- The Upload
- Posts
- The Future of Weight Loss?
Ilya Sutskever is betting billions on a mysterious new approach to AI, Microsoft is trying to untangle itself from OpenAI, and Stanford’s AI just discovered a potential Ozempic alternative that could shake up the weight loss industry. Meanwhile, Meta is gearing up to put AI in WhatsApp, Instagram, and Facebook.
In Today’s Upload:
🗺 Ilya Sutskever’s secretive ASI project is raising billions
👀 Microsoft building its own AI to cut OpenAI dependence
🚀 Meta’s Llama 4 to power AI across its entire ecosystem
🏆 The new AGI "app store" changing AI development
💊 Stanford AI finds a possible Ozempic alternative
Let’s get into it.
Image source: MIT
🗺 Ilya Sutskever’s secretive ASI project is raising billions
Former OpenAI co-founder Ilya Sutskever’s Safe Superintelligence Inc. (SSI) is raising $2B at a $30B valuation without any public product, revenue, or roadmap. He claims to have found an entirely new approach to artificial superintelligence (ASI).
Key Details:
Sutskever says SSI is not scaling existing AI but instead "climbing a different mountain" in AI research.
The startup has just 20 employees and no plans to commercialize anything until ASI is achieved.
Despite the lack of a product, billionaires are lining up to invest in this secretive AI moonshot.
Why it matters:
This is one of the biggest contrarian bets in AI—an attempt to rewrite AI’s foundations, rather than competing with OpenAI, DeepMind, and Anthropic in scaling up existing models.
What it means for you:
If SSI succeeds, AI development could shift dramatically overnight. If it fails, it could be one of the most expensive AI dead-ends in history.
Image source: Justin Sullivan / Getty Images
👀 Microsoft building its own AI to cut OpenAI dependence
Microsoft is developing MAI, a new family of AI models, to eventually replace OpenAI’s models in Copilot and reduce reliance on OpenAI after a series of disagreements.
Key Details:
Microsoft’s MAI models reportedly match OpenAI’s and Anthropic’s latest offerings.
Microsoft is testing xAI, Meta, and DeepSeek as backup AI providers.
OpenAI and Microsoft’s partnership has become increasingly strained, with OpenAI limiting Microsoft’s access to its tech.
Why it matters:
Microsoft’s $13B investment in OpenAI doesn’t mean it wants to stay tied to them forever. AI is becoming decentralized, and companies are hedging their bets rather than being locked into one provider.
What it means for you:
We’re entering a future where AI models are interchangeable, meaning businesses will have more choice—and AI development could become much more competitive.
Image source: Analytics India Mag
🚀 Meta’s Llama 4 to power AI across its entire ecosystem
Meta is preparing to launch Llama 4, a new AI model that will seamlessly integrate with WhatsApp, Instagram, and Facebook.
Key Details:
Meta is combining AI chat with voice, allowing users to interact with AI in real conversations inside its apps.
The company’s scale means hundreds of millions of users could start using AI overnight.
Llama 4 will be free and open-source, making it an alternative to OpenAI and Google models.
Why it matters:
This could be the biggest mainstream AI rollout yet. OpenAI and Google dominate the AI space now, but Meta has the distribution power to get AI into the hands of billions instantly.
What it means for you:
Expect AI-powered customer service, AI-generated content, and smarter assistants built directly into Meta’s platforms—completely reshaping how businesses interact with consumers.
Image Source: Proxy
🏆 The new AGI "app store" changing AI development
Convergence AI just launched Template Hub, a marketplace for pre-built AI agents that can be deployed instantly.
Key Details:
Users can browse, download, and modify pre-trained AI agents for various tasks.
Includes automated research assistants, financial forecasting bots, and content generation tools.
Aims to make AI development as easy as installing an app.
Why it matters:
AI isn’t just a service anymore—it’s becoming a plug-and-play tool anyone can use instantly. This shift could massively accelerate AI adoption.
What it means for you:
If AI becomes modular and customizable, businesses won’t need expensive AI teams—they’ll just install pre-built AI agents and customize them.
The 12-amino-acid BRP peptide (spheres are atoms and sticks are bonds) suppresses appetite and reduces weight gain in mice and pigs without causing nausea or food aversion. Image Source: Katrin Svensson
💊 Stanford AI finds a possible Ozempic alternative
Stanford researchers used AI to discover BRP, a natural molecule that mirrors Ozempic’s weight-loss effects but with fewer side effects.
Key Details:
AI analyzed 20,000 human genes to find the molecule.
Animal tests showed a 50% reduction in food intake with significant fat loss.
A biotech startup is already launching human trials.
Why it matters:
AI isn’t just helping diagnose diseases—it’s finding new treatments in ways humans never could. This could accelerate drug discovery across industries.
What it means for you:
Expect AI-designed medications to hit the market much faster, leading to cheaper, more effective treatments in everything from weight loss to chronic disease.
🎯 Key Takeaways
AI models are decentralizing – Microsoft is building its own, and Meta’s Llama 4 will challenge OpenAI and Google.
AI automation is reshaping industries – Content creation, compliance, and AI deployment are becoming instant and app-like.
The job market is shifting – AI agents are getting better at autonomous work, changing how businesses hire.
AI in medicine is moving fast – Stanford’s discovery shows AI isn’t just diagnosing—it’s inventing new treatments.