MorelandConnect

I know this is late, but I’ve been busy. :) It was a pleasure speaking at the COSE BIG Summit event about how Generative AI is transforming businesses today and what to expect from Artificial General Intelligence in the future. Thank you Megan Kim for the opportunity and Caroline Guscott for the shout out. Multiple people mentioned how excited they were to incorporate ideas they were learning back into their own businesses. I did have the opportunity to hear Tim Marshall talk about Marketing AI Starter Kit for the Small Biz and Amy Martin talk about Building a Technology Mindset and Tech Stack for the Future.?Fantastic presentations with actionable insights. Inspiring stories were also told by Mike and Bill Malley of Malley's Chocolates , Pat Conway of Great Lakes Brewing Company, Scott Mueller of Dealer Tire , Trina Bediako of New Horizons Baking Company and Maria Bennett of SPR Therapeutics, Inc. ?Great stories about innovation and culture powered by people that enabled them to grow. I enjoyed speaking about real world transformations happening across industries powered by AI.? Some of these examples included: In Software: Amazon migrated 30,000 product applications from Java 8 or 11 to Java 17 using AI tools. The results: Saved over 4,500 years of development work and $260 million in annual cost savings from performance improvements In Healthcare: Aidoc uses an AI-powered platform designed to analyze radiological images in real-time and flag urgent cases. Result: Improved detection rate for critical conditions by 96%, and hospitals reported a 60% reduction in turnaround time for life-threatening diagnoses Alphafold is a deep learning model for protein structure prediction, developed by research group at Google’s DeepMind. Result:?Projected to reduce the 12-15 year drug development time by half. In Manufacturing: Adidas has adopted generative AI in its "Speedfactory" initiative, where AI-powered robots automate the design and manufacturing of custom shoes Result: Adidas reduced product development time by 50% In Construction: Bentley uses digital twins technology for Construction firms. Result:?Construction firms using Bentley’s digital twins reported a 30% reduction in project delays by identifying potential bottlenecks and optimizing workflows In Retail: The North Face uses Generative AI to recommend products Result: 60% increase in engagement and a 75% higher likelihood of customers purchasing products recommended by the AI assistant compared to other methods. In Media: The Washington Post developed "Heliograf," a generative AI system used to write short news stories. Result: AI-driven content also helped increase audience engagement, resulting in a 500% increase in output for certain types of coverage, without expanding the newsroom staff. Exciting to think about what comes next as all the Frontier LLM companies work toward Artificial General Intelligence, which Generative AI is just a stepping stone on the path to.

In celebration of #OhioTechDay, MorelandConnect in collaboration with Case Western Reserve University Quantum Computing Club, Entrepreneurs Club and The App Innovators and Developers Club are organizing an AI Hackathon Oct 12-13th. The MorelandConnect sponsored AI Hackathon is a two-day intensive event designed to engage undergraduate students in cutting-edge AI research and entrepreneurship. This hackathon aims to bridge the gap between academic learning and real-world application of AI technologies, with a focus on conversational AI and prompt engineering. By partnering with student clubs at Case Western, we seek to foster innovation, teamwork, and problem-solving skills while providing students with hands-on experience in AI development and business application. Sign-up link and more details coming soon.

How do you define wealth? This weekend was our 30 year high school reunion from St. Edward High School. Two thoughts keep running through my head. The first is how impactful St. Eds was on shaping me as a person and the adventures I would choose to take later in life because of that experience.?The St. Edward community, The Brothers of Holy Cross, teachers, and friends I made there all influenced me.?For me it wasn’t about religion, and they didn’t push it, what they did, is show you what ethics, morals and values actually look like and gave me something to aspire to. The second thing that keeps running through my head, is the concept of wealth.? For most, the accumulation of money is what comes to mind when defining wealth.? As I get older, I increasingly value a measure of wealth based on relationships.? I’m sure this perspective is influence by my 99 year old father, who says things like, “you can’t take it with you,” lived through the great depression, fought in WWII and has had an inordinate amount of loss and love in his life. The establishment and nurture of relationships that enrich your life and sustain you, is, as many in their final years will attest, a much more valuable measure of wealth. Many of the guys in this picture I’ve known since second grade, many others we gained in high school.? The wildest part is, I’ve heard from at least one if not a dozen almost EVERY DAY for 30+ years.? These friendships are one of our greatest blessings and make us all extremely wealthy. ? What a gift St. Edward continues to offer the young men that pass through these halls. Tim Yaroma, Mark Steinmetz, Brian Laux, Patrick Duffy, Jason Levanduski, Drew Gartner, Ken Kleinhenz, Ben Hatta, Samuel Neider, Tyler W., Bob Adams, Patrick Stanton

This is a NeRF ("Neural Radiance Field") of my son's Nerf (gun... see what I did there!?) Despite what it looks like... this is not simply a video I recorded. Nor is it a 3D model rendered in the traditional sense. NeRFs take regular ole 2D images of a scene or object and use neural networks to learn the color and density of points in space. Once the NeRF is trained it can render realistic and novel views that weren't in the original training images. For these "drone-like" visuals of my living room, the source video was actually a simplistic spin view of the room! With a trained "digital twin" NeRF, a user can create virtual cameras to recapture the scene in a new way. "Render" is a keyword here. The new 3D visualization isn't actually backed by a 3D-mesh like traditional photogrammetry and/or LIDAR-based scans. NeRFs essentially skip creating a mesh and "synthesize" what a high-quality render of a 3D model would look like. NeRFs are a relatively new technology (with the original paper published in 2020). Frankly, the tech is rough and frustrating to work with at this point. The promise of using sparse training images or videos of a scene has not been realized in my experience. Even recent breakthrough implementations like Instant NGP and Zip-NeRF still require thorough video capture to get good results and take a long time to train. But I can't shake my fascination with this technology, its potential applications, and its forthcoming evolution. My mind immediately goes to how it can evolve with the integration of generative AI technology (diffusion models like Flux or SD) and/or by introducing more physics-based constraints. Some relevant news caught my attention today: Viggle.AI announced a $19M round led by Andreessen Horowitz. Viggle claims its competitive advantage is that its AI video generator model is "built with a genuine understanding of physics, enabling the creation of realistic character movements." CEO Hang Chu explained, "The model itself is quite different from existing video generators, which are mainly pixel-based, and don’t really understand structure and properties of physics. Our model is designed to have such understanding, and that’s why it’s been significantly better in terms of controllability and efficiency of generation.” We've played around with Viggle a lot and it is impressive. This hybrid approach echoes the potential I see to improve upon NeRF technology.?NeRFs are a foundational technology, but it will be hybrid models that lead to groundbreaking applications.?Just think about the ability to create digital twins with the already available data or very little additional capture required...