Boy, do I have bad news for you.

Edit: To be clear, as a sibling post said, the basic arithmetic is easy enough. It's the tax opinion stuff that is absolutely not deterministic. If your situation is even moderately complex, there's a vast number of ways to describe your deductions, each with different tax implications and multi-year requirements. I'm not talking about being Jeff Bezos, either. Is your spouse an independent contractor? Do you own a home? Do you have stock options? Do you have a home office? These alone are enough to make some pretty creative reporting situations.

Ultimately, taxes is just filling out a spreadsheet and doing basic math... but the hard part of taxes is understanding how to fill out that spreadsheet correctly. Doing that requires answering several questions that many tax filers may simply not have the background to understand--I'm always struck whenever answering the question about "do you need to correct your W-2?" is how would I know when the answer is "yes." I can see how AI could be helpful here... at least were not AI plagued with hallucinations.

That said, Intuit's actual business model is convincing millions of people that their taxes are so complicated they need to spend $60 on a program that is just copy-pasting numbers from one document to another.

I asked an LLM (Gemini) about a calculated field in my taxes that was wrong but I couldn’t figure out why and every time it tried to tell me something like “It’s a common glitch for tax software to calculate like this.”

When I did figure out what was wrong and asked if that made sense, it told me I was absolutely right though.

I think people are lucky the IRS fired all their employees this day and age so this work isn’t getting checked as much.

This. An important property of tax calculations is being consistently reproducible. In the context of an audit, the calculations don’t need to be accurate per se just precise.

I have a formative story where a large industrial company was counting things for tax purposes using my software. Given the same data and same binary, the count was off by one depending on if it was running on Intel or AMD CPUs. They didn’t care so much what the count was, they just needed it to be the same value at all times in all places so they could defend it in an audit.

The inability to consistently reproduce the value was a bigger liability for tax purposes than small variances in accuracy that fell out of the process.

Why? What's wrong with some non determism if Intuit foots any mistakes? I am all for AI auto filling taxes where the filer offering takes the burden.

Intuit has a pretty broad financial software portfolio, not just a tax company.

Also, yes the actual arithmetic at the end should be handled by deterministic code. I doubt anyone, including Intuit, thinks otherwise. But there's a ton of uses for LLMs before you get to 2+2 = 4, explaining concepts, document extraction, understanding the full financial picture, etc.

Kind of feels like you're criticizing a cartoonish idea of AI's place in their products.