Tech news picked on 2026-11-09
- Reflex is a design Shopify created while moving a React Native app to native Swift and Kotlin. It separates the screens from the app's calculation logic, so the logic can be run without a simulator. It was published as an explanatory article, not as a library.
- No code is shared. A single behavior test written in TypeScript is run identically on both platforms. A PR that passes on only one side cannot be merged. Shopify said that more than 1,700 tests take about 4 minutes on both CIs.
- If even one external request goes out during a test, it is treated as a failure, even if the app swallows the error. The author (a Teranovel developer) built a weather app and ran 21 tests, and wrote that it took 48ms on iOS and 270ms on Android (including JVM startup).
- The shared test also revealed differences between the two platforms. When a number was sent as a string, Swift raised an error while Kotlin accepted it, so only Android failed, and the author fixed the Kotlin side.
- Clossy is a service that ships clothes from Korean designer brands directly to overseas consumers. Trenfit said its core goal is to help brands that find it hard to build overseas sales channels and logistics.
- Trenfit said on the 9 that it has simultaneously filed trademark applications in the UK, the EU, the US and elsewhere. It plans to secure trademark rights in key regions first through each country's patent office and the Madrid System (the international trademark filing system).
- The company explained that it combines its experience running a London pop-up store 3 times last year with an AI filtering algorithm to match overseas consumers' tastes with brands. It reportedly operates direct delivery networks in the UK, Europe, the US, Canada, Australia, New Zealand and elsewhere.
- After establishing itself in fashion, it plans to expand into beauty and lifestyle. All of the above comes from statements by Trenfit and its CEO.
Apps for enjoying drawing and radio instead of endless feeds
- Instead of endlessly scrolling social media feeds, here are apps that let you draw, listen to the radio, or learn words in a short time.
- Finch is an app where you care for a virtual pet by completing daily goals such as drinking water or taking a walk. Doodle Draw has you turn a shape given each day into a drawing.
- Radio Garden is an app where you pick a city on a map and listen to live broadcasts from around the world. Drops lets you learn 45 or more languages through short games.
- Doodle Draw is available only on iOS. Radio Garden and Drops are available on iOS and Android, and offer free use and paid plans.
- KetchUp is a free app for drawing 3D models with exact dimensions before building furniture such as shelves, tables, cabinets and workbenches. It is described as usable without an account.
- You draw board outlines with rectangles and lines, pull up faces to give them thickness, and enter lengths directly or add dimension marks to check the assembled shape and proportions.
- Installers are available for Windows (x64), Mac (Apple silicon) and Linux (x64). It also opens in desktop browsers that support WebGPU or WebGL2. It is said to be released under the MIT license and written in Rust.
- Work is saved automatically in that device's browser and can be downloaded as a .ketchup file. It can also import SketchUp (.skp) geometry and export OBJ files.
- llama-modchip, released by developer synchopate, is open source software (Apache-2.0) that compresses AI models further and adjusts CPU computation to let people run large AI models locally without a graphics card.
- According to the developer, DeepSeek V4 Flash shrinks from 76.87GiB to 56.56GiB, and peak memory use is about 57.6GiB even with a 256K context, so it fits on a 64GB PC. The original text says the existing 76.87GiB version stops due to insufficient memory on 64GB Linux.
- README figures show input processing at about 44tok/s and response generation at 9.06~10.74tok/s on an AMD Ryzen 9 9955HX (16 cores, 64GB DDR5). With an auxiliary draft technique, it reaches 11.5~14.7tok/s. These are the developer's own measurements.
- Nemotron 3.5 30B at the 30 billion scale was reportedly reduced to 14.35GiB and achieved 25~60tok/s on a PC with 16~32GB memory. The compressed models use a proprietary format and cannot be opened in standard llama.cpp, Ollama or LM Studio.
- Data shows that the price of tokens, the units AI uses to process text, has fallen, so bills are rising less even with heavy use. According to the AI Index from spending management platform Ramp, usage rose 37% and spending rose 9% over a month.
- The original article interpreted the 26.8% drop in token prices as a sign that AI models are becoming commodities amid price competition among model providers such as Anthropic and OpenAI.
- The Financial Services Commission said it will provide up to 4.5 trillion won in guaranteed loans to regional small and medium-sized businesses and mid-sized companies over 3 years. Banks will contribute 300 billion won, and the Korea Credit Guarantee Fund and Korea Technology Finance Corporation will provide guarantees worth about 15 times that amount.
- The Ministry of Trade, Industry and Energy said 7 foreign-invested companies reported $2 billion in new investments, and 97% were outside the Seoul metropolitan area. The cornerstone investor system takes effect on November 13.