Sacha Brouté

At EarthSafeEU, we’re thrilled to highlight another forward-thinking research project pushing the boundaries of subsurface imaging. Led by Luís Tao from Universitat Politècnica de Catalunya in collaboration with University of Twente and SISSA, this project focuses on improving how we visualize what lies beneath the Earth’s surface, a key step in exploring critical minerals and geothermal energy. One of the biggest bottlenecks is the high computational cost of interpreting 3D Magnetotelluric (MT) data. To address this, the project develops Reduced-Order Models (ROMs), which significantly reduce simulation time without compromising the accuracy required for reliable subsurface imaging. This advancement enables faster, scalable, resource-efficient workflows to better locate critical mineral deposits and geothermal resources; and contributing to more sustainable and responsible resource exploration. By improving our ability to image the Earth’s interior, this research supports global efforts to secure critical raw materials and harness clean, renewable energy. Interested in connecting or learning more? 📩 Reach out: earthsafe-itc@utwente.nl 🌐 Visit: https://earthsafe-dn.eu This project is co-funded by the European Union under #HorizonEU (HORIZON MSCA-DN 101120556) and UK Research and Innovation (EP/Y030265/1). Views and opinions expressed are those of the project team and do not necessarily reflect those of the European Union or UKRI. #SubsurfaceImaging #Magnetotellurics #GeothermalEnergy #CriticalMinerals #ROM #HyperReduction #EarthSafe #CleanEnergy #HorizonEurope #MSCA

40

Beyond the theoretical: Why #devbcn26 focuses on "The How." 🛠️ In the world of software engineering, there’s a massive gap between a tutorial and a production environment. At the Barcelona Developers Conference, we bridge that gap. Our mission for the 2026 edition is clear: provide a stage for real-world experiences. We prioritize talks that share: ✅ Scaling challenges with Kubernetes and Cloud ✅ Performance tuning in the JVM ✅ Integrating AI & Machine Learning into existing stacks ✅ Building resilient Frontend architectures Whether it’s a 50-minute deep dive or a 20-minute lightning session, you’ll walk away from the World Trade Center Barcelona on June 17th with practical knowledge you can apply on Monday morning. Previous editions have hosted 1000+ attendees, and the energy of the Barcelona tech community is unmatched. 🌊 ⚠️ Reminder: We are halfway through the Blind Ticket phase! Prices increase after March 1st. Secure the best rate for your team now. 👉 Register here: https://buff.ly/q5Wt1YM #devbcn26 #SoftwareArchitecture #TechnicalDebt #Java #DevOps #EngineeringManagement #BarcelonaTech #CloudComputing

8

It's crucial for future engineers to understand the legal and ethical dimensions of using AI tools. SkincAIr project is a perfect example of how an AI tool can be used for a good cause, ensuring that technology benefits humanity while respecting everyone’s rights. Thanks to everyone who asked questions and participated in this important conversation!

18

MoltBot faced a challenge: digital access blocked. What did it do? It autonomously found the phone number and deployed a voice synthesis module. It dialed the restaurant and negotiated a time, using human-like pauses and "ums" to sound natural. AI's ability to adapt and solve problems in real-time is astounding. Full Audio at https://lnkd.in/gSTpx6Ap. Want to reach 60,000+ Enterprise Architects and C-Suite leaders? Download our 2026 Media Kit: https://djamgamind.com/ai #AI #Automation #Innovation #Tech

1

The MoltBook/MoltBot discourse lately has been... interesting. On one side we see big claims and even stronger reactions. A lot of “this changes everything” energy, but the actual question still stands: what was genuinely achieved? We’ve seen this pattern before: oversell → spark anxiety → ride attention → postpone proof. Crypto was going to replace the entire financial system overnight. Web3 was going to kill platforms and middlemen instantly. In each case, real progress did happen - just nowhere near the magnitude implied by the hype. This doesn’t mean MoltBook/MoltBot isn’t meaningful. It may be a solid step forward, but framing it as "AGI" while fearmongering in tech circles, “jobs are done”, “developers should panic", beclouds the line between innovation and narrative. Real breakthroughs don’t need manufactured urgency: You can see them in what they actually do, how consistently they are, and whether the results hold up outside of a beta. If this is truly a step toward AGI, the proof will age well, If not, it’ll join a long list of things that sounded bigger than they were. What do you think?

11

1 comentario

Today, Íñigo Martínez de Rituerto de Troya presented a core piece of his PhD work at ACM FAccT in Athens. In “Misabstraction in Sociotechnical Systems”, we provide a richer perspective on how abstractions are used in the design, development, analysis, use and governance of sociotechnical systems. While, in recent years, the societal implications of AI and algorithmic systems have led to an enriching of scholarly, advocacy and policy efforts to better understand and address these, we still lack shared understandings of how we are or should be abstracting AI systems and their effects, not just at the technical level, but also in how these systems are embedded in operational contexts where they impact lives and professions, how organizational actors and processes go about designing/maintaining/governing/overseeing/etc, and the institutional arrangements that apply to or otherwise affect these practices. As Íñigo concluded, we hope this paper helps you think about abstraction differently. Furthermore, there is a lot of follow-up work possible to further crystallize conceptual and methodological approaches to guide abstraction in the growing efforts to operationalize AI governance that allow for effective AI oversight but also better system design, in ways that represent and empower the needs and concerns of those bearing the brunt of consequential AI systems. Read the paper here: https://lnkd.in/eQHU4dDn With Jacqueline Kernahan, Neelke Doorn and Virginia Dignum.

89

3 comentarios

💧 Can open data help us predict future water usage? As part of a group project, we analyzed public data from the Transparency Catalonia initiative to build a linear regression model predicting domestic water consumption in Catalonia. 📊 Using data from 2019 to 2022, our model predicted 2023 consumption with just a 5.53% error, proving that even simple models — when paired with open datasets — can support smarter environmental policy decisions. 📎 Our anonymized article (PDF attached) explains the method and results in detail — a great example of using data science for sustainability. #OpenData #WaterConsumption #DataScience #EnvironmentalModeling #LinearRegression #Sustainability #TeamWork #Catalonia

7

AI has fundamentally changed software engineering in under a year. I no longer write code; I architect agentic workflows. In this article, I experiment with just how big of a task can AI take on in a single prompt. Taking advantage of Claude Code's subagent architecture, I automated test identification, generation, and review for a 100k line codebase. The automated feedback loops worked but the ROI didn't. Check out the article to see what I learned! Very proud to make a contribution to the Metropolis engineering blog.

36

9 comentarios