News & Community Highlights
Dyad 3.4 is Here With a Dedicated Desktop Environment: With the much-awaited 3.4 release, Dyad Studio is now available as an independent standalone desktop application. The entire platform arrives in a unified, ready-to-run environment, complete with a bundled Julia toolchain, native source control, and autonomous AI delegation, designed to get you straight into modeling from day one. Dyad 3.4 allows you to upgrade existing projects easily with one-click migration; opening your workspace in 3.4 prompts an automatic migration that updates and pins the latest library versions across all mechanical, electrical, thermal, and fluid domains, including DyadModelOptimizer v15.4.0. Find out more.
The Best AI Models Fail at Physics: Coding Harnesses are to Blame: What happens when the same frontier AI model tackles the same physics problems inside two different agentic workflows? In a new study, Chris Rackauckas, Anas Abdelrehim, and Anshul Singhvi found that Dyad Agent scored 0.899 compared with 0.533 for Claude Code, at nearly identical cost and runtime. The difference came down to the harness, meaning the tools, documentation, and verification steps surrounding the model. Find out more.
Juia in TIOBE Top 20: Julia has returned to the TIOBE top 20, ranking 18th in October 2026. TIOBE attributes its renewed popularity to demand for languages that combine high performance with accessibility in numerical computing, a combination central to JuliaHub’s work in scientific computing and engineering.View the October index.
Next-Generation System Simulation - Paris Workshop: Join Dr. Michael Tiller and Dr. Michael Hoffman on October 20, 2026, at Cocoon Space, Paris, for a hands-on workshop on AI-assisted system simulation. The morning session covers using AI agents to build models, automate simulation sweeps, and diagnose errors. The afternoon explores Scientific Machine Learning (SciML), combining differential equations with neural networks for surrogate modeling, system identification, and parameter estimation. The workshop offers full-day, morning-only, or afternoon-only registration. Register here.
One HVAC System Model, No Design Tooling Friction: HVAC and refrigeration engineers often juggle four or five separate tools for modeling, steady-state analysis, transient simulation, parameter estimation, and control, and every handoff between them costs time. In this upcoming webinar, we'll introduce HVACComponents, a Dyad-native library for vapor compression cycles featuring tube-fin heat exchangers, compressors, valves, fans, pipes, and conditioned spaces, all backed by fast spline-based refrigerant and air property models. Attendees will see a complete cycle built in Dyad's graphical editor and run as both a transient and a steady-state simulation from the same model, with just a couple of lines changed. Register here.
Save the Date: JuliaCon Global 2027 in Estonia: JuliaCon Global 2027 will take place August 9–14, 2027, at Tallinn University of Technology (TalTech) in Tallinn, Estonia. The annual conference brings together Julia developers, researchers, and users from around the world, marking its first edition in the Baltic-Nordic region. Visit the JuliaCon 2027 website for event updates.
Julia v1.13.0 has been released: The latest Julia release brings faster precompilation (roughly 30% faster than 1.12), quicker startup, and dramatically faster garbage collection. It also adds built-in REPL syntax highlighting, a new fuzzy history search, Pkg performance improvements, and a new graphical interface for Juliaup. Find out more.
Free Online Webinars from JuliaHub: Join JuliaHub staff and other experts for free, one-hour webinars. Space is limited and registration is required, so sign up today!
One HVAC System Model—No Design Tooling Friction with David Dinh, Wednesday October 21 at 12:00 PM Wednesday 12:00 PM Eastern (US)
From Seconds to Milliseconds - Mastering Apple Silicon GPUs in Julia with Tim Besard, Wednesday October 28 at 12:00 PM Eastern (US) .
Slow Simulations, Complex Models? Rethinking HVAC and Thermal-Fluid Modeling with Dr. Avinash Subramanian, Wednesday November 18 at 12:00 PM Eastern (US)
Events and Conferences
Dyad Paris Workshop: On October 20, 2026, Dr. Michael Tiller and Dr. Michael Hoffman will lead a hands-on day at Cocoon Space exploring how AI agents can build and debug models, followed by an afternoon on Scientific Machine Learning with Dyad. Attend the full day or just one session. Register here.
Julia and SciML workshops in Brazil: Chris Rackauckas’s September visit to Brazil connected JuliaHub and the Scientific Machine Learning (SciML) ecosystem with researchers across space science, critical care, and chemical engineering. The trip included engagements at INPE, the ICCAI conference, and the University of São Paulo’s Department of Chemical Engineering. At ICCAI, Chris presented “Toward Solving the Inverse Problem in the ICU,” contributing to discussions on multiscale modeling and ECMO in critical care. The visit highlighted SciML’s relevance across disciplines, connecting mathematical models with observations to investigate complex systems.
Exploring the Next Directions for SciML: Chris Rackauckas delivered the opening keynote at SciML Frontiers: Progress, Evolution, and Future Directions, hosted by Lars Mikelsons at the University of Augsburg. Across two days, four tracks, and breakout sessions, researchers from academia and industry examined current progress in Scientific Machine Learning and discussed priorities for future research and engineering applications.
Insights and Recommendations
Modeling, Control and C Code Generation for a Rotary Pendulum: In this demo, Dr. Fredrik Bagge Carlson shows how Dyad takes a rotary pendulum controller all the way from multibody model and closed-loop simulation to automatically generated C code running on a Raspberry Pi, with a single click. Read it here.
Qiskit Beyond Python: Fortran, C++, and Julia: IBM has opened up Qiskit, the most widely used quantum computing SDK, to Julia, Fortran, and C++. Developers no longer need Python to use it. The new bindings connect directly to Qiskit's high-performance core through its C API, so researchers can build and run quantum circuits in the same language as the rest of their scientific code. For Julia users, Qiskit.jl and QiskitIBMRuntime.jl support a complete quantum workflow, from circuit construction to execution on IBM Quantum hardware, all from a Jupyter notebook and with syntax close to Python's. IBM's tutorial simulates quantum dynamics on 100 qubits and checks the results against Julia tools including OrdinaryDiffEq.jl and tensor-network simulators. Read the full blog post here.
Julia supports lunar landing simulation at ESA: A new paper by Francesco Capolupo and Frederik Markus presents ATLAS, a Julia-based simulation framework developed at ESA for the Argonaut lunar lander programme. It combines detailed spacecraft dynamics, sensor and actuator models, and guidance, navigation, and control algorithms to study descent and landing. The framework supports early prototyping and large-scale Monte Carlo analysis in one environment, highlighting Julia’s potential for computationally intensive simulations and rapid design iteration. Read the paper.
Recent JuliaHub Webinars: JuliaHub provides free one-hour Webinars on topics of interest to Julia users. Nearly 100 past Webinars are available online. Watch them here.
This Month in Julia World: This Month in Julia World is a newsletter from Stefan Krastanov with up-to-date information about Julia events, new releases and more. Read it here.
Nouvelles Julia - Julia News en Français: Nouvelles Julia is a newsletter in French with the latest Julia news. Read it here.
Julia Dispatch Podcast: Julia Dispatch is a Julia podcast from Dr. Chris Rackauckas (JuliaHub VP of Modeling and Simulation) and Dr. Michael Tiemann. Watch it here.
JuliaHub Digital Twin Solutions and Consulting: We help enterprises build deployable and scalable solutions leveraging SciML to create highly accurate and trustworthy Digital Twins. Applications span asset health monitoring, optimization and predictive maintenance, process optimization, model-based control, design optimization, and internal or external simulation tools. We also offer consulting and technical support. Schedule a consultation with our solutions team to discuss your use case.
Careers at JuliaHub: JuliaHub is a fast-growing tech company with fully remote employees in 20 countries on 6 continents. Click here to learn more about exciting careers and internships with JuliaHub.
Julia Expertise Needed at University of Glasgow: Dr. Eric Silverman, Research Fellow at the University of Glasgow, seeks a Research Associate for a 5-year research project on computational modeling for public health using an agent-based modeling framework developed in Julia. Julia experience and a PhD are required for this position. Click here for more information and to apply.
Julia and JuliaHub in the News
Now You Can Build quantum circuits in Fortran, C++ And Julia with Qiskit
A New Cone, FastRényiQKD, Boosts Quantum Key Distribution Speed
Julia Blog Posts
Modeling, Control and C Code Generation for a Rotary Pendulum (Dr. Fredrik Bagge Carlson)
The Best AI Models Fail at Physics: Coding Harnesses are to Blame (Dr. Chris Rakcauckas, Anas Abdelrehim, Anshul Singhvi)
Open-Weight Models Can Do Physical AI, but Only With Specialized Harnesses (Dr. Chris Rackauckas, Venkateshprasad Bhat)
This Month in Julia World (Stefan Krastanov)
Nouvelles Julia (Pierre Navaro)
CUDA.jl 6.4: NVIDIA Jetson Support, and Compiled Code Caching (Tim Besard)
Upcoming Julia and JuliaHub Events
Online: One HVAC System Model - No Design Tooling Friction, October 21, Wednesday 12:00 PM Eastern (US)
Online: Slow Simulations, Complex Models? Rethinking HVAC and Thermal-Fluid Modeling. November 18, Wednesday 12:00 PM Eastern (US)
Paris: Next-Generation System Simulation - From Agentic Engineering to SciML with Dyad, October 20 (Full day workshop)
Contact Us: Please contact us if you want to:
Obtain pricing for Dyad and Julia consulting projects for your organization
Schedule Dyad or Julia training
Share information about exciting new Julia case studies or use cases
Partner with JuliaHub to organize a Julia event online or offline
About JuliaHub, Julia and Dyad
Dyad combines physics-based modeling with scientific machine learning(SciML) for mission-critical engineering. Dyad is fully agentic in its design, making it possible for engineers to carry out complex workflows through natural language interaction. Dyad integrates code, diagrams and agentic workflows in a seamless tool driving 10x productivity. Leveraging the Julia and the SciML ecosystem under the hood, Dyad also benefits from significantly higher performance compared to the competition, often being 100x faster at simulating complex physics. Teams leverage Dyad to build smarter, faster, and more reliable systems without compromising the rigor of traditional engineering, supporting use cases from predictive maintenance to real-time performance tuning and over-the-air updates. The Dyad tool powered by Julia and SciML is free to use. Get started here.
JuliaHub is a fast and easy-to-use code-to-cloud platform that accelerates the development and deployment of Julia programs. JuliaHub users include some of the most innovative companies in a range of industries including pharmaceuticals, automotive, energy, manufacturing, and semiconductor design and manufacture.
Julia is a high performance open source programming language that powers computationally demanding applications in modeling and simulation, drug development, design of multi-physical systems, electronic design automation, big data analytics, scientific machine learning and artificial intelligence. Julia solves the two language problem by combining the ease of use of Python and R with the speed of C++. Julia provides parallel computing capabilities out of the box and unlimited scalability with minimal effort. Julia has been downloaded by users at more than 10,000 companies and is used at more than 1,500 universities. Julia co-creators are the winners of the prestigious James H. Wilkinson Prize for Numerical Software and the Sidney Fernbach Award.
