Second-level thinking is the process of understanding that the obvious narrative for a stock is already priced in, and extrapolating one or two levels further to estimate whether the price accurately reflects reality.
Forcing one’s self to think one step beyond the consensus is the only way for an individual investor to make returns superior to the market.
A contemporary example might look like the following:
As of writing Workday Inc. (WDAY) is down 37% year to date. While there are a swarm of reasons contributing to its underperformance, the most commonly quoted is the threat that agentic AI will remove the need for enterprise HR/Finance applications.
First Level Thinking:
- AI coding agents will be able to “vibe-code” user specific HR and Finance applications that can replace Workday.
- AI coding agents will allow incumbent software providers to produce identical software at lower cost, enticing enterprises away from legacy software solutions.
- The application layer of software will be no longer needed as agents will sit on top of raw data and complete processes ad-hoc.
Second Level Thinking:
- The economic value of cloud-based software providers is not just the end-state application and processes, but the ability to outsource IT infrastructure spend and maintenance costs. Proctor & Gamble does not want to bring software creation and management in-house because that is not the core competency of the business. It is more economically viable to pay for software, AI generated or not.
- Software is extraordinarily cheap in relation to the capability and scale of process complete it allows the customer to achieve. The value in the software is thus it’s reliability, rather than its capabilities. The risk of choosing or switching to a small AI-native software company over an established, resource rich software provider to provide mission critical digital infrastructure is higher than the savings associated with switching.
- What is actually the economic value of this. Customers only care that the software works, not how it is done. AI agentic software will only be more desirable if it allows more work to get done more reliably. Asking AI to solve a problem that occurs each week is directly opposed to the economic value of deterministic code in the first place in which the problem is solved once and then just rerun each time it occurs in the future. The reliably aspect is also in question. Any heavy user of AI will understand that it’s most frustrating limitation is the probabilistic nature of response. The same prompt often produces different outcomes, which while similar, are not identical. Predictability itself has value and deterministic code will always be more predictable than probabilistic neural networks.
This is not an endorsement to invest in Workday or other SAAS companies, but by thinking through the problem at the second-level, we can see the AI fear specifically may be over-valued in the market.
Some questions to ask to assist in second-level thinking:
- What is the range of likely outcomes?
- What does the consensus expect?
- How does my expectation differ from consensus?
- Is the consensus view baked into the current price?
- What happens to price if the consensus is right — or if I’m right instead?
Connections
In Risky Markets, Leaders Provide The Best Risk-Adjusted Returns
Link Explanation: The current situation in software (03-30-2026), is one of extreme fear. Many massive companies are down 30-50% from their highs, including hyper-scalers like Microsoft. While the AI impact discussed above may be true, there are other unanswered questions as to how AI may affect SAAS companies, such as their pricing models which are traditionally seat based. All-in-all, we are in a high risk situation. While that can be a good time to buy, the linked note above, reminds us that it is wise to stick to market leaders if choosing to invest. The established core is there for a reason and thus provides access to upside returns while minimizing downside risk.
Reference
🟢 The Most Important Thing Uncommon Sense for the Thoughtful Investor