AI is not a product but a technology: 315 points HN hot post dismantling Silicon Valley’s real AI dilemma

John Gruber published an in-depth article 'AI Is Technology, Not a Product' (HN 315 points) in Daring Fireball, pointing out that AI is over-packaged as a product rather than the underlying technology. At about the same time, University of Arizona students booed Eric Schmidt’s AI advocacy at their graduation ceremony—Silicon Valley’s AI narrative was losing the trust of ordinary people.

Core conclusion

Two seemingly independent events last week outline the narrative turning point the AI industry is experiencing—the AI revolution that Silicon Valley has been promoting so much is encountering collective skepticism from academia, the technology community, and ordinary users.

On the one hand, John Ternus, Apple’s director of hardware engineering, made it clear in an interview with Wired: “We never consider delivering a technology, we only deliver excellent products and experiences.” This stance was quoted by John Gruber of Daring Fireball and written as this week’s HN hot post “AI Is Technology, Not a Product” (315 points). Gruber’s core point is that Silicon Valley has packaged AI into the next iPhone, but AI is neither a product nor a function—it is just the underlying technology and needs a real product form to carry it.

Elsewhere, the University of Arizona’s Class of 2026 publicly booed former Google CEO Eric Schmidt at his commencement ceremony—his AI-praising speech as a guest speaker was collectively booed by students. The title of The Verge’s report bluntly stated: “Silicon Valley apparently can’t read the atmosphere in the room.”

Key Points

Gruber's Argument (HN 315 points): AI is overly mythologized as a product, when in fact it is just technology
Ternus' Apple position: "We don't deliver technology, we deliver products"
Student Reaction: Eric Schmidt booed at University of Arizona graduation
Underlying signal: The AI bubble narrative is shifting from "everyone believes it" to "more and more people are questioning it"
Meanings for content practitioners: The usefulness of AI tools is being taken seriously, but the "AI omnipotence theory" is fading.

Background: Restoration of two incidents

Event 1: Gruber confronts Silicon Valley’s AI narrative

It all started with an article in Wired. Steven Levy wrote an article "Apple's next CEO needs to launch a killer AI product" after the handover of Apple CEO, believing that AI will "subvert the entire iPhone ecosystem" - by 2030, people will no longer need mobile phones and will only need to express their needs to the "always-on AI Agent".

Gruber's response was unsparing: "It's hard for me to read this without worrying that Levy has lost his mind just like those AI snake oil salesmen." He refuted Levy's assumptions one by one: Will the AI agent know when you have finished your meal and ordered a ride? Perfect execution every time? Don't you think this is creepy? Who will answer these specific questions—what microphone is used to listen for instructions? What speaker is used to reply? What screen do you use to see if the car has arrived?

Gruber finally made a bold prediction: By 2030, the device most used by people to hail a taxi will still be a mobile phone. And those assumptions about watches, headphones, and glasses operating independently of mobile phones—their camera and screen functions ultimately rely on mobile phones.

Apple John Ternus AI philosophy

Incident 2: Eric Schmidt was interrupted by students booing

At about the same time, Eric Schmidt took the stage as the commencement speaker for the University of Arizona’s Class of 2026. When the former Google CEO touted the "infinite possibilities" of AI in his speech, students in the audience continued to boo and boo, interrupting the speech at one point.

This is not an isolated incident. Schmidt has been controversial in recent years over his disputes with female employees over sexual harassment allegations and his radical stance on AI. But the collective booing by students is a sign of something more than personal — it reflects a growing rebellion among younger generations against the “you have to accept AI” narrative from Silicon Valley elites.

Key Impact

Dimensions	Changes	What it means for AI practitioners	Recommended actions
Public trust	The ebb of AI omnipotence is accelerating	Simply saying "AI changes everything" is no longer convincing	Focus on specific scenarios and quantifiable effects, and avoid generalized promises
Product thinking	Technology ≠ product, needs real value carrier	Shell AI products will be eliminated by the market at an accelerated pace	Find user pain points that AI can truly solve, rather than find scenarios for AI
Academic attitude	The younger generation is wary of AI propaganda	Consumer-oriented AI communication needs to be more restrained	Content strategy shifts to "pragmatic tools" rather than "magic solutions"
Narrative Strategy	"AI Anxiety Marketing" is gradually ineffective	The resistance to conversion driven by FOMO increases	Replace anxiety driven with education and practical value to build long-term trust

Adaptation suggestions

For practitioners engaged in AI content and AI tool affiliate marketing, the implications of this trend are clear:

Stop the AI omnipotent narrative: Stop writing “AI will replace everything” type of headlines. Readers have developed antibodies.
Focus on specific output: Don’t say “use AI to improve efficiency”, but “use Zerostack to run the coding agent on 8MB of memory, saving US$20 in monthly server fees”.
Word of mouth takes precedence over traffic: In an environment where trust in AI is declining, real practical cases are more transformative than generalized tutorials.
Refer to Apple's product philosophy: A good AI product should make users perceive "solving problems" rather than "I am using AI".

Extended reading: Key paragraphs of Gruber’s original article

Gruber’s best piece of irony:

Levy imagines that the AI Agent "will automatically call a taxi when you finish your meal, and the taxi will be waiting for you." Every time? Do you find this enjoyable rather than creepy? Will ride-hailing drivers respond to all these requests because requests never go wrong? None of this requires you to take out your phone? And will this all happen within the next four years?

Even if it were possible, I wouldn't want it. But it doesn't sound feasible at all.

Tool entry

Viewpoints and related tools/platforms mentioned in this article: OpenAI, ChatGPT, Claude Code, Apple Intelligence, AI Agent, LLM.

Reference sources

Internal link guidance

How to use AI Agent tools to automate content production? Watch the tutorial: AI Agent Tools 2026 Complete Tutorial: 5 Tools to Build an Automated Pipeline in 30 Minutes
The truth behind the cost of AI Agent scaling: Watch the news: OpenClaw 30 130 OpenAI Token
Real case: How to monetize AI automation? See the case: He Built an AI Automation Stack with Claude + n8n — $4K to $12K/mo in 6 Months