Audio Deepfakes, EU AI Act, Cognition's Devin release & more...

We take a closer look at - MIT Insights on Audio Deepfakes, EU Parliament's AI Act, Cognition release Devin, and more that happened this week in the world of AI

🧠 Synthminds.AI Weekly: The Essence of Innovation and Unity

Join us as we continue our weekly exploration of artificial intelligence's vast landscapes. This week we take a closer look at:

Insights from MIT about Audio Deepfakes, The EU Parliament’s new AI Act and what a top Think Tank shares about it, Devin the AI Software Engineer, new videos from Goda and Joseph, plus the truly exciting Retrieval Augmented Thoughts method which greatly reduces AI “hallucinations”

Read on, and also, visit us online: https://www.synthminds.ai/

đź’ˇ Synthminds in Action:
This Week's Highlights…

Goda reviews Devin, a brand new AI tool from Cognition who promote it as an “AI Software Engineer”. Math prodigy Scott Wu, the creative trailblazer who built it, is definitely causing a stir . See what he’s created and hear where Goda believes this can take us.

15th in the short form video series, Joseph (Professor Synapse himself!) shares his tips for consolidating calendars, better time management, and how to effectively organize tasks using tags and inbox in Obsidian.

Synthminds - Voices in AI: Mira Murati

Ever wondered about the brilliant minds behind AI's biggest breakthroughs? Meet Mira Murati, the CTO of OpenAI. Joe and Eldad discuss why she’s more than a Tech leader, and someone you should know more about…

🌍 AI in Focus: The Stories from the Week we believe are most important for you to know

MIT's latest insights: From deceptive robocalls to safeguarding privacy and enhancing healthcare, Audio Deepfakes are shaping our digital future. What are the risk, and potential benefits?

Understand how mastering the language of machines is crucial in a data-driven world. Explore how Prompt Engineers will shape the strategic use of AI in businesses of all sizes.

EU Parliament passes new AI Act - is this a landmark law?

There’s plenty of buzz about the EU passing their Artificial Intelligence Act and what it could mean for the AI race between nations. Here’s the EU page about it, and the view from Chatham House, the Royal Institute of International Affairs Think Tank. EU laws, like GDPR, have global impact. So, what are the “outright ban on certain applications and significant obligations on high-risk AI use” that just got passed?

Regulators around the world are urgently trying to hire AI experts to help study and regulate powerful new AI models. But will the salaries governments pay be enough to help them keep up with the industry?

Trained and tested across nine different video games in collaboration with eight game studios. SIMA is a new generalist AI agent that can follow instructions in 3D video games, understand natural language instructions, and perform various tasks. The researchers hope SIMA can be used to develop more helpful AI agents for real-world applications.

🌍 Industry Spotlight: HEALTHCARE

Organized by the Healthcare Information and Management Systems Society (HIMSS), this important conference is the industry’s G20 summit. It provides an opportunity for healthcare professionals, vendors, and industry experts to learn about new technologies and discuss emerging trends in healthcare IT.

✍️ Go Deeper with our Thought Leaders🔍️ 

Amy Zolotow, our AI in Healthcare Specialist reports from the floor of the show - “Biggest takeaway from HIMSS 2024: Forget what you know about healthcare today. AI is and will continue revolutionizing care beyond conventional medicine, allowing us to refocus our attention on where it needs to be - prevention, early detection, and decentralization of medical services.

🧾 AI Word of the Week: Computer Vision

Simplifying Complex terms so everyone - beginner to advance - can better understand AI, ML concepts and terminology

Computer Vision - Using AI to process and analyze visual data. Computer vision applies AI to process digital images and video to identify, classify, and understand what is depicted. The process involves CNNs - Convolutional Neural Networks. CNNs are designed to recognize patterns in image data and can be trained using large amounts of labeled data.

Getting more Advanced: Computer vision enables computers to extract high-dimensional data from digital images and video to identify and classify objects, locations, actions, text, and more. Key techniques include neural networks like CNNs, feature extraction, object recognition and image segmentation. The convolutional operation in CNNs helps to extract features from an image by performing a series of filters and pooling operations.

Context/Relevance: Computer vision powers facial recognition, image captioning, medical imaging analysis, autonomous vehicle navigation, security surveillance and other visual perception tasks.

Please get in touch to see more of the AI Glossary that we’re compiling. There’s 117 terms and growing. Share what you want to see added to the list

🚀 #52weeksofAI - with Wes Shields đź’ˇ

Generating the Same Character/Subject Across Multiple Images is Now Here! This week, Wes tried out the new CREF tag that Midjourney released. This new capability helps for when you need to have a consistent character across multiple images. This has been a serious & severe limitation across AI-image generation, and this is a key step making it easier to get consistency.

Join Wes in this weekly exploration as he tests new tools & prompts. Get motivated to experiment with these cutting-edge models yourself.

🎓 Research and Discoveries:
Actionable Insights You Can Use

The top article this week from the latest research and practical applications of AI shared within our community. From theoretical breakthroughs to hands-on applications, this is the one thing to be sure to read about the evolving landscape of AI technology.

Purpose: Introducing RAT - retrieval augmented thoughts, an innovative approach crafted to refine AI's reasoning by integrating a mechanism for accessing relevant external information. ​ The RAT method aims to revise each thought step one by one with retrieved information relevant to the task query, the current and the past thought steps, after the initial zero-shot CoT (chain of thought) is generated. This method significantly curtails AI's propensity for "hallucinations," and serves to enhance its reasoning and generative prowess across a broad spectrum of long-horizon tasks.

Key Findings, Insights/Implications: RAT marks a potential paradigm shift in AI, reducing errors and deepening reasoning capabilities. It exhibits remarkable performance improvements, up to 42.78% in diverse tasks such as coding and strategic planning. This leap in AI utility showcases RAT's potential to revolutionize how AI is applied, moving from mere automation to a rich, informed problem-solving partner that leverages the vast expanse of human knowledge for more accurate and dependable outputs.

Framing the Discussion: RAT's introduction leads to a future where AI aids in decision-making and problem-solving with better accuracy. Its importance cannot be overstated, especially in critical sectors like finance and healthcare, where the cost of misinformation is high. RAT is not just an upgrade; it's a transformative approach that redefines how AI interacts with the wealth of human knowledge . It helps ensure AI-generated solutions are informed and reliable. It looks to herald a new era of AI that's capable of tackling complex, real-world challenges with layered understanding and precision.

Putting it into Daily Workflows: Using RAT in daily workflows transforms AI from a tool of convenience into a cornerstone of strategic decision-making and innovation. For marketing professionals, it means creating more targeted campaigns based on deep data insights. In leadership and operations, it paves the way for more informed policy-making and efficient process optimizations. For finance, RAT enhances risk assessment models with greater accuracy. Developers can produce code not just faster, but with fewer bugs, by leveraging RAT's comprehensive understanding. Healthcare professionals gain a research assistant that offers up-to-date, evidence-based advice. Across all fields, RAT empowers professionals to make decisions not just based on data, but on the depth of human knowledge, making AI a true partner in creation and problem-solving.

🤝 Together, We Innovate

At Synthminds, we believe in the power of collaboration to unlock the full potential of AI. Your insights, experiences, and curiosity are the fuel that drives our community forward.

We very much wish for this newsletter to be what you want to read weekly. Please share your feedback and ideas with us. Thanks for reading and join us next week!

Reply

or to participate.