Stripe Founder Collison: Personal Agents Will Make Consumers More Rational in Interactions with Companies

Stripe founder and CEO Patrick Collison stated that personal agents will make consumers more rational when interacting with companies. Companies, in a sense, are primitive superintelligences that can leverage amortized cognitive surplus to counteract some consumers' attention deficits, information scarcity, and irrationality, benefiting from it, such as the significant markup on insurance add-ons at the end of booking processes compared to regular policies. The concept of bounded rationality proposed by Herbert Simon is often exploited by businesses, and personal agents are expected to change this situation.

Collison pointed out that certain price discrimination tactics will lose effectiveness if each agent meticulously searches for coupons, raising questions about how the coupon mechanism operates. Customer cross-subsidies will also change; currently, users who forget to cancel subscriptions subsidize those who remember to cancel, and the ultimate bearers of these changes in the overall economy remain unclear.

At the corporate strategy level, the relative importance of distribution versus product quality has been long debated. Better products may still be overlooked if no one knows about them, leading consumers to continue purchasing existing brands. Personal agents will act as structural subsidies for product quality; if individuals consistently invest at least 10 hours researching purchasing decisions, the market landscape will be significantly different, making a focus on improving product quality a more effective strategy.

Information related to product quality will become more important and influential. Currently, most people have limited ability to aggregate and integrate such data, but agents will have a continuous demand for these signals. It is widely accepted that data appreciates due to its role in model training, but as a cheaper form of intelligence, its value will also increase.

Collison posed an open question: Will agents reinforce the superstar effect (rational agents converging on the same product), or will they increase dispersion (by better uncovering differentiated preferences and breaking artificial clustering)? Many internet companies monetize through demand routing, which often relies on short attention spans, leading users to click on the first item they see, causing businesses to cater to ranking algorithms and pay various display fees, resulting in some inefficiency, although the willingness to pay itself is also a valuable signal to buyers.

The key assumption behind these judgments is that personal agents stand on the side of consumers. Collison believes this is likely to be a winning strategy, but the specific implications involve more questions. Overall, the market is a large-scale, parallel, decentralized, and highly imperfect value attribution machine, and agents will change the nature of backpropagation, increasing the returns for better products, and companies that create products they are proud of may find the environment more favorable to them.

Market Mechanism: Agents, as cognitive amplifiers on the consumer side, will weaken pricing and distribution models based on information asymmetry and attention arbitrage, with funds shifting from platforms reliant on display fees and short attention routing to entities that can provide high-quality signals and product quality data. Companies with high product quality will benefit from higher conversion rates and lower customer acquisition costs, while those relying on cross-subsidies and price discrimination will face pressure, and the value of demand routing and advertising space will be subject to repricing.

Source: Public Information