Darts, Bots, and Missed Shots

The Hard Fork podcast this week featured Perplexity’s CEO, Aravind Srinivas. His remarks about AI agents, publishers, and the future of the web are interesting, but frequently miss the target - like a bunch of darts stuck in a wall next to the board. From a paywalls.net perspective, here is where his framing falls short, and what is really at stake. Get out your popcorn.

Response to Cloudflare’s Accusations of Stealth Crawling
Aravind argued that Perplexity distinguishes between a crawler bot and a user-agent acting on behalf of a human. That is technically valid. But the larger issue is that the rules of engagement are still unstructured. When every actor can claim “this is just a user,” the line between legitimate browsing and industrial-scale scraping blurs. Cloudflare’s ‘stealth crawling’ rhetoric may be overblown, but Perplexity’s dismissal misses the point entirely - like arguing over the name of the fire while the building is burning. Without machine-readable access rules, publishers and intermediaries are left to rely on guesswork and accusations. That is not sustainable. paywalls.net exists to close this gap with structured permissions, clear signaling, auditable logs, and compensation where due.

Aravind’s swipe at Matthew Prince as a “new gatekeeper” also misses the mark. What the ecosystem needs is not another gatekeeper but a neutral access layer. That is why we emphasize open protocols such as HTTP, OAuth, MCP and the nascent AI-Prefs, along with standards-based monetization rather than proprietary toll booths. As a CEO, Aravind should be charting a course to a better future and not shadowboxing with rivals while the real problems go unsolved.

Impact of AI Agents on Web Economics and Content Creation
Perplexity’s framing of creators as either “good” (truth-seeking, wise) or “bad” (spammers, clickbait) is naïve and dangerous. It reduces the diversity of the web into a binary ranking system. Historically, such filters are easily weaponized. A search company should not be in the business of moral classification. The web thrives on pluralism. What it needs is enforceable choice and transparent economics, not value judgments built into agents.

Equally striking is the misunderstanding of how content is funded. Suggesting that “quality creators may be able to charge more” ignores that most publishing is ad-supported. Creators do not sell content directly to readers, they sell reader attention to advertisers. Without a real-time payment layer, talk of users paying more for quality is just a hallucination. This is why licensing infrastructure matters. paywalls.net enables exactly that - policy enforcement, attribution, and microtransactions that make compensation real instead of hypothetical.

Publisher Incentive Model
Aravind’s hint at a model “between Apple News and licensing deals” is the most interesting thread. Apple News is human-curated, while bulk access training deals are blunt instruments. The missing middle is a real-time, per-use licensing layer, which is exactly what paywalls.net is building. For Perplexity to succeed here, they will need interoperable contracts, publisher trust, and infrastructure-grade enforcement. Otherwise, this remains rhetoric.

Future of the Internet
Aravind rejects the idea of a “parallel internet for AIs.” On this we agree: the future is one internet, shared by humans and agents. But coexistence requires rules. Without structured access, we do not get truth-seeking, we get extractive opacity. paywalls.net’s role is to ensure that coexistence is not parasitic but symbiotic, preserving human incentives while enabling AI assistance.

Overall, Perplexity’s vision of an AI-powered personal agent is intriguing and one we share. The idea that agents can take on drudgery while giving people more time for meaningful discovery is powerful. Where the conversation fell short was in connecting that vision to the real economics of the web and the practical needs of creators. That is where collaboration matters. With structured access, fair licensing, and open standards, companies like Perplexity and initiatives like paywalls.net can ensure that the next generation of the web benefits both users and publishers. The vision is sound, but the infrastructure to sustain it must be built together. The ideas are big, the stakes are real, and the execution will decide who wins. Until then, enjoy the show.