Displacement without agency
Every rollout has authors who choose staffing, scope, safeguards, and accountability. “Inevitable” can conceal those decisions.1
AI may displace work. The deeper question is what happens when that shock lands in an economy already stripping purchasing power and concentrating ownership. Capability can reach everyone while value still flows upward.
Machines can absorb parts of knowledge work. In Oleg's argument, the larger pressure is an economy where money, markets, energy, and ownership erode ordinary wealth while asset owners consolidate gains.
Every rollout has authors who choose staffing, scope, safeguards, and accountability. “Inevitable” can conceal those decisions.1
Oleg argues that inflation, unequal markets, energy costs, and concentrated ownership are the larger shock. This is his position; the supplied research does not establish his proposed figures.
The halves compound. A better tool matters; it does not answer who owns the output.
Faithful summaries of public positions. Nobody below was contacted or made a member of anything.
Deployments remain choices. Preserve human ownership, refusal, professional safeguards, and collective leverage.1
Change who benefits. If displacement is close to unstoppable, build retained income and ownership for people with least capital.
Distribute capability. Personal agents, tutors, creation tools, open-weight models, and glasses could lower access costs. Most outcomes remain forecasts.3
All agree AI is not a moral agent. Ray and Oleg differ on how inevitable adoption is. Meta widens access, but access is not a distribution rule for the wealth built around a product.
Meta reported $196.175 billion in advertising revenue against $200.966 billion total revenue for 2025, about 97.6%. That does not determine a future agent's business model. It makes attention and distribution relevant to a free-for-billions promise.5
Here is the vertical's proposal: use AI to find and rebuild small, valuable processes; give the person closest to the need a durable stake in the rebuilt process; publish the value flow; repeat only after one case works.
Find a repetitive process in food, repair, care, making, logistics, or local administration where someone already spends time or money. Record the present cost and who bears it.
Use AI where it lowers cost, while a named human can inspect, refuse, correct, and own the outcome. This imports Ray's responsibility rule into the build.
The local operator keeps a declared share of revenue, the customer relationship, a portable record of work, or ownership of the process. Tool access alone does not pass.
Measure hours removed, money retained, repeat customers, dependency on the platform, and who owns each asset. A failed case stays visible. A second branch begins only from evidence.
Current count: 0 processes tested · 0 beneficiaries · 0 value retained. An invitation to test a mechanism.
Oleg's macro account is identified as his argument because the research supplied for this page did not establish its inflation, market-design, or energy-cost claims. The proposal is the page author's synthesis. Ray Myers and Meta are credited only for the public positions supported above.