Skip to content
ChatGPT · OpenAI · GeminiEpisode 083 · 9 November 2025 · 57:40

AI Agents Are Not Ready to Replace Work Yet—but They Are Already Changing Prices, GDP, and the Rules of Online Stores

What to watch for

1Compare “How the AI influences US GDP growth” with “Andrei Carpathian: Disappointment in autonomous systems and NanoChat”: they provide different criteria for judging the same issue.
2Test the conclusion from “Delta Airlines and AI ticket prices: optimization or hidden discrimination” in your own use case—what actually changes in the process and what remains a promise.
3Before choosing a product or approach, record the constraint identified in “Amazon against Perplexity: agents who pretend to be human and future e-commerce”.
4Define the owner of the outcome and the quality metric for the situation described in “Deep Research, Ask and UX-bol: where the logic of large models breaks”.
Signals to track afterwards
Watch for actions by Amazon and Apple that confirm or challenge the episode’s central claims.
Compare new launches and policy changes with “Delta Airlines and AI ticket prices: optimization or hidden discrimination”: have access, quality, price, or constraints changed?
Check whether the scenario in “Deep Research, Ask and UX-bol: where the logic of large models breaks” becomes repeatable practice rather than a one-off demonstration.
Most useful for
EntrepreneursExecutives and managersAI usersProduct teamsProcess ownersDevelopers

Key takeaways

00:00AI Agents Are Not Ready to Replace Work Yet—but They Are Already Changing Prices, GDP, and the Rules of Online Stores

In the context of “AI Agents Are Not Ready to Replace Work Yet—but They Are Already Changing Prices, GDP, and,” this criterion applies: the important signal is not one funding number: the next round, available runway, and closure rate show whether a company can survive the new cost of capital.

04:01The market tests it through use: how prohibitions affect the market

For the “How prohibitions affect the market” scene, the decisive point is this: the important signal is not one funding number: the next round, available runway, and closure rate show whether a company can survive the new cost of capital.

05:23The boundary between value and constraint: who is responsible for AI errors: Comparison with

The “Who is responsible for AI errors: Comparison with books, courses and offline experts” issue should be assessed with one constraint in mind: the key test is enforceability: who must prevent the risk, who records the violation, and who is accountable for the consequences.

10:35Who owns the outcome: new button Update in ChatGPT: why did it

The “New button Update in ChatGPT: why did it and how it actually works?” topic becomes clearer once this point is included: an announcement becomes meaningful only when it changes access, quality, price, or user behavior in a real scenario.

13:36What changes in real work: deep Research, Ask and UX-bol: where the logic

The decision in “Deep Research, Ask and UX-bol: where the logic of large models breaks” depends on one criterion: a launch matters only when it changes access, quality, price, or user behavior in a real workflow.

18:03The economic impact is already measurable

The discussion of “How the AI influences US GDP growth” yields a practical test: this section clarifies the mechanism behind the topic and preserves a constraint that would otherwise be lost in an overly simple conclusion.

29:59Andrej Karpathy explains the caution through real development

In the context of “Andrei Carpathian: Disappointment in autonomous systems and NanoChat,” this criterion applies: this section clarifies the mechanism behind the topic and preserves a constraint that would otherwise be lost in an overly simple conclusion.

39:31Delta’s algorithmic pricing shows that agency already exists without a human- interface

The discussion of “Delta Airlines and AI ticket prices: optimization or hidden discrimination” yields a practical test: this section clarifies the mechanism behind the topic and preserves a constraint that would otherwise be lost in an overly simple conclusion.

55:21Amazon’s dispute with Perplexity concerns the same boundary

For the “Amazon against Perplexity: agents who pretend to be human and future e-commerce” scene, the decisive point is this: an agent enters a store a person and acts on that person’s behalf, while the platform wants to know who is really making the request. The future of e-commerce depends on rules for identity, access, and commissions. AI agents are not yet ideal workers, but they are already forcing companies to rewrite economic and legal relationships.

What this episode is about

OpenAI is updating ChatGPT, Karpathy is disappointed with autonomy, Amazon is disputing with Perplexity, Delta uses algorithms for pricing, and home robots are reaching the market. Agency is becoming an economic force before it becomes a reliable digital employee.

Even founders of AI companies admit that today’s agents are far from universal employees. The new Update button and Deep Research modes improve individual scenarios, but the interface still forces a person to determine what to click and why the system asked the wrong question. A million enterprise customers do not eliminate that pain.

The economic impact is already measurable. Models accelerate document preparation, code, search, and support, so some productivity growth enters GDP. But the net effect is difficult to calculate: companies spend on licenses, infrastructure, and verification, while an employee may save an hour only to spend it correcting the answer.

Andrej Karpathy explains the caution through real development. An autonomous system can run for a long time, but errors accumulate, tools change, and infrastructure is expensive. Trying to automate an entire process in advance is therefore often worse than a narrow assistant that completes one step and hands the result back to a person.

Delta’s algorithmic pricing shows that agency already exists without a human-like interface. A system analyzes demand and may assign different prices to different customers. For the business, this is optimization; for the user, a risk of hidden discrimination. The question is not whether the algorithm is “intelligent,” but which decisions it has been permitted to make.

Amazon’s dispute with Perplexity concerns the same boundary. An agent enters a store like a person and acts on that person’s behalf, while the platform wants to know who is really making the request.

The future of e-commerce depends on rules for identity, access, and commissions. AI agents are not yet ideal workers, but they are already forcing companies to rewrite economic and legal relationships.

The future of e-commerce depends on rules for identity, access, and commissions. As a result, AI agents are not yet ideal workers, but they are already forcing companies to rewrite economic and legal relationships.

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 69 segments: 43 identified, 4 mixed, 12 marked with ✓, and 10 unresolved.

Loading…