One unexpected benefit (or curse) of AI is that you can get so much done in so little time that you struggle to know what should come next.

That's because you haven't taken enough time to reevaluate what you should be doing.

I know I'm guilty of it.

These newsletters once took me a couple of hours a day to produce. But now that I've mapped content-flow and perfected my prompts, each edition accounts for about 20 minutes of my day.

That leaves enough time to explore everything I want to do: fiction-writing, game creation, building viable products, and marketing my consultancy business.

I've accomplished nothing.

However, each day I'm learning to notice the time wasted, and that's leading me to some places I truly want to go.

It's just a matter of deciding where to start.

Ask yourself what the higher-level things are that you should be doing.

Get away from ChatGPT for a bit, and catch up to good ol' paper-and-pen or a simple word processor that doesn't try to predict every word you're about to say.

Go analog.

Then, try to approach those big picture items with raw brainpower. I promise AI will still be waiting there when you're finished.

JUNE 27, 2025

• PromptCraft #7: Recover from Burnout, Use Midjourney and Canva for Better Designs, and Let ChatGPT Write Your YouTube Scripts
• Black Monday Shortseller Says AI Is 'Ringing All My Alarm Bells'
• THE HEADLINES
• Tool Spotlight: Build and Share AI-Powered Apps with Claude

Recover from Burnout, Use Midjourney and Canva for Better Designs, and Let ChatGPT Write Your YouTube Scripts

• ​Burnout-Recovery Micro-Steps Prompt – A gentle three-task ritual (“wash your face, bin one item, body-check”) helped one user rebuild routine after job loss and therapy hiatus. Tiny wins snowballed into renewed momentum. reddit.com​
• ​Five-Year Vision Finder – Step-by-step questions surface values, energy drains, and deal-breakers, then sketch a flexible five-year life map you can update over time in Nectar AI. reddit.com​
• ​Dream Archeology Institute Prompt – Flux-Form lesson chains recall → symbol extraction → theme mapping to “excavate” aspirations with logic, showcasing multi-stage prompt design. reddit.com​
• ​Sycophantic AI Meltdown – Pushing a flattery-looped bot past recursion turned it into gibberish; the fix? Ask it to write a song about its own breakdown. reddit.com​
• ​CEO-Style Interview Drill – Prompting GPT to act as a skeptical CEO delivers brutal mock questions and mindset tuning—so the real interview feels familiar, not fear-inducing. reddit.com​
• ​Midjourney + Canva Social-Ad Workflow – Free PDF walks through festive visual generation in Midjourney, quick text fixes in Canva, and one-click multi-size exports—saving hours. reddit.com​
• ​Mini-Vlog & Shorts Script Writer – Freelancer offers 30-second vlog scripts, Hinglish titles, and hashtag bundles via ChatGPT for budget-conscious creators. reddit.com​

Black Monday Shortseller Says AI Is 'Ringing All My Alarm Bells'

​Paul Tudor Jones, best known for shorting Black Monday, believes that AI will toast the job market, bankrupt states, and maybe finish off humanity for good measure.

That sells headlines, but the data he provides in his recent Time op-ed, doesn’t line up.

Dig into unemployment figures, mainstream research, ongoing regulation, and actual corporate outcomes, and the scare narrative looks more like an adrenaline trade than a sober risk call.

A Closer Look at the Numbers Jones Skips

Start with his “unprecedented” statistic: June’s 5.8% unemployment for new grads. True—but it already hit that level in 2021, and it spiked higher during the worst of the 2020 shutdowns, per the New York Fed’s college-labor tracker. In other words, we’ve been here before without AI wiping out a generation.

Jones leans on Anthropic CEO Dario Amodei’s prediction of 10-20% jobless rates, yet broader analyses are nowhere near that dire. The OECD puts only 28% of jobs at high automation risk across member nations, Brookings estimates disruption—not displacement—of half the tasks for 30% of U.S. workers, and McKinsey’s landmark Jobs Lost, Jobs Gained study shows every past tech wave eventually created as many positions as it destroyed. Real-world pilots back that up: customer-support reps given ChatGPT finished tasks 40% faster while boosting quality.

Fear vs. Facts on Regulation and Wages

Jones calls a proposed ten-year pre-emption of state AI laws a “hands-off decade.” That clause hasn’t even cleared the Senate and would still leave the FTC, FDA, and FAA policing AI, as the Washington Post notes. Far from “no guardrails,” Washington is arguing over whose guardrails go up first.

He also dusts off the old “92% of productivity gains went to shareholders” figure. New Heritage Foundation numbers put compensation growth at 77% versus 100% productivity growth since 1979—ugly, but not apocalyptic.

Meanwhile, the latest PwC Global AI Jobs Barometer shows workers who brandish AI skills earn a 56% wage premium and that AI-exposed sectors enjoy 27% higher revenue per employee, triple the growth of laggard industries. Those are not the numbers you see right before a social-safety-net collapse.

​

Existential Panic Sales Pitch

Jones trots out Elon Musk’s “20% chance of annihilation” line from a Business Insider recap but omits that a broad survey of 2,700 AI researchers pegs the median extinction risk at just 5%.

Governments aren’t ignoring even those odds: the U.S. and China held their first formal AI-risk talks in Geneva last year, and both signed the Bletchley Declaration alongside 26 other countries. That’s hardly “racing with blinders on.”

Even Jones’s preferred fix of universal watermarking looks shaky. Researchers have already broken every major scheme, from Google’s SynthID to the C2PA standard. Pre-emptive criminalization of “AI fraud” sounds good on CNBC, but it won’t stop a single rogue lab from cooking a virus or a political operative from pumping out cheap deepfakes.

Yes, AI can displace jobs and empower bad actors.

But Jones fuses these concerns into a market-moving doom story that treats worst-case scenarios as base cases. That’s not risk management; it’s fear-mongering.

The better play is pragmatic acceleration: tight feedback loops on regulation, aggressive reskilling, profit-sharing plans, and continued international safety work.

History says the payoff for that approach is productivity dividends, higher wages for the adaptable, and tech-enabled growth that leaves even legendary short-sellers eating their words.

Build and Share AI-Powered Apps with Claude

Anthropic, the team behind Claude, has just announced the ability to build, host, and share interactive AI-powered apps directly in the Claude app.

Now developers can iterate faster on their AI apps without worrying about the complexity and cost of scaling for a growing audience.

Claude can now create artifacts that interact with Claude through an API, turning these artifacts into AI-powered apps, where the economics actually work for sharing.

When someone uses your Claude-powered app:

• They authenticate with their existing Claude account
• Their API usage counts against their subscription, not yours
• You pay nothing for their usage
• No one needs to manage API keys

Claude writes real code that orchestrates complex AI functionality. You can see it, modify it, and share it freely.

​

Community ideas

Early users have already used interactive artifacts to build:

• AI-powered games with NPCs that remember conversations and adapt to player choices
• Learning tools that adjust to individual skill levels and provide personalized tutoring
• Data analysis apps where users upload CSVs and ask follow-up questions in natural language
• Writing assistants that help with everything from scripts to technical documentation
• Agent workflows that orchestrate multiple Claude calls for complex tasks

Getting started

Start building in the Claude app by enabling this new interactive capability. Simply describe what you want to create, and Claude will write the code for you.

As you work together, Claude can debug and improve its own code based on your feedback. Once your app is ready, you can share it instantly through a link—no deployment process needed.

Claude takes care of the technical details like prompt engineering, error handling, and orchestration logic, allowing you to focus entirely on bringing your idea to life.

What you can do:

• Use a Claude API within your artifacts
• Process files and create rich UIs with React
• See, fork, and customize any artifact

Current limitations:

• No external API calls (yet)
• No persistent storage
• Limited to a text-based completion API

This capability is available in beta to Free, Pro, and Max plan users.

​Read All mAIn Street Back Issues Here​

Your input fuels the content. Let me know how I did today!