iPhone to become AI Phone with iOS 18

Today’s top AI Highlights:

1. Adobe releases Firefly Image 3 AI Model for image generation in Photoshop

2. Amazon Bedrock introduces the new custom model import feature

3. iOS 18 to introduce a number of on-device AI features

4. Stanford University releases the AI Index Report 2024

5. Speaking Copilot for Non-English speaking professionals 

& so much more!

Read time: 3 mins

Exciting Opportunity: Share how you use AI and get featured in Unwind AI! Details below.

Latest Developments 🌍

Adobe’s Image Ideation + Generation AI Model

While not very known for its text-to-image capability, Adobe has released Adobe Firefly Image 3 Foundation Model, its next generation of creative generative AI model. Image 3 not just delivers high-quality images, but also helps you ideate fast with prompt auto-complete suggestions and generates images within seconds. The model is available to use in Photoshop and the Firefly web application.

Key Highlights:

1. Improved Quality: You can now enjoy better lighting, more precise positioning, and a greater variety of moods and expressions in people's rendering. The model also excels in handling complex scenes and structures, with more photorealism, detail, accuracy, and variety.

2. Structure Transfer: Lets you generate images that closely align with the structure of a reference image. This reduces the need for trial and error and gives you more control.

3. Style Transfer: You get finer control over the aesthetic output of your images. This feature lets you precisely match the visual tone and mood of your reference image with the output.

4. Training and IP: The model was trained on licensed content, such as Adobe Stock, and is designed to generate content for commercial use that does not infringe on copyright and other intellectual property rights.

Import Custom Models to AWS Bedrock ♻️

Amazon Bedrock, a fully managed service for building and scaling generative AI applications, has introduced a new feature called Custom Model Import. This new capability allows you to import models you’ve fine-tuned or adapted elsewhere directly into Bedrock’s managed environment. By supporting architectures like Llama, Mistral, and Flan T5, this feature streamlines the development process and enhances model accessibility. You can now focus more on application development without the complexities of managing diverse model sources.

Key Highlights:

1. Bring Your Own Models: Import models you’ve customized outside of Bedrock and use them alongside Bedrock’s existing models.

2. Supported Architectures: Works with popular open-source architectures like Llama 2, Mistral, and Flan T5, giving you flexibility in model choice.

3. Unified Experience: Manage and access your custom models through the same interface and API as Bedrock’s built-in models.

4. Flexibility for Customization: Supports various customization techniques like fine-tuning, domain adaptation, and even models trained from scratch.

5. Scalability and Performance: Leverage Bedrock’s managed infrastructure to run your custom models with on-demand or provisioned throughput.

AI Features in the New iOS18 will Run Entirely On-Device 📱

Apple’s upcoming iOS 18 will prioritize user privacy by processing all AI features directly on iPhones, avoiding the need for cloud-based solutions. This will ensure that your personal data doesn’t have to leave your device to benefit from advanced AI capabilities. According to reports, this includes a variety of enhancements to applications like Messages, Safari, and Siri, which will now utilize Apple's in-house developed large language model, known as "Ajax".

Despite this focus on privacy, certain advanced AI features will still require an internet connection. As Apple integrates AI more deeply into its operating systems, it has also considered licensing technology from Google and OpenAI for more sophisticated tasks. This strategy could help Apple address concerns about AI-generated errors and improve overall user experience without compromising on privacy.

AI Giants Lead the Innovation Race, Leaving Academia in the Dust 🪚

The advancements in AI have brought a wave of excitement and concern. Understanding the complex landscape of AI development, its impact, and the ethical considerations surrounding it is crucial for individuals, businesses, and policymakers alike. Stanford University’s AI Index Report 2024 offers valuable data and insights into the current state of AI, covering areas like research trends, technical performance, responsible AI practices, economic influence, and public perception.

Key Highlights:

1. AI beats humans on some tasks, but not on all: AI models now outperform humans in areas like image classification, basic reading comprehension, and visual reasoning. But they still struggle with more complex cognitive tasks such as advanced mathematics, planning, and understanding nuanced situations.

2. Industry leads the charge in frontier AI research: The development of cutting-edge AI models is dominated by big players like Google and OpenAI, who possess the resources needed for data and compute-intensive training. In 2023, the industry produced 51 notable AI models, while academia contributed only 15.

3. Training costs have skyrocketed: Developing state-of-the-art AI models like GPT-4 and Gemini is expensive, with estimated training costs reaching $78 million and $191 million, respectively. This high cost creates a barrier to entry.

4. The US maintains its leading position in AI: The U.S. remains the primary source of top AI models and patents, followed by China and the EU. However, the gap between the U.S. and other countries is narrowing, particularly in areas like industrial robotics where China is rapidly advancing.

5. Responsible AI practices are gaining traction, but challenges remain: There’s a growing focus on mitigating AI risks related to privacy, security, fairness, and potential misuse. But a lack of standardized benchmarks and reporting practices makes it difficult to systematically compare and evaluate the same.

6. Generative AI is attracting significant investment: Despite a decline in overall AI private investment, funding for generative AI, which includes tools like ChatGPT and DALL-E, increased dramatically to $25.2 billion in 2023. 

7. AI is proving its worth in the workplace: AI can significantly improve worker productivity and work quality. However, overreliance on AI without proper oversight can lead to decreased performance.

8. AI is transforming science and medicine: AI is playing a growing role in scientific discovery and medical advancements. Applications range from developing faster sorting algorithms and optimizing 3D mesh generation to improving medical imaging analysis, predicting viral evolution, and classifying genetic mutations. 

9. Public awareness and anxiety: More people are recognizing AI’s potential to impact their lives, but this awareness is coupled with growing concerns about job security, privacy violations, and potential misuse of the technology. 

Tools of the Trade ⚒️

1. Fluently: It helps non-native professionals improve their English with instant feedback after each Zoom call. Imagine having a personal coach, always listening and providing tips.

2. Unicorn Platform: AI website builder that helps to quickly create websites without design or development skills for SaaS, Apps, Directories, Blogs, Personal pages, and more.

3. Monterey AI 2.0: Lets you effortlessly aggregate, analyze, and manage user feedback across multiple channels like support tickets and social media by integrating with platforms such as Zendesk and Salesforce. It provides proactive AI-driven insights, helping you make data-driven decisions quickly and at scale.

Text within this block will maintain its original spacing when published🌟 Spotlight on You: Share Your AI Use Case and Get Featured!

Hot Takes 🔥

1. My wife just googled me and found out that I’m supposedly an “Internet personality”. It came as a shock to both of us. I didn’t realize that people still used Internet. ~Bojan Tunguz

2. The next big thing in AI is agents, except it’s a decades-old idea with whole conferences dedicated it, it always hits a wall of complexity and goes nowhere, and no one has explained what will be different this time. ~Pedro Domingos

Meme of the Day 🤡

You can’t deny the coherence 🤷‍♀️

That’s all for today! See you tomorrow with more such AI-filled content.

Real-time AI Updates 🚨

⚡️ Follow me on Twitter @Saboo_Shubham for lightning-fast AI updates and never miss what’s trending!

PS: I curate this AI newsletter every day for FREE, your support is what keeps me going. If you find value in what you read, share it with your friends by clicking the share button below!