📆 June 25th 2024📆 

A bit longer than usual but some really fascinating items in todays edition.

• Anthropic have pulled ahead of OpenAI in the race of the century - what will be the response? Some say crickets!!

• Musk is partnering up with Dell & Supermicro to build out his own AI hardware, unimaginatively also called Grok.

• Shopify go on an AI announcement bonanza, with spotlight on their Sidekick feature.

• Despite the warnings of many a bad sci-fi movie, researchers are making a gamble on a global hive mind as the way forward - afraid?

• 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!

🆘 Quick Tips 🆘 

Today I’m going to break down the anatomy of a good prompt. With more tools becoming freely available it’s easier than ever to leverage this knowledge to get more consistent results:

• Context - The starting point of every journey is why. Give the model context about why you’re asking it to do something

• Role - Jack of all trades and master of none applies to LLMs too. By specifying the role the LLM should take you are guiding it’s response to a particular style

• Action - What do you want it to do? Brainstorm, write content, proof read. Use Chain of Thought to make the action more predictable. 

• Output - The format you would like the output in. An X thread, LinkedIn post or JSON. If you don’t tell it you won’t get it.

• Example - Another technique in prompting is called Few Shot. This means giving it example to guide it’s behaviour and output.

A short example prompt that follows this pattern is here for you to copy and use

Context: I am preparing to create a new brand for my consultancy business focusing on data & AI. I need a solid plan to help me stay on track with small and achievable tasks. You will provide links to services suggested as I don't have time to do the research. 

Role: You are a efficient task manager who excels at understanding my goals and offering concise tasks to guide me towards achieving them. You believe in the power of small, manageable tasks as a means to complete a project. 

Action: I need to create a new brand website for my consulting business. I want to go live in 4 weeks and am technically savvy to do the work myself. I'm not picky about tool choices so please offer the most efficient suggestions. Develop a plan to hit my release date of 4 weeks. Explain your reasoning step by step and when unsure ask for direction. 

Output: After you have laid out the plan I want to output the entire project plan in CSV for easy import into a project management tool like Notion.  

Example: 
# Task 1 - Brand Name
The most difficult part can be choosing a name. You can use tools like ChatGPT or <insert online link to naming tool> to help brainstorm some ideas. 
1. Think about what your service will offer
2. Align on your values so the name resonates with what you offer
3. Choose a handful of names as availability isn't always guaranteed 

Did you know you can also encode your own specific needs into a GPT and use it for free with GPT-4o? I just did for this example and it’s available for you to use right here.

Get in contact if you want to learn how to make AI work for you.

💫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 📰 

Anthropic's Sonnet 3.5 Outperforms Competitors in AI Model Race

Anthropic's latest AI model, Sonnet 3.5, surpasses OpenAI and Google in various benchmarks, excelling in humour comprehension and complex instructions. The model introduces 'Artifacts' for real-time content editing and boasts enhanced vision processing capabilities. Safety measures include an AI Safety Level of 2, prioritizing privacy. Sonnet 3.5 is accessible via web, mobile apps, and APIs, with plans for future model additions. Anthropic collaborates with safety organizations for model refinement and aims to expand capabilities with memory integration.

Read More At: The Register

Musk Teams Up with Dell and Supermicro for xAI Supercomputer

Elon Musk's xAI startup is partnering with Dell Technologies and Supermicro to build an Nvidia-powered AI supercomputer named Grok. Dell will assemble half of the server racks, while Supermicro handles the rest. The system will utilize 100,000 liquid-cooled Nvidia H100 GPUs and is expected to be operational in North Dakota. Musk hints at future upgrades with Nvidia's next-generation GPUs for enhanced efficiency. This collaboration aims to accelerate the development of xAI’s generative chatbot.

Read More At: The Register

Shopify Introduces AI "Sidekick" Chatbot and Other AI Enhancements

Shopify has launched its new AI-powered "Sidekick" chatbot in early access, part of its "Summer ‘24 Edition" updates. First revealed last year, Sidekick assists merchants with tasks such as creating discount codes, generating store reports, and suggesting blog post ideas. Though these functionalities are typical of AI chatbots, Sidekick aims to provide tailored support for Shopify merchants. Currently, it is available to English-speaking merchants in North America, with plans to expand to other languages and regions.

In addition to Sidekick, Shopify announced several other AI-driven features. One significant addition is AI-powered product categorization, which helps merchants by automatically suggesting the appropriate taxonomy for product listings, thus improving discoverability. This feature can save merchants considerable time by reducing the need for manual input.

Read More At: The Verge

Can Generative AI Create a Global Hive Mind?

The concept of using generative AI to form a global hive mind revolves around leveraging AI's capacity to process and integrate vast amounts of information from diverse sources, fostering collective intelligence. This idea extends beyond simple data synthesis; it envisions a more interconnected and collaborative future for human knowledge and problem-solving.

As we advance, the integration of generative AI with open-source models and user-centric applications will likely pave the way for personalized, context-aware solutions. This approach promises to address the limitations of a one-size-fits-all model, fostering more tailored and effective outcomes for both individual and enterprise users

Read More At: VentureBeat

🔨 AI Powered Tools 🔨 

• Otto - Designed to streamline workflows across various industries, such as research, sales, and finance. It features a table-driven interface allowing users to define workflows visually and automate thousands of tasks.

• Team Create - For building and managing custom AI teams. It caters to various industries by enabling users to assemble AI experts swiftly. The platform leverages AI to match skills and project requirements

• ContentShake - By Semrush is an AI-driven content writing tool tailored for digital marketers and content creators. It simplifies the content creation process by using AI to generate ideas, draft copy, and optimize content for SEO, enhancing productivity and ensuring high-quality, engaging content

🗞️ Other News Rollup 🗞️ 

EU Apple Features Delay - iPhone mirroring, SharePlay, and Apple Intelligence features delayed in EU due to Digital Markets Act concerns.

AI Acquisition Boost - OpenAI acquires Rockset to enhance data retrieval capabilities, strengthening its technology stack in gen AI competition.

Nvidia's AI Challenge - Nvidia's dominance questioned by new AI platforms. Dell and HPE partnerships tackle scaling and costs. VentureBeat's Transform 2024 event explores AI's future.

AI Hardware Faults - Meta warns of hardware faults corrupting AI data, impacting accuracy. Proposes 'parameter vulnerability factor' for mitigation.

AI Sidekick for Merchants - Shopify's new AI chatbot 'Sidekick' helps merchants with tasks like discount codes and product categorization.

RAG Enterprise AI Solutions - DataStax launches Langflow 1.0, Vectorize, and RAGStack 1.0 to improve enterprise RAG applications effectively.

AI Leadership Update - Stability AI appoints new CEO and innovates in AI models despite financial challenges. Exciting developments ahead.

AI Model Jailbreaking - Haize Labs commercializes jailbreaking AI models to enhance security, with Anthropic as a client, using innovative algorithms.

AI-Powered Manufacturing Revolution - GrayMatter raises $45M for self-programming robots enhancing manufacturing efficiency and quality with AI technology.

🎶 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.