The Atlantic just published "Everything You Do Is Being Recorded" — and it hit 157 points on Hacker News with 132 comments. The article describes a near future where AI-enabled wearables record everything, and the countermeasures people are building to fight back.
This is one of the most important privacy stories of the year, and it's flying under the radar.
The Problem: AI Wearables Are Becoming Ubiquitous
Apple is rumored to be developing an AI pin or pendant that serves as an iPhone's constant eyes and ears. Multiple startups are shipping AI-enabled recording devices — glasses, necklaces, lapel pins — that continuously capture audio and video. These devices are marketed as:
- Silent notetakers — recording conversations and generating summaries
- Personal assistants — always-on AI that can answer questions about your surroundings
- Therapists — analyzing your conversations and emotional states
- Memory aids — creating a searchable record of everything you've ever experienced
The technology is impressive. The implications are terrifying.
Why This Is Different From Smartphones
A smartphone in someone's pocket isn't recording your conversation. It could, but the social norm is that it's not. AI wearables flip this assumption:
- They're always on — no need to pull out a phone and start recording
- They're invisible — a pin on a lapel or a necklace is far less obvious than someone holding up a phone
- They're AI-enhanced — not just recording, but transcribing, analyzing, and summarizing in real-time
- They're networked — recordings go to the cloud for processing, creating a permanent, searchable record
As the article notes, until recently, only spies and criminals needed to worry obsessively about electronic surveillance. Now, the average person having a casual conversation at a bar needs to think about it.
The Countermeasures Industry
A startup called Deveillance has built the Spectre I, a hockey-puck-shaped device designed to prevent others from recording you. It's a physical countermeasure — the wearable equivalent of a Faraday cage for conversations.
But the HN discussion revealed a deeper debate about what the right response is.
The Policy Response: Make PII Toxic
The top-voted comment on HN offered a policy framework:
"We need PII to be so toxic that companies run screaming from the liability unless they have explicit consent and a legit reason to house it. We don't need the government to be antagonistic toward corporations; we just need a government that makes the corporations internalize the risk they're making us take."
This is the Pigouvian approach: make the cost of storing personal data reflect the risk it imposes on individuals. If a company storing your conversation recording faces HIPAA-level liability, they'll think very carefully about whether they need to store it.
The commenter also raised a critical point about Terms of Service:
"The fact that ToS are enforceable when every court in the country knows that only a vanishingly small percentage of people ever read them, much less understand them, means that they're not doing their job."
The Separation of Corporation and State
Another popular comment took a more structural view:
"We desperately need a separation of corporations and state the same way we have a separation of church and state. We need a government that is antagonistic towards corporations that are abusing us."
This frames the problem as one of power dynamics. When corporations have more surveillance capability than governments, and governments are dependent on those corporations for data, the traditional checks and balances break down.
The Counter-Argument: Don't Create Protection-Seeking Tyranny
A nuanced response pushed back:
"The entire civilized world has been built on the uncomfortable alliance between power/capability and governance/legitimacy. Protection-seeking is the primary support system for tyranny which is happy to induce the chaos they protect you from. We just need enough negative feedback for any kind of abuse that it stops. The main part of that negative feedback is effective competition so people can vote with their feet."
The argument: instead of asking government to be antagonistic toward corporations (which can backfire), create conditions where people have real alternatives. If you don't like Apple's always-on recording, there should be a privacy-first alternative that's just as good.
What Developers Can Do
As developers and technologists, we have a role to play:
1. Build Privacy-First Alternatives
If the market for AI wearables is dominated by always-on recording devices, build the alternative:
- Local processing only — no cloud uploads, all AI inference on-device
- User-controlled retention — recordings deleted by default unless explicitly saved
- Transparent indicators — a physical LED that shows when recording is active
- Peer-to-peer — no central server, no company holding your data
2. Support Strong Privacy Legislation
The EU's GDPR was a start, but it doesn't go far enough on:
- Implied consent — wearing a recording device shouldn't count as consent from everyone around you
- Data portability — you should be able to export and delete your data from any wearable platform
- Liability for breaches — when a company's recording database is breached, the penalties should reflect the sensitivity of always-on recordings
3. Build Countermeasures
The Spectre I is one approach, but there are others:
- Audio jamming — ultrasonic interference that disrupts microphones
- Detection apps — scan for nearby Bluetooth/WiFi signals from known recording devices
- Social protocols — a standardized way to request "no recording" in a conversation
The Deeper Question: Do We Want Total Recall?
The appeal of AI wearables is real. A device that remembers every conversation, every meeting, every interaction — that's powerful. For people with memory issues, it could be life-changing. For professionals, it could eliminate note-taking forever.
But the cost is a world where privacy requires active countermeasures. Where every conversation is potentially recorded, transcribed, analyzed, and stored forever. Where the social norm shifts from "this conversation is private" to "this conversation is recorded unless someone objects."
The Atlantic article asks: "Is there any way of fighting back?" The answer is yes — but it requires action on multiple fronts: technical countermeasures, policy changes, market alternatives, and social norms.
The window to shape this technology is closing. Once always-on AI recording becomes as normalized as smartphone cameras, the fight shifts from prevention to mitigation. We're at the decision point right now.
What's your take on AI wearables? Would you wear one? Would you want to be recorded by someone else's? The comments are open.
Top comments (0)