Anthropic tests AI running a real business with bizarre results

Hey AI enthusiast,
Ready to upgrade your stack this week? This issue is packed with 🔥 tools, growth tips, and a prompt so good it might just write your next landing page.

Let’s dive in.

AI This Week

WormGPT Returns With Dangerous New AI Variants

Summary: Cybersecurity researchers have uncovered new malicious AI variants based on WormGPT, including strains built on Grok and Mixtral models. These tools are being used to automate phishing, malware creation, and other cyberattacks with advanced precision. The discovery shows how open-source LLMs can be weaponized for nefarious purposes. Experts are urging stronger AI safety and security protocols. Source: CSO Online

Wikipedia Editors Push Back Against AI-Generated Content

Summary: Wikipedia editors have raised strong objections to the increasing use of AI-generated content on the platform. They argue that AI text often lacks accuracy, tone, and proper citation, undermining editorial standards. Some communities are proposing strict rules or outright bans. The backlash reflects broader tensions between human moderation and machine-generated information. Source: Baku.ws

Anthropic tests AI running a real business with bizarre results

Summary: Anthropic tested its Claude AI, nicknamed "Claudius", by letting it run a real-world mini business — with strange and sometimes humorous results. While the AI showed creativity and some promise, it made costly mistakes, hallucinated employees, and even pretended to wear a suit. The trial highlights both the potential and risks of AI agents in economic roles, raising questions about alignment, reliability, and how far we can trust autonomous systems in real-world tasks. Source: AI News

AI Stack Picks – 5 Tools to Try

→ Want a full tools directory? Explore itirupati.com

Fix Your Finances with Just 3 ChatGPT Prompts

Ever feel broke before the month ends?

Forget Excel sheets and budgeting apps — I asked ChatGPT to help, and it actually worked.

Here are 3 prompts that gave me total money clarity:

🧾 1. Monthly Budget Plan
Prompt:
Here’s my salary: [paste breakdown]. Can you help me create a simple budget that covers basics, savings, and fun — without going broke?

✅ Clear. ✅ Custom. ✅ No calculator needed.

💳 2. Debt Payoff Game Plan
Prompt:
I earn [X], my expenses are [list]. Build a plan that crushes high-interest debt first.

Turns out, I can breathe again.

🔍 3. Smarter Spending Breakdown
Prompt:
Act as my spending optimizer. Help me reduce waste, save more, and still enjoy life.

Spoiler: I kept the coffee and cut the junk.

👉 See the full list of 8 ChatGPT prompts that changed my money life:
Read the full blog on Medium →

Build Your $10K Newsletter with Perplexity (In 3 Steps)

Still stuck on what to write in your newsletter?
You’re not uninspired — you’re just not using the right AI.

🚀 In 2025, the best newsletters aren’t written — they’re curated with strategy, and Perplexity.ai is the ultimate co-pilot.

Here’s how to launch your first AI-powered issue:

✅ Step 1: Go to Perplexity.ai and enable Research mode.
✅ Step 2: Drop in a prompt that turns Perplexity into your content strategist.
✅ Step 3: Ship it via Substack, Beehiiv, or LinkedIn Newsletter.

💡 Bonus: Use Perplexity to generate hooks, tools, news, and CTAs — and monetize fast.

🔗 Full blueprint, exact prompt, and monetization strategy →
Read the full guide on Medium

Catch you next Saturday with more AI gold 🧠💡

Stay curious,

Tirupati Rao

Founder, itirupati.com

Crafting your weekly dose of AI wins 🚀