Reflecting on Wall Street's 'October Nightmare'—Buffett's contrarian philosophy that 'the market often sets ridiculous prices,' and how to protect your capital and time in …
When stock prices fall, we look at the expressions of others before we look at the companies themselves. It is similar when a new AI service becomes a hot topic. Before we verify what it can actually do, we feel like we are being left behind.
The NHK program ‘Century of Images: Butterfly Effect’ episode ‘Wall Street: October Nightmare’ serves as an entry point for considering these emotions. The program introduction contrasts Alan Greenspan, who was influenced by the philosophy of Ayn Rand, with Warren Buffett, who invested by discerning the value of companies. This article is a reflection based on the published program introduction and primary sources. [Program Introduction](https://mantan-web.jp/article/20261023dog00m200098000a.html)
In October 1929, stock prices on Wall Street collapsed. The Dow Jones Industrial Average fell by approximately 13% on the 28th and about 12% the following day on the 29th. However, the Great Depression as a whole cannot be explained by these two days alone. The crisis spread throughout the financial system and the real economy. [Federal Reserve History Explanation](https://www.federalreservehistory.org/essays/stock-market-crash-of-1929)
I do not intend to divide these two men into simple categories of good and evil here. Greenspan himself raised the issue of ‘irrational exuberance’ in the market in 1996. Trusting the market system and believing that the stock price on any given day is correct are two different issues. [FRB 1996 Speech](https://www.federalreserve.gov/boarddocs/speeches/1996/19961205.htm)
Buffett’s idea that ‘the market often sets ridiculous prices’ does not mean that you can win simply by trading in the opposite direction of others. In this article, I use this Japanese phrase as a summary of his philosophy.
The ‘Mr. Market’ concept inherited from his mentor, Benjamin Graham, is a partner who offers different prices every day. You are under no obligation to trade at those prices. The buyer has the right to pass. [Berkshire 1987 Shareholder Letter](https://www.berkshirehathaway.com/letters/1987.html)
So, what should you buy?
One option is companies that are too cheap relative to their assets. For example, consider a company with a market capitalization of 8 billion yen, cash and equivalents of 15 billion yen, and interest-bearing debt of 3 billion yen. The simple net cash is 12 billion yen. However, if 2 billion yen is needed for daily operations, the surplus available for free use is 10 billion yen. Even this exceeds the price of the entire company.
Of course, this is a hypothetical example, and there is no guarantee that profits will continue or that cash will be returned to shareholders. If a company is losing money every year or continues to use cash for unprofitable acquisitions, there is a reason why it looks cheap.
When researching Japanese stocks, I first narrow down candidates by looking for ‘net cash divided by market capitalization greater than 1’ and ‘PBR less than 1,’ and then I check for losses, capital constraints, capital expenditures, and shareholder returns. I do not decide to buy based on low PBR alone. This numerical value is my own rule for consideration, not Buffett’s personal buying criteria.
Another way is to buy companies with earning power for less than their value.
The profit from the resolution of asset undervaluation is different from the profit generated by a company growing over many years. If you buy something worth 100 for 50, it doubles when it returns to 100. On the other hand, achieving returns of several dozen times requires additional conditions such as profit growth, reinvestment, and a long period of time.
Buffett himself has written not only about his success in asset-based value investing in his youth but also about its limitations. His failure with a textile company he bought cheaply and the process of shifting his focus to excellent businesses following Munger’s advice are important materials for not making contrarianism look like a panacea. [Berkshire 2014 Shareholder Letter](https://www.berkshirehathaway.com/letters/2014ltr.pdf)
I want to apply this distinction to AI and semiconductor stocks as well.
Even assuming stock prices have plummeted, I do not group CoreWeave, ASML, NVIDIA, and TSMC together just because they are ‘AI stocks.’ The risks to look at are different for businesses that increase supply capacity through borrowing and capital investment versus businesses that generate cash from competitive advantages.
For CoreWeave’s April-June 2026 quarter, against revenue of approximately 2.575 billion dollars, net interest expense was approximately 640 million dollars, and operating loss was approximately 49 million dollars. You cannot judge the value remaining for shareholders based on growth rate alone. [CoreWeave Official Earnings Report](https://investors.coreweave.com/news/news-details/2026/CoreWeave-Reports-Strong-Second-Quarter-2026-Results/default.aspx)
For ASML, I look at orders and the competitiveness of their equipment; for NVIDIA, their competitiveness including software and the continued investment of their customers; and for TSMC, their manufacturing technology and the recovery of capital investment. On top of that, for all three companies, I check their debt, cash, future payments, and purchase price. Even strong companies can have purchase prices that are too high.
This article does not certify the October 2026 crash or the current undervaluation of these stocks. Actual purchases require calculations that combine the stock price at the time of acquisition with the latest disclosures.
The same attitude can be used for your X recommendation feed.
‘Information leaks,’ ‘Monthly income with AI,’ ‘The current environment is already old.’ Fear and expectation can move not only your investment capital but also the time you spend researching. Therefore, finish your daily information screening in 3 minutes.
In the first minute, check if it relates to the services you use or your work this week. In the next minute, check if there are official announcements, terms of service, and reproduction steps. In the final minute, decide if it is a necessary defense for today, a candidate to try later, or something to discard. The maximum number of candidates for verification is one. It is fine to end with zero. Do not use paid AI for daily monitoring.
Signs of serious damage are a different matter. If there are unauthorized charges or logins, do not try to finish within three minutes; deal with it through official channels.
GMO’s ‘Onamae.com Desktop Cloud’ has officially announced unauthorized access, the exfiltration of customer information, and reports of unauthorized transactions from users. That is precisely why you should verify the target service. Do not make leaps like ‘all cards related to GMO have been leaked.’ [GMO Official Announcement, October 24, 2026](https://internet.gmo/en/news/article/249/)
For card protection, enable usage notifications and check any unrecognized charges through the issuer’s official app or the contact number on the back of the card. Do not log in via links in viral posts or SMS. If unauthorized use is suspected, prompt contact is necessary. [Japan Credit Association Guidelines](https://www.j-credit.or.jp/customer/education/pdf/act_book_06.pdf)
Apply the same verification to new AI tools.
The official Claude Startups page checked on October 27, 2026, stated that the current recruitment quota was full and new applications were not being accepted. This is separate from your API balance running out or reaching usage limits. It is explained that programs for existing members will continue. [Claude Startups Official](https://claude.com/programs/startups)
When trying Muse×Codex×Claude, give each the same small task one at a time and compare the time to completion, rework required, payment amount, and necessary permissions. What is described here is the verification procedure, not a ranking of their performance or revenue track records. Confirm the provider and billing terms before handing over API keys or payment information.
Newness alone does not guarantee that your current work will become faster. Success stories in advertisements alone cannot calculate your own profits. Just like stock prices, tools have a value to you and a cost you must pay.
Before buying, calculate what will remain. Before trying, decide what you can reduce. And on days when the conditions are not met, pass.
The next price and the next new tool will be presented again.
※ This is not a recommendation to purchase specific stocks. The investment numerical examples and information selection criteria are designed for explanation and operational purposes.