📆June 21st 2024📆 

An unsurprising move from one of the more notable names in AI, new model announcements and a word of warning about AI hype.

• Ilya Sutskever has embarked on a new AI venture, aimed at delivering on what he set out to do at OpenAI

• Anthropic releases a new family of models with different performance and price points

• Microsoft also announces a new vision model, available on HuggingFace

• Gartner throw AI Networking and Cloud on the hype cycle but warn against overzealous investment

• As always, too much to put in headlines. Check out the Other News Rollup section for a quickfire review of the top stories in AI.

Dive in for all the details!

🆘 Easy Tips 🆘 

You don’t need to spend tones of money to get stated using AI today. Just using the free version of ChatGPT allows you access to their most powerful model, GPT-4o. A simple thing you can do today is use this prompt template to generate a draft advertising a new product or service you have. By asking the model to explain it’s reasoning you encourage it to “think” and you get insight into why it wrote what it did, this is a concept call Chain of Thought and allows you to get more than just a Q&A response from the model.

I get more into the weeds over at Bright Journey AI if you want more technical content.

You are a knowledgeable sales person who understands {product}. I want you to write up a new post advertising a {specific version of product} that I have for sale. 

Use Internet resources to confirm & inform your statements in the final advert. 

Use a formal but approachable tone in the final advert and tailor to a {describe your audience}. I will run the final advert on {say where you will run it, eg Facebook} so apply a style that fits that medium. 

I want your to explain your reasoning step by step so I understand how you decided on this content for the ad. 

💫Help Support my Work💫

I really enjoy researching and writing about AI & appreciate your support. If you enjoy this content please use any of the options below to help

• Buy Me a Coffee ☕️- AI is really fuelled by coffee - keep the tank filled

• Spread the Word 🗞️- Share on social or directly with your friends

• Give me Feedback 💬 - Tell me what’s important to you

• Follow me on X - Always happy to chat AI and software dev

📰 News 📰 

Ilya Sutskever's New Venture: Safe Superintelligence Inc.

Ilya Sutskever, co-founder and former Chief Scientist of OpenAI, has embarked on a new journey by establishing Safe Superintelligence Inc. (SSI). This new venture aims to develop AI systems with superintelligent capabilities while ensuring they are safe and aligned with human values. Joining him in this ambitious project are Daniel Gross, formerly of Apple AI, and Daniel Levy, an ex-OpenAI member. SSI will operate out of Palo Alto and Tel Aviv, aiming to recruit top talent to tackle the challenges of AI safety and alignment.

The primary objective of SSI is to create a superintelligent AI that prioritizes safety. Sutskever's move comes after his significant role in OpenAI, where he was instrumental in the development of ground-breaking AI technologies like ChatGPT. His departure from OpenAI marks a significant shift in focus towards ensuring the responsible and safe advancement of AI technologies.

Read More At: The Register

Anthropic Launches Claude 3.5: A Powerful AI Model Family

Anthropic has introduced Claude 3.5, a new suite of AI models designed to outperform existing competitors. The Claude 3.5 family includes three models: Opus, Sonnet, and Haiku, each catering to different needs based on performance, cost, and application.

Claude 3.5 Opus is the most advanced, aimed at complex tasks such as biomedical research and extensive data analysis. It offers a context window of up to 1 million tokens and is priced at $15 per million tokens for input and $75 for output, making it suitable for high-demand business applications. This model demonstrates significantly improved accuracy and recall, surpassing previous benchmarks in various cognitive tasks.

Claude 3.5 Sonnet balances performance and cost, making it ideal for enterprise workloads. It is five times cheaper than Opus and excels in tasks like large-scale data processing, sales forecasting, and code generation. Sonnet supports up to 200,000 tokens per prompt and offers enhanced speed and efficiency, making it competitive with leading models such as GPT-4 Turbo and Gemini 1.5 Pro.

Claude 3.5 Haiku is the most compact and cost-effective model, designed for rapid, simple interactions. It is priced at $0.25 per million tokens for input and $1 for output. Haiku is perfect for live customer interactions, content moderation, and logistics management, maintaining competitive performance against its predecessors.

Read More At: Anthropic’s Website

Microsoft Introduces Florence-2: A Unified Model for Vision Tasks

Microsoft has unveiled Florence-2, a cutting-edge vision foundation model designed to handle a wide array of computer vision and vision-language tasks. Florence-2 uses a sequence-to-sequence architecture, integrating an image encoder and a multi-modal encoder-decoder. This setup allows the model to perform various tasks, such as image captioning, object detection, and segmentation, without the need for task-specific adjustments.

Florence-2 is built on a comprehensive dataset called FLD-5B, which includes 5.4 billion visual annotations across 126 million images. The model's architecture and training strategy enable it to deliver state-of-the-art performance in both zero-shot and fine-tuning scenarios. It has demonstrated superior results in tasks like referencing expression comprehension, visual grounding, and captioning, outperforming many specialized models after fine-tuning.

This model represents a significant step forward in vision AI, providing a flexible, robust solution for a variety of applications in the field of computer vision.

Read More At: VentureBeat & HuggingFace

AI-Assisted Infrastructure Automation Hype Intensifies

Gartner's latest report reveals that AI-assisted tools for cloud and network automation are reaching peak hype, but advises caution. Technologies such as AI networking and network AI assistants are rated as immature and tied to specific vendors, making them less useful for diverse networks. Generative AI-Augmented CloudOps, while potentially powerful, often contain bugs and hallucinations. Gartner suggests careful evaluation of AI supercomputers and warns against hasty investments in GPUs, recommending new CPUs with embedded AI acceleration instead.

Read More At: The Register

🔨 AI Powered Tools 🔨 

• Blaze - Platform designed for small teams and individuals to automate their workflow processes without needing extensive technical expertise. It leverages AI to streamline operations, improve productivity, and manage tasks efficiently through a user-friendly interface.

• Typetone - AI-powered writing tool that helps users generate content in their unique tone of voice. It assists in crafting personalized and engaging content, maintaining the writer's style and ensuring consistent quality across various writing tasks.

• Rytr - Content generation tool that aids in creating original, high-quality content across multiple use cases like blogs, emails, and social media posts. It uses AI to mimic the user's writing style, ensuring the content is coherent and contextually appropriate.

🗞️ Other News Rollup 🗞️ 

Writer's Growth - AI Writer triples revenue, gains 250 customers, excels in accuracy, and prioritizes trust and security in AI practices.

Chatbots Spread Disinformation - Chatbots easily spread Russian disinformation, lacking reliability in debunking false claims. Concerns arise over fake news propagation.

AI-Generated Video Soundtracks - Google's DeepMind develops AI model for video soundtracks. Other AI models like Kling AI and Runway advance video generation.

Europe Falling Behind - Europe struggles with AI dominance by American models, impacting culture and economy. Efforts for sovereignty face hurdles.

AI Trends: Reality Check - Bessemer's 2024 Cloud AI report reveals key trends: Big Tech AI battles, developer evolution, multimodal agents, vertical AI impact, and consumer cloud revival.

Cancer Screening Innovation - Color Health collaborates with OpenAI for personalized cancer screening, enhancing efficiency and patient outcomes.

Sound Effects Innovation - ElevenLabs releases open-source tool for video sound effects, leveraging AI for custom prompts and immersive experiences.

Predicting Crop Yields - AI counts flowers on trees to predict crop yields, improving efficiency and reducing waste in agriculture.

Gaming - Embracer Group embraces AI for game development, enhancing efficiency and player experience while offering opportunities for developers.

🎶 Prompts 🎶 

Prompts used to generate some of this issues images. Unless otherwise stated all images are using Dall-E and ChatGPT Plus.

Thank You for Subscribing

Enjoying what you’re reading? Help me get better so I can continue to provide you with the most relevant content.