Make Money With Your AI Skills!

GM, my AI-obsessed friends.

Welcome back from the long weekend. This is the latest edition of /Imagine, your go-to source for the latest news and developments in AI. In this edition, we bring you some of the most exciting news and tools since Friday. Let’s dive in…

High Level: Hive3 AI - The World’s First Competitive AI League Where You Can Earn Real Cash! 💰

The Bytes: 

✣ Weekly AI Challenges: They drop weekly AI image and video challenges sponsored by brands and anyone can enter. Creators are given a brief and usually a few days to make something special with any available AI tools.
✣ Rewarding Talent: Winners of these challenges receive epic cash prizes in the thousands of dollars. My wife and I have dabbled in a few contests thus far and already won $4,000. Not half bad for playing around with AI.
✣ Brand Collaboration: Brands actively participate, offering unique opportunities for AI creators to engage in real-world applications and potential private project hires.
✣ The League: If you place 1st, 2nd, or 3rd in their competitions you are entered into The League Playoffs for a chance at $250,000 in cash prizes.

Why You Should Care: Hive3 AI isn't just a competition; it's a nexus where AI creativity meets commercial interest. This platform provides a unique opportunity for AI talents to showcase their skills, gain recognition, and potentially collaborate with leading brands. For anyone interested in the future of AI, Hive3 AI represents a fascinating glimpse into how AI can be applied, developed, and celebrated in a competitive yet collaborative environment. By the way, this wasn’t sponsored and I’m not affiliated with them in any way. I just think it’s badass and more people need to know about it and participate in it — especially our /Imagine tribe.

Join Hive3’s Latest Challenge

High Level: AI in Healthcare — Transforming Diagnosis and Patient Advocacy

The Bytes:

✣ Google's AMIE (Articulate Medical Intelligence Explorer): An awesome new AI chatbot developed by Google Research and DeepMind, AMIE is revolutionizing medical diagnostics with its ability to engage in detailed, differential diagnosis through simulated dialogues. It outperforms human doctors in many aspects of diagnostic accuracy and consultation quality in simulated environments.
✣ House GPT in Patient Advocacy: Mirroring this technological leap, a tweet went viral on Friday by Patrick J. Blum where he discusses leveraging his House GPT's capabilities in managing rare diseases. His House GPT assists in analyzing vast amounts of medical data, uncovering insights that often elude traditional consultations.
✣ Empowering Patients with AI: Both AMIE and House GPT exemplify how AI can act as a powerful ally for patients, has no time limit, and won’t dismiss your questions . While AMIE excels in diagnostic precision, House GPT enhances patient agency, enabling them to navigate their health journeys with more informed and empowered decision-making.
✣ Future of Healthcare AI: AI advancements indicate a shift towards more personalized, accurate, and patient-centered healthcare. They not only demonstrate AI's potential in enhancing medical consultations but also in supporting patients in understanding and managing their health conditions.

Why You Should Care: Obviously, don’t let this replace your healthcare provider, and this isn’t intended to be medical advice. But, the integration of AI like AMIE and House GPT in healthcare represents a significant leap in medical diagnostics and patient advocacy. They offer new possibilities for accurate diagnoses, comprehensive health analysis, and empowered patient engagement. This evolution is pivotal for anyone interested in the future of healthcare, especially for those navigating complex health conditions, underscoring the potential of AI as a critical tool in both clinical and personal health management contexts.

High Level: Pioneering AI with Historical Insights — Brian Roemmele’s Initiative

The Bytes:

✣ Extraordinary Data Source: Roemmele, and AI enthusiast (and fantastic & interesting person to follow) started a project that involves approximately 385,000 vintage pieces from his personal collection, including magazines, newspapers, and publications, many of which are exclusive and not digitally available.
✣ Holistic Training Approach: The AI model is being trained comprehensively on all elements of these materials - text, imagery, and even ads - providing a rich and multifaceted dataset entirely different from the internet driven models we’ve come to know.
✣ Remarkable Preliminary Outcomes: According to Roemmele, preliminary tests of this AI model show distinct and intriguing results, revealing a distinct "can-do" and "do-it-yourself" ethos, and language reminiscent of the late 19th to mid-20th century.

Why You Should Care: Roemmele's endeavor goes beyond creating a new AI model; it's about instilling a new worldview in AI, and also preserving the past. This unique approach equips AI with the wisdom and optimistic outlook of earlier generations. According to Brian, this AI model doesn't just process queries; it interacts with a depth of understanding and practical know-how that could revolutionize how we interact with AI, offering insights and solutions grounded in a rich historical context. I’m looking forward to giving it a try when he makes it public.

High Level: Consider Running AI Locally on Apple Silicon

The Bytes:

✣ Shift Towards Decentralized & Local AI: Phixtral exemplifies a growing trend and interest towards local AI processing. This shift prioritizes data privacy, reduced latency, and independence from cloud-related vulnerabilities, marking a significant change in the AI landscape.
✣ Mixture of Experts: Phixtral innovatively combines 2 to 4 fine-tuned Microsoft Phi-2 models, to create an MoE model that is lean and efficient. This configuration surpasses the performance of individual experts, showcasing a significant advancement in MoE design.
✣ Leaderboard Dominance: On the 'Yet Another LLM Leaderboard' (YALL), Phixtral stands out, performing impressively against base phi-2 models and nearly matching the top-ranked Zephyr2-7B model.

Why You Should Care: 

Phixtral's on-device AI processing capabilities signal a major shift in the AI field, moving towards more personalized, secure, and efficient AI experiences. The model's ability to run sophisticated AI processes locally offers enhanced privacy and real-time performance, independent of continuous internet connectivity. This transition not only makes AI more accessible but also aligns with sustainable computing practices by potentially reducing the energy demands of large data centers. Worth paying attention to!

All the credit to Maxime Labonne, a talented machine learning scientist at JPMorgan for creating the model, sharing it, and teaching people about it.

/Imagination of The Day: Let's end on a creative note. Right now, I'm giving Hive3’s latest Beachclub Hotel competition a go! Can't decide between two submissions—help me choose which I should do!

Prompt Share: a Mexican resort that blends indoor and outdoor, beautiful woman in a bathing suit looking outward, striking vivid pink limestone walls, and bronze as accents, color theme in the style of wes anderson, cinematic lighting, hyper realistic photograph --ar 7:4 --stylize 750 --v 6 

Depending on when you read this, there may be time to enter! Check out the latest competition here!

Stay Curious: Don't forget to follow me on Twitter at @jake_joseph and the brand on its fresh new page @slashimagineai for the latest AI news, my two cents, and discussions. Share your thoughts and join the conversation.

Share The Insights: Know someone else who's obsessed with AI? Forward this newsletter and let's expand the community.

Feedback: Your perspectives help shape the content, and hopefully helps me make the newsletter suck less every day. Share your feedback on this issue or what you'd like to see next by directly replying to this email or by reacting to the poll below.