👨‍🚀 AI Power User Checklist – Your Practical Guide

Hello AInauts,

Welcome to the latest issue of your favorite newsletter!

We get a little queasy when the bosses of the big AI companies predict a 20% unemployment rate... Or announce that the Singularity will be a reality in a few hundred days. Which also means that robots will be able to do (almost) everything better than us.

This raises the question: what does that mean in for you, for all of us?

We are not panicking. But we do take a closer look. Critically and solution-oriented. That's why you'll also find practical tips on how to take action now. Because one thing is certain: waiting is not an option!

Here's what we have in store for you today:

• 🚨 Anthropic sounds the alarm - 20% unemployed due to AI!

• 🔥 Using AI like the pros: a helpful practical guide

• 😡 Digital flood of dirt: How AI slop is making us all stupid . ..

Here we go!

🚨 Anthropic sounds the alarm - 20% unemployed due to AI!

Imagine you're standing on a dirt road in the year 1900 - and someone hands you an iPhone. That's exactly what's happening right now. Only with AI. And very few people understand what that means.

The new reality: AI is already killing jobs - newcomers are having a hard time

Just a few months ago, the warnings about AI-related job losses were more of a theoretical nature.

But now the creators of the technology themselves are sounding the alarm. Loudly and unmistakably!

Dario Amodei, the CEO of Anthropic and maker of the Claude chatbot, predicts an unemployment rate of 10-20% within the next five years! This is not some distant dream of the future, this is the equivalent of the Great Depression - only this time it won't be triggered by a stock market crash, but by code.

According to Axios, traditional entry-level office jobs are quietly but systematically disappearing. Law firms no longer need armies of junior lawyers to sift through files - AI can do it in minutes. Marketing agencies are having AI create their first campaign designs. Entry-level jobs in financial analysis and software development are being systematically replaced by AI assistants.

Particularly heartbreaking: the lowest steps of the career ladder are being sawn away. For an entire generation, entering the world of work is becoming a real hurdle. And if the ground beneath the labor market shakes, the whole system will eventually collapse…

Economic Futures Program: Red alert at Anthropic

It becomes clear: the architects see the tsunami coming - and make the issue a top priority. Amodei's reaction? Not wait and see, but act.

Last week, Anthropic launched the Economics Futures Program. It is a multidisciplinary program to fund research grants, policy development and data infrastructure to help society understand and manage economic change through AI.

This is proof that this is not an empty threat: it's like a weapons manufacturer launching a peace research program…

What does that mean for you?

Forget "AI as a tool". Think "AI as a player" from now on. Today, the latest models are already generating legal opinions, analyzing financial data and writing better code than most developers. And tomorrow, a robot will build your shelves, do the housework and drive you around in your car. For 20k, one time.

Let's not kid ourselves, the message is clear: everything that is repeatable, rule-based and scalable will be automated. Maybe not tomorrow, but the day after tomorrow - the change is coming.

That doesn’t meant that you have to become an AI engineer. But you do need to understand what's happening and use the opportunities to your advantage! (... keep reading ... we have put together a few important tips on how you can position yourself.)

🔥 Using AI like the pros: a helpful practical guide

Many type prompts. Some speak them. But only a few work with AI.

The difference? Context, features and a bit of nerd knowledge. You need to know what works and how. You have to use features. And you have to experiment in everyday life.

(Sure, you don't have to do something - but we think it's the most worthwhile way).

From our point of view, prompting + knowing the chatbot features will help you turbo-charge your business.

So today we're looking at an updated overview of how best to use ChatGPT, Claude and Gemini.

1. The Big 3 - and why you should invest $20/month

All state of the art models have web access, can analyze images and understand context via files or language. If you want to use AI seriously, choose at least one of these tools:

• ChatGPT (OpenAI) - great for text, images, structure

• Claude (Anthropic) - strong for long texts, privacy-friendly, no images

• Gemini (Google) - flexible, good at voice & video

Cost: approx. 20 $/month - in contrast, the free versions are only a demo with limited functionality, but not a reliable workflow tool.

Pro tip: By default, you are often using the fast, slimmed-down mode (e.g. ChatGPT 4o). For real power, you have to manually switch to a more powerful model. This may sound trivial, but it is the first really important lever!

2. It's not the tool that matters - it's how you use it

All three chatbots offer:

• Fast model (e.g. GPT-4o, Claude Sonnet)

• Powerful model (GPT-o3, Claude Opus, Gemini Pro)

• Many features such as voice, vision, deep research, coding, file uploads

By default, you are most efficient in the fast model. Switch manually for more in-depth tasks!

You can also use all-in-one tools such as ChatLLM or Genspark - these combine the best models and features under one roof.

3. Deep Research - the underestimated super feature

Once you've tried how Deep Research works, you won't want to go back. The feature enables structured analyses with source information - and yes, the details are almost always correct.

Deep research tools are extremely useful because they can create very high-quality reports. And not just in the 08/15 style, but in-depth and detailed.

The use cases are diverse: market analysis, research, professional (second) opinion, vacation planning with extra requests, ...

And you can also use the data to create a quiz, podcast or infographic. Deep research may not yet replace experts in every case. But you'll save hours of googling and copy-pasting and make much faster progress. I promise!

4. Voice Mode is not Siri - it's your co-pilot with a camera

Most people use Voice Mode like Siri. If you ask something, you get something back in return. But that's not the point.

The real highlight: share the camera or screen!

Show the model your problem. That doesn't just feel futuristic. It' s the future of UX! ChatGPT and Gemini currently have the best voice implementations.

Use cases: converting a recipe while you're cooking, solving a math problem with a camera shot, identifying a plant while hiking, showing a screen for on-the-fly help, …

5. Feeding AI correctly: context, clarity, variants

A good prompt is no longer rocket science.

The checklist for better results:

• Provide context: include files, links, text, PowerPoint, etc. - everything helps

• Be specific: say exactly what you want - not just the topic

• Ask for formatted output: Have slides, apps, images, reports built

And don't be modest. AI has no feelings and no breaking point.

• ❌ Wrong: "Give me 5 ideas for a blog article."

• ✅ Correct:"Generate 50 radically different ideas for a blog article on the topic of the 'future of work'. Divide them into categories: provocative, practical, humorous and visionary."

What you should always do: Think for yourself. Criticize. Follow up. Have the thought process shown to you ("show reasoning") or give one model the answer of another.

Pro tip: Change your questions after the first output. This creates branches - alternative paths that you can compare. A kind of "what if?" simulation in real time that leads to new perspectives.

6. Misconceptions that cost you time

• "AI hallucinates too much" - Yes, but with a power model + web search this is much rarer

• "Prompts have to be perfect" - Nope, clarity > formulation

• "I use voice like Siri" - You're leaving 80% of the potential untapped

7. What you should test now - 3 ideas in 60 minutes

Enough reading. Time to act. This is your to-do list for the next hour:

1. Start a complex work challenge. Take a real problem from your job. Upload the relevant document, switch to the strongest model (Opus, Pro, etc.) and have a real, interactive discussion. At the end, have a concrete result created (e.g. an email, a table, a concept, a graphic, ...).

2. Test deep research. Choose a topic in which you are well versed in order to be able to assess the quality. Have a detailed report created.

3. Experiment with the voice/vision mode. Go for a walk or make yourself a coffee and use the camera function for a real-life task.

You'll soon notice: The difference is huge when you activate power user mode! And if you want to delve deeper into the topic, take a look at Ethan Mollick's article.

😡 Digital flood of filth: How AI slop is dumbing us all down ...

It's a trend that's clogging up feeds and brains: low-quality, AI-generated content - known as "AI slop". Ever more blatant, ever wilder, ever wilder - the main thing is to produce masses of clicks for cash ...

Algorithms from Meta & Co. pump such content through the digital data highways - with serious consequences: Fake news, copyright theft and the loss of a collective understanding of reality.

Comedian John Oliver gets to the heart of the problem with humor and examples: AI slop is not only annoying, but also potentially dangerous.

… but: of course you can also do really cool things with AI video! And you should know how it works yourself.

The new Google Veo 3 model can even create sound. This video should provide some interesting tips to try out.

We made it! But no need to be sad. The AInauts will be back soon, with new stuff for you.

Reto & Fabian from the AInauts

P.S.: Follow us on social media - it motivates us to keep going 😁!
X, LinkedIn, Facebook, Insta, YouTube, TikTok

Your feedback is essential for us. We read EVERY comment and feedback, just respond to this email. Tell us what was (not) good and what is interesting for YOU.