What Survives

State the bear case fairly, because it is the right question. If every desk can read everything, all the time, with perfect memory, then information advantage goes to zero and this entire series is an obituary.

It is a good argument. It is also wrong, and the reason it is wrong is the oldest lesson in the industry: scarcity migrates. It never disappears. When attention becomes abundant, the edge moves up the stack. What remains, in order of ascending value:

The questions

When alternative data arrived, everyone with a budget bought the same feeds, and the edge went to whoever asked the data better questions. Agents repeat that lesson at a higher level. The prompt is the new scuttlebutt. Taste, hypothesis, the instinct for which footnote matters and which metric lies, becomes more valuable, not less, because asking is now the only expensive part. A workforce makes asking cheap and answering free. It says nothing about what to ask. That was always the rare thing. It is simply visible now.

Variant perception

Being right when consensus is wrong requires a thesis, and a thesis is not a summary. Agents can show you what the world said. They cannot tell you what the world is wrong about. Consensus is, by construction, the average of everyone's attention; when everyone's attention gets better, consensus gets better too, and the payoff for departing from it correctly rises. That remains a human sport. It is where the money is.

Judgment under uncertainty

Sizing. Conviction. The drawdown held or cut. The moment to admit error, which is the most expensive moment in the business. Agents inform risk. Humans own it, legally, financially, and psychologically. The last layer of the stack, the one that signs, was never going anywhere. What changes is that it finally gets to spend its whole day on the part only it can do.

Process

Renaissance read the same markets as everyone else for decades and won anyway, because the moat was never the data. It was the cleaning, the testing, the discipline: the process. The same holds here, amplified. What your workforce is tuned to notice, and just as important what it is tuned to ignore, becomes the proprietary asset. Two desks with identical agents and different processes will not get identical results. They will not be close.

The dispersion claim

Here is the counterintuitive part. Agents are leverage, and leverage amplifies whatever it levers. A sharp process with a workforce pulls further ahead. A sloppy process with a workforce produces confident nonsense at scale and finds out faster. The likely future is not a flattened industry. It is a more dispersed one, where the gap between good and mediocre widens precisely because the tools are equal. Equal tools have never produced equal investors. They have only ever removed the excuses.

The treadmill

None of this is optional for long. When your competitors run workforces, coverage becomes table stakes, then marketing, then a line in the allocator's questionnaire. Standing still means falling behind at machine speed. This is the uncomfortable half of every repricing: the terminal was optional too, until it wasn't.

The good news sits at the top of the stack once the dust settles. Judgment, the thing this industry has always claimed to prize, becomes the only scarce input left. For the first time in the history of the business, it gets the full working day.

The objections deserve more than a wave of the hand. The next essay takes them seriously.