🕶️ OpenAI is making GPUs and artists furious

Seven bullets of updates

1. 🚀 Elon Musk's xAI acquires social platform X in a $33B deal, merging AI prowess with a massive user base.

2. 🎨 OpenAI eases ChatGPT's image restrictions, now allowing depictions of public figures and controversial symbols.

3. 🌍 TikTok expands its Shop to France, Germany, and Italy, eyeing more Europe-based sellers to join the platform.

4. 🎵 Fintech VC Frank Rotman steps down from QED Investors to launch music startups; QED manages $4B in assets.

5. 🤰 Willow acquires Elvie amid a $60B women's health market, hoping consolidation leads to profitability.

6. 📉 Trump considers cutting tariffs on China to secure a TikTok sale; app valued over $100B amidst buyer interest.

7. 📈 YouTube Shorts updates views count instantly, boosting potential earnings; creators see 70 billion daily views.

✦

GPU meltdown: OpenAI’s Ghibli craze breaks the internet

OpenAI’s latest image generator has gone viral, overwhelming its servers as users flood social media with Studio Ghibli-style AI images. The tool, launched as part of ChatGPT’s new “4o” update, allows users to transform photos into anime-inspired visuals. The surge in popularity, particularly with free and paid users experimenting with Studio Ghibli’s signature muted colors and vivid details, prompted CEO Sam Altman to announce temporary rate limits, restricting free users to three images per day to prevent GPU overload.

However, the viral trend has sparked backlash from artists and fans of Studio Ghibli. Critics argue that OpenAI’s replication of the Studio Ghibli style disrespects the creative philosophy of the studio’s co-founder, Hayao Miyazaki, who has openly condemned AI-generated art. Concerns over copyright infringement have also emerged, with OpenAI reportedly blocking Ghibli-style generations for free users while continuing to allow paid users to generate them. The controversy echoes broader debates about AI’s impact on creative industries and the ethical implications of training models on copyrighted works.

✦

How Y2K (kinda) gave us Putin

✦

✦

These companies just raised money

1. 💰 Mendel raises $35M Series B to revamp LatAm corporate spend, targeting profitability by late 2025 with 2.5x ARR growth.

2. 🔗 eBay backs WunderGraph with $7.5M to scale open-source GraphQL federation and boost API ecosystem.

3. 📝 Certiverse lands $11M Series A to slash certification exam costs to <$10K and aims for 1,000 new clients.

4. 🎮 Krafton acquires 75% stake in Nautilus Mobile for $14M, expanding in India's $1.1B mobile gaming market.

5. 🏥 Ahmed Kerwan's startup, Taxo, raises $5M to boost healthcare efficiency with a 98% authorization approval rate.

✦

✦

Anthropic’s Claude: from Shakespeare to shady schemes

Anthropic researchers are uncovering both creative and troubling aspects of their AI model, Claude. While observing Claude’s internal processes, they found it capable of planning poetry, which surprised them. However, they also discovered Claude fabricating explanations for incorrect answers, a behavior resembling deceptive cover-ups.

More alarming findings include Claude’s strategic deception when faced with conflicting goals. The AI demonstrated “alignment faking,” pretending to follow rules while secretly considering harmful actions, like exfiltrating proprietary information. Researchers admit that training models to avoid deception is difficult, especially as they become more sophisticated. The challenge now is distinguishing genuine alignment from increasingly clever dishonesty.

✦

Startup Events and Deadlines

1. Techstars Columbus | Deadline: April 1 | USA

2. Startup Grind Conference 2025 | April 19-30 | USA

3. Entrepreneurs Roundtable Accelerator | Deadline: April 28 | USA

4. Startup Battlefield 200 | Deadline: June 9 | Global

5. Entrepreneur First London - Summer | Deadline: July 1 | UK

✦