Archives for May 2026

Get new articles by email:

Oblivious Investor offers a free newsletter providing tips on low-maintenance investing, tax planning, and retirement planning.

Join over 24,000 email subscribers:

Articles are published every Monday. You can unsubscribe at any time.

Financial Planning Roundup: New Book Recommendation

Over all the years I’ve been writing this blog, Ben Carlson is among the top few writers I’ve linked to most often (Jim Dahle and Christine Benz are probably the competitors).

Ben recently released a new book, which I just finished reading: Risk and Reward: How to handle market volatility and build long-term wealth. In short, it’s excellent. It would be on my short list of books to recommend to a new investor, and I think it’s a fun read for any experienced investor who enjoys good investment writing.

The book drills home the “…and now stick with the plan” lesson, in a variety of ways: stories, history, and (as you’d expect from Ben’s blog) a whole bunch of great charts for anybody who is a visual learner.

Other Recommended Reading

Thanks for reading!

Using AI for Tax Research

I recently had an email exchange with an Oblivious Investor reader about the (new for 2026) deduction of $1,000 ($2,000 if married filing jointly) for charitable contributions, for people who don’t itemize. The question was whether the deduction reduces adjusted gross income (AGI) or not.

The short answer is that, no, it does not reduce adjusted gross income.

But what’s perhaps more interesting is the conversation the reader was having at the same time, with Gemini AI, about the same topic. The reader had asked whether the deduction is an “above the line” or “below the line” deduction.

And the AI gave an answer that was frankly nonsensical. That’s likely because there’s conflicting writing on the internet about this topic, in part because the terms “above the line” and “below the line” are no longer sufficient to describe the categories of deductions that we have.

We used to have only:

  1. Adjustments to gross income (i.e., things that reduce adjusted gross income), which were known as above-the-line deductions.
  2. Itemized deductions, which were known as below-the-line deductions.

These days, we have a third category of deductions, which do not reduce adjusted gross income but which are not itemized deductions. Included in this category are:

  1. The $1,000 ($2,000 if MFJ) deduction in question for charitable contributions.
  2. The deduction for qualified business income (i.e., pass-through income).
  3. The deduction for qualified tip income.
  4. The deduction for qualified overtime income.
  5. The deduction for vehicle loan interest.
  6. The enhanced senior deduction.

Some people are calling these deductions “above the line” because they aren’t itemized. And some people are calling them “below the line” because they do not reduce adjusted gross income. And so human readers reading these different articles — as well as AI chatbots being trained on these articles — get confused.

When Gemini was asked to look specifically at the law itself and help interpret that, Gemini answered that it’s an above the line deduction, reducing AGI. Here’s Gemini’s explanation:

The specific language in Section 70424 of the Act that executes this change is as follows:

“(b) DEDUCTION ALLOWED IN COMPUTING ADJUSTED GROSS INCOME.—”

“Section 62(a) of the Internal Revenue Code of 1986 is amended by inserting after paragraph (21) the following new paragraph:”

“(22) CHARITABLE CONTRIBUTIONS OF INDIVIDUALS WHO DO NOT ELECT TO ITEMIZE DEDUCTIONS.— In the case of an individual who does not elect to itemize deductions for the taxable year, the deduction allowed by section 170(p).”

Gemini’s reasoning makes perfect sense — or it would, if that wording actually existed in the legislation. But it doesn’t. The entire wording Gemini is quoting is completely hallucinated. It doesn’t appear in the actual text of the law at all.

It’s remarkable though that the wording Gemini is providing is exactly what the law would say, if it were trying to make the deduction an adjustment to income. The AI is being very clever here — it just fundamentally misunderstands the assignment. It found wording from other pieces of legislation, in which a new deduction was made an adjustment to income. And it modified the wording exactly as needed, to fit the new context.

It’s exactly the sort of thing you’d want if you were using it to help write software code: finding something that worked in a very similar situation, and then modifying it in just the right way to fit this new situation.

But that’s precisely the opposite of what we want in tax research. We don’t want it to create new wording. We need it to specifically look at the wording that is there, understand that wording, and explain it to us.

AI chatbots can be very helpful, for a range of different tasks. But they can still get things wrong — and in ways in which no human ever would. (A human may very well answer the same tax question incorrectly, but no human in their right mind would completely fabricate a bunch of legislative wording as part of their answer.)

If you want to use AIs for tax research, there are a few things that can help improve the results:

  • In your permanent account settings, give it some information along the lines of “Correctness is always a critical concern. If you are ever unsure of anything, please let me know and/or ask questions. Please never give me an answer of which you are unsure.” (Note though that this can cut down on hallucinations but will not eliminate them, as the model can sometimes be confident but incorrect.)
  • In the actual prompt you give, do not only ask it to base its answer on the law, first ask it to find the text of the law itself. Confirm that it has read the text. And then ask your question. For example, start your conversation with: “Please look on the law.cornell.edu website for the IRC section about health savings accounts. Please find and read that.” And then go from there, with whatever questions you have about HSAs.
  • And for any points that are absolutely critical (i.e., points where it would be very problematic if the information you received was wrong), ask it specifically where it is finding the information in question. And then go confirm that for yourself. Acceptable sources include Internal Revenue Code sections, Treasury Regulations, Revenue Rulings, Revenue Procedures, and judicial decisions. IRS publications or articles anywhere (including IRS.gov) are not sufficient.

It’s critical to understand that AIs do not behave the same way that humans do. The situations in which AIs make mistakes, and the ways in which they make mistakes, are different than for humans. So if you’re going to rely on AI answers for critical topics, you have to be thorough/explicit about confirming the information on your own.

I often think of an AI chatbot as an employee that is: smart, extremely eager to please, hardworking, and literally an alien from another planet. Imagine that that’s who you have sitting on a stool next to you. It’s ready to work! But you can’t assume that it will understand you in the same way that a human would. They can not only get things wrong, they can sometimes fundamentally misunderstand the assignment in a way that no human would (as in the case above, in which it decided that making up tax law was what the user wanted).

For More Information, See My Related Book:

Book3Cover

Taxes Made Simple: Income Taxes Explained in 100 Pages or Less

Topics Covered in the Book:
  • The difference between deductions and credits,
  • Itemized deductions vs. the standard deduction,
  • Several money-saving deductions and credits and how to make sure you qualify for them,
  • Click here to see the full list.

A testimonial from a reader on Amazon:

"Very easy to read and is a perfect introduction for learning how to do your own taxes. Mike Piper does an excellent job of demystifying complex tax sections and he presents them in an enjoyable and easy to understand way. Highly recommended!"

Investing Blog Roundup: Household Finances as an Early Indicator of Cognitive Decline

A Planet Money podcast episode last week directed me to a paper released earlier this year: Dementia and Long-run Trajectories in Household Finances. The authors explain their findings in pretty plain English, so I’m just going to quote their work:

“Our results point to impaired financial decision-making beginning about six years prior to clinically recognizable dementia.”

“A growing body of work links cognitive ability to financial outcomes such as portfolio allocation, credit use, and susceptibility to fraud. […] Most recently, Mazzonna and Peracchi (2024) show that older adults are often unaware of their own cognitive decline, and that such misperception can lead to persistent financial mistakes. We extend this literature by showing that the economic consequences of cognitive decline appear well before the onset of dementia symptoms (or clinical diagnosis) and manifest in sustained losses in household wealth.”

“Our findings underscore that dementia-related financial risk should be viewed as a universal and forward-looking household risk, not just a concern for those who are already symptomatic or diagnosed. Because financial mistakes often appear before individuals or families become aware of any cognitive decline, relying solely on diagnosis or subjective awareness is unlikely to be sufficient. Safeguards that can be implemented well in advance—such as designating trusted financial contacts, simplifying financial arrangements, or allocating part of retirement wealth to products with guaranteed payouts—may help reduce exposure to losses from impaired financial decision-making.”

Their findings are exactly in line with the stories I keep hearing stories from readers: situations in which, conversationally, Mom/Dad/Uncle Steve still seems fine. No obvious red flags. And yet, they’re making financial decisions that make no sense or falling victim to scams that they would not have fallen for just a few years ago.

Of course, dementia (or other significant cognitive decline) does not happen to everybody. But we don’t know in advance to whom it will happen. And waiting until there are obvious personality-related symptoms before putting protective financial policies in place doesn’t work very well, because the financial problems often come first.

I think the authors suggestions are excellent: simplified portfolios (and simplified finances in general), designated trusted financial contacts (to allow the financial institutions to share information with that contact), options for guaranteed income (most notably: delaying Social Security). And I’ll add: adopting security practices that are resistant to scams/phishing (i.e., passkeys and security keys).

A key point here is that all of the above suggestions are generally good ideas regardless of the possibility of cognitive decline.

Other Recommended Reading

Thanks for reading!

For More Information, See My Related Book:

Book6FrontCoverTiltedBlue

Household Cybersecurity: A CPA’s Guide to Protecting Your Accounts and Personal Information

Topics Covered in the Book:
  • Why passkeys and physical security keys are the biggest security upgrade most households haven't made yet.
  • How to choose and safely use a password manager.
  • The account settings, device habits, and network policies that actually reduce risk.
  • How to build a personal household security checklist, so good habits don't get abandoned.
  • What to do if the worst happens: a clear, step-by-step response plan.
  • How to help aging parents or other vulnerable loved ones, as well as a research finding that may surprise you about when to start.
  • Click here to see the full list.

Keeping Vulnerable Loved Ones (and Yourself) Safe from Fraud

“Have your parents been scammed yet?” Jim Dahle asked me, in a brief one-on-one conversation at the recent White Coat Investor Conference.

The implication, of course, is that it’s a matter of when it happens rather than whether it happens.

At the conference, three other attendees happened to share stories with me of their parents being scammed. (And, to be clear, I was not walking around polling people on this topic.)

And we recently discussed the stats from the FBI showing explosive growth in the rate at which people age 60+ are becoming victims of cybercrime.

Frankly, we have a situation on our hands.

Technology keeps moving. If anything, it’s speeding up. But as humans, at some point, we slow down. That’s just the nature of things.

That’s a major part of the context for this whole discussion of cybersecurity. It’s not just about how we can keep ourselves safe, it’s also about how we can keep our elders and less tech-savvy loved ones safe. And the combination of tech advancements (which older people are less likely to have kept up with) plus the possibility of some degree of cognitive decline makes for a challenging situation.

And to be clear, when I mention cognitive decline here, I’m not referring to full-blown dementia (although that obviously does dramatically increase the likelihood of being victimized). I’m talking about the more mundane but common situation in which a person is still capable of functioning normally in day-to-day activities, yet their ability to promptly detect a fishy situation — their “bullshit radar,” if you will — has started to become somewhat less sensitive.

What follows are important steps you can take to make yourself safer. And after implementing these policies for yourself, you can help your vulnerable loved ones implement them as well.

An important thing to note here is that if there are multiple people in the household, everything is only as secure as the weakest link. Each person should be following these policies.

Communication Policies

If you receive any inbound communication (whether email, text, or phone call) that purports to be from a company or government agency with which you have any sort of account:

  • Do not reply.
  • Do not give them any information whatsoever.
  • Do not click on any links.

Essentially, don’t interact with inbound communications. If you need to interact with any organization with which you have an important account, make sure it is you reaching out to them, via a trusted means. That could be:

  • Calling the number on the back of your credit/debit card.
  • Typing the company’s URL directly into your browser.
  • Visiting their website via a previously-saved browser bookmark (or navigating there via your password manager).
  • Using an app on your phone (e.g., using the Bank of America mobile app, to contact Bank of America).

And the same thing goes for situations in which you’re contacted by (ostensibly) a loved one asking for money. Don’t engage with the inbound communication. Instead, reach out to the person in question via the contact information that you already have saved.

Credential Management Policies

To be clear, by far the largest risk for most individuals is getting tricked, rather than actually being hacked in some way. A major portion of cybercrime incidents are cases in which the victim:

  1. Accidentally gives a thief their multi-factor authentication code (e.g., over the phone or by text) because they thought the thief was somebody else, or
  2. Accidentally enters their login credentials on a fake website.

If you can eliminate those two possibilities, you have dramatically reduced the risk. Policies that work toward reducing or eliminating those possibilities include:

  • Use passkeys when a website allows you to do so.
  • Use security keys (i.e., YubiKey or similar) as the multi-factor authentication option when allowed to do so.
  • Use a password manager (so that when a website forces you to use password-based login rather than a passkey, your password manager will not populate the password if you end up on the wrong website).
  • Remove less-secure sign-in methods and account-recovery methods, to the extent you’re allowed to do so.

Essentially, you can’t get tricked into sharing the password with a malicious party if signing in with a password isn’t even an option (i.e., because the only sign-in option is to use a passkey). And you can’t get tricked into sharing the multi-factor authentication code with a malicious party if there is no such code (i.e., because the only accepted MFA option is a physical security key).

Other important policies regarding login credentials include:

  • Use strong passwords for your various accounts. (Your password manager should have a tool for generating such.)
  • Do not reuse passwords across different accounts.
  • Use strong passcodes for mobile devices and tablets.
  • Use an authenticator app as the multi-factor option (rather than SMS or email), if security keys are not an option and the website doesn’t allow passkeys.
  • Activate SIM-swap protection with your mobile carrier.
  • Use fake answers to “security” questions whenever an account asks you to set up such things (e.g., your third grade teacher’s last name was “i6c@vNShpJy$$9oV”), and store those answers in your password manager.

Other General Policies

Some of the other highest-impact security policies you can implement would be to:

  • Freeze credit at all the major credit bureaus, and implement a security freeze with ChexSystems.
  • Create an IRS Identity Protection PIN (and enroll to get a new one each year).
  • Take advantage of any security features your brokerage firms offer such as money transfer lock at Fidelity or ACATS lock at Vanguard.
  • Keep your operating system and apps up to date on all devices. (It’s unfortunate, but in many cases this will mean replacing devices at a point at which they’re still functional, in order to keep them on a current operating system.)

It’s Time to Upgrade Our Practices

To be clear, the policies mentioned in this article are not meant to be an exhaustive list. There are 100+ additional practices you could adopt, which would further improve your security. But the policies mentioned above are some of the highest-impact things you can do to keep yourself and/or your vulnerable loved ones safe. And if you implement all of the above, you (or your vulnerable loved one) will be a much harder target than most people.

Admittedly, for many people, this will mean learning to use some new tools (passkeys, security keys, password manager, authenticator app). But now’s the time. It’s not going to be any easier to start learning these things a few years from now.

And an important thing to note (and an important thing to mention, if you’re helping a loved one implement these practices) is that passkeys, security keys, and password managers all provide significant convenience benefits, relative to the alternatives. That is, in addition to being more secure:

  • Logging in via passkey is generally more convenient than logging in via password + MFA code.
  • Using a security key as the MFA method is generally more convenient than SMS, email, or authenticator app as the MFA method.
  • Using a password manager is more convenient than trying to remember your passwords, record them on paper, or record them manually in some other electronic format.

The thieves have better tools available to them than they did just a handful of years ago. We have better tools available to us as well. But we have to actually use them. If your (or your loved one’s) security practices are no stronger than they were 10+ years ago, it’s time to upgrade those practices.

For More Information, See My Related Book:

Book6FrontCoverTiltedBlue

Household Cybersecurity: A CPA’s Guide to Protecting Your Accounts and Personal Information

Topics Covered in the Book:
  • Why passkeys and physical security keys are the biggest security upgrade most households haven't made yet.
  • How to choose and safely use a password manager.
  • The account settings, device habits, and network policies that actually reduce risk.
  • How to build a personal household security checklist, so good habits don't get abandoned.
  • What to do if the worst happens: a clear, step-by-step response plan.
  • How to help aging parents or other vulnerable loved ones, as well as a research finding that may surprise you about when to start.
  • Click here to see the full list.
Disclaimer: By using this site, you explicitly agree to its Terms of Use and agree not to hold Simple Subjects, LLC or any of its members liable in any way for damages arising from decisions you make based on the information made available on this site. The information on this site is for informational and entertainment purposes only and does not constitute financial advice.

Copyright 2026 Simple Subjects, LLC - All rights reserved. To be clear: This means that, aside from small quotations, the material on this site may not be republished elsewhere without my express permission. Terms of Use and Privacy Policy

My Social Security calculator: Open Social Security