🕶️ $100/hr to bully a chatbot

The latest AI job listing looks like satire, but it’s really a roadmap. Memvid is offering $800 for an eight‑hour remote session where one “professional AI bully” hammers chatbots on camera, systematically probing every dropped detail and hallucinated fact (pays $100 an hour). No CS background required; the only credential is a long personal history of being let down by technology.

For Memvid, which pitches itself as “better AI memory,” this is structured user research disguised as a stunt. It’s also turning rage into R&D and a promotional video asset in one move, as models still break in real-world memory and long-context flows.

Zoomed out, it matches a broader incentive pattern: companies now paying people to engage deeply with AI. With workers anxious about automation, serious AI adoption will include a behavioral acquisition cost, founders will pay not just for GPUs and tokens, but for people willing to push these systems until they snap.

✦

In this episode, we break down the quiet crisis inside America’s power grid: transformer shortages, 7-year interconnection waits, exploding component prices, and a supply chain the country outsourced decades ago. We explain how substations actually work (yes, with a hose), why voltage matters more than current, and how a seemingly boring piece of infrastructure is driving up your electricity bill.

Did you know… There’s a little-known way to get paid amazingly well to tell a unique kind of story? It’s only 2 pages at most. Thousands of companies are paying writers $1,500 on average to write it for them.

It’s all detailed in a short guide you can download immediately.

China just turned an open-source AI toy into a national token sink. OpenClaw, a DIY agent framework, is driving a full-on hype-fueled agent craze: nontechnical users rent Tencent cloud boxes, subscribe to Kimi, and burn through tokens so their “lobsters” can half-manage stock portfolios and content farms.

For most, the UX stops at “configure API port” and never recovers, but the cashflow is crystal clear. Cloud vendors, LLM providers, and installers are monetizing every misconfigured agent, while local governments bankroll subsidized agent workshops and grants as a new proxy for “innovation.”

Because it’s open source, every major platform has shipped copycat claw variants that integrate natively and quietly lock users into their ecosystems. The pattern is familiar: retail FOMO, heavy infra spend, thin real utility so far—but a decisive proof that a mass market will pay real money for always-on agents, even when they barely work.