Nilay Patel, speaking on The Vergecast, argued that existing AI products are insufficient to convince users of their value relative to the costs imposed. The discussion, collected by Simon Willison and posted on 6th November 2026, emphasized that the benefits are not real for most people.reportedA podcast guest says AI tools are too expensive and not useful enough for most people right now.The commentary asserts that the cost-benefit ratio of current AI deployments is negative, with agents lacking genuine utility.
The figures
The source material does not provide specific financial figures, model parameter counts, or performance metrics. It references general concepts such as "more than they're making" and "too high" costs without numerical data.
Behind the figures
No measurement methodology is described in the source material. According to the source material, The claims are qualitative opinions expressed by Nilay Patel during a podcast episode titled Big Tech lost big in the 2026 election.reported
What it means for your stack
The commentary suggests that the current generation of AI agents lacks genuine utility. Patel states that agents "aren't real" and may not be used in ways that align with user desires. This implies that developers building agent-based workflows may face adoption hurdles if the underlying technology does not deliver tangible improvements to daily life.
Building on this
Developers should focus on creating products that demonstrably improve user life without requiring ad campaigns for validation. The source indicates that until users can see clear benefits without persuasion, the industry will remain in a difficult position. Building on documented capabilities requires moving beyond "slop" and deep fakes to tools that offer real value, as the current offerings are deemed not worth the cost.
Why it matters
- Inference engineers must consider that current agent architectures may not deliver the promised utility, risking low adoption rates.
- The high cost of infrastructure, described as data center noise, may not be justifiable if end-user products remain low-quality.
- Companies are currently operating at a loss on these products, suggesting a need for more efficient or valuable use cases.
- The lack of clear user benefit means developers cannot rely on hype to drive integration of AI features.
Topics: ai, agents, commentary, vergecast, nilay-patel
Updated
Then / now / next
AI companies were spending heavily on products without clear user benefit.
A prominent commentator argues these products are not worth the cost or infrastructure impact.
The industry must create better products that visibly improve life without needing ad campaigns.
Story timeline
- Then AI companies were spending heavily on products without clear user benefit.
- 06 Nov 2026 Quoting Nilay Patel
- Now A prominent commentator argues these products are not worth the cost or infrastructure impact.
- Next The industry must create better products that visibly improve life without needing ad campaigns.
The strongest doubt
Counterpoint from the sources
The source material presents a single perspective from Nilay Patel and does not include opposing views or data from AI vendors.
For readers
Developers
- Focus on building products that offer clear, undeniable value to users.
- Avoid relying on hype to justify the integration of AI features.
Business
- Current AI products are costing companies more than they are making.
- The industry needs to demonstrate tangible life improvements to justify costs.
Consumers
- AI overviews and deep fakes may not be worth the infrastructure noise.
- Products must prove their worth without needing ad campaigns to convince users.
Ask the article
Local keyword search of this article; no generated answers.Did this article give you the evidence you needed?
Your choice stays on this device. No public tally.Terms in this story
slop
Low-quality or excessive AI-generated content, often described as hilarious deep fakes.
agents
Autonomous AI systems that the source claims are not real or used as intended.
Vergecast
A podcast hosted by Nilay Patel where this commentary was recorded.
Questions readers ask
Are AI agents currently considered effective by industry commentators?
No, Nilay Patel stated that agents are not real and may not be used as anyone wants them to be yet.
What is the main criticism of current AI products in this report?
The products are not good enough to sell people on the costs, and the benefits are not real.
When was this commentary published?
Simon Willison posted the quotation on 6th November 2026.
Connected stories
Reading receipt
- Current AI products do not justify their costs to consumers, according to the source.
- The speaker states that AI agents are not real and may not be used as intended.
- Companies are spending more on these products than they are generating in revenue.
Still open: What specific product improvements would make AI worth the infrastructure costs?
Still open: How will companies adjust spending if products remain unprofitable?