-
Posts
-
By Mighty Pen · Posted
Go ahead -
By zikalify · Posted
Microsoft's Aurora 1.5 weather model could make hurricane predictions a lot better by Paul Hill Microsoft has just launched a major update for its open-access Aurora Earth-system foundation model, Aurora 1.5. This extension adds 22 new weather variables that are relevant to energy, agriculture, transport, and climate risk. Additionally, it provides hourly temporal resolution and probabilistic ensemble forecasting. If you have ever tracked a hurricane or typhoon, you might be aware that there are various forecasting models available. The best of the bunch are generally considered to be the European Centre for Medium-Range Weather Forecasts (ECMWF) ensemble models. Well, with Aurora 1.5, Microsoft claims its model outperforms on 88.9% of evaluated targets. This model also performs better against its previous version, with testing showing that the ensemble median achieved a one-third lower track error for tropical cyclones, such as Hurricane Helene. While open access to Aurora 1.5 is great news for researchers, agencies, companies, and civil society, Microsoft will also be integrating it into its commercial products such as Microsoft Weather. We all know that AI can give some dodgy outputs, so it wants to couple Aurora 1.5 with physics-based models, such as ECMWF, rather than replace them. The best cyclone forecasts are those that draw from multiple models, and drawing in Aurora 1.5’s predictions will just add more depth to the composite forecast, potentially helping to make people safer. You can read the model’s paper or head over to GitHub to download it. -
By Mindovermaster · Posted
I started using Vivaldi again after maybe 2-3 years. I dunno what I didn't like about them. Maybe Floorp was just a shade better. -
By zikalify · Posted
Google launches LiteRT.js to make AI and ML workloads faster in the browser by Paul Hill Google has just launched LiteRT.js, a new library that enables machine learning models to run locally within the web browser, bypassing the need for server-side processing. The firm said this will bring native AI performance to web browsers via its mobile-focused LiteRT runtime, though LiteRT.js also works on desktops. The new library uses WebAssembly and hardware acceleration such as WebGPU and WebNN to replace the slower TensorFlow.js, which uses a slower JavaScript-based kernel. The LiteRT runtime has been reserved for Android and iOS until now. With today’s update, Google is exposing the runtime via WebAssembly to turn the browser into a more capable platform for AI and ML workloads. The search giant claims that this new runtime delivers 3x greater speeds over existing solutions on current hardware. Specifically, it was tested on a 2024 Apple MacBook Pro with M4 Silicon. In the real world, for users on older hardware or using browsers with different engines, performance could vary significantly. For developers looking to switch over from TensorFlow.js, the process is straightforward. If you already have a .tflite file, you just need to switch your JavaScript runtime to LiteRT.js. However, if you have a TensorFlow/Keras SavedModel, then you can use the LiteRT Converter built into the Python TensorFlow package. You can read more about the conversion process on Google for Developers. Going forward, it will be interesting to see whether Google eventually sunsets or de-prioritizes TensorFlow.js in favor of LiteRT.js. It will also be interesting to see how it really performs on other hardware besides a MacBook. Source: Google -
By Copernic · Posted
AltSendme 0.5.0 by Razvan Serea AltSendme is a minimal, cross-platform application designed for fast, secure, and private peer-to-peer file transfers. It allows users to send files or entire directories directly between devices without relying on cloud servers, accounts, or any personal information. Everything is encrypted end-to-end using modern protocols like QUIC and TLS 1.3, ensuring both strong security and low-latency performance. Transfers are verified with BLAKE3 for data integrity, and interrupted downloads automatically resume, making the experience reliable even on unstable connections. You can transfer anything—images, videos, documents, and more. Integrity checks are performed on both ends, so your files are automatically verified for correctness during both sending and receiving. AltSendme works seamlessly across local networks or long-distance links, capable of saturating multi-gigabit connections for extremely fast delivery. With built-in NAT traversal and encrypted relay fallback, it connects devices almost anywhere. The app integrates with the Sendme CLI and will soon support mobile and web platforms. Fully free and open-source, AltSendme offers a lightweight, privacy-first alternative to traditional cloud-based services, removing size limits, upload costs, and unnecessary data exposure. AltSendme 0.5.0 changelog: Web version available at app.altsendme.com Zero analytics, please raise issues as you come across them Harden app trust boundaries More test coverage - added end-to-end test cases verifying receiver temporary directory lifecycle and resume behavior Added disarm mechanism to AutoCleanupDir to support preserving partial downloads for resume Bug fixes Download: AltSendme 0.5.0 | ARM64 | ~9.0 MB (Open Source) Download: AltSendme for MacOS | Android Links: AltSendme Home Page | GitHub | Screenshot Get alerted to all of our Software updates on Twitter at @NeowinSoftware
-
-
Recent Achievements
-
BenTuckett1997 earned a badge
Dedicated
-
SRODERICK earned a badge
One Month Later
-
Case_f went up a rank
Community Regular
-
Jeroen Wilms earned a badge
One Month Later
-
BrianMunton earned a badge
First Post
-
-
Popular Contributors
-
Tell a friend
.thumb.png.b7089ebbdf00d96b3b92dddbd04bac02.png)
Recommended Posts