Capital de Tokens
What lies behind the headline
A story distributed by MSN and published by Futurism frames an Egan-Jones warning as a possible destruction of the economy by AI. The coverage attributes to the agency a scenario of pressure on professional services, venture capital returns, and housing prices tied to office jobs.
Semafor also reports on the commentary, titled It’s Over. The original link has been identified, but its body could not be retrieved during this review: specific claims here are attributed to news coverage rather than a complete reading of the document.
The useful question is how the value of expert work would change if producing some of its outputs becomes much cheaper.
The mechanism: from hours sold to outcomes delivered
The following example is hypothetical. A consultancy prices a project requiring 100 hours. A tool reduces draft preparation to 40 hours while preserving final quality after human review.
That can open different paths:
| Company decision | Possible outcome |
|---|---|
| Keep the price and sell the outcome | Higher margin, if the client accepts that price |
| Cut the price to compete | Client savings and pressure on competitors |
| Take on more projects | More output with the same team |
| Reduce staff | Lower employment if additional demand fails to compensate |
None of those paths follows automatically from the model. Contracts, competition, demand, and implementation costs matter. Expertise remains necessary to define the problem, check the result, and take responsibility.
A cheaper startup still needs capital
As an economic interpretation, reducing the cost of developing a prototype could let a small team negotiate funding with less urgency. That can change who captures value.
But a prototype is only one part of a business. Distribution, customer support, infrastructure, security, and compliance may still require substantial resources. The total cost of delivering the service, including review and errors, should be assessed before assuming a company needs very little money.
How the effect could reach housing
The connection is indirect and conditional: if a local market loses well-paid jobs, households’ ability to buy or rent homes could weaken. Turning that possibility into a price forecast requires data on income, migration, housing supply, and credit conditions.
An improvement in AI capabilities therefore cannot, by itself, quantify a housing decline or establish its timing.
Technological exposure and lost employment are different measures
The 2025 ILO–NASK study estimates that a quarter of global employment is potentially exposed to generative AI. It clarifies that exposure does not mean observed layoffs and considers job transformation more likely than complete replacement.
That study provides context, but it does not predetermine what later models or individual companies will do.
Which signals to watch
Assessing the scenario means tracking outcomes: productivity after review, prices charged, hiring, wages, and demand. An automation announcement provides less evidence than a sustained change in those measures.
For professionals and businesses, a practical response is to identify specific tasks, test tools with quality controls, and measure the full benefit. Verification skills, customer knowledge, and accountability for decisions may become more valuable when drafts are abundant.
The warning deserves analysis; certainty about collapse requires evidence the headline does not provide.