← All Posts
You're Not Tired of AI. You're Tired of Renting It.
May 27, 2026 · 5 min read
On May 27, 2026, a post titled "I'm Tired of Talking to AI" hit #1 on Hacker News with over 1,700 upvotes and 850+ comments. The thread struck a nerve that's been building for months.
But here's the thing almost everyone in that thread missed: you're not tired of AI. You're tired of being a tenant.
The Real Source of AI Fatigue
Read the comments carefully and a pattern emerges. People aren't frustrated that AI exists. They're frustrated by three specific things:
- AI being used at them — colleagues, bosses, and strangers forwarding AI-generated responses without reading them
- Loss of agency — feeling like they're operating someone else's tool on someone else's terms
- Paying to be dependent — $20, $100, $250/month for tools that can change pricing, remove features, or cancel access overnight
The GitHub Example
In the thread, someone described three different people posting the exact same AI-generated response to a GitHub issue. None of them read it. None of them thought about it. They didn't use AI — they
outsourced their judgment to it.
Source: HN #48292224, 1,764 points, 852 comments
This isn't an AI problem. It's an ownership problem. When you rent your intelligence from a platform, using it thoughtfully isn't incentivized — dumping outputs is. The friction between you and good work isn't the model. It's the landlord.
The Rent Cycle: Why Cloud AI Creates Fatigue
Cloud AI operates on a fundamentally extractive model:
| The Rent Cycle |
What Happens |
| 1. Subscribe |
$20-250/month for access to AI capabilities |
| 2. Overuse |
Token-based billing incentivizes volume over quality |
| 3. Fatigue |
AI noise accumulates — yours and everyone else's |
| 4. Price Increase |
Vendor raises rates (Copilot token billing June 1, Claude $200/mo, Gemini $250/mo) |
| 5. Lock-In |
Your workflows, data, and habits are now platform-dependent |
| 6. Repeat |
Pay more, get louder AI, get more tired |
Copilot Token Billing — June 1, 2026
Microsoft is transitioning GitHub Copilot from flat-rate to token-based billing starting June 1. Teams currently paying $19-39/user/month will see costs increase 3-10x with zero additional features. The only change: you pay more for the same tool.
Source: GitHub billing documentation
What "Owning" Your AI Actually Means
The alternative isn't "no AI." The alternative is AI that runs on your terms, on your machine, with your data staying yours.
"When AI runs on your machine, you're deliberate about when it participates. No one's forwarding your prompts. The interaction stays intentional."
Here's what the ownership model looks like:
| Dimension |
Rented AI ☁️ |
Owned AI 🏠 |
| Pricing |
$20-250/mo per tool, rising |
Flat rate, no surprises |
| Data |
Stored on vendor servers |
Stays on your machine |
| Access |
Can be revoked anytime |
Permanent — you own it |
| Bill shock |
Token billing = unpredictable |
Same price every month |
| Agency |
Vendor decides your limits |
You decide when AI participates |
| Fatigue |
Built into the model |
Cured by ownership |
The Fatigue Test
Ask yourself honestly:
- When was the last time an AI tool suggested a response that actually matched your voice?
- How many AI-generated messages have you received this week that clearly weren't read by the sender?
- If your AI subscription doubled in price tomorrow, could you keep your work?
If question 3 made you pause, you're renting. And that rental agreement is the source of the fatigue.
The Evidence Is Everywhere
Microsoft is moving Copilot to token billing (June 1). Claude Max is $200/month. Google AI Ultra is $250/month. GitLab's CEO publicly frames "$100s-$1000s/user/month" as the new normal. Every major vendor is raising prices while simultaneously locking you deeper into their ecosystem.
Sources: GitHub billing docs, Anthropic pricing, Google I/O 2026, GitLab CEO
The Path Forward
You don't need less AI. You need AI you actually own — that lives on your machine, respects your agency, and doesn't send you a surprise bill at the end of the month.
The "I'm Tired of Talking to AI" thread proved something important: thousands of people are feeling this. The fatigue is real. But the cure isn't turning AI off. It's turning ownership on.
Related:
What Happens June 2: Copilot Token Billing ·
Flat Rate AI vs Token Billing ·
Copilot Alternatives 2026
Related Posts
- Claude Opus 4.8 and the Effort Tax: Why Your AI Vendor Controls How Hard Your AI Thinks — Claude Opus 4.8 adds effort control — your vendor now decides how much thinking your AI does. Fast mode costs 3x less, p
- Copilot Alternatives in 2026: Flat-Rate AI Tools That Won't Surprise-Bill You — GitHub Copilot token billing starts June 1, 2026. Compare 7 flat-rate AI coding alternatives — no tokens, no metering, n
- Copilot Token Billing Starts June 1: What Changes, What It Costs, and Why It Matters — GitHub Copilot adds per-operation token billing on June 1, 2026. Here's what your bill will actually look like — and why
- Day One: The Rent Already Arrived — June 1, 2026. GitHub Copilot's token billing went live today. Bills jumped 10x-50x. The safety net is gone. Two vendors,
- The Three Tiers of AI Access: Own, Rent, and Mythos — Anthropic's Claude Mythos (NSA, 150+ orgs, cyberweapon-grade) reveals the real three-tier model. Most developers are stuck in the middle: renting subsidized AI that can be repriced, throttled, or locked at any moment.