Jian Huang

Over the holiday break, I spent some time revisiting some of the foundational work shaping the field of scientific AI. One paper that stands out—published back in 2021—was co-authored by Damien Guého, now a Senior R&D Engineer at Geminus. The paper tackles a timeless challenge: how do you uncover the true physics of a system when all you have is sparse, noisy, imperfect data? Damien and his co-author's approach is both clever and elegant. Instead of a brute force deep learning strategy, they leveraged clever math to uncover a physics-grounded structure with interpretable dynamics—from messy, nonlinear signals. The result is a robust, reliable and generalizable solution that uses a fraction of the computational effort. They first solved the problem using a deep learning approach that required 7,720 parameters. They then applied their method that used just 3 terms and was 10,000× more accurate. This mindset is core to Geminus' mission to simplify engineering and operational decision making in complex systems where uncertainty abounds and data is sparse and messy. Work like that of Damien and his colleagues continues to inspire where we’re headed next. Proud of the team at Geminus Labs—and excited for what’s coming. Check out his paper here. https://lnkd.in/eHt-5tXX Watch this space - be sure to follow Geminus to be part of the conversation.

32

xAI’s reported $20B investment in Southaven is an exciting milestone for Mississippi and a strong signal of the state’s growing role in the AI economy. Investments at this scale create opportunities not only for infrastructure and jobs, but for building long-term talent and innovation capacity. Through the Mississippi Artificial Intelligence Network (MAIN), we’re focused on ensuring Mississippians have access to no-cost pathways to gain the skills needed to participate in and benefit from this momentum statewide. #xAI #AI #Mississippi #MississippiAI #MAIN

32

Yael Grushka-Cockayne leading an interesting panel on Building AI Infrastructure in Virginia at VirginiaAI Summit hosted by Virginia Academy of Science, Engineering and Medicine. Panelist: Lingjia Liu, Milos Manic, Nathan Mellis, and Manish Parashar. Key takeaways • Virginia is emerging as a national leader in AI infrastructure. • Community and environmental impacts must be designed in from the start. • Cybersecurity is driving real-world AI use cases—and the infrastructure required to support them. #AI #AIInfrastructure #Cybersecurity #Sustainability #Virginia #VirginiaAI

48

2 Comments

DeepSeek AI published a technical paper a few days ago, co-authored by co-founder Liang Wenfeng that argues for a new architecture to train foundational models. Essentially moving away from 'muscle head logic'. As the writer of the article I linked below by Larry Dignan states, The DeepSeek paper is a bit wonky. But remember what DeepSeek did to the market a year ago when they released their LLM - it was very performant and low cost. The AI market to date has been driven by muscle head logic. The answer for everything in the last two years has been more muscle. In AI's case that approach means more GPUs, more farm land going away, more energy usage, more water and by all means more data centers, debt and capital expenditures. This approach and consumption is likely to change with deep technical architectural change including at the chip level. Price and performance matter. LLM giants are diversifying their chip bases. Google's TPUs and Amazon Web Services' Trainium are going to be threats to Nvidia as inference becomes the main event. It will be really interesting to see how the industry shapes up here - lots to follow and very interesting! News Article here with link to DeepSeek paper: https://lnkd.in/eN38mSAV

6

3 Comments

🔥 Innovation or Injustice? 🔥 Elon Musk’s AI company, xAI is under scrutiny—not for model hallucinations, but for real-world pollution. In Memphis, a new AI supercomputing facility—Colossus—has drawn concern from community leaders and the NAACP who’ve raised alarms about its environmental impact. The facility relies on 35 methane gas turbines, emitting pollutants in a historically underserved area—prompting questions around air quality, permitting, and public health. 💡 This isn’t just about AI—it’s about how and where we build, and whether communities are part of the process. ⚡ Training a single large AI model like GPT-3 consumes over 1,287 MWh—enough to power 130 homes for a year. Multiply that across global infrastructure, and the environmental stakes become clear. 🤨 Organizations like the AI Now Institute have long called for greater scrutiny around the societal impacts of large-scale AI systems. Meanwhile, local stakeholders like the Greater Memphis Chamber are navigating the tradeoffs between innovation and public accountability. As AI infrastructure scales, are we asking: • Who’s consulted in site decisions? • What environmental and community guardrails are in place? • How do we ensure innovation uplifts rather than burdens? 📌 The future of AI isn’t just technical—it’s civic, ethical, and local. ❓What does responsible AI infrastructure look like to you? #AIethics #EnvironmentalImpact #LeadershipInTheLoop #ResponsibleAI #AIgovernance #InfrastructureEquity

3

The first of the FrontierMaths Open Problems to be solved entirely by GPT 5.4 PRO was based on Ramsey hypergraphs and was set by mathematician Will Brian, an Associate Professor at UNC Charlotte. The framework was orchestrated by humans who cheekily wrote authored by GPT 5.4 PRO. Consequentially, the future of maths and physics is set to become more and more artificially intelligent.

111

4 Comments

Ohio State University's AI Fluency initiative (https://lnkd.in/etB8EyMB) sets an ambitious goal: every undergraduate will graduate with practical generative AI skills by 2029. Let's unpack this: ✅ Strong Points: ▪️Bold, clearly articulated institutional vision. ▪️Multi-layered integration (Gen Ed courses, discipline-specific modules, innovation opportunities). ▪️Dedicated resources and infrastructure (faculty support via Drake Institute). ▪️Commitment to hands-on innovation (hackathons, prototyping workshops). ⚠️ Areas of Concern: ▪️Faculty workload: Realistic expectations for widespread course redesign and adoption? ▪️Standardization & Assessment: What does "AI fluency" mean in practice across diverse fields? ▪️Curriculum Overload: Risk of student pushback if mandatory modules aren't streamlined or clearly valuable. ▪️Equity and Faculty Incentives: What's the motivation for faculty, especially those skeptical or hesitant about AI integration? What if they're not on board with using AI in the classroom, or it doesn't make sense for their course? ▪️Soft Skills and Ethical Frameworks: How thoroughly are ethics, interpretability, and responsible AI practices embedded? OSU's vision is exciting, but effective execution will be key. I'm curious: What does successful "AI fluency" look like in your discipline? And how should institutions balance ambition with realistic implementation? Let's talk!

28

11 Comments

Thrilled to announce a major milestone for Layer10! We've officially signed a strategic partnership with Notre Dame University - Louaize (NDU) and A.N.Boukather Holding to build Lebanon's next generation of AI talent. This isn't just another MOU that sits in a drawer. Selected NDU students will join us as full-time interns, working on real AI projects in automation, data engineering, and applied machine learning—specifically within the automotive sector where ANB has been pioneering AI adoption across Lebanon and the UAE. What makes this special: we're combining ANB's century-long business legacy with Layer10's technical expertise built through projects in the US and UAE, and NDU's commitment to connecting students with real-world professional opportunities. Students get mentorship, access to real operational environments, and hands-on experience solving actual business challenges. Big thanks to Anthony Boukather and Nicolas Boukather for believing in this vision and for ANB's commitment to investing in Lebanese talent rather than watching it leave. And to my cofounder Carl Nehme for making this happen. Lebanon has no shortage of brilliant young minds. What we need are more bridges between universities and industry. This is one of them. Excited for what's ahead. https://lnkd.in/eRUVUvbg

80

3 Comments

The new frontier for AI is tabular data and predictive AI. Recent advances in tabular foundation models, combined with graph-based approaches to aggregating and compressing relational datasets, are starting to look very promising. https://lnkd.in/gm6bd_25

4