• AI Academy
  • Posts
  • 🤖 GPT-4o's Ghibli Effect Melts Servers

🤖 GPT-4o's Ghibli Effect Melts Servers

Plus: OpenAI's $40B Funding Gamble and Amazon's Browser-Based AI Agents

Hello AI Enthusiast,

OpenAI's new image generator sparked a Ghibli-style phenomenon that melted their servers and netted them a million new users in just an hour. Meanwhile, OpenAI secured a historic $40 billion in funding (with some pretty serious strings attached), and Amazon jumped into the AI agent race with Nova Act for browser automation. Let's break down what's happening and why it matters.

Before diving into the news, know that we're launching our first-ever AI Avatar Bootcamp starting next week on April 9th! This edition is exclusively available to our Members, making this the perfect time to join if you've been on the fence. Membership gives you 12 months of access to this and many other programs we run throughout the year. Learn how to create professional avatars that automate your work.

The Big Picture 🔊

OpenAI's Ghibli Images Break the Internet

OpenAI's new ChatGPT-4o image generator went viral with Studio Ghibli-style transformations, overwhelming servers and forcing rate limits. Seattle engineer Grant Slatton's family photo post gained 46 million views, helping drive ChatGPT to 150 million weekly users with a million new sign-ups in just one hour. The trend sparked backlash from artists and raised copyright questions, with some citing Ghibli founder Miyazaki's previous criticism of AI art.

Gianluca Belloni
Gianluca BelloniCMO and Marketing Nomad

Sometimes the simplest things drive the most adoption. No fancy reasoning models or technical breakthroughs - just the ability to make cute Ghibli-style images sent ChatGPT usage through the roof.

The copyright debate remains stuck: Can you copyright a style? OpenAI’s distinction between living artists (protected) and studios or deceased artists (fair game) feels arbitrary. It’s a thorny issue with no clear answers.

The real question: Will these new users stay? Will they explore ChatGPT’s other features, or fade away once the novelty wears off?

Have you jumped on the Ghibli-style AI image trend?

Login or Subscribe to participate in polls.

OpenAI's Record $40B Funding Round

OpenAI has secured a $40 billion funding round, the largest private tech funding in history, valuing the company at $300 billion. SoftBank is leading with a $30 billion investment, alongside Microsoft and other backers. However, up to $20 billion of the funding depends on OpenAI successfully restructuring into a for-profit entity by December 31st. About $18 billion will support the Stargate AI infrastructure project announced by President Trump in January. OpenAI now reports 500 million weekly users and expects to triple revenue to $12.7 billion this year.

James Varnham
James VarnhamCEO and Rainmaker

This funding deal is a house of cards built on future promises. SoftBank is putting in $30B but could pull back $20B if OpenAI doesn't convert to for-profit by year-end. The $300B valuation seems detached from reality when you consider they're still losing money despite $12.7B expected revenue.

The whole arrangement depends on California's Democratic Attorney General approving the restructuring, adding political complexity given Trump's involvement with Stargate. After DeepSeek showed you can build competitive AI cheaply, OpenAI needs to justify these astronomical numbers.

Amazon Launches Nova Act for Browser-Based AI Agents

Amazon has released Nova Act, an AI model designed to perform actions in web browsers. The system allows developers to build agents that can complete tasks like submitting forms and managing calendars. Amazon claims Nova Act achieves over 90% accuracy on challenging browser interactions and outperforms competitors on specific benchmarks.

Gioele Mottarlini
Gioele MottarliniCOO and Image Addict

Amazon's Nova Act feels like a solution searching for a problem. The examples they showcase (ordering salads, calendar tasks) are surprisingly basic, and having AI navigate websites like humans seems inefficient when APIs exist for machine-to-machine communication.

The fact you need to anticipate issues ("don't accept the insurance upsell") undermines the automation promise. Plus, there's the trust factor - most people might forgive their own payment errors but would be far less forgiving of an AI making mistakes with their money.

Bits and Bobs 🗞️

  • Elon Musk has announced the sale of his social media platform, X, to his AI company, xAI, for $45 billion. This move aims to integrate AI more deeply into X with xAI's Grok chatbot already part of the platform.

  • OpenAI is set to release a new open-weight language model with enhanced reasoning capabilities, their first since GPT-2.

  • Runway's next-generation Gen-4 AI model allows for precise media generation by maintaining character, object, and location consistency across scenes.

  • Dartmouth researchers conducted the first clinical trial of Therabot, an AI-powered therapy chatbot, finding significant symptom improvements in users with depression, anxiety, and eating disorders.

  • The recent Anthropic’s report on Claude's usage reveals increased application in coding, education, and sciences since the launch of the Claude 3.7 Sonnet model.

From Our Founder’s Channels 🤳

Here’s a video of Gianluca explaining why it’s important to discuss both the positive and negative impacts of AI.

@gianluca.mauro

AI and intellectual honesty #Ai #learnontiktok #artificialintelligence #business #machinelearning #product #ux #entrepreneur

LOLgorithms 😂

Memes turned into Studio Ghibli are on a whole other level! OpenAI’s GPUs will keep melting.

That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:

Catch you next week! 👋