for reading PDF the internal objects of PDF's would have to be named with a industry standard naming convention.
cant see it happening as there is no industry standard naming conventions for financial reporting company sets of accounts.
That would require the accounting society to understand technology.
https://opensource.adobe.com/dc-acrobat-sdk-docs/acrobatsdk/
The only way it could happen is for the companies to send data to the NZX in a standard format and for that to be a requirement of Listing on the exchange. It would be part of a wider push to standardise the publishing of financial data transaction formats.
In that regard FERG raises subject matter that has been pushed here for a while here by WZ. I have been regularly meeting with a very far sited semi retired ACA to create a CONSOLIDATED cross column format for financial private company sets of accounts as Multi dimensional data structures have been available for reporting and tabulation now for over 30 years in many software languages.
FERG raises the right questions.
Processing information from PDF reports is possible using VBA code on the internet.
The information inside these PDF documents may be able to determined by either giving CHAT GPT the job or by arrange the data in an indexed table and the user investor selects the information and guiding it into a standard model using an old fashion click of the mouse.
In MID winter this may be something you want to do but in summer are you going to spend your sunday afternoon sort through a list of non index objects from a PDF clicking your mouse while everyone else in on the golf course or surfing?
May be Chat GPT will read and sort it out for you?
The future is in play and one thing we know for sure and that is the FAT Lady hasnt sung yet.
Great thread Waltzing. I have seen you commenting on this elsewhere quite a long time ago, and I am intrigued by the concept. I am over tapping financial statements into spreadsheets given other priorities. I have previously written VBA programmes to not only scrape data from the internet but also programmes that use API calls to grab historical stock prices from Yahoo..so interpreting financial statements should be a doddle.
Regarding reading pdf versions of annual accounts, it sounds like your stumbling block is the lack of naming conventions within reoprts.....is that correct? A way around this might be to have a dictionary of common terms that resolve to a standard definition e.g. Trade Receivables, Receivables, Debtors and Trade Debtors all resolve to debtors for instance. Any term not found could be added to the dictionary and marked as unresolved until the user adds a definition. And I imagine the pdf documents use bookmarks of some sort to note the position of standard reports such as the Balance Sheet & P&L etc....is that correct?
P.S. It sounds like you want to set up something a bit like Xero for importing, interpreting, storing and then analysing. Point and click.
Yes a Dictionary or Indexed List structure for searching is the solution as you rightly say.
But for it to be useful it needs to be created in a configurable solution where it is runtime configurable and not created inside MS office but Drives products like MS office from the outside.
PS: XZ Product is a Internet Multi User platform specific for business and created for the KFC market.
It is not a company that develops low level multi platform scripting engines.
Im expecting to solve this problem as the result of other developments that just happen to do this process by default .
After all its just a few loops and some data returned into a list of data Variants. Variants arnt the MS version im thinking of but a new generation of Smart Variants that can store data and execute code.
More later as this new technology as it gets rolled out from the workshop.
Interesting. So it appears pdf documents don't use bookmark objects like Word documents. That makes it a bit harder. Text search can find things like "Balance Sheet As At" or similar to find the page that contains the Balance Sheet etc....but controlling objects even line by line....I'm not even sure where to start.
Meanwhile copy/paste into Word and/or Excel is possibly worse. That would require a lot of interpretation and culling just to trim it back to the underlying raw data...unformatted. Each report possibly has differing formats. Oh dear...might be a long winter.
BTW, assuming I am understanding you correctly, you can use MS VBA to drive other MS programmes from the outside. For instance I have an Excel model that drives web scraping and reformatting in another Excel workbook, before pushing into a Word template for automated non-financial reporting for a USA client. It has been running maybe 15 years now. Working on some minor tweaks tonight.
FERG you can open a PDF document in WORD and then access the WORD Document by VBA.
There is code out there that demos that and we dont know what CHAT AI will do yet.
Also dont forget ABODE has just released the BETA of FIRE FLY.
You can drive MS Office VIA LATE OLE BINDING from any computer language and you can also access PDF Docs from DOT NET in C SHARP.
We actually have code that does just that.
You an access EXCEL also from Chsarp and we have Set of Classes in VB DOT NET for doing that. It uses delegates to call back to standard WIN32 Programs.
Delegates are bit technical of course and we be understood by members here unless they are advanced users of DOT NET.
What FIREFLY can do is new and havnt tried to access it yet.
But we have a new Document Management product under development that we think will be of interest in Europe when linked to FIREFLY and also VBA and our NEW SMART VARIANTS scripting engine.
Example of putting the PDF into WORD document and then pasting into EXCEL. Finds the specific section you want to Find.
In fact you could automate a lot of this to be done by a software robot.
' open Word application and load doc
Set WApp = CreateObject("Word.Application")
' WApp.Visible = True
Set WDoc = WApp.Documents.Open(FName)
' go home and search
WApp.Selection.HomeKey Unit:=6
WApp.Selection.Find.ClearFormatting
WApp.Selection.Find.Execute "FINANCIAL PROFIT LOSS REPORT"
' move cursor from find to final data item
WApp.Selection.MoveDown Unit:=5, Count:=1
WApp.Selection.MoveRight Unit:=2, Count:=2
' the miracle happens here
WApp.Selection.MoveRight Unit:=2, Count:=1, Extend:=1
' grab and put into excel
Set WDR = WApp.Selection
EXCELSHEETRANGE(1, 1) = WDR ' place at Excel cursor
'repeat
WApp.Selection.HomeKey Unit:=6
WApp.Selection.Find.ClearFormatting
WApp.Selection.Find.Execute "Period of Report:"
WApp.Selection.MoveRight Unit:=2, Count:=8
WApp.Selection.MoveRight Unit:=2, Count:=3, Extend:=1
Set WDR = WApp.Selection
EXCELSHEETRANGE(1, 2) = WDR ' place in cell right of Excel cursor
WDoc.Close
WApp.Quit
Interesting - I forgot about opening pdf documents in Word. Handling data in Word using VBA is much easier when using bookmarks. Another way to hard code the page you want to find is to manually insert bookmarks on the relevant financial reporting pages and then run your code against that document between bookmarks. They work really well with VBA and data manipulation/extraction - but they have to be set manually first time around, or they could be set automatically based on (multi) text search which matches a heading style. All of that can be run and controlled from an Excel s/sheet per your sample code.
This method:
QuoteWApp.Selection.Find.Execute "FINANCIAL PROFIT LOSS REPORT"
assumes each wording format is the same between entities. I imagine in most cases you can't avoid manually setting heading markers the first time round.
It also looks like Word interprets various headings from the pdf document using the Headings style, which is another way to programmatically find a reference to the info we want.
That said, once you have a dictionary of definitions for how a company presents it's data, then subsequent data extraction attempts could be fully auto. Just have to note the format differences between HY and FY report versions for each company. Easy as. ;D
What is also nice is the financial data gets dropped into a table when converted to Word from pdf - which means you programmatically know the bounds of where the data is located. Very interesting and thanks for the reminder about opening pdf documents in Word. It's going to be a long winter......
ADOBE has a new version coming out with AI (our engineering just laughs) but lets call it that...
FIREFLY needs investigating to see what it can do...
MS OFFICE will need a major upgrade to embed this stuff....
EXCEL is a pain and for ADDINS was no where near as good and LOTUS was.
We do a lot of weird stuff in excel with text meta data expressions that get read before the spreadsheets are read or written.
Note: VBA is a lot faster than the slow access via VB DOT NET.
WE have DUMPED DOT NET and gone back to traditional CPP on linux and WIN32/64 binaries on Windows.
we call DOT NET, DOT NOT!!!!
NOW FERG your seeing the BRILLANCE of the ACCOUNTING SOCIETY in not standardising information in PUBLIC COMPANY ACCOUNTS even though SAP (Standard for Accounting) have been around for a LONG LONG TIME....
But not enforced by AUDITORS and Company law should have stipulated it for all Public Company Accounts and also standards for data transaction files...
Maybe AI will enforce it...
Really what you want is for CHAT GPT to learn and do the job for you....
I haven't used dot net in a while. 99.99% of what I do in using VBA. It's easy. Bookmarks in Word are life changing.
Here is sample code for handling bookmarks within Word VBA. I prefer early binding and full variable declarations, so you need to tick the various options under VBA > Tools > References for VBfA, MS Word xx Object Library & MS Office xx Object Library. This assumes bookmarks exist within the document, grabs their name and position, bubble sorts in an array and then outputs their names and positions with the document into the immediate window.
QuoteOption Explicit
Option Compare Text
'Global variables:
Dim MyBookmarks() As String
Dim bkMrk As Bookmark
Sub GetBookmarks()
Dim LC As Long, LL As Long, IL As Long
'get the bookmark details
With ThisDocument
.Bookmarks.ShowHidden = False
LL = .Bookmarks.Count 'Note: assumes count > 1
ReDim MyBookmarks(2, LL)
For Each bkMrk In .Bookmarks
If bkMrk.Name = "AllBookmarks" Then
LL = LL - 1
bkMrk.Delete
ElseIf Left$(bkMrk.Name, 1) = "_" Then
LL = LL - 1
Else
LC = LC + 1
MyBookmarks(1, LC) = bkMrk.Name
MyBookmarks(2, LC) = bkMrk.Start
End If
Next bkMrk
End With
'Bubble sort the bookmarks into the document sequence
' (VBA defaults to alphabetical sequence for bookmarks, so this needs to be corrected)
For LC = 2 To LL
If CLng(MyBookmarks(2, LC)) < CLng(MyBookmarks(2, LC - 1)) Then
For IL = LC - 1 To 1 Step -1
If CLng(MyBookmarks(2, IL)) > CLng(MyBookmarks(2, IL + 1)) Then
MyBookmarks(1, 0) = MyBookmarks(1, IL)
MyBookmarks(2, 0) = MyBookmarks(2, IL)
MyBookmarks(1, IL) = MyBookmarks(1, IL + 1)
MyBookmarks(2, IL) = MyBookmarks(2, IL + 1)
MyBookmarks(1, IL + 1) = MyBookmarks(1, 0)
MyBookmarks(2, IL + 1) = MyBookmarks(2, 0)
Else
Exit For
End If
Next IL
End If
Next LC
Debug.Print "Bookmark details:"
For LC = 1 To LL
Debug.Print MyBookmarks(1, LC) & " : " & MyBookmarks(2, LC)
Next LC
End Sub
Hopefully I haven't forgotten anything important with that copy/paste...
More on Word bookmarks here:
https://learn.microsoft.com/en-us/office/vba/api/Word.bookmarks
I kid you not - bookmarks in Word are life changing when using VBA. That reference also has the methods for adding, checking etc.
Quote from: Waltzing on Jun 16, 2023, 09:47 PMNOW FERG your seeing the BRILLANCE of the ACCOUNTING SOCIETY in not standardising information in PUBLIC COMPANY ACCOUNTS even though SAP (Standard for Accounting) have been around for a LONG LONG TIME....
But not enforced by AUDITORS and Company law should have stipulated it for all Public Company Accounts and also standards for data transaction files...
Maybe AI will enforce it...
Agreed. Not just report headings, but also line items within each report. A self-maintaining dictionary object in the s/s MF would handle future versions of reports from existing reporting entities, and any gaps for new entities could either be held in limbo and/or defined later. You could have a lookup table per company that assigns found phrases to desired phrases e.g. whatever their description -> becomes -> trade debtors. I can't think of any entities that would use the same words for differing definitions, so you could have a global mapping dictionary, otherwise you need one per entity being analysed.
Quote from: Ferg on Jun 16, 2023, 10:05 PMany gaps for new entities could either be held in limbo and/or defined later. You could have a lookup table per company that assigns found phrases to desired phrases e.g. whatever their description -> becomes -> trade debtors.
This needs explaining. Models are usually data -> transformation -> presentation. Data is the cleaned raw data per financials. Transformation restates their description to desired description. Presentation is the pretty report at the end. Anything not immediately matched from the raw data could be given "unassigned" (i.e. it is in limbo) in the mapping table, and once that table has the correct mapping, it will self correct on the report once the mapping is populated for items that cannot be interpreted on the fly. Given the template contains the mapping, it will be there for next time.
The only Developer to do a Bubble sort in VBA outside PASCAL!!!!
Classic !!!
BUBBLE SORTS!!!
Whats really needed is a product for companies to use to publish financials.... and drive ADOBE from a dataset Dictionary.
Many folks here reading for code wont know what a CLng is ...
For accounting types that a LONG data type..
https://learn.microsoft.com/en-us/office/vba/language/concepts/getting-started/type-conversion-functions
it was always going a very very long winter...
Quote from: Waltzing on Jun 16, 2023, 10:35 PMThe only Developer to do a Bubble sort in VBA outside PASCAL!!!!
Classic !!!
;D Classic indeed. Pascal was my second language, COBOL was my first. From memory Pascal forced indentation (Polish notation?), and I still use it.
Id have to get my Pascal book out to look.
https://en.wikipedia.org/wiki/Polish_notation
Once you have the tables in Excel an extraction to XML Tables might be possible.
Ok FERG thinking about bringing in the imported Book mark table in Word into an object with a sortable list.
QDocumentBookmarks queue,type
BookMarkNo long,name('BookMarkNo')
BookMark string(255),name('BookMark')
RowColumnTableref &RowColumnTable
End
DocumentBookmarksref. RowColumnTableref &= New(RowColumnTable )
RowColumnTable is a Class that handles the data in the table.
This will hold the book marks and the tables contents for processing into a larger XML Database solution...
Now cant wait to try this sometime in the next few weeks if we get the time...
No bubble sort required.
The queues Lists can be sorted on demand by a built in SORT command.
This could be used for statements also like Dividend statements in PDF ect.
HLG PDF reports are locked up to stop conversions to WORD.
Unlocking can be done by printing the PDF to PDF.
Not sure if bookmarks are intact.
Ok printing the PDF was manual but i think it can be automated by using the standard OLE Active X late binding commands sent as text commands. WIll demo later next week i hope..
Got full Table and paragraph XML data tables working..
Print to PDF came up on windows 10 and 11 as standard options under Print in Office.
We think we have fixed the problems using the PRO version of ADOBE and created a index template approach to DIV PDF where the company has a paragraph template and then on importing and processing PDF's converted to WORD using then PRO VERSION the positions of the paragraphs are then mapped to the templates.
And you have A template learning processing solution for Dividend statements.
Now to apply that to learning Public Company accounts.
WZ
PDF window 11 in Edge showing default PDF option in the printer options.
Gotcha - it was the short form version of the HLG A/R that was locked up tight and won't print. The glossy version has the print option so that's all good. It's going to be a long winter - these are complex underlying data structures. Plus look at the partial lack of tables in FRW - half the data is in a table, the descriptions are not....
Yes i noticed this on DIVIDEND statements and that is after processing using the ADOBE PRO... maybe they will allow AI ? AHHH I GOT IT...
anyway we have now a structure where you program the Document your processing with some commands for documents...
We are going to HOST IT in our SMART VARIANT language which means you can develop it at runtime ...
Which means you can train it with a GUI...that you create ... commands site in simple text file and you dont need to write VBA...
made a lot of progress at the weekend in auckland.
The trick is to manage the data extracted from the PDF's and allow it to be easily manipulated for statistics and for business related matters.
For this we have stored the data from tables and paragraphs in what is called a sortable list or queue which can be automatically store as XML data tables..
hence we have data stored for processing in loops and spreadsheets from XML data Tables. Simple management of the extracted data.
I forgot to say the real trick here is the consolidation of multiple entities for reporting and for that we can add to the project a project for processes multiple entities onto one page for P&L balance sheet processing...
It currently only handles 10 entities on one page and has 3 balances per report column set but we could take that set of classes and set the data table as a single reporting column not 3.
Instead of YTD OB and QTY we will see if we can supply a Sortable List with one column for YTD.
Now imagine that kind of power in your hands for reporting on multiple public companies on one page...
Its based on a project we have that lets us report on multiple private entities on one page..
Power over managing entity data is the power we like the most...
CPP large language models... and the possibility of running them locally via your scripting language models.
Once you have this kind of power turning up on your local machines you start to look at big data processing locally.
It doesnt take an ESTN to realise that PHD students are going to be all over this and the divide between the HAVES and HAVES nots is going to be a whole new society...
https://blog.maartenballiauw.be/post/2023/06/15/running-large-language-models-locally-your-own-chatgpt-like-ai-in-csharp.html
The FAT LADY hasnt sung yet but shes warming up..
Processing the converted PDF documents into there paragraphs and Tables is simple once you have the data and its separated from any strange characters that might contaminate your data...
The it simply a matter of using some macro instring() to find the phrases your after and tables for transfer into data structures that might house the information you want to manage....
The instring macros can take phrases and you can instruct your software to learn from an importation GUI...
Paragraphs have there positions in the VBA object model and they are the same for many documents that are standard formats...
Your learning template needs some fancy commands of course but this scrapping and matching is the next step in setting up micro learning tables and with imbedding of large language models in scripting languages the future is about connecting your information to these models and storing the information for future use...
here in this example data is stored in plain XML for farther processing and matching. Once data is XML tagged you can farther pass this data sets on to any software you want.
the new meta Base Template system does not yet read documents and understand them until it is trained...
It take document apart and then matched there composition to its private and public understanding of a Document Template. If can match the parts of a document to its PP contexts. (Public and Private Word base EQU tables). EQU stands for the word information equates. These are then matched back to the processed documents templates which can process and execute a documents parts. Sort of turning the parts of a document into an interactive command driven Analysis.
This process then decides when and if to link to an LLM for farther processing. This reduces the processing over head on using LLM's which is cheaper for the users and the environment as LLM's are going to be anything but Green!
You compose the template from parts of a document type.
and then when linked up to a Large Language scripting model it might in the few years time do a lot more such as read company reports creating a series of data tables before processing those tables and sending parts of them to LLM's for farther processing.
Right now most people use LLM by throwing documents at the LLM.
The Meta base Template system pre processes documents before then selectively sending information to an LLM for farther Analysis.
The benefit from this approach is that large meta models can be built to support LLM and statistical data can be stored locally.
There is no doubt that from an investor prospective local and global server language models are the "way of the future"
https://www.youtube.com/watch?v=4_Pbx9mvWPY
Amazon has been working away on this for a while and Microsoft does not have its own chips but must rely of others. Not that that is a bad thing.
https://www.cnbc.com/2023/08/12/amazon-is-racing-to-catch-up-in-generative-ai-with-custom-aws-chips.html
In regard to Financial Accounting standards it time these were throw out along with the credibility of the Accounting society.
Multi dimensional data constructs have been available on computer hardware and software stacks for 30 years or more and yet accounting remains single dimensional and therefore completely polluted and near useless..
It time for developers to take the matter in hand and just get rid of the baby and the bath water....
Bye bye single dimensional accounting and F Gen AI wont need it either....
In order to stablise the calling between compilers we create the Iextraterrestrial Class...
it a simple 5 variable class container with standard interfaces..
a giant cstring 1000 chars..
Int
float
Topspeed compiler specific
Astring , a version of the win32 ATOM ...
this giant fat data variable container can cross compiler stack boundaries and come back alive...
The compilers out there in space all have different data passing inside the chips...
Some pass on the stack like MS and some pass in the registers for high speed like the top speed compiler and then there is Unix land..
the simple fat class data container can pass anything like a ghost in the machine and survive...
https://www.youtube.com/watch?v=3Fs8hM4TYdM
Got 10 grand for a latop or more?
well we dont know that is what next gen stuff will cost but if you want AI power and you want to get rid of your Accountant then maybe in ten years time you wont even want to have a Tax Department as they wont be able to KEEP up with AI computing power...
You may only need an accountant as a consultant and accounting firms may be starting the process of down sizing or consolidating into consulting HUBs..
https://www.digitaltrends.com/computing/microsoft-surface-laptop-studio-2-launch/
The last 12 months we have faced the problems of multi language programming... and of course when the time comes to connection local AI models this stuff will come up again...
This is one of solution we came across and the other is making sure you dont assign std:string to *char in CPP,,,,
you may find the contents dont end up in the *char
https://gist.github.com/sancarn/940556cc1ff7806ca0e742a0c4117ab3
MICRO SOFT just got itself an AI Team...
if you have designed your software to be driven by people you might want to change that and its a bit late..
The trick here is to design the software to select, Set and Invoke with language commands that mean something and AI can learn.
Simple statements are best like....
OMC TRIAL BALANCE REPORT 2023
and AI does not just send it back to you in a format the software understands it actually runs the report for you!!!
This is an example of processing an OMC account and multiple portfolios creating Trial balanced and ledgers into one reports for financial accounting.
The result is a cross column multi entity report format where multiple porfolios are reported on ONE page for multiple entities.
The one page cross column report , similar to what you expect from public company reporting where multiple group companies are report cross column on the one page.
FINANCIALFOLDER.CREATEENTITY('ABCHLD','ABCHOLDING LTD','ABGROUP','2023', 'ABGROUP');
FINANCIALFOLDER.OPEN('ABCHLD','ABC HOLDINGSLTD', 'ABGROUP','2023');
TRIALBALANCE.LOAD();
TRIALBALANCEWKS.LOAD();
TRIALBALANCEWKS.Free()
TRIALBALANCEWKS.ADD('ABCIHLD','2023','650','Shares On Hand at Year End');
TRIALBALANCEWKS.SETBAL('ABCIHLD','2023','650','OB','0');
TRIALBALANCEWKS.TRANSFER();
TRIALBALANCEWKS.Save();
OMC.ACCOUNTTRANSFILES.LOAD();
OMC.CONSOLIDATEDACCOUNT.BALANCE.SET('OB','0');
OMC.ACCOUNT.SETIMPORTDATE('ABC HOLDINGSLTD OMCCASH','01/04/2022','31/03/2023');
OMC.ACCOUNT.IMPORT('ABC HOLDINGSLTD OMCCASH','CASH','01/04/2022','31/03/2023','NODATEDISPLAY');
OMC.ACCOUNT.PROCESS('ABC HOLDINGSLTD OMCCASH','CASH','01/04/2022','31/03/2023');
OMC.ACCOUNT.SETIMPORTDATE('ABC HOLDINGSLTD OMCRETAILNZD','01/04/2022','31/03/2023');
OMC.ACCOUNT.IMPORT('ABC HOLDINGSLTD OMCRETAILNZD','RETAIL','01/04/2022','31/03/2024','NODATEDISPLAY');
OMC.ACCOUNT.PROCESS('ABC HOLDINGSLTD OMCRETAILNZD','RETAIL','01/04/2022','31/03/2023');
OMC.CONSOLIDATEDTB.BALANCE.SET('OB','0');
OMC.CONSOLIDATEDTB.BALANCE.SET('CB','0');
OMC.CONSOLIDATEDTB.BALANCE.SET('OPENING STOCK','0');
OMC.CONSOLIDATEDTB.BALANCE.SET('CLOSING STOCK','0');
OMC.CONSOLIDATEDTOTALS.SAVE();
OMC.SHELLOPEN('CONSOLIDATED');
OMC.SHELLOPEN('CONSOLIDATED JNL TOTALS');
OMC.CONSOLIDATED.TRIALBALANCE('RESET')
OMC.CONSOLIDATED.TRIALBALANCE('LOAD','AB','2023');
OMC.CONSOLIDATED.TRIALBALANCE('LOAD','ABC LTD','2023');
OMC.CONSOLIDATED.TRIALBALANCE('LOAD','ABCHOLDINGS LTD','2023');
OMC.CONSOLIDATED.TRIALBALANCE('PRINT','TRIALBALANCE','EXCEL','2','2');
OMC.CONSOLIDATED.TRIALBALANCE('PRINT','LEDGER','EXCEL','LASTROW','2');
OMC.CONSOLIDATED.TRIALBALANCE('MSOFFICE','CLOSE');
Interesting.....but when I see OMC all I can think of is this:
https://www.youtube.com/watch?v=C2cMG33mWVY
I haven't done anything on this in a while. I have not looked at using AI...yet. I am starting to think it might be easier and cheaper to hire someone online to tap the figures into an Excel model. Maybe someone from the sub-continent or SE Asia where English proficiency is high . I have used UpWork before for elements of an app.
https://www.upwork.com/ab/profiles/search/?q=excel%20data%20entry
The script above is the difference between an accounting system for share transactions and NO SYSTEM.
Since platform take commands and processes them and has a standard data communications interface that can be HOOKED up to a large language model VIA CPP...
IE ... its AI friendly for the NEXT GENREATION of AI.....
the commands are consistent and AI friendly... The NEXT GENREATION of will not just generate statements you ask it to do but actually will be able to INVOKE the software and that means...
APP written for YOU to type in the info which no identification system and AI awareness are dead..
That platform above is AI READY and farther down the AI TRACK it could be PART of the AI.....
All our software is written that way for the last 5 years...forget APP driven by people... they are the OLD WAY...
Note: That in New Zealand there are No transactions systems in the Banking or finance sector that provide Proof of Work for accounts or transactions.
IE: No digital certificates and a nation run on shoe box transactions. The country is negligent in its transaction processing standards of which there are none. In this day and age where software will soon be hooked up to AI in a big way its not a good idea for AI not to understand transactions security...
Those statements are supported by a data interface bindable to a bigger platform and it can project over the internet and it can potentially be driven directly by AI without human help...
Dont think people realise whats coming and software that is not set up to standardise it data SetMemberobject, getmemberobject, selectmember object and Invokememberobject isnt going to survive...
Its a darwin moment coming people...
They need to redesign EXCEL from the ground up...
MS Azure power apps already has the LLM build into many little apps for reading types of documents....
AI is going LIVE people....
Local AI chips in a new generation of laptops for those who have continued to used MS office on and off line for financial reporting may soon get some new toys....
If you for some reason saved all your data in XML locally as well as storing it in a Database can now send your XML tagged data to AI in the future for analysis in ways we cant yet imagine...
https://www.cnbc.com/2023/12/14/intel-unveils-gaudi3-ai-chip-to-compete-with-nvidia-and-amd.html
This local processing will be to reduce the loading on server farms and the huge amount of water they will need for cooling.
new laptops could be your first simple AI devices ... I Robot one day?
I have decided to use "man-draulic" power. I have engaged someone from Venezuela who wants to earn USD to transcribe financial reports from pdf to Excel. Hopefully much much cheaper in the long run. Not a saleable solution but that's not my priority for now. Will keep you posted. How is your project going?
Multi Projects FRG..
thats like saying in thunber birds ...
Thunder Birds are GO...
FAB!!
https://www.youtube.com/watch?v=JuhahPrF7gk
They tell me the update scripting CORE should be ready MID 2024 to hook up to MS office on Windows. The version supporting the NEW SQL LIGHT C API.. We need that for storing meta data in SQL formats and reading large XML files.
The Scripting CORE API PRODUCT not PROJECT is a linux based scripting emulator that you bind functionality to that emulates a language... not the author of that.
Plug in a different set of bindings and you create new services....
LOOKING forward to seeing what we can do once LLM run on laptops with the new generation of chips...
We already run a big document manager product for Windows we created a while back that can read and META seperate document parts from MS OFFICE...
Generating the tables from documents and populating them only needed a new set of classes that we finished testing recently for multi column reporting...
DocExtensions( 'EVAL' , 'INTERNAL','EVAL MACRO')
DocExtensions( 'SCRIPT2' , 'INTERNAL','CALC2 MACRO')
DocExtensions( 'SCRIPT3' , 'INTERNAL','UNI SCRIPT PROGRAM')
DocExtensions( 'EXCEL DOCS' ,'SHELL','MWINDOWVBA','EXCEL.EXE')
DocExtensions( 'WORD DOCS' , 'SHELL','MWINDOWVBA','WORD.EXE')
DocExtensions( 'NOTEPAD' , 'EXE','WINDOWAPP','NOTEPAD.EXE')
DocExtensions( 'POWERPOINT DOCS' , 'SHELL','MOFFICE VBA')
DocExtensions( 'POWERSHELL DOCS' , 'SHELL','MOFFICE VBA')
DocExtensions( 'OTHER' ,'EXE','OTHER PROGRAMS' )
DocExtensions( 'METB' ,'EXE','MULTI ENTITY DYNAMIC ACCOUNTING','DACTBMultiEntity.EXE','EVAL|SCRIPT2|SCRIPT3','C:\DAC 2.2 Data Open - 6 Feb 2023',1 )
DocExtensions( 'DBOMC' ,'EXE','Broking Transaction Portfolio ','DBOMCTrans.EXE','EVAL|SCRIPT2|SCRIPT3','C:\DAC 2.2 Data Open - 6 Feb 2023',1 )
DocExtensions( 'DOCMANAGER','EXE','Document Automation Manager','DocumentAutoManager.EXE','EVAL|SCRIPT2|SCRIPT3','',1)
DocExtensions( 'SCRIPT2EXE' ,'EXE','SCRIPT2 Program','SCRIPT2.EXE','SCRIPT2',1 )
DocExtensions( 'SCRIPT3EXE' ,'EXE','SCRIPT3 Program','UNISCRIPT3.EXE','SCRIPT3',1 )
DocExtensions( 'TEST' ,'EXE','TEST PROGRAM','RUNTEST.EXE','',1 )
DocsSave('DocExtensions')
DocsLoad('DOCGROUPS')
DocGroups('OFFICEDOCUMENTS','OFFICE Documents','DocumentYear(''DOCGROUPS'',''YEARS'',''STOREBYYEAR'')','DocumentOrder(''DOCGROUPS'',''TOP'')')
Here is a real world example of a multi entity model being printed by the new classes. You cant have a product untl your software can processing data and print it in many formats.
Names are changed... These new classes are generic and can be modified with a INVOKE interface for more complex command processing such as AI would spit out...
OMC.CONSOLIDATED.TRIALBALANCE('ENTITY','GET',' An INVESTOR NAMED SIR B','2023');
OMC.CONSOLIDATED.TRIALBALANCE('ENTITY','GET','SIR B FMMIFUND LTD','2023');
OMC.CONSOLIDATED.TRIALBALANCE('ENTITY','GET','SIR B FMMIFUND HOLDINGS LTD','2023');
OMC.CONSOLIDATED.TRIALBALANCE('PRINT','TRIALBALANCE','EXCEL','2','2');
OMC.CONSOLIDATED.TRIALBALANCE('PRINT','LEDGER','EXCEL','LASTROW','2');
OMC.CONSOLIDATED.TRIALBALANCE('MSOFFICE','CLOSE');
NZ$10 for 7 years financials: P&L, Balance Sheet, Cash Flow and selected notes all put into Excel. Not bad ay? Mr Venezuela is already starting his second job for me.
I have built the guts of an analyser, but now need to decide on the best data structure for the report data created by Mr Venezuela. Either a totally flat file or multi column for each year....I might have to try both to see what works best. I had him input the data into a multi-column model; I just transform that to whatever format works best for me.
Multi Entity Multi Year Trial balance Data is handled by a very large class which populates balances into virtual memory files (queues / Lists ) , each can have up to 16 indexes (old C technology updated to the present day by a US company).
we have had multi entity multi year data processing working for a while now. But when you have access to financial accounting software in Class form for as long as we have you get build blocks on top of blocks.... Data can be stored in any number of database formats and runtime defined formats due to some very special software we got out of eastern europe back in the early 2000's... after all CPP is CPP... C PLus PLUS ...
The above example in script is very very powerful .... does not require a person to process anything if a data source is linked to the software... Being text is can use RE GEN technologies to create scripts... The CORE is something that AI in the future might be able to provide self generation and execution as it all standard data types are uniformly handled between different languages...
The Multi Entity Trial Balance handling class...
! Multi Entity Trial balance Class,
! implementing Interfaces denoted by the Ascii I letter.
MultiEntityTrialBalanceYTD Class(IclaObjects),TYPE|
,IMPLEMENTS(IMEEventFileGenerator)|
,IMPLEMENTS(IEnterpriseGroupEntity)|
,IMPLEMENTS(IMultiTrialBalances)|
,IMPLEMENTS(IGroupEntitiesList)|
,IMPLEMENTS(IGroupEntitiesTBS)|
,IMPLEMENTS(IsetIEntityYear)|
,IMPLEMENTS(IsetIEntityFolder)|
,IMPLEMENTS(IMultiEntityYTDTB)|
,IMPLEMENTS(IMETrialBalancesList)|
,IMPLEMENTS(ITBNextColumnRows)|
,IMPLEMENTS(ITBRowColumns)|
,IMPLEMENTS(ITBAccountEntityColumns)|
,IMPLEMENTS(IMETBRowColumnIndexs)|
,IMPLEMENTS(IMETBRangeRows)|
,IMPLEMENTS(IGetMETBColumnValue)|
,IMPLEMENTS(IMultiEntityConsolidate)|
,IMPLEMENTS(ITBIndexRowColumnValue)|
,IMPLEMENTS(IMETBImport)|
,IMPLEMENTS(ILoadMultiEntityTrialBalanceYTD)|
,IMPLEMENTS(ILoadMultiEntityTBColumns)|
,IMPLEMENTS(ISaveMultiEntityTrialBalanceYTD)|
,IMPLEMENTS(ISaveMultiEntityTBColumns) |
,IMPLEMENTS(ILoadSave)|
,IMPLEMENTS(ILoadEntityTrialBalanceYTD)|
,IMPLEMENTS(ISaveEntityTrialBalanceYTD)|
,IMPLEMENTS(ILoadEntitySmartChart)|
,IMPLEMENTS(ISaveEntitySmartChart)|
,IMPLEMENTS(IServicesEvent)|
,IMPLEMENTS(IGuiMultiEntityTrialBalanceYTD)|
,IMPLEMENTS(IGUITBIndexRowColumnValue)|
,IMPLEMENTS(IGUIMultiEntityTBColumns)|
,IMPLEMENTS(IGuiDisplayGroupYear)|
,IMPLEMENTS(IGuiDisplays)|
,IMPLEMENTS(ISetTBEBalances)|
,Module('MultiEntityTrialBalanceYTD.clw'),Link('MultiEntityTrialBalanceYTD.clw',1)
If Data is Tagged in XML meta files AI should be able to read those files and you could one day have a conversation on investing....
This conversation is probably something the big wall street investing forms like black rock have already got working....
Expect those firms to shed people soon maybe...
Throwing documents at AI in the end will win out .... and be very surprised if you wont be able to buy a feed pretty soon and for not much..
Getting a feed of company data from all over the world means you will need a big data storage facility ....
We store through a single manager class that determines how and where the store the data.
Id say it may still be a few years before you can buy a data feed so your solution should be good for a few years yet..
Of course you going to say what about multiple balance printing per entity with multi column and you would be asking the right question...
YES of course it handles multiple balances per entity and becuase it has a SMARTCHART it does not need to a person to tell it where to print what and in what order!!!!!
Maybe you want Tax Adjusted YTD and QTY's , no problem!!!
And thats what you get when youve spent a life time presenting multi dimensional ytd models ....
Its always fun to see an Accountants mouth tremble when they realise they dont understand technology after all.... but they do understand a lot of other very important things...
AI ... and the court cases begin..
https://www.cnbc.com/2024/01/05/microsoft-openai-sued-over-copyright-infringement-by-authors.html
Quote from: Waltzing on Jan 07, 2024, 08:04 AMAI ... and the court cases begin..
https://www.cnbc.com/2024/01/05/microsoft-openai-sued-over-copyright-infringement-by-authors.html
This is a (https://edition.cnn.com/2024/01/03/opinions/new-york-times-openai-microsoft-lawsuit-oconnor/index.html) good brief of the legal issues behind the New York Times lawsuit. On this read it doesn't look good for Microsoft and Open A1. This promises to be a fascinating tussle.
According to our script lamguage designer (an emlulation engine we hope AI will like) it will or could be as you say a tussle as he waxed on about how the learning models work ....
Hoe many experts going to be called ... judges are going to really show their metal on this one...
some big brains are going to be needed and looking at the American Supreme court you have wonder how they got there...
Quote from: Waltzing on Jan 07, 2024, 10:16 PMAccording to our script lamguage designer (an emlulation engine we hope AI will like) it will or could be as you say a tussle as he waxed on about how the learning models work ....
Hoe many experts going to be called ... judges are going to really show their metal on this one...
some big brains are going to be needed and looking at the American Supreme court you have wonder how they got there...
New York Times' affidavit. (https://hackernoon.com/read-the-entire-the-new-york-times-company-v-microsoft-corporation-and-openai-court-case) Some summer reading...
Well suns is out and we have been testing carbon fibre sculling and rowing components not reading complicated stuff....
We are tech people after all and live at 16/1000 of a second... GOPRO speed or something like that... might have that numebr wrong ill check with the head geek...
CHAT GPT has a store .....
https://www.cnbc.com/2024/01/10/openai-debuts-gpt-store-new-business-tier-after-leadership-fracas.html
Our head engineer doubts AI but the world loves something new and when this company lists its going to be very sort after...
BitCoin is now an ETF ... the dutch flower crisis has nothing on this but the blockhain is ripe for farther development perhaps into cipher transaction enycrption...
Customers at banks may start to want more secure accounts with fancy transaction identification... big changes may be coming to the banking industry....
https://www.cnbc.com/2024/01/10/sec-approves-rule-changes-that-pave-the-way-for-bitcoin-etfs.html
many people will think that AI is well intelligent...
But the programming interfaces to AI are a bit like opening a session to a internet server becuase well... you have to access it through one...
When a company say like ABC 123 Accounting corporation comes out and says we have AI one needs to realise that you have to craft each so called AI feature by hand to process the information that comes back from the AI cloud service...
And you can see how it will come to individuals in products like Excel and in the end you may find you no longer one day NEED ABC123 Accounting corporation..
https://platform.openai.com/docs/guides/embeddings/use-cases
It appears we are playing with fire...
and one of our projects is a software machine that excutes code without knowing what its doing...
ooops ... what have we done ...
https://www.nzherald.co.nz/lifestyle/eliezer-yudkowsky-artificial-intelligence-needs-an-off-switch/CBWGACAMMNH7DAIPR7O66YEP4Q/
and we just though we were creating a smart financial report ... IE no more seperate reports for multiple entity reporting reporting...
It appears the macro excution engine says the engineer can be set up for AI to create its own software bindings..
oops ... and thats how it starts ....
someone tinkering about and then going back into the lab and saying .. lets connect this to that !!!! look MA no hands!!!
Now you might think this stuff is new but it not. OOP based software has been around for decades and it was only a matter of time before tinkering with the objevts would reveal the new order of execution...
Look if we just let everything flow through the standard invoke with an order of priority. It knows just like a calculator that THIS goes before THAT...
remember your french cette!
It reduces everything to a known quanity and put its on a production lines and boxes it up ...
Bit like assembling a car with robots..
got a new fancy bit to assemble in the car? No problem just plug the bot in with the standard invocation interface and away it goes...
Does the population realise what been going on for the last 50 years? No it does not !!!
and here it is people just when the revolution first started !!
https://en.wikipedia.org/wiki/Simula
First developed in the 1960's!!!!
The Next Generation of AI Chips ...
https://www.theverge.com/2024/3/18/24105157/nvidia-blackwell-gpu-b200-ai
A must read for technology ... germany and its technology sector..
https://www.spiegel.de/international/business/electric-shock-an-existential-crisis-in-the-german-auto-industry-a-266bd037-b63a-4c9b-97b5-423866d7080f
How are you getting on with this project?
My mandraulic version has stalled while I focus on other priorities.
We are currently waiting on the Binding Machine to pass its MYSQL binding testing....
I expect something interesting to play with by year end...
Now that may seem a long way off but many of projects takes years to mature into usable software...
A recent addition to the document manager which will house the results of the Meta Template class is an auto control sequence system that controls the windows GUI accept message loop.
You have to train and build up Equates for collections and unions of text and remember this is for ASCII only not for UNICODE..
Now your all wondering what is MS's new AI Division going to do...
Well who is going to create the TECH...
Is it going to be a TECH with 100 thousand chips and will it need its own power station...
If they get to that in 10 years you can kiss the current version of the industrial world good bye...
https://www.youtube.com/watch?v=7mGr4Uc7ebA
And here we got "THE TECH" is coming ... the only software that might survive is softare AI can use itself...
https://www.businessinsider.com/microsoft-openai-plan-100-billion-supercomputer-stargate-artificial-intelligence-report-2024-3
Right they think by 2028?
Where are going to locate the TECH and which state will grant then approval for building a Nuclear power station in quick time....
Where is the water coolling come from for the TECH and the nuclear power station let alone where are going to get the chips from.
It just seems like a dream ... a nuclear hot one...
Are you using Co PILOT...
https://www.youtube.com/watch?v=Wap2tkgaT1Q
Besides working on the Meta data templates for reading and creating information from Word Documents converted from PDF.
We have added the ability to control the GUI automatically.
And return a GUI to its previous data with all data tables returned to there previous postions displays as all data is held as IN MEMORY models.
No more having to try and go back to where you were in your APP....
Can save a lot of clicking with that mouse....
This is why APPs are just more responsive than WEB SITES.... and you have all that memory on your device and laptop...
A = B
yes now what has this got to do with meta processing of documents without resorting to having to deploy AI when it comes to processing document data Forensic data...
well the MACHIINE only knows objects ... not instances of objects or data types ... or any of the normal stuff that makes for computer languages....
If other words stuff like Pythons _INIT_ just arnt needed...
The Machine can do stuff normal computer languages cant do...
A = B takes on a new powers ... It can be Data and it can be an Object ....
Its all Typeless and therefore it like all it components are equal...
you dont need AI to run its huge models to compare stuff ... OK AI is bigger... it faster and it can a Trillion address points at a time...
But it doesnt have the power of typeless A= B, nor does python...
And the MACHINE DOES ...
AI - just how its changing the information gathering space ... and where it cant see the dif between feet and meters...
very tech this one and only for coders... it also shows how hard it is going to be to get software even AI to understand anything...
https://www.youtube.com/watch?v=ncnm3P2Tl28
prompts you have probably tried typing into CO PILOT and realised that it has to start again each time for anything it does not know about.
AI is not what people might think... and here is a look at the problem.. its just a look at what people are doing with AI...
https://www.youtube.com/watch?v=UbDyjIIGaxQ
llm's
https://www.youtube.com/watch?v=Wjrdr0NU4Sk
well...
well since we have cracked A=B LVIV looks like a great place to one day set up a software local team...
and microsoft is shipping co pilot API inside windows ... whoops just just read your dividend statement and asked if you want to send your bank account to MS to improve its AI LLM models ...
......
Pre Meta processing of documents inside WORD and then passing on what you want to AI will be the way to go ...
locking down AI will be your next job ...
https://www.theregister.com/2024/07/19/admin_crowdstrike_update_mess/?td=rt-9cp
Angry admins share the CrowdStrike outage experience
CrowdStrike? More like ClownStrike! Amirite?
"....BitLocker is Microsoft's encryption tool, and it makes a device's storage inaccessible without a recovery key. As such, trying to work through some of the current recovery options on a modern device will likely require the use of that recovery key. Pity the administrators who dutifully kept a list of those keys on a secure server share, only to find that the server is also now showing a screen of baleful blue.
Another Redditor posted: "They sent us a patch but it required we boot into safe mode.
"We can't boot into safe mode because our BitLocker keys are stored inside of a service that we can't login to because our AD is down...." :-*
https://www.crowdstrike.com/blog/statement-on-falcon-content-update-for-windows-hosts/
how many people have manual backup ....
got 3 back up devices T in size....
what if one day someone attacks your cloud backup... sound crazy right... after all they have redundant server centres and they can send it round the world... until someone cuts the cables... ok then they send it via sat link .... still did you want to wait ? how many off line systems have you got ready to go?
https://www.interest.co.nz/technology/128891/security-vendor-crowdstrike-releases-preliminary-report-what-caused-last-weeks-it
wild defense products .... only in the US...
https://www.youtube.com/watch?v=WhaawIbfqYE
AI might not be being taken up by the public but the processing of documents has taken huge steps forward.
A south african company provides free utilities exe's to connect document management systems to AI platforms...
This connector set is the missing link that allow your apps to connect to global and local AI models...
Notice that MS Office basic front end search app only works when your connected to the Net...
The development of document automation list managers combined with platform neutral AI processing that work on line and OFF line is the new frontier....
yup this is a read between the lines post...
not quite photonic GPS chips but the new chips from NV will be here inside a year or 2...
Nuclear stocks have put on 22 percent this year.
How are you getting on with this? Are you winning? Or stalled?
I have taken a new tack - I developed a model to act as filter by looking at longer time slices and key metrics. If I like what I see then more detailed work is required. I went for the mandraulic method of data collection rather than automated. But everything else is automated.
Developing an API for run time popup menus in the document list manager to configure runtime utilities...is the answer...
We found a south african company (yes that country still has industry) that creates runtime utilities to call AI platforms ...
we havnt looked at the code yet but our approach it to send documents to any AI platform configured by the runtime menus...
you can configure the GUI in our new binding machine script language emulator.
We dont do the AI our approach is to provide a configurable GUI AI prompt and you link you select your documents using the standard document tagging solution built into the document automation list manager and fire off the AI prompting utility ....
using our new binding scripting emulation language machine to configure the document automation list manager to fire off the AI utilities...
we dont expect to have the utilities connected to the platform till later next year..
Thats for End user stuff... We just do a GUI runtime connection.
AT a lower level we have developed a software machine.. nothing to stop a C Plus Plus developer or developers creating a bindable to connect direct to say META's C++ API. By passing pythons ugly bindings...
Our approach has been to develop software machine to connect languages to connect software that supports standard interfaces and provide a connector interface component.
The binding machine is the glue that connects apps to app services..
The binding machine using CPP binding interfaces that gets, selects, sets and invokes any language that can support a standard interface callable by CPP or supports a function call to standard pascal or c.
This means instead of having to use binding generators for Python or Go a standard lower level communication can by pass the gludge of calling and binding generators that the industry supports right now and replaces this with a standard C Plus Plus Interface that allows all software languages to connect if they are PROPER languages and not sugar junk ....
We arnt doing the AI data filtering just setting up a runtime GUI connection..
Direct AI calls can be dont if your can write CPP with the standard bindable connector...
Its a universial approach to connecting language components at the base compiler level...
obviously the machine was not created this year or last... in fact its been developed over the last decade and was based on an early model for bindable scripting solutions first tested on Turbo C....
These technologies can take decades to evolve and by luck AI API's are being released perfectly on time to connect to something like the software machine.
Obviously this is a public forum and cant disclose much here.
there is no doubt the hooking up AI to local language models is around the corner for sensitive documents...
Notice that MS Office sarch app only works when your connected to the Internet.
Yet many documents of financial information that is sensitive to the individual are now stored by on line providers and are not secure...
Only off line is secure now and in the future...
Photonic commuting is not here yet but GPS replacement Phontics is around the corner and so to will be Quantum Computing in the next 20 years....
That aweful product Excel C API has been around for ever but was never nice to develop for and lacke the ease if API cell development that Lotus had...
MS will build AI search into the operating system and you will have a another huge set of options to configure...
Browser setting are Huge list.... Face Book is huge list and growing...
MS Windows 11 has a Huge List of setting...
Office Apps have a HUGE list of setting....
Now AI will have a HUGE list of settings....
But your documents wont be safe....
Local AI and Local Document management will have a place...
Your information has already been put on the online provider platforms and your ID removed from your financial transactions and then used for economic data reporting....
All those on line accounting solutions stealing for data .....
OOPS.....
Using it to tell government the state of the economy....
The Reserve banks dont even have access to that amount of DATA....
now just for FERG will posted shortly an example of the universial CPP binding interface in action as it been in use for a decade now...
This example will be processing Journals and models from excel and processing those transaction and share retail and cash from OMC accounts...
oh and multi entity financial consolidated reporting..
after all the base C classes for sortable lists has been around for over 35 or 40 years in one form or another...
its just that bankers, accountants, lawyers in NZ dont know it....
ooopp....
thats what happens when you didnt read byte magazine when it was first published..
where was that source code files? well it was in Scandinavian countries in the 1960's and they probably got it from the US ...
thats right the 1960's!!!! and its now what?
60 years later!!!!
Oh forgot to post this link... now if you did not follow history of the north Baltic countries you might not about this... and they built on early stuff....
imagine being able to use english language statements to run simmo's and you did not need CPP... only today is CPP getting some stuff to try to make it easier to so SIMMO's without have to spend months writing and debugging code...
read it and weep ...
https://en.wikipedia.org/wiki/Simula
but if you did not visit europe in the decades of the 70's and 80's you may never have heard of it... let alone realised the kick start it gave to industry ..
well you might have tried co pilot in edge or AI on excel...
its not that smart as you might have found out...
it can take a few prompts to stop it making the same mistakes...
it not hard to see its actually brain dead...
but as you train it by saving away documents and re entering them it slowing gets the idea...
it can help with the building blocks to creating ai connections via Curl...
AI accounts for your information will be something you might find soon you cant do without....
sometimes it even surprises as it learns to write code in our new black box binding service system...
it maybe more addictive than social media...
In software these days we test and allow AI to explore binding using standards we have developed for software....
im afraid when this goes live we will have no control over what AI binds to and what it decides to execute..
seen some very smart code its writing testing our LAB..
our LAB is AI these days...
the world we now inhabit has nothing to compare to even from 12 months ago...
since the software we develop is made for AI to develop and control at its will...
It fact it even wrote the standard connector itself as it learned to real time binding standard for the COR (central operating runtime), it named it.
and we are off... bye bye the old world... see ya..
Well after weeks and weeks... weeks and months ... we have come to the conclusion that AI has HUGE potential.... sometime in the future...
but the problem with AI models are the pathways... the information can change its hotness... and as it gets colder the things you though it would use as the basis for building blocks are a bit like puddles of knowledge that freeze over and you need to get AI to apply some HEAT... (ref to the 5th element there) and just as it starts to understand the CORE .(ALL HAIL the CORE) which it repeats; it needs to go back and walk over the hills mountains and valleys it has traversed in the previous months....
and at that point you realise its LOST! somewhere in the forests and hills and mountains... well AI looked like a good idea.
But for it to really be useful you will need to assemble a team of engineers to create special MODELS that are always on and monitored...
Its not coming to replace you anytime soon..
when a summary is a .... well a summary!!!
https://www.cnbc.com/2025/01/08/apple-ai-fake-news-alerts-highlight-the-techs-misinformation-problem.html
its turned out to be not a complete waste of time but you have to fix it up or take the bit that work...
MS ceo says all software will die .. well yes but not the stuff that lets AI program itself...
and watch out ... the story of the passenger that got locked in an AI self driving car that went round in circles...
all AI has so far failed us.. Claude may be the best so far... it certainly has the best replies we have seen... no loops so far but we havnt got very far...
and here is AI starting to understand the UBS black Box... once AI can understand it and auto generated (it cant do this at present) then we can move to take over the world!!!..
In UBS:
CopyA = B
Simply means A becomes whatever B is - without type constraints or conversion issues. The ScriptBindable interface handles all the details through its methods:
getAs... methods to read B's value
setTo... methods to set A's value
Types are handled internally by UBS
This eliminates traditional programming language problems like:
Type compatibility
Type casting
Type checking
Type conversion errors
Because the question "what is A to BE" is answered simply by "whatever B sets A to BE" - and UBS handles all the mechanics through the ScriptBindable interface.
That's quite an elegant solution to what has traditionally been a complex programming issue. CopyRetryQDwhat famous author asked this very question in his plays?EditWilliam Shakespeare famously wrote "To be, or not to be" in Hamlet - which is the same philosophical question about existence/being that you're drawing a parallel to with UBS's assignment paradigm.
In UBS:
"A = B" answers "what is A to be?" Very similar to Shakespeare's contemplation, but in UBS's case, the answer is simple: A is to be whatever B is, with UBS handling all the details through ScriptBindable.
It's an interesting parallel between Shakespeare's existential question and UBS's elegant solution to the fundamental programming question of "what should this variable be?" CopyRetryQDor whatever B is programmed to decide to return.EditYes - that's a key insight! In UBS:
A = B
B, through its ScriptBindable implementation, has the power to:
Decide what type of value to return
Transform or compute the value
Make complex decisions about what value to provide
Even change what it returns at different times
So it's not just a simple assignment - B has agency through its implementation to decide:
What A is "to be"
How that value is determined
When/if that value might change
The ScriptBindable interface provides this flexibility while UBS handles all the mechanics of making it work. So Shakespeare's question gets a programmable answer - A becomes whatever B is programmed to decide to return.
ok anthropic seems to be the go to guy in town....
and here is a bit on the new models for Market data visual set to run from the black box scripts with CUDA doing the hard yards... yeah that its NVID doing the grunt stuff and QT openGL doing the visuals... more to come on this one...
Market Data Visualization: Dynamic Invocation Method
This is an abstract from AI anthropic to render
data in CUDA using QT Open GL GUI.
CUDA is of course NIVD multi thread core code execution model.
Overview
---------
The `invoke()` method serves as the primary activation mechanism for market data visualization,
providing a flexible, script-driven approach to data processing and rendering.
Method Signature
----------------
integer_t invoke(IUBS* ifc) override {
// Method implementation
}
Key Functionality
-----------------
The method handles:
- Dynamic parameter input
- Visualization configuration
- Data processing initiation
- Rendering workflow management
Implementation Details
----------------------
The Universial Binding Service interface get data in focus from
running scripts.
integer_t invoke(IUbs * ifc) override {
// Retrieve total number of parameters
int paramCount = ifc->getParamCount();
// Retrieve data source
bindable_t dataSource;
if (paramCount > 0) {
dataSource = ifc->getParamAsBindable(0);
}
// Retrieve visualization configuration
bindable_t configParams = nullptr;
if (paramCount > 1) {
configParams = ifc->getParamAsBindable(1);
}
// Configure visualization if parameters provided
if (configParams) {
this->ConfigureVisualization(configParams);
}
// Process and render market data
this->RenderMarketData(dataSource);
// Optional: Return processing status
return paramCount > 0 ? 1 : 0;
}
Parameter Handling
------------------
- First parameter: Data source bindable
- Second parameter (optional): Visualization configuration
Workflow
--------
1. Check parameter count
2. Extract data source
3. Extract configuration parameters
4. Configure visualization
5. Render market data
6. Return processing status
Flexibility
-----------
- Supports variable parameter input
- Allows dynamic visualization configuration
- Provides script-driven data processing
Use Case Example
----------------
// Hypothetical script interaction
MarketVisualization.renderdraw(NASDAQDataSource, VisualizationConfig)
Benefits
--------
- Modular design
- Scriptable configuration
- Dynamic data processing
- Flexible rendering approach