We hooked up a multi-agent physics engine to a bounded 3D lattice to see if it could survive a massive audio drop. To keep the math from exploding, the agents dynamically calculate phase shifts to distribute the kinetic energy, pushing the system past a non-linear threshold where chaotic ripples suddenly phase-lock into a stable, synchronized geometric standing wave. resonant.sh --> see if you can break the grid! We're also rolling out data export pipelines this week so you can generate training data for multi-agent alignment that can run be stacked on top of your existing training runs. #PhysicsSimulation #CreativeCoding #MultiAgentSystems #NonLinearDynamics #AudioReactive

Yes, I'd like to have a learned physics simulator (gravity, fluids, collisions, lighting, object permanence, etc.) that is 10-1000x faster than classical rule-based solvers, while being accurate enough for engineering use. All emergent from data, not coded. "Move 37 moment" for physical superintelligence. #GoogleDeepmind

The value of a structured and robust computational engine to engineering innovation cannot be overemphasised, considering the values it brings to efficiency, speed, and scalability. This Series 2 of SEPAL Solver explores complex, real world engineering behaviour through advanced computational modelling: ✔️ Multi-body dynamics with a six-linked bar mechanism ✔️ Marine stability analysis through ship roll simulation ✔️ Nonlinear chaos in a double compound pendulum ✔️ Behavioural divergence driven by mass variation ✔️ Electrical machine simplification using the park transformation SEPAL Solver delivers a unified scientific computing environment engineered within C# ecosystem, enabling scalable simulations, stable numerical execution, and seamless integration into production systems. From mechanical dynamics to electrical transformation models, SEPAL Solver transforms complexity into clarity. 📩 ng@cyphercrescent.com 🌐 www.cyphercrescent.com #CypherCrescent #SEPALSolver #AdvancedComputation #EngineeringSimulation #Computing #Algorithm

How would you prove you can build something great with a small-scale model? That is the essence of Project M-2, led by our director Marc Sanchis. Recently, we faced a fundamental challenge: our previous prototype was functional, but it failed to convey the true magic of a particle accelerator. It lacked that visual component that allows complex concepts to be understood intuitively. 𝗧𝗵𝗲 𝗣𝗿𝗼𝗯𝗹𝗲𝗺 The previous version did not effectively demonstrate the actual physics behind an accelerator. If you can not see it, it will be difficult to understand how it works. 𝗧𝗵𝗲 𝗦𝗼𝗹𝘂𝘁𝗶𝗼𝗻 Develop a new model capable of generating a real particle beam visible to the naked eye that intuitively demonstrates how particles behave inside a particle accelerator. This shift is a critical step in validating our technology and proving we are ready for the next level. As Marc says: "𝘐𝘵’𝘴 𝘢𝘣𝘰𝘶𝘵 𝘣𝘶𝘪𝘭𝘥𝘪𝘯𝘨 𝘴𝘰𝘮𝘦𝘵𝘩𝘪𝘯𝘨 𝘴𝘪𝘮𝘱𝘭𝘦 𝘵𝘩𝘢𝘵 𝘱𝘳𝘰𝘷𝘦𝘴 𝘸𝘦 𝘢𝘳𝘦 𝘤𝘢𝘱𝘢𝘣𝘭𝘦 𝘰𝘧 𝘣𝘶𝘪𝘭𝘥𝘪𝘯𝘨 𝘴𝘰𝘮𝘦𝘵𝘩𝘪𝘯𝘨 𝘣𝘪𝘨." We were inspired by the work of Kramnik Daniel for this model. Video source: https://lnkd.in/edyFnTuF #LINAC #UPV #GeneracionEspontanea

Why High-Speed Imaging Is Always a Compromise High-speed imaging is not about having the fastest camera. It’s about making the right compromises to capture the right physics. You cannot maximize everything at once. Every setup is a trade-off between: • Frame rate • Exposure time • Light • Resolution • Recording duration What this means in practice To capture a shock wave: • You push frame rate → 100,000+ fps • You shorten exposure → microseconds ✔ You freeze motion ❌ You lose light ❌ You reduce resolution ❌ You shorten the recording time This is not a camera problem—it’s physics. The part most people miss: geometry Your Field Of View(FOV) is governed by: FOV = (Sensor Size × Distance) / Focal Length At ultra-high speeds: 👉 The sensor effectively shrinks (cropping) 👉 Your field of view collapses 👉 You may completely miss the event Common mistakes ❌ Expecting detail from wide lenses ❌ Ignoring sensor crop at high FPS ❌ Underestimating lighting needs ❌ Trying to maximize all parameters At PETS, we focus on turning these principles into practical tools and training for engineers, blasters, and researchers. If you’re working with: • Shock waves • Detonation physics • Fragmentation • High-speed diagnostics Let’s connect—or share your experience below 👇 #HighSpeedImaging #Engineering #ExplosivesEngineering #Optics #Research #STEM #DefenseTechnology #PETS

New Post: Real‑Time Thermal Contrast Enhancement for Low‑Cost Infrared Imagers Applied to Early‑Stage Crack Detection in Industrial Pipeline Welds - https://lnkd.in/gu4gw2Ca Thermal Contrast Enhancement for Low‑Cost Infrared Imagers Applied to Early‑Stage Crack Detection in Industrial Pipeline Welds **Abstract** We introduce a commercially viable, end‑to‑end pipeline that leverages physics‑based radiometric modeling, deep convolutional reconstruction, and reinforcement learning‑guided sensor calibration to achieve sub‑millikelvin thermal contrast in ultra‑compact infrared cameras. The system is validated on a dataset of \[…\]

Part 2/4: The Post-Silicon Era: Photonic Processing ⚡ If pushing electrons through copper wires is generating too much heat and hitting physical limits, what if we compute with light instead? In my second MATLAB hardware simulation, I modeled a Photonic Integrated Circuit (PIC). Instead of discrete binary logic, this hardware computes using the physics of light interference. Watch the simulation: We take two input lasers, alter their phase, and compute the output based on how the waves physically interact (Constructive/Destructive Interference). Because this is analog and optical, it processes continuous data almost instantly with near-zero heat generation. This is how we achieve massive bandwidth without melting the processor. Up next: What if we make memory act like a human brain? 🧠 #Photonics #Optics #ElectricalEngineering #MATLAB #HardwareDesign #DeepTech #FutureTech

📱 A phone on a hoverboard… measuring real physics. In this experiment, my student attached a mobile phone to a hoverboard to investigate circular motion using real-time data 📊. Using the Phyphox app and the phone’s built-in sensors, we measured centripetal acceleration while the system was rotating, and compared it with angular velocity. The result? The data clearly confirms the relationship: centripetal acceleration ∝ ω². This is what physics should look like: not memorizing formulas, but testing them, measuring them, and proving them 🚀. #Physics #STEMeducation #IGCSE #ALevel #ScienceTeaching #EdTech #Phyphox

Started a reservoir simulation run this morning before my coffee was done brewing. ☕ ECHELON from Stone Ridge Technology continues to impress — clean startup, GPU acceleration and off to the races. Multithreaded well solve, compositional physics, no drama. In a field where simulator setup can eat half your day, that kind of simplicity is a real competitive advantage. Fast software = more time for actual engineering. Highly recommend giving it a look if you haven't. #ReservoirEngineering #ECHELON #GPUSimulation #Subsurface #EnergyTransition

For those in Reykjavík this Friday, I’ll be speaking at the Intelligent Instruments Lab Open Lab about our ongoing "radial drum sequencer" prototyping (see my previous post). Beyond the prototype, I’ll also share the motivation behind a broader idea: that sound objects may be better understood as dynamic, relation-based phenomena, shaped by interaction and context, rather than as fixed entities described only through classical-physics metaphors. Hope to see some of you there. Here's the link: https://iil.is/openlab/119