directly teach you how to program an 8008 may be quite difficult, using AI to try and find it may not be a good strategy.

Expecting the AI to do your reasoning for you will ultimately makes you extremely dependent on the AI, which is a very bad long-term plan.

That is especially true when you consider that the objective of the folks developing it is, as usual, to part you from as much of your money, attention, time, etc as they can.

P.S.

A big part of "X is hard to find pn the internet" has to do with the reality that it has become more and more cluttered over time with low-quality information and useless nonsense like advertising. This is almost entirely due to corporations getting involved in using it as an e-commerce platform and product/information distribution system.

I hear your argument and I understand that if you were to use AI, you’d become dependent on it. Everyone learns and comprehends differently. Without AI I wouldn’t have been able to keep on time target with trying to reverse engineer the 8008 given all the resources out there on the net. It increased velocity several fold and I still had to go through three iterations. I’m sure there are others like me who can use the tool as a force multiplier without losing the edge and ability to hand code and understand the systems under design/test/build.

I would have to disagree with ease of being able to find good information. Using Google can only get you so far from trying to find specifics. I found myself trawling Google Groups, this forum, old sites, data sheets and all the usual go to to find the information I needed all while using an LLM.

I can also see from your arguments and statements that you have your stand and those who can use AI efficiently have theirs.

I’m not trying to defend anything, other than the fact people are different, can work differently, enjoy different things and ultimately take their own journey on their own hobby, with their own time and their own money. I just get caught off guard at the amount of gatekeeping and misperception there is. I do get it though, there are those we are adverse to change and take their frustration out on others.

I liked @Slob’s 8008 code and he is justified in feeling a sense of accomplishment utilizing tools to create something “new” from something old.

 

[Istarian]

I hear your argument and I understand that if you were to use AI, you’d become dependent on it. Everyone learns and comprehends differently. Without AI I wouldn’t have been able to keep on time target with trying to reverse engineer the 8008 given all the resources out there on the net. It increased velocity several fold and I still had to go through three iterations. I’m sure there are others like me who can use the tool as a force multiplier without losing the edge and ability to hand code and understand the systems under design/test/build.

I would have to disagree with ease of being able to find good information. Using Google can only get you so far from trying to find specifics. I found myself trawling Google Groups, this forum, old sites, data sheets and all the usual go to to find the information I needed all while using an LLM.

I can also see from your arguments and statements that you have your stand and those who can use AI efficiently have theirs.

I’m not trying to defend anything, other than the fact people are different, can work differently, enjoy different things and ultimately take their own journey on their own hobby, with their own time and their own money. I just get caught off guard at the amount of gatekeeping and misperception there is. I do get it though, there are those we are adverse to change and take their frustration out on others.

I liked @Slob’s 8008 code and he is justified in feeling a sense of accomplishment utilizing tools to create something “new” from something old.

You are literally just playing a game of sitting on fence posts here.

If people want to use AI, I'm hardly going to waste my time trying to beat it out of them.

But I fundamentally object to the assertion that they are actually learning much at all in the process, let alone how to write a program on their own.

If that was the point of the exercise, they're going about the whole business the wrong way.

-----

When I say they are likely to become dependent on it, I am expressing a general
concern for the kind of hole my fellow human beings may be digging for themselves.

Kindly keep your projections to yourself.

Some of us actually enjoy the thinking, reasoning, digging for information, etc.

Having some of it handed to me might be nice in the moment when I also want the practical result, but it kinda defeats the point. Might as well just find a person that wants to do it and compensate them for doing so.

 

[rico]

You are literally just playing a game of sitting on fence posts here.

If people want to use AI, I'm hardly going to waste my time trying to beat it out of them.

But I fundamentally object to the assertion that they are actually learning much at all in the process, let alone how to write a program on their own.

If that was the point of the exercise, they're going about the whole business the wrong way.

-----

When I say they are likely to become dependent on it, I am expressing a general
concern for the kind of hole my fellow human beings may be digging for themselves.

Kindly keep your projections to yourself.

Some of us actually enjoy the thinking, reasoning, digging for information, etc.

Having some of it handed to me might be nice in the moment when I also want the practical result, but it kinda defeats the point. Might as well just find a person that wants to do it and compensate them for doing so.

Appreciate the response. Not trying to play the fence post game.

Just trying to allow someone who had the courage and want to share their experiences and outcomes to have a positive experience with their hobby. A lot of AI posts I’ve seen here get the same treatment, same exact arguments, same exact phraseology as you’ve expressed to someone using AI. I’m hoping that some openness and acceptance will allow others to feel more confident about sharing and to not let other people let them be shamed for using AI.

You can have your assertions and beliefs all you want, it’s just not factually true for everyone. I will always challenge those whose believe AI hands you things and you don’t have to actually think. It is a common misconception and I’m just trying to say out loud, that it’s not true.

That being said, @Slob keep up the good work.

 

[Istarian]

You can have your assertions and beliefs all you want, it’s just not factually true for everyone. I will always challenge those whose believe AI hands you things and you don’t have to actually think. It is a common misconception and I’m just trying to say out loud, that it’s not true

The point isn't that zero thinking is involved, but that it involves much less use of your brain than doing all the work yourself.

It would be just as true if you relied heavily on one or more human partners, but they wouldn't generally be able to hand you even a partially complete emulator without having done prior work.

When rank newbies claim to have produced something in mere days or weeks that would have taken an experienced programmer months or years their claims are immediately suspect.

The moment they mention that they used AI it becomes questionable what they actually did or learned about anything other than prompting the AI.

All I see here is your emotional reaction to what other people think or communicate. What proof do you have for your assertions? Because there is otherwise no reason for me to accept what you have to say over what I think is the case.

 

[Istarian]

lot of AI posts I’ve seen here get the same treatment, same exact arguments, same exact phraseology as you’ve expressed to someone using AI. I’m hoping that some openness and acceptance will allow others to feel more confident about sharing and to not let other people let them be shamed for using AI.

That's because a lot of people on here were born quite some years ago and have lived experience in a world without LLM or ML based AI.

We know, from personal experience, what it takes to learn things on our own with certain kinds of resources and acquire the skills to produce a particular kind of result.

As a result we tend to be skeptical about claims that there are shortcuts to getting to the same outcome.

There are probably even some people who understand a lot more about how this "AI" works than most people who are actually using it. They also know that it doesn't start from scratch and gets a ton of existing information and work produced by other people for training data.

 

[rico]

All I see here is your emotional reaction to what other people think or communicate. What proof do you have for your assertions? Because there is otherwise no reason for me to accept what you have to say over what I think is the case.

I don't mean to have a perceived emotional reaction to the use of, defending or saying that using AI offers the same amount of learning as doing it "by hand". This is akin to high level programming. I know people who can program a for loop in JS, but wouldn't know where to start or finish that same loop an an assembly language. If the emotions are leaking through the forum boards, it's more because it's a bummer finding a hobby with elitist gatekeepers who just have no positive attitude towards new people with new ideas. It's unfortunate, but I think I am just finding that you all exist, and it's just the ugly part of a public forum. I do find excellent people with open minds and excellent communication, which is what keeps me here. I am starting to regret even opening this door. It's akin to arguing politics. People will believe what they believe and want to believe. I am just trying to support the newcomers.

RE: Proof of my assertions.

We know, from personal experience, what it takes to learn things on our own with certain kinds of resources and acquire the skills to produce a particular kind of result.

There are probably even some people who understand a lot more about how this "AI" works than most people who are actually using it. They also know that it doesn't start from scratch and gets a ton of existing information and work produced by other people for training data.

I am well within the both of these communities, I just existed a bit later than the 60s and 70s when I'd have seen these things in my home. I guess I just don't understand the skepticism part and that's the fundamental misunderstanding. When I see the new posts, my gut reaction isn't "Oh, AI - then it's fake or not real / worthy / they didn't work on it right. WELL I DID it before AI, etc."

That's all. Everything you said is logical and I get it. I just don't see the need to make a post telling someone what they did is "less than". But then again... it is a public forum. So you are all well within your rights to spew what you will.