In softgoods development, the digital model should follow the physical prototype—not the other way around. From a guide on designing smart textiles, an interesting read about how the development process for soft goods (think smartwatch bands, AR/VR headstraps) differs from traditional mechanical components. Documenting the architecture and details in CAD may be necessary, but the digital representation needs to follow the real textile behaviour.
Another nuance of interest: prototyping is either much faster or much slower compared to traditional hardware, mainly because off‑the‑shelf fabrics already use the same production-ready methods in prototyping, while custom textiles need full-scale manufacturing runs.
awesome-opensource-hardware is a curated list of resources for designing chips using open-source EDA software. Unlike PCB design where you wire together off-the-shelf components, chip design involves the actual creation of those components from the ground up. That means defining logic gates, timing paths, and physical layout at the transistor level. The process can be crudely summarized as RTL (describe what the circuit should do in a hardware description language like Verilog) → logic (convert code into gates) → layout (place and route the gates on silicon) → physical silicon (manufacture the final chip at a foundry). This guide from Vik’s Newsletter is a good way to get started with the tools mentioned in the original repository.
When you press a mechanical switch, the contacts don’t always make a clean connection. Ever tapped a remote and had the volume switch two steps when you thought you only pressed once? That’s bounce, rebound from the mechanical contacts that produce a burst of on-off signals instead of a single, stable one. Debouncing ensures only one keypress is registered per physical press, using either hardware (RC circuits) or software (polling for stable states). Most systems rely on software debouncing which involves periodically sampling the input and only acting once the state has remained stable for 10–20ms. In matrix keypads, debouncing is often built into the row-column scanning routine, helping avoid ghost presses while keeping input responsive.
In light of OpenAI’s $6.5B (!) acqui-hire of Jony Ive’s design studio, it’s worth revisiting What Do Prototypes Prototype?, a late 1990s paper that gives a glimpse into Apple’s early design philosophy, particularly around interaction design. The core message reads as follows: stop judging prototypes by fidelity or tools and start evaluating them based on the design question they’re meant to answer. Houde and Hill argue that every prototype should focus on one or more of three axes: role (the job the product plays in someone’s life), look and feel (how it looks, sounds, and feels), or implementation (how the technology actually works). Engineering teams often default to the look and feel or implementation axis.
The fastest way to unlock capacity in the US power grid is to replace century‑old steel‑core wires on existing transmission towers and string modern composite‑core “advanced conductors” in their place. Advanced conductors work by swapping out the steel core in traditional wires for a lighter, stiffer composite that allows for higher operating temperatures and a more conductive aluminum to fit within the same diameter. This ends up doubling power capacity compared to traditional aluminum conductor steel reinforced (ACSR) wires. Grids are increasingly becoming the bottleneck of the energy transition, with over 1,200 GW of renewable energy (projects in the United States, and over 3,000 GW globally, awaiting connection to the grid.
Startup News 🚀
OpenAI has acquired io, the AI hardware startup co-founded by Apple’s former Chief Design Officer Jony Ive, in a $6.5B all-stock deal. The acquisition brings in his 55-person team including former Apple design leads to develop AI-native consumer devices. The first device is expected in 2026 and was described as "a new device that sits alongside your iPhone and MacBook." From details leaked from an internal call, the device will be pocket-sized, contextually aware, screen-free, and specifically not another pair of smart glasses.
Throne raised $4M in seed funding to develop its AI-powered toilet device that monitors gut health with a pre-production prototype set for launch in January 2026. The company’s device mounts onto existing toilets, using computer vision to analyze health indicators like chronic conditions, hydration, and urological function, with privacy controls to anonymize data. Funny enough, what started as a poker night joke in 2021 gained real traction after a few serendipitous connections to researchers at UW and UChicago, who helped validate the technology. The funding round was led by Moxxie Ventures.
Rivan, a London-based climate tech startup, raised $12.7M in Series A funding to scale its modular synthetic fuel plants from 100kW to 1MW by 2026. The company’s tech. combines direct air capture, green hydrogen electrolysis, and Sabatier reactors to produce a carbon-neutral synthetic natural gas for heavy industries like steel and aviation. The 100kW pilot current runs out of a 10,000 sq ft factory and produces fuel using just air and water. The funding round was led by Plural.
Lumina is seeking $20-$40M in Series A funding to produce five more electric autonomous bulldozers by March 2026. The company’s 32-ton Moonlander is a zero-emission bulldozer that matches Caterpillar’s D6 footprint but rivals the larger D9’s performance, aiming to disrupt construction with lower-cost excavation services. We usually skip open fundraising asks, but this one came with a great datapoint: for Lumina’s $8M seed, the founder had messaged 3,000 angels with only around 10 responses.
Glass Imaging raised $20M in Series A funding to advance its AI-powered GlassAI technology for enhancing digital image quality across devices like smartphones, drones, and wearables. Founded by former Apple engineers Ziv Attar and Tom Bishop, the company uses AI to correct lens aberrations and sensor imperfections, delivering sharper, true-to-life images with up to 10x performance improvement. The funding will support refining GlassAI integration with camera manufacturers and developing new optical hardware designs to rival professional DSLR quality. The round was led by Insight Partners, with participation from GV, Future Ventures, and Abstract Ventures.
From some old friends working on a new startup: Ground Control has launched through YCombinator building vertical AI tools for regulated manufacturing. Their first product automates AS9102 First Article Inspection Reports (FAIRs) for regulated manufacturing, and can generate inspection reports in minutes. Worth a look if you deal with aerospace, defense, or medical parts!
