🤖 Microsoft Launched the Phi-3 Mini?!

PLUS: Apple Quietly Acquires Datakalab, Adobe’s Firefly Image 3 Foundation Model

Author

James Roesser
April 24, 2024

Welcome back AI enthusiasts!

In today’s AI Report:

  • 🚀Microsoft Launched the Phi-3 Mini?!

  • 🍎Apple Quietly Acquires Datakalab

  • 🔥Adobe’s Firefly Image 3 Foundation Model

  • 🛠5 Trending Tools

  • 💰Venture Capital Updates

  • 💼Who’s Hiring?

Read Time: 3 minutes

🗞RECENT NEWS

MICROSOFT

🚀Microsoft Launched the Phi-3 Mini?!

Image Source: @KMERON/OFFICIAL LEWEB PHOTOS

Microsoft launched the Phi-3 Mini, a Small Language Model (SLM) that can easily be fine-tuned to meet the specific needs of companies with limited resources.

Key Details:

  • Phi-3 Mini was only trained on 3.8 billion parameters. In comparison, OpenAI’s GPT-4 Turbo was trained on more than 1 trillion parameters.

  • Phi-3 Mini leverages model scaling and training data curation to outperform language models up to 25X larger.

  • SLMs are designed for applications that can run locally on smartphones or laptops. They’re also perfect for performing tasks that don’t require extensive reasoning.

  • SLMs require less computational power and memory, making them faster to train and deploy.

  • SLMs can be fine-tuned for specific domains or tasks, resulting in better performance within those areas.

Why It’s Important:

  • Phi-3 Mini’s compact size allows researchers to isolate variables for mechanistic interpretability (i.e., reverse engineering neural networks), safety improvements, or fine-tuning task experimentation.

  • Because it’s efficient and easy to fine-tune, Phi-3 Mini is more accessible to companies and organizations with limited computational power or expertise.

APPLE

🍎Apple Quietly Acquires Datakalab

Image Source: Ryan McKnight/Apple Store Georgia/Flickr

Apple quietly acquired Datakalab, a French startup that compresses computer vision neural networks, in a move to strengthen the company’s on-device AI ambitions.

Key Details:

  • Datakalab focused on low-power, high-efficiency deep learning algorithms that function on local hardware rather than cloud-based systems.

  • Apple is exploring on-device AI features for iOS 18 to create a proactive iPhone experience. Imagine an AI assistant that anticipates your needs, automating mundane tasks like developing a grocery list, managing emails, or responding to texts.

  • Apple acquired over 32 AI startups in 2023, the most of any Big Tech company. This acquisition was finalized in December 2023 and only recently came to light.

Why It’s Important:

  • Apple CEO Tim Cook explained: “It’s very important to be deliberate and thoughtful in how you approach these things.” Apple is slowly implementing AI add-ons to improve the iOS ecosystem. “We view AI as huge, and we’ll continue thoughtfully weaving it into our products,” Cook said.

  • Before its acquisition, Datakalab helped Disney develop facial recognition software to monitor real-time audience reactions during cinema screenings.

🩺 PULSE CHECK

Why do you think Apple acquired Datakalab?

Vote Below to View Live Results

Login or Subscribe to participate in polls.

AI RESEARCH

🔥Adobe’s Firefly Image 3 Foundation Model

Adobe released the Firefly Image 3 Foundation Model, a text-to-image AI model that converts user queries into photorealistic images.

Adobe highlights improved text rendering, advanced prompt understanding, and enhanced illustration capabilities. These unique features elevate the AI model, offering more quality, control, and versatility.

Adobe trained the Firefly Image 3 Foundation Model on licensed content designed to generate images for commercial use that don’t infringe on Intellectual Property (IP) rights.

🛠TRENDING TOOLS

🗣Feedbuck instantly generates a feedback widget for your website.

📱Dify builds GenAI apps with an advanced open-source stack.

🎨Recraft creates visually consistent digital illustrations.

📓Grimo is the notebook that helps you find and organize your best insights.

🕵️‍♂️Crypto Co-Pilot elevates your crypto trading with AI-driven insights, trends, and news updates.

🔮Browse our always Up-To-Date AI Tools Database.

💰VENTURE CAPITAL UPDATES

  • Perplexity is raising at least $250M at a $3B valuation to develop an AI search platform.

  • Metsera secures $290M to develop medicines for obesity and metabolic diseases.

  • Rivos secures a $250M Series A-3 to create power-optimized AI chips for data analytics.

💼WHO’S HIRING?

  • Ramp (New York, NY): Software Engineer Intern, Backend, Summer 2024

  • TikTok (San Jose, CA): Android/iOS Software Engineer Intern, Video-on-Demand (VOD), Summer 2024

  • Databricks (Mountain View, CA): IT Data Engineering Intern, Fall 2024

  • NewsBreak (Bellevue, WA): Machine Learning (ML) Engineer, New Grad

  • AE Studio (Los Angeles, CA): Sales/Revenue Leader

🤖PROMPT OF THE DAY

CUSTOMERS

📌Customer Pain Points

What are the common paint points experienced by customers in [Niche or Industry], and how can market researchers effectively identify and address them?

Niche or Industry = [Insert Here]

📒FINAL NOTE

If you found this useful, follow us on Twitter or provide honest feedback below. It helps us improve our content.

How was today’s newsletter?

❤️TAIP Review of the Day

“It’s always a great read, with clear and concise sections.”

-Jetson (⭐️⭐️⭐️⭐️⭐️Nailed it!)
REFER & EARN

🎉Your Friends Learn, You Earn!

You currently have 0 referrals, only 1 away from receiving ⚙️Ultimate Prompt Engineering Guide.

Refer 9 friends to enter 🎰May’s $200 Gift Card Giveaway.

Reply

or to participate.