Verdict: In 2026, the cost of building with AI has shifted from "API bills" to "attention management." Founders can now monetize developer downtime with Kickbacks.ai, automate earned media through Newsjack.sh, and reduce token overhead by 90% using Headroom—effectively making the "AI tax" a thing of the past.
Last verified: 2026-06-19 · Key Trend: Monetizing agent wait-states · Top Pick for PR: Newsjack.sh · Best for Token Savings: Headroom Proxy
How can you get paid to use Claude Code?
Answer: By installing the Kickbacks.ai extension, developers can earn a 50% revenue share from ads displayed in the terminal or IDE during AI agent "wait states."
Launched on June 11, 2026, by Andrew McCalip, Kickbacks.ai turns the loading spinner of tools like Claude Code and Codex into a micro-ad marketplace. While your agent "discombobulates" or runs tests, a non-intrusive sponsored status line appears. If you click, you get half of the advertiser's bid. Early adopters report that the earnings can effectively cover a $20–$30/month AI subscription, turning a cost center into a self-funding workflow.
- Supported Tools: Claude Code (CLI & Extension), OpenAI Codex, Cursor.
- Payouts: Via Stripe or PayPal (typically $10 minimum threshold).
- Privacy: Aggregated, anonymous data only; no code context is shared with advertisers.
How do you get free PR using AI agents?
Answer: The open-source project Newsjack.sh automates the entire "newsjacking" cycle—monitoring trends, identifying reporters, and drafting personalized contributor briefs.
Earned media has traditionally required expensive PR firms or deep industry relationships. Newsjack.sh (released by Connor Showler in June 2026) democratizes this by providing a set of agent skills that monitor live news feeds (X, RSS, News APIs). When a trending story matches your expertise, the agent drafts a brief suggesting you as a contributor and sends it to the relevant reporter.
| Feature | What it does |
|---|---|
| Trend Scout | Monitors X and RSS for 2026 trending keywords. |
| Angle Drafter | Brainstorms a unique, "newsjackable" angle for your brand. |
| Media Outreach | Finds journalists' contact info and drafts a cold PR pitch. |
How can you reduce AI token costs by 90%?
Answer: Use Headroom, an open-source proxy that compresses 70–95% of redundant context—like logs and database outputs—before it hits your LLM provider.
As AI agent operating systems become more complex, token bloat from tool outputs and RAG (Retrieval-Augmented Generation) metadata can crush budgets. Developed by Netflix engineer Tejas Chopra, Headroom acts as a transparent proxy. It intercepts requests, strips boilerplate, and keeps the original data locally so the LLM can "retrieve" full details only if needed.
- Lossless Compression: Same answers, but with ~87% fewer input tokens.
- Drop-in Proxy: Works by simply setting an environment variable (
ANTHROPIC_BASE_URLorOPENAI_BASE_URL) tolocalhost:8787. - Compatibility: Optimized for Claude Code, Cursor, and Hermes Agent.
Is Firecrawl still the best free scraper in 2026?
Answer: Yes, following its June 2026 update, Firecrawl allows users to search, scrape, and parse PDFs into markdown for free without even requiring an API key for initial use.
Standard LLM web-fetchers often hit bot-walls on sites like X or YouTube. Firecrawl circumvents these by handling rotating proxies and JS-rendering behind a clean API. It now offers 1,000 free credits per month, making it a staple for AI agents for everyone who need real-time data from the web.
Why is Midjourney branching into medical imaging?
Answer: On June 17, 2026, Midjourney Medical unveiled a "Full Body Ultrasonic CT" scanner designed to provide 60-second body scans with no radiation for "a few dollars."
In one of the most unexpected pivots of 2026, David Holz announced that the reconstruction math used for AI image generation is fundamentally the same as medical imaging reconstruction. The Midjourney Scanner uses a ring of nearly 360,000 ultrasound transducers to create 3D maps of muscle, fat, and bone. While not yet FDA-cleared for diagnosis, the vision is to deploy these in "Midjourney Spas" worldwide.
What this means for you
For founders and small business owners, the "AI Tax" is disappearing. You no longer need a massive budget to run a marketing ecosystem. By layering monetization (Kickbacks) with efficiency (Headroom) and automated outreach (Newsjack), you can scale a lean, agent-first company for the cost of a few coffee subscriptions.
FAQ
Q: Does Kickbacks.ai share my code with advertisers? A: No. Kickbacks.ai only reports anonymous impression data. It does not read your source code or agent prompts.
Q: Can I run GLM-5.2 locally? A: Yes. GLM-5.2 is an open-weight model released under the MIT license by Z.ai (ZAI) in June 2026. It features a 1M context window and is 1/6th the price of GPT-5.5.
Q: Does Headroom work with all LLMs? A: It is compatible with any provider supported by LiteLLM, including OpenAI, Anthropic, and Google DeepMind.
Q: Is Newsjack.sh free to use? A: It is an open-source project available on GitHub. You only pay for the LLM tokens used by the agents.
Discussion
0 comments