How the Dyad Harness Turns Silent AI Model Failure into Scientific Rigor

Checkout Dyad's newest features and see why it is the most innovative model-based engineering solution.

Eliminate design tool friction. Model, run transient, and solve steady-state HVAC & refrigeration systems in one unified Dyad platform.

See how Dyad takes a rotary pendulum from multibody modeling and control design to automatic C code generation and deployment on a Raspberry Pi.

“This is really all about reducing that time that the engineers spend doing repetitive work…and replacing them with agentic tools that will do it automatically.”

Julia approaches top 20 in the September TIOBE Index while MIT News explore JuliaHub's evolution from an MIT research project to a global platform and a cool new launch with Dyad 3.4

Highlights of the Julia 1.13 release.

Explore JuliaHub 26.4, featuring improved collaboration with project history, branching and peer reviews, enhanced security controls, smoother deployments, and zero-downtime updates.

Explore Pluto notebooks in VS Code with AI agents, Julia tooling, and interactive workflows for modeling, simulation, and analysis.

September 2026 TIOBE Index keeps the top 10 unchanged, as Python falls below 18%, C++ widens its lead over Java, and Julia nears the top 20.

Explore Pluto notebooks in VS Code with AI agents, Julia tooling, and interactive workflows for modeling, simulation, and analysis.

Julia, the programming language that originated as a research project at MIT, has gained a loyal following among scientists, engineers, mathematicians, and others. The free and open-source language co...

Explore 3D multibody dynamics, URDF-based modeling, Modelica MultiBody workflows, and AI-assisted engineering for complex mechanical and control systems.

Learn the foundations of acausal, multi-domain physical modeling, from connector design and conservation laws to reusable system components.

Deploy a rotary pendulum controller from simulation to real hardware using Dyad, automatic C code generation, and a Raspberry Pi—without manual reimplementation.

Explore Physical AI benchmarks, Dyad 3.3, JuliaCon 2026, and the latest JuliaHub news, research, engineering insights, and upcoming events.

See how the Dyad Agent turns NASA HL-20 specs into a validated simulation model, using compiler-verified AI for autonomous aerospace modeling.

Explore how the Dyad agent supports modelling, simulation, analysis, and testing workflows through live demos and engineering best practices.

Build portable CPU/GPU code in Julia with JACC.jl. Learn vendor-neutral parallel computing from a single codebase for modern HPC workflows.

See how the Dyad Agent uses natural language to simulate activated sludge models, optimise aeration, and balance energy use with effluent quality.

Learn more about the impressive lineup of speakers and experts from JuliaHub at JuliaCon 2026.

See how the Dyad Agent turns NASA HL-20 specs into a validated simulation model, using compiler-verified AI for autonomous aerospace modeling.

We tested the latest frontier models in the Dyad agent on five modeling and simulation problems, comparing accuracy, cost, time, and work style.

See how the Dyad Agent uses natural language to extend an activated sludge model, run parameter sweeps, and explore the tradeoff between effluent COD and aeration energy.

Find out more about Dyad 3.2, our upcoming conferences and discovering missing physics with the Dyad agent and SciML.

Advance your simulation and modeling with Dyad 3.2. Design multi-model AI agents (GPT-5 & Claude) using enhanced visual canvas tools.

Explore cuTile.jl in Julia to write high-performance GPU kernels for AI, linear algebra, and HPC using NVIDIA’s CUDA Tile model

Discover how Dyad powers software-defined machines with faster simulations, living digital twins, and AI-driven engineering workflows.

Learn how Dyad uses symbolic manipulation to solve DAEs, speed up simulations, and build hierarchical system models with greater accuracy.

This blog examines how the traditional way we model physical systems is inadequate, how acausal modeling changes the game, and why the Dyad platform represents the future of engineering design.

The co-creator of the Julia programming language discusses physics-aware AI, global enterprise scaling and the launch of the industrial platform Dyad

After six years, it’s time to release Pluto version 1.0! With this release, I want to celebrate all the progress over the past years, and to symbolise that Pluto is ready. I am proud of what we have a...

Learn how Dyad uses symbolic manipulation to solve DAEs, speed up simulations, and build hierarchical system models with greater accuracy.

Viral Shah, architect of Aadhaar and co-creator of Julia, on what AI-ready public infrastructure really requires, and the risks no government should take with probabilistic AI.

Learn how Scientific Machine Learning helps engineers uncover missing vehicle physics using test data while keeping simulation models robust and reliable.

Dyad 3.1 elevates your modeling workflow with Claude Opus 4.8, 1M context limits, and configurable reasoning effort. Plus: automatic JuliaC-trimmed FMUs and local library imports.

Build portable CPU and GPU code in Julia with JACC.jl. Join ORNL scientists to learn vendor-neutral parallel computing from a single codebase.

In this newsletter: Dyad 3.0 GA release, Pluto 1.0 release, JuliaHub 26.3 release and much more from the world of Julia

Learn how Dyad uses symbolic manipulation to solve DAEs, speed up simulations, and build hierarchical system models with greater accuracy.

See how Boeing and AE Ventures use Dyad and Julia to unify modelling, simulation, code generation, and verification in aerospace engineering.

Explore how Dyad and Scientific Machine Learning help engineers uncover missing vehicle physics and improve ride model accuracy with test data.

See how Dyad HVAC unifies thermo-fluid design with faster simulation, agentic workflows, and real-time model creation for HVAC and refrigeration systems.

Build portable CPU and GPU code in Julia with JACC.jl. Join ORNL scientists to learn vendor-neutral parallel computing from a single codebase.