SoftBank brings OpenAI to Japan

Plus: Anthropic's bug bounty challenge.

Author

Amandeep Singh
February 04, 2025

Together with

Welcome Humans🤖,

Here is what we have Today:

  • 🎸 Beatles' AI song wins Grammy — History made

  • 💰 SoftBank bets $3B on OpenAI's future — Japan exclusive

  • 💪 AI tried to break Anthropic for 3,000 hours... and failed

  • 🌍 Meta's AI rulebook: Don't destroy the world

  • 💼 New Job Opportunities

OpenAI & SoftBank

SoftBank is making a game-changing move in AI. It just committed $3 billion annually to OpenAI while launching Cristal Intelligence, a joint venture to bring custom AI solutions to Japan.

Image Source: Ideogram

What’s in the deal?

🔹 SoftBank will spend $3B per year on OpenAI tech across its subsidiaries like Arm and PayPay.
🔹 Cristal Intelligence will offer a Japan-exclusive business version of ChatGPT and OpenAI’s API.
🔹 The joint venture focuses on secure, tailored AI for enterprises, challenging traditional consulting firms.
🔹 This follows SoftBank’s role in Stargate, a $500B AI data center initiative with OpenAI and Oracle.

Why does this matter?

Japan is doubling down on AI dominance. This investment—$3B every year—cements SoftBank as a key OpenAI partner, expanding AI adoption in Asia’s business world. With Cristal Intelligence and the Stargate project, SoftBank is setting the stage for a massive AI infrastructure boom. 🚀

Anthropic

AI safety just got serious. Anthropic has unveiled Constitutional Classifiers, a cutting-edge security system designed to stop jailbreak attempts. And after surviving 3,000+ hours of attacks, they’re throwing down the gauntlet—inviting the public to try and break it.

Image Source: OpenAI

The Challenge:

🔹 The system generates training data in multiple languages and styles, making it adaptable against diverse threats.
🔹 It blocked 95.6% of jailbreaks in tests against 10,000 attacks—compared to just 14% for unprotected Claude.
🔹 183 expert hackers spent weeks trying to breach it for a $15,000 reward—and failed.
🔹 Now, Anthropic is letting anyone test it until February 10.

Why It Matters:

As AI grows more powerful, so do the risks of manipulation. While most companies focus on training safeguards, Anthropic is taking a bold new approach: using AI to fight AI. By opening it to public testing, they’re proving they take AI security seriously.

Think you can outsmart the system? Game on. 🚀

Together With The Rundown AI

Start learning AI in 2025

Everyone talks about AI, but no one has the time to learn it. So, we found the easiest way to learn AI in as little time as possible: The Rundown AI.

It's a free AI newsletter that keeps you up-to-date on the latest AI news, and teaches you how to apply it in just 5 minutes a day.

Plus, complete the quiz after signing up and they’ll recommend the best AI tools, guides, and courses – tailored to your needs.

Trending Today

  • Bioprinting for hearts and livers refers to a cutting edge technology where scientist uses 3D technologies to create living organs is still in the process of becoming reality but the 3D printing has worked well in making skin cells and any organs that are not that complex to make.

  • OpenAI expanded ChatGPT's WhatsApp integration globally, adding image uploads and voice message support simply by texting 1-800-CHATGPT (1-800-242-8478).

  • Meta has released a new policy document—called the ‘Frontier AI Framework’—which is a three-stage process that outlines its approach to developing AI systems involves anticipating risks, evaluating them (eg. a "critical risk" would result in an “unmanageable catastrophic outcome”), and deciding whether to release, restrict, or halt the model, as a result. 

  • The Beatles have secured their 8th Grammy for ‘Best Rock Performance’ for their 2023 track “Now and Then”—which was made with AI—was put together using a demo recorded by John Lennon in the ‘70s, with Paul McCartney and Ringo Starr finalizing the track in 2021 using an ML technology that cleaned up Lennon's recording.

  • Gemini quietly got upgraded to a stable version of 2.0 Flash and Imagen 3. The 2.0 Flash model for devs is still under the “experimental” label but I assume a stable version is coming soon.

Recommended Reading
If WE had to recommend other newsletters

AI Tool Report is an Newsletter, we enjoy reading daily. AI Tool Report delivers top techniques on how AI can transform your business filled with practical tips, real-world examples.

A Smart Bear is for people who love thinking about strategy, startups, product, marketing, decision-making, and founder psychology.

Job Opportunities

  • Databricks - Machine Learning Engineer, Gen AI & LLM - Apply

  • Anyscale - Technical Curriculum Developer - Apply

  • Shield AI - Senior Integration Lab Specialist - Apply

  • Sanctuary AI - Senior Software Engineer, Robotics Simulation - Apply

  • Palantir Technologies - Embedded Software Engineer - Apply

AI News

Source: Ideogram

Ideas? Comments? Complaints?

We read your emails, comments and poll replies daily.

Get the most important AI, tech, and science news in a free daily email.
New Here? Subscribe!
Sponsorship Slots Open for January and Reach over 800+ active readers. (Now 40% off) 🤯

What`d you think of today`s edition?

Login or Subscribe to participate in polls.

Until next time, Stay Informed!

Reply

or to participate.