@Slomo im watching a lot of this stuff nowadays lol, this one was one of the most clear explanations, to my small brain anyway

robert smith from vista..the middle part is useful

https://www.youtube.com/watch?v=KEX0q4yO4xE

@Solvetheriddle, that interview of Robert Smith was great – the perfect level for investors looking into the impacts of AI on their investments with a “mostly” nonpartisan and practical take on what to think about.

Lots to take away, but the big ones for me were:

• The distinction between Deterministic and Probabilistic and how that impacts AI use cases. Business processes are Deterministic – guarantee an outcome that is expected, but AI process use Probabilistic which is not guaranteed.
• Proprietary Vs Public data, the need to bring AI to your data rather than effectively giveaway your data by taking it to the AI. Companies that don’t have proprietary data or processes are most exposed to AI being used to spin up competitors.
• TAM expansion is possible with Agentic AI when used by a software that has proprietary data but is able to expand via AI to take economic rent (labor/services) of some other part of the market. WTC and it’s productivity (labor saving) processes is the live case study of this for me – they are now able to extract some of the value of their customers internal workforce beyond just the services provided by CargoWise!

These points I had inferred to some extent but was still unsure how to think about them, it just provided a lot more clarity around them.

Thank you for sharing @Solvetheriddle.

@Tom73 funny thing, there is nothing like very weak SPs, for companies and supporters to open up with info. I've found s/w a bit of a black box, not being a tech guy, i could see the numbers but little understanding of what was going on behind those. I've learned more about s/w in the last few months than in the previous (many) years. lol

i am adding to into the weakness

Very interesting @Solvetheriddle, and well summarized @Tom73.

I especially liked his insight that agents can eat services that are adjacent to software.

I've been thinking about different layers to the AI cake.

Seems that vibe coding is really around cost reduction - that the big play here. But also speed of dev, delivery, etc.

A lot of incumbents will no doubt be trying to tap into the apparent cost-out capacity of AI.

This will be a necessity for many as lower cost competitors will soon be knocking on the door - classic innovators dilemma.

A bigger impact could come from the revenue side - as Robert Smith discussed with TAM expansion / organic horizontal integration.

This likely requires a more agentic approach.

Apparently the biggest impact from AI integrations is to put it at the centre of your business (disrupt yourself) - or start that way with AI native businesses.

Otherwise you are tinkering at the edges, making old processes more efficient, adding AI to products, etc.

This could be a slow death unless you have genuine competitive advantages that AI cannot eat (proprietary data sets, etc).

@Solvetheriddle, many thanks for the Robert Smith link. This dude really talks pragmatic sense from the coalface of enterprise software development.

Robert put out another shorter summary 7 minute+ at: https://www.youtube.com/watch?v=-M_026Esvf0. It is good up to about 6 mins+, the rest is a pitch for Vista Equity Partners. Suggest listening to the longer interview version from @Solvetheriddle's post first, as it provides a lot of context, then watching the summary to wrap it all up.

I would add a few points to @Tom73's notes which were quite key for me:

1. The key is to bring the LLM models to the data, not the data to the LLM models

• Less than 1% of enterprise data actually exist in these foundational models and what they have been trained on - IBM Model
• If you have data in your infrastructure, in your company, in the industry, bringing the models and utilising this technology to enable you to use that more effectively is going to create massive EV
• If you leech your data into those foundational models, you have now diffused a lot of EV into other organisations 

2. The concept of "Sovereignity and dominion over datasets and workflows"

If a company has it, it has the right to dominate. If you don't have it, you don't have a right to exist as an enterprise software company

3. Enterprise software will end up in 1 of 3 categories: 

• Agentic Enterprise Solutions - platforms where agents execute real work, software becomes digital labour and where growth opportunity is the most dramatic
• Margin Expansion Companies - not every workflow needs execution, AI drives internal efficiency across coding, sales, customer support, enabling massive margin expansion
• Fail to Evolve - If all of your data and your workflows exist outside of your enterprise, you may not have a right to exist

4. 3 Qualities that Enterprise solutions in the AI era must have to win:

Context - transition to AI for most companies is hard, AI agents cannot adjudicate outcomes based on a companies workflow/policies/rules etc. AI models with access to only GENERAL intelligence cannot do these things because they lack context. Less than 1% of enterprise data is what these AI foundational models have been trained on. Software platforms have accumulated deep industry-specific knowledge, proprietary dynamic data, embedded workfows and rules. Context matters because generic AI models cannot reliably execute complex processes without it. And that is why the players that own the context have the long-term, sustainable advantage  

Trust - when an agent executes a regulated workflow, enterprises need to know who is accountable and how those outcomes can be explained and corrected. Regulated industries will not trust autonomous workflows to systems without these controls. Enterprises will trust long term partners that they have worked with to create and reinforce these essential workflows 

Ability to Operate at Scale - partners that Enterprises trust them to run their businesses are often the same partners best positioned to deploy agents at scale, not because they are established, but because they alredy operate inside the rules that enterprise live by

Despite the incumbents having these qualities, the biggest risk is speed to execution. 

Hi Folks

Following on from recent discussions around AI disruption & moats - came across this article in Bloomberg earlier this week I thought I'd share.

It takes a look at some US tax payers who have started using AI to lodge tax returns &/or help with assessing their eligibility for government grant/funding.

Few interesting take aways for me:

1) It isn't quite there yet - potentially error prone or giving in correct advice - often by way of unusual exemptions or other nuances of tax law that the AIs miss. Personaly I'd be nervous relying on an AI blindly - however for simpler tax returns - I guess the errors would be insignificant & cost savings probably outweigh any small tax savings missed.

2) Interesting noting Elon Musk has been using paid expert accountants to help train his xAI. Talk about doing yourself (or the next generation) out of a job!

I guess we'll be seeing cases soon where people have relied on AIs for preparing their tax to their detriment. Guess as always the buck will have to stop with the person regardless!

Cheers @Randy.

******************************************************************************

Claude and ChatGPT Tax Prep Is Here. Use Caution - Bloomberg

People Are Using Claude to Do Their Taxes (But Maybe They Shouldn’t)

Accountants warn of the potential for costly mistakes using ChatGPT and other AI tools.

Photo Illustration: Ben Mendelewicz; Photos: Getty Images

By Ben Steverman and Charlie Wells

March 18, 2026 at 9:30 PM GMT+11

Martijn Lancee has an accountant, but he still hates doing his taxes. They’re complicated: He and his wife have a mortgage on a home in the Bay Area, she owns a small business, and he makes money consulting for companies on artificial intelligence. Each year his tax adviser sends him a long list of tedious requests for information on business expenses, bank statements, bills. In February, Lancee had an idea. “Hey,” he typed into Claude Code, “how can you help me file my taxes?”

At the chatbot’s instruction, Lancee downloaded a bunch of his tax documents as PDFs on his computer and dropped them into a folder on the desktop for Claude. He asked it to create a spreadsheet organized with several tabs the way his accountant likes. Lancee checked the work, sent the file to his adviser and then spent the rest of his night playing Mario Kart with his kids. He didn’t tell his adviser the file was AI-generated. “I’m sure he noticed,” Lancee says, “because the output was a lot better than last year.”

Lancee. Photographer: Talia Herman for Bloomberg Businessweek

Interviews with more than a dozen tax professionals offer a clear takeaway: This is the year AI has come for tax prep. Artificial intelligence has finally improved to the point where it at least seems capable of navigating the US tax system as leading AI developers release more features for financial services. About a quarter of US workers say they plan to use AI to help file their taxes this year, up from 11% last year, according to a new survey by Adobe Inc. What better way to untangle the tax code—something so sprawling and complex that even the most experienced professionals can’t master every detail—than to use computer code?

The problem, according to tax pros, is that the chatbots keep messing up. They can give bad planning advice, sometimes informed by outdated rules, and tend to make mistakes reading digits off tax documents, especially on less standard forms such as a K-1, for so-called pass-through entities, or a 1099, for various kinds of payments. “Tax is incredibly nuanced,” says April Walker, a senior manager at the American Institute of Certified Public Accountants.

Recent tax changes, if anything, have added to contradictions nested in the code. “There are so many exceptions,” says Misty Erickson, content manager at the National Association of Tax Professionals. “You can even have an exception to the exception.”

Benjamin Cox, a business owner in Beaverton, Oregon, has tried using AI to prepare for conversations with his accountant. He asked ChatGPT whether he’s eligible for tax breaks such as the federal deductions for qualified business income or those for state and local taxes, known as SALT. The bot, according to Cox, was oblivious to the law approved last year making each provision more generous. “I kept on trying to correct it,” he says. “I’m sitting here arguing with AI, which is sort of like arguing with a pig. You feel stupid, but the pig’s just having fun.”

One technique software engineers use to get the best code out of AI is to deploy multiple chatbots to work on a project at once. Zhiyao Pei, a credit-analytics consultant, asked three AI models to offer their assessment of how California’s business-franchise tax would apply to his new business. “They actually had a fight,” he says. Google Gemini derided the work of another model as a hallucination. “That’s a typical AI mistake,” Gemini wrote. In the end, Pei just paid California the highest amount the chatbots suggested.

Joshua Youngblood, whose business specializes in assisting people who have run-ins with the IRS, sees an opportunity in all this. “I hate to be that person,” he says, but “it’s definitely going to create some issues.” One customer is already in trouble with the IRS for failing to report a cryptocurrency transaction after an AI tool erroneously told the client that income of less than $3,000 didn’t need to be reported. “ChatGPT is very good at telling you what you want to hear,” he says.

The most effective users of AI understand what it’s good at—and what it’s not. Large language models on their own aren’t skilled with numbers or adept at navigating a complex branching system like the tax code that’s supposed to lead to a single answer. A Loyola University Chicago study in November tested LLMs on a common tax question—how much of a home sale is exempt from the capital gains tax?—and found the bots failed two-thirds of the time. “It makes numbers up,” says Joshua Scott, whose Greensboro, North Carolina, firm specializes in the returns of health-care professionals.

But modern LLMs have capabilities that can help compensate for their shortcomings. Some of the accountants who complain most loudly about AI are the ones who use it the most—including Scott, who says he’s careful to check all its work. He also relies on AI to develop his internet marketing content. Nayo Carter-Gray, owner of 1st Step Accounting in Baltimore, uses it to walk clients through difficult concepts, such as “breaking down what depreciation is,” she says. Given the weaknesses of LLMs, Michael Geller, a partner at Gursey Schneider in Los Angeles, says, “I’m not afraid of it taking my job.” Still, he urges colleagues to adopt AI as “a new tool to do things quicker than before.”

Several software startups offer AI services for tax preparers but say they put careful limits around what the AI is allowed to do. It can be “dangerous,” says David Yue, chief executive officer of two-year-old Accordance, which helps preparers with research and other tasks. Tax returns contain a lot of numbers, and the IRS doesn’t have much tolerance for errors, says Dave Haase, founder of Juno, another tax software startup. “If you have a 99% accuracy rate,” he says, “you have a mistake on every tax return.”

TurboTax, the dominant tax filing software in the US, has been using AI to answer customers’ questions and help them navigate the app for years, but the implementation is deliberately limited, says Keela Robison, vice president for product management at parent company Intuit Inc. “As of right now, we do not feel that LLMs can process taxes, certainly not complicated taxes,” she says.

Some of the big AI companies recognize the potential. Felix Rieseberg, an engineering lead on Claude Code and Claude Cowork at Anthropic PBC, posted on X that he’d used Claude Cowork to read his own financial documents and fill out forms on a tax-filing site. The owner of X, Elon Musk, advertises his chatbot as a tax adviser. Musk’s xAI has been hiring accountants, seeking those with experience at Big Four accounting firms, to train its AI models, according to the company’s website.

The hype around AI and the annual feeling of dread associated with tax season together make this combination hard to resist. Andrew Pierno, a software entrepreneur in Los Angeles whose company has acquired multiple startups, has a complicated tax picture. Tax prep usually takes an entire weekend, “a three-day epic adventure where you’re unhappy and drinking a lot,” he says.

This year he’s been obsessed with OpenClaw, software that allows him to control his computer by sending messages to a bot through WhatsApp. Between games of pickleball in February, he texted his Claw asking it to build a tool that would pull together his personal and professional financial information into a file for his accountant. The whole thing took about three hours and was, in his words, “wonderful and fun.” He couldn’t immediately verify the level of fun his accountant experienced.