Re: XAVVI and app/code copyright
All that it would take is a simple statement from LC ltd to clear it up. But, I am not really worried about it. I am just curious. Linda ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Can Xavvi generate sensible text output as well as code?
My day to day work is in medical education – and assessment specifically. I’m interested in using LLMs to (1) help generate assessment items (well draft versions of them at least) and (2) individualised descriptive feedback based on assessment performance. But, rather than trying to get LC/Xavvi to do it all I think I’ll likely end up using LC draw on the relevant data to construct an appropriate query, passing that to the LLM (ChatGPT, BioMedLM or whatever) via API to generate and return the response, and then having LC deal with it from there (e.g. upload to assessment item bank for further work; assemble and distribute feedback report). A few hours yesterday spent crafting and refining inputs and queries suggests that’s all entirely possible. Should be fun! > How to integrate the AI part with the LC part? Again, one could acquire > just the list, and then let LC generate the output string. But at least all > would end up in a handler, fully in LC, where I assume additional “real” > work would then be done. > > What I mean is, how can one best integrate the “outside” AI work with the > “inside” LC work? That is what I have to get my head around. > > Lurking in the background,I do not want AI to put LC out to pasture. I did > that once with HC, and still have not gotten over it. ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Latest video -- integration of AI into script editor
On 5/18/23 4:14 PM, Richard Gaskin via use-livecode wrote: J. Landman Gay wrote: > I had to make a form yesterday with lots of fields and while it isn't > hard, it took 20 minutes to name everything and align the labels and > entry fields. If you need to do that again: http://fourthworld.net/lc/4W_FormMaker.rev Cool. That does help with alignment and field creation. It would still need some tinkering to get the sizes and properties right but it's a quick start. -- Jacqueline Landman Gay | jac...@hyperactivesw.com HyperActive Software | http://www.hyperactivesw.com ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Latest video -- integration of AI into script editor
J. Landman Gay wrote: > I had to make a form yesterday with lots of fields and while it isn't > hard, it took 20 minutes to name everything and align the labels and > entry fields. If you need to do that again: http://fourthworld.net/lc/4W_FormMaker.rev -- Richard Gaskin Fourth World Systems Software Design and Development for the Desktop, Mobile, and the Web ambassa...@fourthworld.comhttp://www.FourthWorld.com ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Livecode enhancements
I ran into another time today where having the ability to define data structures (or templates) in Livecode would be a significant enhancement (for me at least) to the language. I see these 2 enhancement requests that cover defining data structures/templates, although they are suggesting different things: https://quality.livecode.com/show_bug.cgi?id=4183 and https://quality.livecode.com/show_bug.cgi?id=9033 I may have missed some other enhancement for the same or similar feature. If you know of one, please let me know. I realize most work on livecode enhancements will be to support Xavvi and Livecode on the web and the things already planned for LC10 and 11, but someday, I really like to see a way to define a data structure and make instances of it in LC. ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Can Xavvi generate sensible text output as well as code?
I used ChatGPT to generate some test data for me. It took several tries before it was able to give me a dozen lines of useable data. I told it to separate each item with a comma and each line with a carriage return. Then I just copied and pasted the data into a spreadsheet program, created a .csv file from it, and imported into my postgresql database that my LiveCode server uses. In your case, you could just copy and paste the data into an LC text field or table. Enjoy your day! Rick > On May 18, 2023, at 9:25 AM, Craig Newman via use-livecode > wrote: > > In Jacque’s “test” AI first had to create a list of ten “people”, garnered, I > assume, from research into our planet’s knowledge base. It knew what a “name” > should be. That is cool. > > That list could have been a list of actual humans, derived from, say, the > staff of a particular bank. > > If the list alone was in hand, it would take just a handful of LC code to > create the string associated with each name. > > How to integrate the AI part with the LC part? Again, one could acquire just > the list, and then let LC generate the output string. But at least all would > end up in a handler, fully in LC, where I assume additional “real” work would > then be done. > > What I mean is, how can one best integrate the “outside” AI work with the > “inside” LC work? That is what I have to get my head around. > > Lurking in the background,I do not want AI to put LC out to pasture. I did > that once with HC, and still have not gotten over it. > > Craig > >> On May 17, 2023, at 8:26 PM, J. Landman Gay via use-livecode >> wrote: >> >> On 5/17/23 5:38 PM, Terry Judd via use-livecode wrote: >>> Let’s say I have a whole lot of input statements/info, some of which is >>> generic but some of which varies according by user. Any thoughts on whether >>> it might be possible to leverage Xavvi to process and combine that info >>> into useful prose summaries for each user in a way that I can’t already do >>> in Livecode - ie generative AI text rather than merging into template text)? >> >> Well, I'd expect so since Bard can do it, if I understand what you mean. >> Here's my test. (Bard thinks it's code.) >> >> My prompt: Make a list of 10 people with first and last names. Combine each >> one into this sentence: "Hello firstname, I'm exploring the family tree of >> your lastname relatives." Substitute "firstname" with each person's first >> name and substitute "lastname" with their last name. Return a list, one >> person per line. >> >> Bard says: >> >> Here is a list of 10 people with first and last names, along with the >> sentence "Hello firstname, I'm exploring the family tree of your lastname >> relatives." with each person's first and last name substituted in: >> Code snippet >> >> Name | Sentence >> --- | >> John Smith | Hello John, I'm exploring the family tree of your Smith >> relatives. >> Mary Jones | Hello Mary, I'm exploring the family tree of your Jones >> relatives. >> Michael Brown | Hello Michael, I'm exploring the family tree of your Brown >> relatives. >> David Williams | Hello David, I'm exploring the family tree of your Williams >> relatives. >> Susan White | Hello Susan, I'm exploring the family tree of your White >> relatives. >> James Green | Hello James, I'm exploring the family tree of your Green >> relatives. >> Sarah Black | Hello Sarah, I'm exploring the family tree of your Black >> relatives. >> Peter Brown | Hello Peter, I'm exploring the family tree of your Brown >> relatives. >> Jessica Smith | Hello Jessica, I'm exploring the family tree of your Smith >> relatives. >> >> -- >> Jacqueline Landman Gay | jac...@hyperactivesw.com >> HyperActive Software | http://www.hyperactivesw.com >> >> ___ >> use-livecode mailing list >> use-livecode@lists.runrev.com >> Please visit this url to subscribe, unsubscribe and manage your subscription >> preferences: >> http://lists.runrev.com/mailman/listinfo/use-livecode > > > ___ > use-livecode mailing list > use-livecode@lists.runrev.com > Please visit this url to subscribe, unsubscribe and manage your subscription > preferences: > http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Can Xavvi generate sensible text output as well as code?
If you watch some of the videos posted on the challenge page it becomes clearer. LC AI will be integrated into the script editor and will create handlers as expected. You can accept, reject, or revise its output. LC AI knows all about what's available and how to use it. Bard, on other hand, doesn't know LC and can only offer generic output though it does know 20 other coding languages. Xavvi is built on top of LC and the team has reassured us many times that the LC we know is not going away. There will be regular updates and you are free to ignore the AI if you want. But given what I've seen so far, I'm betting most of us will use at least some of its features to save time or to expedite certain tasks. We made the challenge! I'm happy. -- Jacqueline Landman Gay | jac...@hyperactivesw.com HyperActive Software | http://www.hyperactivesw.com On May 18, 2023 8:28:42 AM Craig Newman via use-livecode wrote: How to integrate the AI part with the LC part? Again, one could acquire just the list, and then let LC generate the output string. But at least all would end up in a handler, fully in LC, where I assume additional “real” work would then be done. What I mean is, how can one best integrate the “outside” AI work with the “inside” LC work? That is what I have to get my head around. Lurking in the background,I do not want AI to put LC out to pasture. I did that once with HC, and still have not gotten over it. ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Latest video -- integration of AI into script editor
Thanks Steven, That sounds like a sensible approach. Mark > On 17 May 2023, at 12:59 pm, Steven Crighton via use-livecode > wrote: > > Hi Mark > > I’m really pleased to hear you enjoyed the video. What I would say is our > testing so far has been extremely impressive, it was the inspiration to make > that video. But you’re right, AI won’t get the script right every time and we > have to handle that carefully. If the script that comes back won’t compile > then Xavvi will intervene to make sure you are not actually accepting code > that won’t run. However there may be instances where it supplies code that > does compile but is still not what you hoped for, in the first instance we > are providing an opt in to training feature where you can provide a thumbs up > or thumbs down to the code you have been supplied as well as give a > description on what was good or bad about it. Additionally there will always > be options to accept or decline the result and try again with a slightly > tweaked prompt. I hope this helps. > > ___ > use-livecode mailing list > use-livecode@lists.runrev.com > Please visit this url to subscribe, unsubscribe and manage your subscription > preferences: > http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Can Xavvi generate sensible text output as well as code?
In Jacque’s “test” AI first had to create a list of ten “people”, garnered, I assume, from research into our planet’s knowledge base. It knew what a “name” should be. That is cool. That list could have been a list of actual humans, derived from, say, the staff of a particular bank. If the list alone was in hand, it would take just a handful of LC code to create the string associated with each name. How to integrate the AI part with the LC part? Again, one could acquire just the list, and then let LC generate the output string. But at least all would end up in a handler, fully in LC, where I assume additional “real” work would then be done. What I mean is, how can one best integrate the “outside” AI work with the “inside” LC work? That is what I have to get my head around. Lurking in the background,I do not want AI to put LC out to pasture. I did that once with HC, and still have not gotten over it. Craig > On May 17, 2023, at 8:26 PM, J. Landman Gay via use-livecode > wrote: > > On 5/17/23 5:38 PM, Terry Judd via use-livecode wrote: >> Let’s say I have a whole lot of input statements/info, some of which is >> generic but some of which varies according by user. Any thoughts on whether >> it might be possible to leverage Xavvi to process and combine that info into >> useful prose summaries for each user in a way that I can’t already do in >> Livecode - ie generative AI text rather than merging into template text)? > > Well, I'd expect so since Bard can do it, if I understand what you mean. > Here's my test. (Bard thinks it's code.) > > My prompt: Make a list of 10 people with first and last names. Combine each > one into this sentence: "Hello firstname, I'm exploring the family tree of > your lastname relatives." Substitute "firstname" with each person's first > name and substitute "lastname" with their last name. Return a list, one > person per line. > > Bard says: > > Here is a list of 10 people with first and last names, along with the > sentence "Hello firstname, I'm exploring the family tree of your lastname > relatives." with each person's first and last name substituted in: > Code snippet > > Name | Sentence > --- | > John Smith | Hello John, I'm exploring the family tree of your Smith > relatives. > Mary Jones | Hello Mary, I'm exploring the family tree of your Jones > relatives. > Michael Brown | Hello Michael, I'm exploring the family tree of your Brown > relatives. > David Williams | Hello David, I'm exploring the family tree of your Williams > relatives. > Susan White | Hello Susan, I'm exploring the family tree of your White > relatives. > James Green | Hello James, I'm exploring the family tree of your Green > relatives. > Sarah Black | Hello Sarah, I'm exploring the family tree of your Black > relatives. > Peter Brown | Hello Peter, I'm exploring the family tree of your Brown > relatives. > Jessica Smith | Hello Jessica, I'm exploring the family tree of your Smith > relatives. > > -- > Jacqueline Landman Gay | jac...@hyperactivesw.com > HyperActive Software | http://www.hyperactivesw.com > > ___ > use-livecode mailing list > use-livecode@lists.runrev.com > Please visit this url to subscribe, unsubscribe and manage your subscription > preferences: > http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Latest video -- integration of AI into script editor
I did the same a few days ago. ;) Heather told me to purchase a2nd 2 year (now 10 months) and she would take care that it would be combined to 2 years. Von meinem iPhone gesendet > Am 18.05.2023 um 11:48 schrieb Jacques Clavel via use-livecode > : > > Hi all and Steven, > > I had pledged a few days ago to "Xavi Business 1 year" (499$) and now try > to extend it to "Xavi Business 2 Years" (999$). > But there is no option where I pay 500$ only to upgrade my pledge. > Please, how to do that ? > Jacques CLAVEL > >> Le mer. 17 mai 2023 à 14:00, Steven Crighton via use-livecode < >> use-livecode@lists.runrev.com> a écrit : >> >> Hi Mark >> >> I’m really pleased to hear you enjoyed the video. What I would say is our >> testing so far has been extremely impressive, it was the inspiration to >> make that video. But you’re right, AI won’t get the script right every time >> and we have to handle that carefully. If the script that comes back won’t >> compile then Xavvi will intervene to make sure you are not actually >> accepting code that won’t run. However there may be instances where it >> supplies code that does compile but is still not what you hoped for, in the >> first instance we are providing an opt in to training feature where you can >> provide a thumbs up or thumbs down to the code you have been supplied as >> well as give a description on what was good or bad about it. Additionally >> there will always be options to accept or decline the result and try again >> with a slightly tweaked prompt. I hope this helps. >> >> ___ >> use-livecode mailing list >> use-livecode@lists.runrev.com >> Please visit this url to subscribe, unsubscribe and manage your >> subscription preferences: >> http://lists.runrev.com/mailman/listinfo/use-livecode >> > > > -- > Jacques Clavel > ___ > use-livecode mailing list > use-livecode@lists.runrev.com > Please visit this url to subscribe, unsubscribe and manage your subscription > preferences: > http://lists.runrev.com/mailman/listinfo/use-livecode ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
Re: Latest video -- integration of AI into script editor
Hi all and Steven, I had pledged a few days ago to "Xavi Business 1 year" (499$) and now try to extend it to "Xavi Business 2 Years" (999$). But there is no option where I pay 500$ only to upgrade my pledge. Please, how to do that ? Jacques CLAVEL Le mer. 17 mai 2023 à 14:00, Steven Crighton via use-livecode < use-livecode@lists.runrev.com> a écrit : > Hi Mark > > I’m really pleased to hear you enjoyed the video. What I would say is our > testing so far has been extremely impressive, it was the inspiration to > make that video. But you’re right, AI won’t get the script right every time > and we have to handle that carefully. If the script that comes back won’t > compile then Xavvi will intervene to make sure you are not actually > accepting code that won’t run. However there may be instances where it > supplies code that does compile but is still not what you hoped for, in the > first instance we are providing an opt in to training feature where you can > provide a thumbs up or thumbs down to the code you have been supplied as > well as give a description on what was good or bad about it. Additionally > there will always be options to accept or decline the result and try again > with a slightly tweaked prompt. I hope this helps. > > ___ > use-livecode mailing list > use-livecode@lists.runrev.com > Please visit this url to subscribe, unsubscribe and manage your > subscription preferences: > http://lists.runrev.com/mailman/listinfo/use-livecode > -- Jacques Clavel ___ use-livecode mailing list use-livecode@lists.runrev.com Please visit this url to subscribe, unsubscribe and manage your subscription preferences: http://lists.runrev.com/mailman/listinfo/use-livecode
