2025 Began Not With AGI, but With AI Moving From the Screen Into the Physical World
Why did 2025 begin not with AGI but with AI moving from screens into robots, vehicles, and physical infrastructure?
Evaluate OpenAI and Google as a physical and industrial system rather than only as software; the assessment must account for chips, energy, cloud capacity, manufacturing, and the cost of scaling.
What to watch for
Key takeaways
In the context of “AGI as a source of $100 billion in profit,” this criterion applies: in the OpenAI–Microsoft agreement the moment of reaching AGI is defined financially — as a system generating one hundred billion dollars in profit; the definition of the shared goal turns from scientific into contractual.
The discussion of “The OpenAI–Microsoft agreement and how it affects the future of AI” yields a practical test: OpenAI and Microsoft were revisiting the partnership terms and arguing over how to measure the moment the shared goal is reached — Microsoft needs the lab not to become an endless research project on its money, while OpenAI wants to preserve freedom and valuation.
The boundary of the “Elon Musk's lawsuit against OpenAI” case is defined by this point: Musk keeps suing OpenAI over commercialization, but the published agreement and the move of part of the structure to Delaware show the direction is set; the intrigue is what changes once Musk gains more power under Trump — the first quarter of 2025 will tell.
The “Humanoid robots: the start of a new technology era” scene leads to a working conclusion: the model gains a body, while the robot gains the ability to learn from data and instructions. This is where the promise of AI begins to affect warehouses, factories, and household labor.
The boundary of the “The trend of investing in loss-making companies” case is defined by this point: the market regularly writes losses off and comes back — Volkswagen wrote off billions in Rivian before new money flowed in again, and Mail.Ru sold HeadHunter for a modest sum only to bite its elbows when it listed on NASDAQ at a multibillion valuation.
The practical meaning of “Smart homes: new technologies and limitations” is that an Amazon courier can receive temporary access to a door or garage, making delivery independent of the owner's presence — but a trust question arises: who issues the key, how is the action recorded, and what happens if something goes wrong.
The practical meaning of “Autonomous vehicles” is that Waymo already carries passengers in San Francisco and Phoenix, including airport routes — not a futuristic video but a working service; Tesla promises a different scale, yet it still has to prove safety and obtain approvals.
In the context of “Technology breakthroughs in 2024 and trends for 2025,” this criterion applies: real progress is visible where AI receives permission to move a vehicle, open a door, work with a physical object, or make a decision inside infrastructure. At those points, technology stops being entertainment and begins changing the economy—together with responsibility.
What this episode is about
OpenAI and Microsoft are debating the future of their partnership, Google is returning to robotics, Waymo is expanding autonomous rides, and the smart home is gradually opening doors for couriers. The year's central trend is not another model, but the connection of AI with machines, infrastructure, and real actions.
At the end of 2024, the market was talking about AGI again, but the most important changes were not in the language. OpenAI and Microsoft were revisiting the terms of their partnership and arguing over how to measure the moment when their shared goal is achieved. Microsoft needs to prevent the lab from becoming an endless research project funded with its money, while OpenAI wants to preserve freedom and valuation.
Google DeepMind, meanwhile, is partnering with Apptronik and returning to humanoid robots after selling Boston Dynamics. The model gains a body, while the robot gains the ability to learn from data and instructions. This is where the promise of AI begins to affect warehouses, factories, and household labor.
Cars already show a more mature version of that transition. Waymo operates in San Francisco and Phoenix, including routes to the airport. This is not a futuristic video; it is a service people use. Tesla promises a different scale, but it still has to prove safety and obtain approvals.
The smart home is also developing through small but meaningful scenarios. An Amazon courier can receive temporary access to a door or garage, making delivery independent of the owner's presence. At the same time, a trust question appears: who issues the key, how is the action recorded, and what happens if something goes wrong?
The year 2025 should therefore not be evaluated by one AGI date. Real progress is visible where AI receives permission to move a vehicle, open a door, work with a physical object, or make a decision inside infrastructure. At those points, technology stops being entertainment and begins changing the economy—together with responsibility.
Together with responsibility; at those points, technology stops being entertainment and begins changing the economy.
Episode transcript
The episode is in Russian; below is an English reading guide to the transcript (the full EN transcript is a machine translation). Voice matching applied to 108 segments: 64 identified, 3 mixed, 21 probable, and 20 unresolved.
Read transcript on a separate page
Loading…