Is AI coming to Quicken?

What would you like this AI tool to do for you?

There are certain things that AI does very well, but you also have to be very aware of its limitations just like any tool.

What it does really well is "search and summarize" like you see your example. But it is also "very trusting". I once asked it about a question I had in a forum. I was looking for confirmation that what the person said was true. Since that person's answer was basically the only thing it could find on the subject it literally gave me his answer back slightly changed. AI states things in a way that people take as facts when they might not be. To its credit (CoPilot) it did give a link to its sources, and when I clicked on it, it took me to the original post in that forum.

With your financial data you can certainly see where this kind of searching of your Quicken data and stating things based on that data as it is absolutely correct might get "worrisome" like @Jon suggests.

On the other hand, different bugs that have been reported at least in Quicken Windows have proven that people can't always take what Quicken reports as absolute either.

Saw this about AI:

AI cannot fully automate the conversion of a 32-bit application to 64-bit; human expertise is essential for manual code adjustments. While AI can assist with tasks like code analysis and identifying issues, developers must manually address problems with data types, pointer arithmetic, and memory structures to ensure a successful conversion. The most reliable method for converting an application is to modify the source code for a 64-bit architecture and then recompile it, rather than trying to convert a compiled binary. How AI can help

• Code analysis: AI can help analyze the code to identify areas that need modification.
• Data type conversion: It can assist in identifying and converting data types that differ between 32-bit (ILP32) and 64-bit (LP64) architectures, such as int and long.
• Memory issues: AI can help pinpoint memory-related issues caused by the change in data size and pointer handling.
• Dependency resolution: It can assist in resolving dependencies on external libraries or processes.
• Testing and debugging: AI can help with the testing and debugging process to find and fix bugs after the conversion. 

Why human intervention is necessary

• Manual code adjustments: Many changes, especially those related to pointer arithmetic and memory structures, require human understanding of the application's specific logic to implement correctly.
• Complex data structures: Developers must manually adjust data structures that were designed with specific assumptions about data types and sizes.
• Nuanced issues: AI may not fully grasp the complex interactions and potential side effects of the conversion, so human oversight is needed to ensure the application behaves as expected.
• Reverse engineering is often required: Converting a compiled 32-bit binary directly is extremely difficult and often requires reverse engineering, a process that is also time-consuming and requires expert human knowledge. 

As I envision that AR would be optional, but it would be enormously helpful for building a budget. The hallucinations have been greatly reduced in the last year. With the right prompting and the right model, it's No longer really a thing.