AI = Actually Indians

PLUS: Disney and Universal sue Midjourney

Author

Barsee
12 Jun

Together with

Howdy! It’s Barsee again.

Happy Thursday, AI family, and welcome back to AI Valley.

Today’s climb through the Valley reveals:

  • AI = Actually Indians

  • Disney and Universal sue Midjourney

  • Sam Altman thinks AI will have ‘novel insights’ next year

  • Plus trending AI tools, posts, and resources

Let’s dive into the Valley of AI…

PUBLIC

Image credit: Public

Multi-asset investing platform Public has just launched Generated Assets, enabling you to turn any idea into an investable index using AI.

  • Just enter your prompt into the chat-based interface in as much or as little detail as you like.

  • In seconds, the AI will pull a custom list of stocks and help you build a Generated Asset tailored to your vision.

  • You can analyze your index's historical performance, compare it to the S&P 500, and fine-tune the holdings to your liking.

  • Then, share your Generated Asset with the world and see how it stacks up against the rest.

*This is sponsored

MAIN UPDATE

AI = Actually Indians

Builder AI, a $1.5 billion startup

Microsoft-backed UK tech unicorn Builder.ai, which claimed its "Natasha AI" could build apps "as easily as ordering pizza," has filed for bankruptcy after audits revealed its AI technology was primarily powered by 700+ human contractors in India who worked overnight shifts to maintain the illusion.

Here’s what you need to know:

  • Former CTO Robert Holdheim’s lawsuit exposed that the company’s claim of apps being "80% built by AI" was completely fabricated. Internal documents show only 3-5% of code was actually AI-generated, mostly for basic template selection.

  • Engineers were instructed to simulate AI-generated typing during demos, and sales teams used pre-recorded "AI coding" videos that weren’t live systems.

  • Multiple former employees described the company as "all engineer, no AI," with most development work performed manually by staff in India.

  • At its peak, Builder.ai raised more than $450 million from the likes of Microsoft, SoftBank’s DeepCore, and IFC to reach a valuation of $1.5 billion.

How did it work?

Behind the scenes, the operation relied on a carefully orchestrated illusion:

  • Clients paid premium prices for "AI-built" apps ($50k–$500k/project).

  • Projects were quietly routed to low-cost engineering teams in Noida and Bangalore.

  • The "AI" interface functioned mainly as a project management dashboard.

How did the financial scheme unravel?

  • Bloomberg reports Builder.ai and VerSe (a social media startup based in India) routinely billed one another roughly $170M from 2021 to 2024, a practice known as “round-tripping,” in which companies create the illusion of growth to mislead investors. 

  • Lender Viola Credit seized $37 million after discovering real revenue was just $50 million vs. the claimed $220 million. Meanwhile, $115 million in unpaid cloud bills to Amazon and Microsoft came due.

Why it matters:

This collapse reveals the dangerous disconnect between AI marketing and reality in the startup world. At least 17 startups now face scrutiny for similar scams, with 38% of "AI" companies reportedly relying on hidden human labor. As regulators ramp up investigations into "AI-washed" companies, Builder.ai serves as a warning: in the rush to capitalize on AI hype, some startups are gambling with careers, capital, and credibility. The fallout could accelerate calls for mandatory AI audits, turning this scandal into a landmark case for accountability in tech.

THROUGH THE VALLEY

Midjourney characters and Disney copyrighted characters

Disney and Universal have filed a major lawsuit against Midjourney, accusing the AI company of mass copyright infringement by generating images of iconic characters like Darth Vader, Elsa, Iron Man, and Shrek. The suit, filed in California, slams Midjourney’s tool as a “virtual vending machine” that illegally churns out and markets Disney and Universal IP without permission or payment. With Midjourney reportedly training a video generator, the studios fear future violations are imminent. They're pushing for a jury trial, calling it a clear-cut case of infringement, marking the first major Hollywood legal battle against generative AI.

Image credits: Yuki Iwamura/Bloomberg

OpenAI CEO Sam Altman’s new blog post, “The Gentle Singularity,” claims we’ve already entered the AI singularity, but instead of a chaotic takeover, it feels more like upgrading internet speed. He says the hardest scientific work is done, and what's left is scaling. Altman envisions ultra-personalized AI, where “idea people” thrive and intelligence becomes nearly free. His subtle roadmap predicts agentic AI by 2025, AI scientists by 2026, and humanoid robotics by 2027. But he acknowledges big risks remain, including alignment, regulation, and infrastructure limits. If those fail, his gentle vision could still turn turbulent.

Meta has unveiled V-JEPA 2, a new open-source AI “world model” designed to understand the 3D physical world, including how objects move and interact. Unlike traditional AI that relies on labeled data, V-JEPA 2 uses simplified internal simulations to reason and predict outcomes, like knowing a ball will fall if it rolls off a table. Meta says this could power smarter delivery robots and self-driving cars. The announcement comes as Meta doubles down on AI, planning a $14 billion investment in Scale AI and reportedly bringing on its CEO, Alexandr Wang, to accelerate its broader AI strategy.

Manus AI, the startup behind the Manus assistant, has launched a new “chat mode” that’s completely free and unrestricted for all users, a bold move aimed at challenging giants like OpenAI and DeepSeek. This chat feature provides instant responses for everyday and complex tasks, with seamless integration into its advanced agent mode, previously paywalled. Built on custom Claude 3.7 Sonnet and Alibaba Qwen models, it offers impressive performance. Manus now boasts 2M+ users on its waitlist and $75M in funding. While its free strategy fuels rapid adoption, its long-term success hinges on converting users into paid plans and expanding globally.

The Browser Company has launched a new AI-first web browser called Dia, signaling a shift away from traditional tools like Arc as AI continues to reshape how we interact with the internet. Unlike Arc, which struggled to gain mainstream traction, Dia integrates AI directly into the browser experience; its URL bar doubles as a chatbot interface that can search, summarize files, draft content based on open tabs, and respond to user preferences like writing tone or coding style. Built on Chromium, Dia is familiar yet powerful, offering features like “History” for contextual answers and “Skills” for creating custom browser shortcuts.

Wikipedia has paused its AI-generated summary trial just two days after launch, following intense backlash from its human editors. The test, which began June 2, displayed “simple summaries” atop articles for 10% of opted-in mobile users, generated by Cohere Labs’ Aya model and marked “unverified.” Editors criticized the summaries for potential inaccuracies and undermining Wikipedia’s credibility, with some calling it a “very bad idea” that risks undoing years of trust. Despite halting the trial, Wikipedia says it will continue exploring tools to make content more accessible, though human editors will still control what appears on the platform.

TRENDING TOOLS

  • Perplexity Finance: Download Excel models directly from finance pages for seamless financial modeling.

  • Dia: The AI-powered browser assistant that enhances your browsing experience by helping you write, learn, plan, and shop by using context from the pages you're visiting.

  • Nourri AI: AI that takes the guesswork out of nutrition, understanding your plate to offer personalized insights.

  • Chat4Data: Think of it as ChatGPT for web scraping, effortlessly extracting and processing data from any website.

THINK PIECES / BRAIN BOOST

THE VALLEY GEMS

What’s trending on social today:

1/ Everything is changing. We need to fundamentally think differently about entire industries.

2/ The state of vibe coding in June 2025.

Credit: Matt Palmer

 3/ What software engineering will feel like in 2027.

THAT’S ALL FOR TODAY

Thank you for reading today’s edition. That’s all for today’s issue.

💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee

👍️ New reader? Subscribe here

Thanks for being here.

REACH 100K+ READERS

Acquire new customers and drive revenue by partnering with us

Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.

If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.