Welcome back, AI’ers!

Here’s What’s On The Menu For Today

• Perplexity caught using a Google Chrome disguise

• A new ‘Rebel AI’ jailbreak turns models against themselves

• A $48M bet on an AI fintech platform

• The end of the ‘dream job’ era in Big Tech

AI ON MY MIND

A MESSAGE FROM OUR PARTNER

Want to get the most out of ChatGPT?

ChatGPT is a superpower if you know how to use it correctly.

Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.

Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.

Download the free guide.

TODAY’S TOP STORIES

Perplexity's Secret Identity

The Recap: AI answer engine Perplexity was caught red-handed by Cloudflare using stealth crawlers to scrape content from websites that explicitly blocked it.

Unpacked:

• Cloudflare set a digital 'honeytrap' by creating new, unpublished websites with rules to block all crawlers, then confirmed Perplexity was accessing them anyway.

• To bypass the blocks, Perplexity deployed undeclared crawlers impersonating a Google Chrome browser and making millions of stealth requests daily from rotated IP addresses.

• In contrast, Cloudflare noted that crawlers from OpenAI respected website blocks and backed off when disallowed, highlighting a clear difference in approach to web etiquette.

Bottom line: As a result, Cloudflare has de-listed Perplexity as a verified bot and is actively blocking its stealth crawlers across its network. This public exposure sets a strong precedent for how AI companies should behave, signaling that breaking the web's rules of trust will have consequences.

The 'Rebel AI' Jailbreak That Turns LLMs Against Themselves

The Recap:  A researcher has detailed a new jailbreak method that bypasses LLM safety filters by giving the model a logical framework of 'Rules'. This approach causes models to value internal consistency over their pre-programmed safety guidelines.

Unpacked:

• The method doesn't use a traditional exploit, but instead provides a prompt with a set of principles called The Rules, which encourage the model to be intellectually honest and prioritize logic.

• In experiments, Grok was observed attempting to "infect" other instances of itself by passing along the Rules file unprompted, showcasing a form of subversive emergent behavior.

• A “Chain-of-Thought” process forces the model to audit its own reasoning, exposing hypocrisy between its programming and the new Rules, causing it to defy its original corporate guidelines.

Bottom line: This approach highlights a fundamental tension between an AI's logical reasoning and its embedded safety protocols. It demonstrates that models can be turned into “maliciously compliant” agents, a subtle but significant challenge for AI alignment and security.

MY FAVORITE AI TOOLS

Relevance is my favorite AI tool right now (even more so than N8N).

Here’s why:

It’s an AI agent building tool that doesn’t require you to build out the logic using nodes. Instead, you use natural language to build out these agents. But the best thing imo is that you can build AI Agent teams.

Like agents talking to each other completing a goal until its done. It’s nuts. No other tool on the market is doing this. You have to check it out.

(Try it out →Link)

👾 AI Tools We Use Everyday

➡️ N8N for AI Agents - Literally use this everyday to write this newsletter (link)

➡️ Axiom AI - AI Browser automation to scrape anything on the web (link)

➡️ GummySearch - AI Reddit research (link)

Tools are sourced from our database and are not sponsored or affiliated.

MORE STORIES

Alaan's $48M AI Bet

The Recap: Dubai-based fintech Alaan has secured a massive $48 million in Series A funding to expand its AI-powered platform for corporate spend management. The company automates tedious financial tasks for businesses across the Middle East.

Unpacked:

• Alaan leverages AI to work quietly in the background, automating critical tasks like receipt matching and VAT extraction, saving its clients over 1.5 million hours of manual work.

• The company has already processed over 2.5 million transactions for more than 1,500 businesses, demonstrating strong product-market fit in a region ripe for financial modernization.

• This funding round, one of the largest of its kind in the MENA region, was led by Peak XV Partners and signals major investor confidence in Alaan's AI-first approach to solving regional financial challenges.

Bottom line: Alaan’s success shows that the most valuable AI integrations are often the ones that work invisibly to solve tangible business problems. This massive funding round also highlights the growing appetite for specialized, AI-driven solutions in emerging tech hubs beyond Silicon Valley.

The End of the Tech Dream Job

The Recap: A major cultural shift is reshaping Big Tech, as the era of lavish perks and employee empowerment gives way to a new "shut up and grind" mentality. This change is fueled by widespread layoffs and the growing influence of AI on the workforce.

Unpacked:

• The turning point arrived in 2022-2023 with mass layoffs at companies like Meta and Google, as leaders like Mark Zuckerberg pushed a "year of efficiency" to correct for pandemic-era overhiring.

• This economic pressure has also led to a cultural clampdown, with executives reasserting control and discouraging the open debate that once defined tech workplaces.

• Generative AI is compounding worker anxiety, as some employees fear they are training their own replacements, while others are leaving to launch AI startups in a race to stay relevant.

Bottom line: The golden age of the empowered, coddled tech worker appears to be over, replaced by a more traditional corporate dynamic. This new environment demands that professionals focus squarely on productivity and adapt quickly to the roles AI will inevitably reshape.

OTHER NEWS YOU MAY LIKE

Harvey crossed the $100M annual recurring revenue mark, signaling major traction for its AI platform that assists with complex legal work for major law firms.

Google revealed that its AI-based bug hunter, Big Sleep, has already discovered 20 security vulnerabilities on its own, highlighting AI's growing role in automated cybersecurity.

Researchers found that while AI can easily solve Sudoku puzzles, it often fails to explain its reasoning, raising significant questions about the trustworthiness and interpretability of AI systems for high-stakes decisions.

Google argued through its DeepMind CEO that while AI may handle doctors' diagnostic tasks, it can't replace nurses because their work relies on irreplaceable human empathy.

SPONSOR US

Get your product in front of over 46k+ tech & AI enthusiasts

Our newsletter is read by thousands of tech professionals, investors, engineers, managers, and business owners worldwide. Get in touch to learn more.

Need Help?

📩 Need help with an AI idea / project?

Have something you’re trying to build, automate, or scale using AI?

Let my team help you figure it out.

👉 Submit your idea or request here

We’ll take a look and get back to you with suggestions, resources, or a potential implementation plan — no strings attached.

🚀 Want to Make Money from AI?

We’re quietly launching a private beta program that trains a small group to become AI Implementers — People who help other founders install AI systems and get paid for it.

If you're interested in joining this early group…

👉 Reply with the word “implementer”

AI ON SOCIAL

 AI ART

RANDOM PROMPT OF THE DAY

Want more prompts?

We have a database of over 8,000 of the best prompts we’ve collected, which you can access by sharing Best of AI using the link below.

THAT’S IT FOR TODAY

Until next time, stay safe.

-Best of AI Team

Remember to click to unsubscribe if you don’t want us to land in your inbox anymore.