🤖 Apple’s ReALM Model Outperforms OpenAI’s GPT-4

Subscribe | Contact | Meet The Team

Welcome back AI enthusiasts!

In today’s AI Report:

• 🍎Apple’s ReALM Model Outperforms OpenAI’s GPT-4

• 📸DALL-E 3 Gets Mind-Blowing Editing Powers

• 💨AI-Powered Robot Reads Braille at Record Pace

• 🛠5 Trending Tools

• 💰Venture Capital Updates

• 💼Who’s Hiring?

Read Time: 3 minutes

🗞RECENT NEWS

APPLE

🍎Apple’s ReALM Model Outperforms OpenAI’s GPT-4

Apple researchers introduced ReALM, an AI model that enables Siri to monitor on-screen tasks, conversational context, and background processes.

Key Details:

• Reference Resolution as Language Modeling (ReALM) offers a new approach to converting on-device screen information into text inputs, allowing ReALM to bypass bulky image recognition parameters when processing user requests.

• In other words, the AI model considers what’s on the user’s screen and what tasks are active.

• Apple researchers outlined four sizes of ReALM:

  1. ReALM-80M

  2. ReALM-250M

  3. ReALM-1B

  4. ReALM-3B

• The “M” and “B” represent the number of parameters in millions and billions, respectively. In comparison, OpenAI’s GPT-3.5 has 175 billion parameters, while GPT-4 boasts 1.5 trillion parameters.

• Apple’s larger ReALM models outperform OpenAI’s GPT-4, despite having fewer parameters.

Why It’s Important:

• Digital assistants like Siri struggle to understand references in conversations. For example, if an iPhone user states, “Call the restaurant I visited last week,” Siri struggles to recall the restaurant.

• Apple’s ReALM tackles this issue by converting all contextual information into text, including open apps (e.g., iMessage or Apple Maps). So, Siri can understand references in requests.

🩺 PULSE CHECK

OPENAI

📸DALL-E 3 Gets Mind-Blowing Editing Powers

OpenAI unveiled a new feature that allows users to edit images generated by DALL-E 3 directly within ChatGPT.

Key Details:

• The DALL-E editor enables users to edit images by selecting an area of the image to edit and describing the changes in ChatGPT.

• The editor interface offers a wide range of options to alter photos:

  1. Undo and Redo Buttons

  2. Add, Update, and Remove Highlight Features

  3. Color Change Chat Options

• The DALL-E editor is accessible through the website interface or ChatGPT mobile app.

Why It’s Important:

• Integrating native image editing capabilities into DALL-E 3 improves creative exploration and promotes a simplified workflow. Users can now effectively alter the details of visual outputs without restarting the image generation process.

• Midjourney has a “Very Region” button that lets users edit specific segments of generated images. Elon Musk is exploring a partnership between X and Midjourney to elevate xAI’s Grok-1.5 product offerings.

AI RESEARCH

💨AI-Powered Robot Reads Braille at Record Pace

Researchers at the University of Cambridge leveraged machine learning algorithms to teach a robotic sensor to quickly slide over braille text lines.

An onboard camera captures blurred images and sharpens them with optimized machine-learning algorithms. These algorithms analyze the images to explore, analyze, and find meaning in complex data sets.

The robot could read the braille at 315 words per minute with nearly 90% accuracy, which is twice as fast and accurate as human braille readers.

🛠TRENDING TOOLS

🦄Magic Hour is an all-in-one video creation platform that streamlines content production from ideation to production.

📑SemanticPDF automatically scans, converts, and embeds PDFs to find answers.

🏜Canyon streamlines your job search by creating resumes, tracking applications, and helping you prepare for interviews.

🦸‍♀️Elicit analyzes research papers at superhuman speed.

⚙️Ellipsis reviews pull requests and converts GitHub comments into working, tested code.

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

💰VENTURE CAPITAL UPDATES

• Hailo lands $120M to keep battling Nvidia as most AI chip startups struggle.

• Mike Lynch-backed legal technology startup Luminance raises $40M to build a legal personal assistant.

• StealthMole secures a $7M Series A to develop an AI-powered dark web intelligence platform.

💼WHO’S HIRING?

• Uber (San Francisco, CA): Safety and Insurance Actuarial Analyst Intern, Summer 2024

• Nutanix (San Jose, CA): Product Management Intern, Summer 2024

• PayPal (Austin, TX): Mobile Dev Ops Intern, Fall 2024

• Universal Electronics (Santa Ana, CA): Software Engineer New Grad

• Open Asset (New York, NY): Support Engineer, Entry-Level New Grad

🤖PROMPT OF THE DAY

SEIZE OPPORTUNITIES

🎣Hook Investors and Clients

Write me a compelling, concise, and unique 30-second elevator pitch for [Business].

Business = [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?

• ⭐️⭐️⭐️⭐️⭐️Nailed it!

• ⭐️⭐️⭐️Average

• ⭐️Epic fail.

❤️AI Pulse Review of The Day

🎁NOTION TEMPLATES

🚨Subscribe to our newsletter for free and receive these powerful Notion templates:

• ⚙️150 ChatGPT prompts for Copywriting

• ⚙️325 ChatGPT prompts for Email Marketing

• 📆Simple Project Management Board

• ⏱Time Tracker