https://chat.openai.com/share/ea5fb0f2-eec5-4d7e-acd0-f2416776e19c

I used the TOT method to get chatGPT to look at forth and see how it could improve… there is plenty of nuggets in the chat.

Cheers!

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In article <a3368d75-8b6a-4e05-8b6f-83b4a3e5c53cn@googlegroups.com>, SpainHackForth <jemo07@gmail.com> wrote: >https://chat.openai.com/share/ea5fb0f2-eec5-4d7e-acd0-f2416776e19c

I used the TOT method to get chatGPT to look at forth and see how it
could improve… there is plenty of nuggets in the chat.

It is exactly what a lower to middle management in ICT would say.
Duely impressed, but what follow up action is inspired by this?
Not much, I'm afraid.
Compared to what we hear in the news, she has her facts right upon
which to be base conclusions on.

Cheers!

Groetjes Albert
--
Don't praise the day before the evening. One swallow doesn't make spring.
You must not say "hey" before you have crossed the bridge. Don't sell the
hide of the bear until you shot it. Better one bird in the hand than ten in
the air. First gain is a cat spinning. - the Wise from Antrim -

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Sunday, June 11, 2023 at 2:15:59 PM UTC+2, none albert wrote:

I used the TOT method to get chatGPT to look at forth and see how it
could improve… there is plenty of nuggets in the chat.

It is exactly what a lower to middle management in ICT would say.
Duely impressed, but what follow up action is inspired by this?
Not much, I'm afraid.

Very disappointing. It's like what you always hear when the "what to do to improve the user base" question is asked. I'd like to hear some technological ideas, some original ones. Completely lacking from this dialog.

People ask me "why don't you use ChatGTP for work". For this reason. I
find the texts it produces rather mediocre - let alone the code it produces. The ideas it comes up with are plain weird or unoriginal. It completely fakes facts if it sees fit.

I can use it to summerize stuff (like throwing in some definitions, now make
a consistent one of out them), some code analysis is amazing (I gotta say that) and its translations are quite adequate.

But it's not a Greek Oracle. Not by a long shot.

Hans Bezemer

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Sunday, June 11, 2023 at 4:35:15 PM UTC+2, Hans Bezemer wrote:
Ok, I'll bite - Chat GPT here:

Language Features and Syntax:

Language features and syntax encompass the capabilities and structure of the programming language. Forth's simplicity and extensibility have been its defining characteristics, but to keep up with modern programming practices, it may need to incorporate
additional features:

Object-Oriented Programming (OOP): Object-oriented programming is a widely adopted paradigm in modern software development. Forth could benefit from incorporating OOP concepts such as classes, inheritance, encapsulation, and polymorphism. These
features enable modular, reusable, and maintainable code, making Forth more appealing for large-scale projects.

Functional Programming (FP): Functional programming promotes immutability, higher-order functions, and declarative coding. By incorporating functional programming constructs like lambda functions, closures, and immutable data structures, Forth can
enable functional programming styles and benefit from the associated advantages, such as concise and expressive code.

Modern Data Structures: Forth traditionally relies heavily on stacks and arrays, which can limit its suitability for certain applications. By introducing modern data structures like lists, dictionaries, trees, and hash maps, Forth can offer more
flexible and efficient data manipulation options, improving its ability to handle complex data scenarios.

Concurrency and Parallelism: With the prevalence of multi-core processors and distributed computing, languages that provide built-in support for concurrency and parallelism have an advantage. Forth could incorporate mechanisms for concurrent
execution, such as lightweight threads, coroutines, or message-passing primitives, to facilitate efficient utilization of modern computing resources.

Standard Libraries and Frameworks:

Standard libraries and frameworks provide pre-built modules and components that simplify common programming tasks and enable developers to focus on higher-level application logic. Enhancing Forth's standard library and framework ecosystem can have
several benefits:

Expanded Functionality: A comprehensive standard library can provide ready-to-use modules for networking, file handling, date/time manipulation, regular expressions, and more. These libraries save developers time and effort by offering tested and
optimized solutions for common programming tasks, promoting code reuse and productivity.

Domain-Specific Libraries: Forth can benefit from the development of domain-specific libraries tailored to specific industries or application domains. For example, libraries for scientific computing, embedded systems, or game development can open up
new opportunities for Forth adoption in niche areas where its strengths align with the requirements of those domains.

Frameworks and Toolkits: Frameworks provide a structured foundation for building applications, handling common tasks like UI development, database integration, or web application management. By supporting or developing frameworks for Forth,
developers can benefit from established patterns, conventions, and tools that streamline application development and maintenance.

Integration with External Technologies: Forth's standard libraries can be extended to integrate with external technologies and APIs. This includes support for interfacing with databases, web services, hardware devices, or other software systems. Such
integrations enable Forth to be used in a broader range of applications and improve its compatibility with modern software ecosystems.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Sunday, June 11, 2023 at 5:42:36 AM UTC-5, SpainHackForth wrote:

https://chat.openai.com/share/ea5fb0f2-eec5-4d7e-acd0-f2416776e19c

I used the TOT method to get chatGPT to look at forth and see how it could improve… there is plenty of nuggets in the chat.

Cheers!

With Google search have to wade through much
fluff to find data instead of peddlers.
So a little while back I decided to ask chatGPT
for data.
Got good looking replies, just the data and not
a bunch of links to services and it was nice to
interact with it to hone in on particulars.
Later I took my
list of answers to check them out. When it
got down to actual facts, none of them held.
The answers all sounded good but all the facts
were flawed. Sorry it won't replace Google
search but it should be a good marketing tool.
--
me

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

S Jack schrieb am Sonntag, 11. Juni 2023 um 17:36:21 UTC+2:

On Sunday, June 11, 2023 at 5:42:36 AM UTC-5, SpainHackForth wrote:

https://chat.openai.com/share/ea5fb0f2-eec5-4d7e-acd0-f2416776e19c

I used the TOT method to get chatGPT to look at forth and see how it could improve… there is plenty of nuggets in the chat.

Cheers!

With Google search have to wade through much
fluff to find data instead of peddlers.
So a little while back I decided to ask chatGPT
for data.
Got good looking replies, just the data and not
a bunch of links to services and it was nice to
interact with it to hone in on particulars.
Later I took my
list of answers to check them out. When it
got down to actual facts, none of them held.
The answers all sounded good but all the facts
were flawed. Sorry it won't replace Google
search but it should be a good marketing tool.
--

https://www.latimes.com/world-nation/story/2023-06-09/chatgpt-lawyers-cite-bogus-case-law

ChatGPT fabricates things, it is a potential lies producing machine.
To be handled with care.

Many mediocre journalists will first love it, then it will kill their jobs...

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Sunday, June 11, 2023 at 7:33:32 PM UTC+2, minforth wrote:

S Jack schrieb am Sonntag, 11. Juni 2023 um 17:36:21 UTC+2:

On Sunday, June 11, 2023 at 5:42:36 AM UTC-5, SpainHackForth wrote:

https://chat.openai.com/share/ea5fb0f2-eec5-4d7e-acd0-f2416776e19c

I used the TOT method to get chatGPT to look at forth and see how it could improve… there is plenty of nuggets in the chat.

Cheers!

With Google search have to wade through much
fluff to find data instead of peddlers.
So a little while back I decided to ask chatGPT
for data.
Got good looking replies, just the data and not
a bunch of links to services and it was nice to
interact with it to hone in on particulars.
Later I took my
list of answers to check them out. When it
got down to actual facts, none of them held.
The answers all sounded good but all the facts
were flawed. Sorry it won't replace Google
search but it should be a good marketing tool.
--

https://www.latimes.com/world-nation/story/2023-06-09/chatgpt-lawyers-cite-bogus-case-law

ChatGPT fabricates things, it is a potential lies producing machine.
To be handled with care.

Many mediocre journalists will first love it, then it will kill their jobs...

I don’t disagree, it’s a language model, not a source of knowledge per se, it’s a great for commenting on text and extend a set of notes and ideas into a meaningful document, it does not create novel ideas, it’s a sophisticated parrot in some
sense. I’m my posting if you look at the interaction,, it creates most of the text on it’s own using the ToT methodology, so the inference if looking back and reassessment from know possibilities will improve the accuracy of the response, most of the
time I find that it’s not accurate at all. I have not been able to try the Bing part of if, but I have been testing Orca, which is a private model that I can extend that seem to be processing way to interact with you documents and code…. I personally
find it fascinating to prompt it and warm it up with info and then get some meaningful insights into your own data. I think the race is on now to create Domain Specific Models for each of the datasets (business) as such a broad model just is not enough
to have an inference meaningful enough to generate value. Te potential is endless and tokenizing data and then running it through a tensor and transformer is the new era in data treatment. While we will “at this point” not replace relational database,
I think data schema companies “ those that sell a bunch of premade such as and reports” might be loosing their weight in the next 5 years, maybe sooner.


The interesting thing is that some of the conclusions it came to are very align with my view of things that could make FORTH better, IMHO.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 12/06/2023 12:35 am, Hans Bezemer wrote:

...
But it's not a Greek Oracle. Not by a long shot.

Advice without threat of calamity for non-compliance lacks imperative.
Forth will die if it doesn't have a standard - is equally lacking in
substance. It gets attention by preying on fear. It's only a matter
of time before AI learns it :)

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

This is just a lot of regurgitated nonsense, which the Internet is full of, so, I'm not expecting anything to improve as long as non-peer reviewed Internet content is supplied as the database the ML (it's not really an AI... that's just marketing spin)
algorithm learns from. In fact, things are going to get much worse as ChatGPT et al will enable unoriginal thoughts to propagate exponentially onto the Internet. It's possible large language ML models will have the opposite effect to what is being
predicted. Instead of enhancing the Internet, they might make it so full of weeds, that people abandon it.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

In article <837870e8-286c-4b07-b3dd-312f1e9b8d2fn@googlegroups.com>,
Rob Kellock <creditscorenz@gmail.com> wrote:

This is just a lot of regurgitated nonsense, which the Internet is full
of, so, I'm not expecting anything to improve as long as non-peer
reviewed Internet content is supplied as the database the ML (it's not
really an AI... that's just marketing spin) algorithm learns from. In
fact, things are going to get much worse as ChatGPT et al will enable >unoriginal thoughts to propagate exponentially onto the Internet. It's >possible large language ML models will have the opposite effect to what
is being predicted. Instead of enhancing the Internet, they might make
it so full of weeds, that people abandon it.

That is just wishful thinking, that AI only repeats the common knowledge. Remember that e.g. wikipedia is a tertiary source. You can add only
knowledge that is present in scientific papers, once thay make it into textbooks, then you have to cite the text books. It is only months and
ChatGpt is on the level of wikipedia. And wikipedia is useful.
I tried comparing versions of Forth. The information I got was known
to me, but should eminently useful for a non-expert.

The next step is make that AI systems aware of cheating. I love to see
a probability attached to each news-item for it being generated by an
USA misinformation system. The ai systems deployed by the military
probably do just that, or they are working on it.

Groetjes Albert
--
Don't praise the day before the evening. One swallow doesn't make spring.
You must not say "hey" before you have crossed the bridge. Don't sell the
hide of the bear until you shot it. Better one bird in the hand than ten in
the air. First gain is a cat spinning. - the Wise from Antrim -

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

none albert schrieb am Montag, 12. Juni 2023 um 09:21:37 UTC+2:

In article <837870e8-286c-4b07...@googlegroups.com>,
Rob Kellock <credit...@gmail.com> wrote:

This is just a lot of regurgitated nonsense, which the Internet is full
of, so, I'm not expecting anything to improve as long as non-peer
reviewed Internet content is supplied as the database the ML (it's not >really an AI... that's just marketing spin) algorithm learns from. In
fact, things are going to get much worse as ChatGPT et al will enable >unoriginal thoughts to propagate exponentially onto the Internet. It's >possible large language ML models will have the opposite effect to what
is being predicted. Instead of enhancing the Internet, they might make
it so full of weeds, that people abandon it.

That is just wishful thinking, that AI only repeats the common knowledge. Remember that e.g. wikipedia is a tertiary source. You can add only
knowledge that is present in scientific papers, once thay make it into textbooks, then you have to cite the text books. It is only months and ChatGpt is on the level of wikipedia. And wikipedia is useful.
I tried comparing versions of Forth. The information I got was known
to me, but should eminently useful for a non-expert.

The next step is make that AI systems aware of cheating. I love to see
a probability attached to each news-item for it being generated by an
USA misinformation system. The ai systems deployed by the military
probably do just that, or they are working on it.
Groetjes Albert

Too late. AI is already heavily used for tax cheating.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 12/06/2023 6:04 pm, minforth wrote:

none albert schrieb am Montag, 12. Juni 2023 um 09:21:37 UTC+2:

In article <837870e8-286c-4b07...@googlegroups.com>,
Rob Kellock <credit...@gmail.com> wrote:

This is just a lot of regurgitated nonsense, which the Internet is full
of, so, I'm not expecting anything to improve as long as non-peer
reviewed Internet content is supplied as the database the ML (it's not
really an AI... that's just marketing spin) algorithm learns from. In
fact, things are going to get much worse as ChatGPT et al will enable
unoriginal thoughts to propagate exponentially onto the Internet. It's
possible large language ML models will have the opposite effect to what
is being predicted. Instead of enhancing the Internet, they might make
it so full of weeds, that people abandon it.

That is just wishful thinking, that AI only repeats the common knowledge.
Remember that e.g. wikipedia is a tertiary source. You can add only
knowledge that is present in scientific papers, once thay make it into
textbooks, then you have to cite the text books. It is only months and
ChatGpt is on the level of wikipedia. And wikipedia is useful.
I tried comparing versions of Forth. The information I got was known
to me, but should eminently useful for a non-expert.

The next step is make that AI systems aware of cheating. I love to see
a probability attached to each news-item for it being generated by an
USA misinformation system. The ai systems deployed by the military
probably do just that, or they are working on it.
Groetjes Albert

Too late. AI is already heavily used for tax cheating.

It won't be the AI that'll be doing jail time :)

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

It won't be the AI that'll be doing jail time :)

The discussion is centered around identifying areas of improvement for the FORTH programming language with the aim of making it more mainstream. The question is, how many people do you come across in a week who are familiar with FORTH? Or even with other
languages like C, Python, or Rust? The goal is to increase awareness and usage of FORTH, which many dedicated users believe is a superior choice, yet many people aren't even aware of its existence.

The proposal is for FORTH to be both a language and an embedded CPU Virtual Machine. There's a need for more efforts to make systems like moveFORTH ANSI compatible if we want to see significant progress. The issue of code portability is indeed a major
one. There's a need for consensus on how to make code cross-compatible. Without this, the incentive for newcomers to adopt FORTH remains low, and the language's adoption rate will continue to lag behind others.

The evidence is clear: if we continue on the current path, FORTH risks becoming obsolete. With the emergence of technologies like MOJO, which aims to make Python a systems language capable of competing with Rust and C in terms of performance, the odds
are increasingly stacked against FORTH. Despite its merits, the language is being adopted by only a small number of users for serious use. Something different needs to be done to change this trajectory

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Monday, June 12, 2023 at 8:57:42 PM UTC+2, SpainHackForth wrote:
[..]

The evidence is clear: if we continue on the current path, FORTH risks becoming obsolete.

I've heard that line for the past 40 years. Now, like then, Forth lets me control modern hardware exactly like I want (actually I have way more
control nowadays).

If there are *really* interesting programming experiments going on, I
can implement them in Forth and make them a first-class citizen.

Most well-educated people speak more than one language, and use them
when most appropriate (even greek has its use). The same ideas hold for programmers. There used to be documentation and financial restrictions,
but these hurdles are largely non-existent in today's programming world.

-marcel

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 13/06/2023 4:57 am, SpainHackForth wrote:

It won't be the AI that'll be doing jail time :)

The discussion is centered around identifying areas of improvement for the FORTH programming language with the aim of making it more mainstream. The question is, how many people do you come across in a week who are familiar with FORTH? Or even with

other languages like C, Python, or Rust? The goal is to increase awareness and usage of FORTH, which many dedicated users believe is a superior choice, yet many people aren't even aware of its existence.

How do people find anything? One has certain interests, explores them,
and finds what's on offer. Crowds delivered through evangelism inevitably dissipate. But why should one need them? If it's to confirm one's own
choices or to fulfil a lack, what does that say?

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)