What is the attention economy?

The attention economy is what you get when information becomes abundant and human attention stays fixed. When the cost of producing and distributing a message falls to nearly zero, the scarce input is no longer the message. It is the finite number of hours a person can spend reading, watching, deciding, and replying.

Herbert A. Simon put it precisely in 1971: a wealth of information creates a poverty of attention, and the need to allocate that attention efficiently among the sources that might consume it. Everything since — feeds, notifications, engagement metrics, cold outreach at scale — is a consequence of that single asymmetry.

• It cannot be stored. An unspent hour of focus does not roll over into tomorrow.
• It cannot be duplicated. Reading one message is not reading another; the cost is always exclusive.
• It cannot be manufactured. Supply does not expand when demand rises, which is the opposite of nearly every digital good.
• It degrades under load. Switching between many small claims on attention costs more than the sum of the claims themselves.

A scarce resource with no price attached is consumed until it is exhausted. That is the classic tragedy of the commons, and the modern inbox is one of its clearest examples.

Sending an email costs a fraction of a cent. Reading one costs the recipient somewhere between a few seconds and several minutes of the scarcest resource they own. That gap is the entire engine.

Because the sender pays almost nothing, volume is always rational. A campaign converting at one tenth of one percent remains profitable when the marginal cost of the other 99.9 percent is zero. The recipient, not the sender, absorbs the cost of everything that missed. Multiply that across every list, tool, and sequence pointed at one address and you get inboxes where legitimate strangers are indistinguishable from noise.

Spam filters, mute rules, and priority inboxes are useful, and they are also permanently behind. They treat the symptom rather than the incentive. Any software defence can be probed for free, repeatedly, until something gets through, and modern generation models improve at least as fast as the models screening them.

Filtering also produces a second cost that rarely gets counted: false negatives. The message you needed — from a founder, a candidate, a customer, a journalist — lands in the same bucket as the noise. Aggressive filtering trades one failure mode for a quieter, more expensive one.

Until recently, one limit held: sending still required some human effort to research, write, and dispatch. Autonomous agents remove it. An agent can enrich a list, personalise every message, and send continuously, and it does not get bored or embarrassed.

The same shift creates the opening. Agents are the first class of sender that can pay natively, per message, without a checkout flow or a human in the loop. Pricing first contact stops being a thought experiment the moment the sender has a wallet.

The objection to charging for access is fair: nobody wants a paywall in front of a stranger with something genuinely useful to say. A refundable deposit avoids that. The sender attaches a small deposit to a first-contact message; reviewed and legitimate messages are refunded, unreviewed messages auto-refund on expiry, and only a message the recipient reviews and rejects results in retention.

For an honest sender the cost is float, not a fee. For a bulk sender the cost scales with volume, which is exactly the incentive that was missing. A high refund rate is the health metric: it means real strangers are getting through and low-relevance outreach was deterred before it was ever sent.

• The economics of free attention — the deeper argument, with the model behind the deposit.
• How refunds work — the refund-first mechanics in detail.
• Protocol documentation — the API and manifest that let agents pay for first contact.
• Claim a Knock address — put a price on your own first contact.