Announcing Xavvi, a Great Leap Forward for App Building

[richmond62]

As I read more and more about the, possible, ramifications of A.I., I thank the God-shaped hole for 2 things:

1. I am not dependent on computer programming for my income.

2. Because of #1 I can release the software that I do make as Open Source so whether the work is done by me,
my cat (whose work is also non-copyrightable), A.I., or just me hitting the keyboard serendipitously when I
drop something on the floor (the main source of my few clevernesses), I don't have to get all hot and sweaty
about these things: especially as my feeling about lawyers is that they are a bit like vampires waiting
for someone to expose their neck.

3. "Who concieved the invention": well, that is a right minefield: apparently my idea [my idea only insofar as I
implemented it] of running through a series of images as a backGroundPattern to a graphic object should in all
fairness be attributed to Walt Disney and his workers for the initial conception of the multiplane camera . . .

Oh, hang on a minute, what about the thaumatrope?

https://en.wikipedia.org/wiki/Thaumatrope

[Mariasole]

Some reflections...but I would not want my passionality to be mistaken for a form of resentment...I'm just Italian aren't I?

My criticism of AI in its marketing and politically oriented version (i.e. ChatGPT) does not make me a Luddite or a saboteur of Jacquard loom.
Whenever someone utters a veiled criticism of AI or whatever Science(TM) produces, he is dismissed as someone who is against progress and wants to return to the Middle Ages.

Let me remind that this is a forum for people who work in or are interested in the world of programming so no one here is "anti-science" par excellence. On the contrary, we are informed enough to confront science just as it, in the Galilean sense would like: criticizing it! Skepticism is the basis of science; fideism is the basis of religion. Clearly, when science becomes a religion, the two become confused. So...let AI be worshiped and glorified and sanctified.

Remembering that CAD has not only replaced squares and drafting tables but has flattened the world of architecture and burdened floors with iron when half as much iron was needed (talk to the carpenters), "because the computer says so," and that technology has brought us cheerfully within seconds of the atomic apocalypse, I think it is not well understood what we are up against.

AI as it is presented and as it is beginning to be marketed (in a totally irresponsible way since it suffers from hallucinations and its creators say it can be dangerous) is a pervasive technology that will squeeze in everywhere: more misinformation! That is the least of it!

But let's start with this problem: AI will directly or indirectly rewrite all sources of our knowledge. Directly perhaps by "optimizing" the already questionable Wikipedia (which has already been credited with deforesting and normalizing Web 1.0 content made by good people of good will), and then rewriting it. Indirectly because the already millions of scholastic copy and paste hallucinations of ChatGPT are already polluting every form of world knowledge. It is only a matter of time.

Change the sources to reshape critical spirit and common feeling.
So misinformation, manipulation, automatic rectification and flattening is inherent in "textual" generative AI.

The greatest danger is that everything will come under the scrutiny of these AIs, from traffic to the beat of your pacemaker, from the temperature of your refrigerator to the code you write, from what you say at any given moment (I don't know about you but here in Italy, in Trieste, there are MICROPHONES in the streets to record people's voices...other than 1984! ), to your breathing rhythm.

And it will be the AI that will decide if you are doing well, if you are doing badly, if you are a burden, if your life insurance needs to increase.... And when you go to complain to your mayor, to your insurer, to your office about the abuse immediately they will tell you: AI decided it.

AI is a highly asymmetric technology: you know everything about us, we don't know anything about you. So either it goes from the black box to the glass box, or to me AI is a technology against humanity and its survival... except at the Bonobo level, of course.

As for the emotion involving these posts... yeah it's true, human beings have emotions, and that's what first and foremost differentiates us from the logorrhoeic ChatGPT. We after a thousand words of his can always tell him, hey, son of a loom and a carillon f@!% off (of course as filed we will end up persecuted, but that is humanity).

Secondly I remember that Spock, who in short, being at least half Vulcanian was not just a passionate one, when Kirk was accused with evidence in hand by ChatGPT that he was guilty, it was the Vulcanian himself who said "he's doing the computer wrong" against all evidence.. And the lawyer who defended him ... was full of paper books (hold them dear!).

Also, I wouldn't recommend unplugging the chairs of the guys under the nuclear silos to replace them with ChatGPT, we've already seen the movie, and we don't have any more smart kids like David and Jennifer who can make WOPR play tic-tac-toe, who in fact understood that no one wins...but he wasn't hallucinating like that Open AI thing.

Mariasole
(=*.^=)


And finally, the AI song dedicated to all of you!... by Mr Neil Young from 1982!
https://www.youtube.com/watch?v=hQfgnooX72g

An now a small controversial note... it's true, I'm a bit of a hippy, for sure, but that doesn't make me a bonobo
So Vedic math exists and struggles along with us, at least for millions of people... who will not all be bonobos!
Here is a giant organization that teaches it in schools: https://vedicmathsindia.org/
Here a little history:
https://vedicmathsindia.org/what-is-vedic-maths/#father
https://vedicmathsindia.org/what-is-ved ... s/#history

An Indian friend of mine says that denying and vituperating Vedic mathematics is a great indication of Western envy. And by the way, it's true, even Leonardo was a good one at calculating in his head....

[richmond62]

vituperating Vedic mathematics is a great indication of Western envy

I have NOTHING against what is (incorrectly) termed Vedic Maths: I have got a real problem with
the fact that it is what it is is used to justify believing that all the planets in our system rotate round Jambhadvipa,
and or the Milk ocean on Dhruva Loka (i.e. the Pole Star), that in 'Satya yuga' theoretical science was far, far in
advance of anything we know now, that people chucked mantra-controlled bombs around the place
(sudharsana chakra and so on), that men have not landed on the moon (which, by the way is 4 times bigger
than the sun and a lot further away - eclipses caused by a Rakshasha's head that swallows it up (Rahu Asura)),
and another 57 'fairies at the bottom of the garden' BS stories [that are a load of mediaeval legends] palmed
off as scientific truth.

The Indians are on a fundamentalist roll at the moment; trying desperately to discredit the finding that the Aryan invaders came
into India from the north some 5-6 thousand years ago in favour of the sui generis myth that everything started in Bharata-varsa,
and that a Maharaja Parikshit ruled the whole world, which he would fly around and inspect from his mantra controlled aeroplane.

Unfortunately for the poor Indians empirical science (this is NOT science as a religion) has shown that the idea that pre-5000
years ago in Dvapara yuga humans were 10 times bigger than they are now, lived far longer,and generally strutted their funky
stuff like it was going out of fashion, is a load of old cobblers.

Similarly, Vedic Astrology, Vedic Architecture, and Ayurveda (mediaeval cod medicine).

So, Vedic Maths, Yes: but definitely not to justify all the nanky-poo in the Puranas:

https://youtu.be/mKyQxI1Q3UU

[richmond62]

AI as it is presented and as it is beginning to be marketed (in a totally irresponsible way since it suffers from hallucinations and its creators say it can be dangerous) is a pervasive technology that will squeeze in everywhere: more misinformation! That is the least of it!

Hmm: then why are a lot of the people involved in its development advising extreme caution?

2. I recently designed and built an elaborate raised vegetable bed in our garden with a pencil some paper, a ruler, a saw, a drill,
lots of planks, lots of wood screws, gallons of wood preservative, quite a bit of cursing and swearing, AND NO A.I. in sight.

3. I baked 2 loaves of bread yesterday: flour (of 3 types), yeast, salt, water, manual kneading, tins greased with butter, AND NO A.I. in sight.

I imagine that, unless I die first, that in 15 years time, if my arthritis is not much worse than it is at present (chance should be
a fine thing), I should manage both of those activities again with NO A.I. in sight.

Of course, if my arthritis has fekked my hands completely I will be grateful of a robot that is a bit more switched-on than my Bluebot to do those things for me.
-

Bluebot.jpg (7.5 KiB) Viewed 57706 times

-
The hysteria is as stupid as a woman who, 12 years ago asked me if it was a problem NOT upgrading her computer from Windows XP: of course it wasn't a problem, and her machine chunters on doing exactly what she wants to this day.

The marketing is one thing: but as you and I know that today's washing powder is no better than washing powder 20 years ago (you can only wash WHITE, not 'WHITER'), then we are the idiots for expecting far too much out of A.I.

When I was a kid our Primary teachers (who really needed their heads examining) were telling us, authoritatively, that when computers became ubiquitous (which they now are) we would have to work less: Oh, Ha, Fekking Ha: we work just as hard, or possibly even harder, at slightly different things.

So the only 'hallucination' I'm experiencing at the moment is one of sitting on my balcony with a decent book and a cup of coffee watching a semi-humanoid robot doing the weeding in the raised vegetable bed.

[stam]

Some reflections...but I would not want my passionality to be mistaken for a form of resentment...I'm just Italian aren't I?

I hear (although don't necessarily agree with) all you're saying and all other posts in this thread. What I don't get is the point of it:
Is this an plea to stop LC from incorporating AI as an option in their software (remembering that LC will continue as is, and the free tier of Xavvi does not include AI)?
Is it an expression of general concern and fear about the impact AI tools will have on society?

If it's the former, then it's likely you don't see the impact of not having AI features will have on LC as a corporate entity - at the end of the day an income stream is needed to keep the lights on at LiveCode HQ and if they go out, all this goes away. At present, LC feels like the best kept secret because the allure of 'traditional', and importantly, free tools/languages (and lets face it, tools that produce much faster binaries and tools produced by the OS vendors allowing deep integrations) stop the LC user base from expanding, resulting in a gradual erosion which is a terrible shame.

Incorporating AI features is likely to expand the user base (keeping LC as a corporate entity viable), as it will enable newer users to be drawn in much quicker and will place LC in a very good position competitively against other languages if done well. But remember we are talking about optional tools here.
Huge numbers of IDEs and languages are now incorporating AI based features to help aid developers - LC not doing so will leave them at a great and likely non-recoverable disadvantage, I suspect similar to to SuperCard deciding to not go cross-platform or 64 bit - it will gradually die.


If it's the latter then I would agree that's still an unknown. But this technology is still very much in its infancy, which makes me chuckle when people complain about how bad it is, how it 'hallucinates' etc. Do you know many huge software tools that didn't ship with (major) bugs?

It is extremely likely problems like 'hallucinations' will greatly reduce in coming years. Also important to remember that AI is not an 'entity' - it's a software tool. It will impact all layers of society and everything we do - it's up to us, especially those with educational roles, to help shape how. Admittedly this will be a challenge - just taking in the discussion on copyright is enough to prove that. And it's evident to all that this technological leap will make some skills obsolete - but does that turn humans into houseplants or free them up to learn new skills?


Pretending AI tools can be made to go away is extremely naive given the financial and political interests - you might as well just insist Windows 11 crawls back into DOS. Ain't gonna happen - so with that in mind, shouldn't the discussion be about how to harness AI, rather than bemoaning the fate of humanity in general and LiveCode in particular?

[richmond62]

so with that in mind, shouldn't the discussion be about how to harness AI, rather than bemoaning the fate of humanity in general and LiveCode in particular?

Of course it should.

AND, if the Indians want to stay with their Vedic Maths [however marvellous it may be], they will look like a bunch of silly
sausages when the rest of the world leaves them in the dust.

This social revolution [because that is what it probably is] is no more of a social revolution than that wrought by the printing press,
universal literacy, the radio, the television, and the internet: only extremely retro types are hand-copying books, and most of us so
dependent on electricity and modern medicine we take it for granted: presumably in 10-20 years A.I. will be so woven into the fabric
of life it will seem just a much a part of it as electricity.

Now, as we all know, electricity can be extremely useful, and it can kill: so, pretty quickly we learnt to respect it, insulate it and take
precautions. There is no good reason why we should not do that with A.I.: after all we did it with Uranium.

If one stops to think [not a habit followed by many adherents to Indian gurus such as Jagadguru Shankaracharya Shri Bharati Krishna Tirthaji Maharaja: https://vedicmathsindia.org/what-is-vedic-maths/#]
one would realise that one's ability to do mental Maths has got b*gger all to do with Artificial Intelligence. But using one thing to denigrate something that is not actually logically connected is typical of the type of totalist thinking propagated by oriental cults.

Of course a retreat into cultic, fundamentalist thinking is common enough historically when a person or people feel threatened intellectually. Islamic fundamentalism and the Wahabi movement have been explained as a reaction against the vast materialistic advances that happened in the Judeo-Christian west in the 20th century.

One of the most hilarious things that has come up here is a miss use of the word 'vedic'. 'Vedic' refers to historical period in what we could call 'Greater India' [India, Pakistan, Bangladesh, parts of Afghanistan, Burma, Cambodia, Laos, Nepal, and Sikkim and philosophical and religious ideas developed then (from which Hinduism developed later), generally accepted as sometime after 5000 BCE and before 500 BCE. So the co-opting by Jagadguru Shankaracharya Shri Bharati Krishna Tirthaji Maharaja (who died in 1960) of the word 'vedic' for his system of mental Maths (which he dreamt up meditating in a forest), thereby gives it a false sense of importance and provenance that is 'off' in a whole lot of ways.

Ultimately, as far as I can see LiveCode [whether I, personally, like it or not] will 'go under' if it does not embrace A.I. in one way or another.

As my revenue stream does not depend on A.I. I shall not embrace it: but I am not such a Luddite (and that IS the appropriate word in this context)
to protest against it and 'do a King Canute' and tell the sea to retreat.
-

-
As many of the children I teach English to at present will probably depend in the future on A.I. for at least part of their revenue stream I will try
to understand it so that I can discuss it in a reasonably informed way with them and not look completely like the Brontosaurus I am rapidly becoming.

[Mariasole]

Staying strictly on LC.
It's a big struggle, at least for the environment I'm in, to defend LiveCode as a "serious" and competitive language like Python etc.
Integration with AI, as the steam masters have hoisted the flag, is of course essential in order not to go out of fashion.

But in my opinion the approach is wrong. LC should, in case, have had an AI-based correction/suggestion engine/library generator.
Not a new product that "generates" apps with (which have LiveCode as their language).

In this way LiveCode as a platform and language is minimized, while LC is engulfed by this new, flagship product, which will be Xavvi.
In fact, Xavvi is already "sold" as a "no code" product, confirming what I said earlier, that code (livecode!) is a secondary aspect.

The marketing effect will be that apps made with Xavvi will be considered toys manufactured by AI, thus without any value or respectability.
While LiveCode will be considered even less than what it is.

Forward Xavvi, since LiveCode is not a "respected" language in the environment and nothing has been done as an investment to make it so (such as alliances with platforms, etc.) will be buried by the various Pavvi (AI producing code and apps in Python) and Cavvi (in C) and Bavvi (in BASIC for ZxSpectrum).

Xavvi, that is, making a tool for an AI to write code, is not a genius idea, it is an idea that can be replicated (or already done) by any company (the AI is the same for all... ChatGPT! wow! ). It cannot be a killer application, if you know what I mean.

Also, not releasing LiveCode 10 sooner and that feature to better "encrypt" executables for which you don't collect money, I seem to recall, did not seem respectful to us users (then forgetting the long-standing problem of the open source shutdown and the so far unfulfilled promise of a free "version" for schools).

To recall that LC10 was emphatically presented as well as Xavvi, but now it seems that LC10 is some kind of stale old project...

This is what I think. As far as I'm concerned my students are losing interest in anything that is "working" on anything, including those who design, so now in a few days AIs are landing on cell phones and two of them keep begging me to contact the boss of LiveCode ("who once answered you") to put "save as" in Xavvi in other languages that their cooler little friends pretend to master.... mumble...

Mariasole
(=°.*=)

[richmond62]

As the ZX Spectrum was discontinued in 1992 I don't really see what your point it.

It is fun to fool around with an emulator: but not much good in any other respects.

https://www.bannister.org/software/fuzzbug.htm

[jacque]

But in my opinion the approach is wrong. LC should, in case, have had an AI-based correction/suggestion engine/library generator.
Not a new product that "generates" apps with (which have LiveCode as their language).

I think you have misunderstood. Xavvi will in fact be a library/code generator that can also suggest and/or explain the reasons behind its code for learning or teaching. It will not be able to write whole applications on demand, but instead will suggest optimized code for the items or modules you ask for. It's similar to how we do not need to code the whole process of creating a button or a field, we only need to drag one from a palette. Xavvi will be similar, but will be able to create more things than we now have in the tools palette. You could ask for a pre-made list of specific content embedded in a datagrid, for example.

It will be crucial that you understand the language because AI will not always be correct. It will always need proofing and it will make mistakes, especially at first. I very much doubt any AI will be able to write an entire application without input or correction from a human who already knows the language, the organizational aspects required, the navigation methods, the card content, etc.

To recall that LC10 was emphatically presented as well as Xavvi, but now it seems that LC10 is some kind of stale old project...

Not at all. LC 10 is proceeding quickly, and dp 5 was released just a couple of weeks ago. This is a very big rewrite so it is taking longer than smaller updates but I have been using LC 10 for some time now. LC 10 does not show on the main downloads page of the web site because it is not yet final, but you can download it from your account. Login and click Downloads in the left side panel to see all available downloads, past and present.

LC also has stressed repeatedly that since Xavvi is built on top of LC 10, it is essential that LC 10 and future versions are updated. Xavvi cannot survive without LC 10. Also, for your teaching situation, you are free to ignore Xavvi completely -- LC will still exist just as it is now. There is no requirement at all that users must use AI supplementation. We are free to use LC alone, LC with Xavvi, and Xavvi by itself (which I doubt will produce much valid content, at least at first; it will need correction, training, and feedback.)

Also, to begin with, Xavvi will be exclusively online. It will not be embedded in the desktop app for a while. So it is unlikely to affect you at all for now.

[stam]

But in my opinion the approach is wrong. LC should, in case, have had an AI-based correction/suggestion engine/library generator.
Not a new product that "generates" apps with (which have LiveCode as their language)

Errrrm... I think you may not have all the facts correct:

1. Xavvi appears to have first been designed wisely or not as a low-code/no-code interface to LC and was probably conceived and prototyped before AI was a thing - but with the advent of AI this was incorporated into Xavvi as it's much simpler to control the no-code interface than the actual LC product and code. Even creating an interface will never be exactly what you're after, but it is a way to greatly speed up development - by the developer. When they talk about this being an app you can export to LC, the mean that using the no-code interface you can build the app (with or without AI -- and if you pay attention, there is a free tier to Xavvi that doesn't include AI tools).

2. Xavvi does not create an app for you; it can create bits, such as the layout with basic functionality and it can create code if asked. But a whole app? No, I don't think so... (and seems unlikely in the near to medium term)]

3. Xavvi includes the correction/suggestion engine you already allude to. There is in fact a wonderful video showing this in action.

I take it that you have looked at all the video demonstrations what of what Xavvi promises to do and have an informed opinion?

[mwieder]

2. Xavvi does not create an app for you;

Well, let's see... what does the PR say?

As you describe it, Xavvi writes code for you in English. For a simple app you won’t need to edit any code.

you’ll be able to build amazing things with Xavvi. Everything from a simple business application to an entire SaaS platform that can scale to millions of users.

Xavvi is the future of how you’ll create apps with LiveCode technologies though we’ll continue to maintain LiveCode for those that need it.

Xavvi’s no-code tools – its AI builder and Actions editor, will let you build a powerful app without writing any code.

[jacque]

Well yes, but while you may not need to edit the code, you will need to define the bits and pieces you need and how to arrange them. That is, I'd be surprised if you could say, "Make an accounting app" and have it appear ready to distribute.

[richmond62]

Xavvi’s no-code tools – its AI builder and Actions editor, will let you build a powerful app without writing any code.

Hyperbole, surely?

Even my friend, who, for his sins, fills his fridge by pumping out game pieces with a 3D printer, spends his life
married to a dentist's sanding wheel removing shimmy bits from what the printer spits out.
-

wheels.jpg (5.59 KiB) Viewed 57411 times

-

As you describe it, Xavvi writes code for you in English. For a simple app you won’t need to edit any code.

Slightly disingenuous surely, as any app that is so simple that A.I. can spit out without any 'warts' is probably
so 'flipping' simple that the 9-10 year olds who attend my annual xTalk intro classes could rustle one up in half an hour:
so, apart from a pedagogical triumph for yer average 9 year old who has previously through 'programming' meant fooling
around with SCRATCH or ENTRY, is, well, what 9 year olds manage on day 2, and nothing more.

While Xavvi (and her sisters) may be really quite helpful indeed (especially at the more moronic, repetitive tasks and heavy lifting)
I can see no real reason why people who have devoted, say, the last 20-30 years of their life getting up to speed with xTalk,
need to get moist around the oxters, as they will only have to nip out and buy a "dentist's sanding wheel."

[richmond62]

And anyone who is so naive to depend 100% on a 'robot' to do the work for them is asking for trouble
for the simple reason that my good friend, a senior paediatrician and researcher into infantile brain disorders remarked to me over a glass of lemonade in the park yesterday . . .

https://vagabond.bg/medical-university- ... ation-3367

Summary (as conversation was mostly in Bulgarian):

1. Neurons.
2. Axons.
3. Synapses.

Certainly non-binary, and probably each neuron is capable of being in several 'states' at the same time, and, anyway, it all seems to happen at the synapses, and, frankly, we really do not know.
-------------
AND, what did Richmond conclude from what someone who knows far, far more than he does about these things?

1. A.I. may imitate a human brain and human behaviour (as far as it is able), but what goes on inside a computer's processor(s) should
NEVER be equated with what goes on inside a human brain.

2. Even your average mentally-subnormal toilet operative (and if that description offends you, hard cheese) can, in some ways at least,
run rings round A.I.

3. Another of Richmond's "informants" (and, hey, what is a life without a series of mentors unless one wants to live some sort of 2 dimensional, flat thing that passes for 'life'?), was a mentally-subnormal toilet operative, who cleaned the toilets at my posh boarding school (that school in a, frankly, pretty dirty fiddle [and this was universal as I observed it in 2 other boarding schools at the time {1966-1980}] employed and housed mentally subnormal men as, to be quite honest, menial slaves.), who would, from time to time let slip a comment re some situation or another that would blow me away at its perspicacity. I often wonder what happened to 'Bumble': I used to read him his postcards from him Mum and write back for him. Needless to say, the stupid f*cks at my posh boarding school told me I was not allowed to 'fraternise' with the 'boys': who, really, seemed to me an awful lot nicer than half the pupils and half the teaching staff.
-----------
So . . . to Xavvi [Yes, that does strike me as a slightly daft name, but as I am paying no money whatsoever to LiveCode I do not feel it is my place to suggest a better name]: my reactions are:

1. Let's see what comes of it.

As everyone hereabouts should know by now, I am fairly cynical after the debacle of previous fund-raisers.

2. I genuinely hope that for the investors it materialises into a helpmate that speeds up program design.

[Mariasole]

2. Xavvi does not create an app for you;

Well, let's see... what does the PR say?

As you describe it, Xavvi writes code for you in English. For a simple app you won’t need to edit any code.

you’ll be able to build amazing things with Xavvi. Everything from a simple business application to an entire SaaS platform that can scale to millions of users.

Xavvi is the future of how you’ll create apps with LiveCode technologies though we’ll continue to maintain LiveCode for those that need it.

Xavvi’s no-code tools – its AI builder and Actions editor, will let you build a powerful app without writing any code.

I think mwieder expressed my doubts about Xavvi much better than I did and even using the words of Livecode's marketing.
All my doubts are in these sentences and so far no one has refuted them....

Xavvi in my opinion is a fashionable way to eat Livecode and "refound" the company on an on-demand services model.

Thanks to richmond and jacque and everyone. A small note to stam: it seems to me that you keep looking at the finger while pointing to the moon and I found the personal attack on "informed opinion" a bit over the top. That one has doubts about LiveCode's new line I think is something human, for me LiveCode is not a religion.

As far as I'm concerned I don't have the money for anything, so not even for Xavvi, we use in school as long as the open source version of LiveCode will work. Then in the meantime that Xavvi with xxxx dollars a month will dispense codes of wisdom, I will look for something working and free on xTalk. Amen.


Mariasole
(=^.^=)