Tag
Snaptrude 正式推出 Snaptrude Agent，一款面向建筑设计的 AI 代理，可从场地或设计简报直接生成含结构、MEP 管线及完整图纸的施工级 LOD 400 模型，同时允许建筑师全程主导设计流程。
作者认为 AI 辅助物理产品设计的基础组件已经齐备：AI CAD 近四个月取得巨大进步，3D 打印、电子设计与按需组装也在不断提升，很快将迎来“vibe-coded matter”——用自然语言直接生成实物的时代。
Linus Ekenstam demos Ply, a skate ramp builder tool grounded in reality that lets users design ramps, verts, features, or full parks and outputs bills of materials, building plans, material-optimized cuts to minimize waste, and cost estimates.
A developer introduces FieldGeo, a set of AI models that encode geometry, physics, and design for manufacturing into a latent space, aiming to enhance generative design with better spatial understanding beyond traditional AI CAD approaches.
Arc Institute announces that their popular 2025 preprint on a framework for generating and evaluating AI-generated genomes is now published in Science. The work, from Brian Hie's lab, used Evo 1 and Evo 2 to design viable bacteriophage genomes.
TypeStax is a type scale generator that wraps a functional design tool in a vintage audio hardware-styled interface, making typography scaling feel tactile and playful.
Introduces GenTO, a diffusion model-based framework that steers topology distributions for unified design of architected metamaterials, achieving diverse design tasks with reusable topology priors and experimental validation.
This paper presents TANS-FO, a research prototype for closed-loop generative design of customized foot orthoses using text-aligned neural surrogates and graph neural networks to predict plantar stress. The system achieves surrogate-predicted peak-pressure reduction of 34.7% over parametric CAD but requires human review and has not received regulatory clearance.
The Czinger 21C hybrid hypercar uses 3D-printed components and generative design from parent company Divergent Technologies, achieving exceptional performance and lap records.
A new tool enables rapid generation of parametric, manufacturable 3D models in seconds, streamlining design and production workflows.
This paper presents a discrete autoregressive transformer that generates planar mechanisms from target coupler curves, using variational autoencoder latents and tokenized joint coordinates to achieve diverse, accurate designs across multiple topologies.
CadX Studio teases an AI CAD model for 2026, demonstrating a workflow from typing to building to sectioning, with a model release planned for Monday.
This paper proposes a memory-augmented reinforcement learning framework for CAD generation agents that integrates geometric kernel toolchains, dual-track memory, and dynamic utility retrieval to handle complex CAD models with long operation sequences and geometric constraints, achieving improved success rate and geometric consistency.
A user conducted a stress test using GPT-5.5-Pro and ForgeCAD to generate a helicopter model, highlighting current limitations in surface development and model capabilities.