Graham Dumpleton, author of mod_wsgi, is developing with the help of AI a new Rust server called Muru for running Python ASGI/WSGI applications, and is documenting his development process in detail. The project was highlighted by Simon Willison.
A research team published the OceanCoder model on arXiv, a fine-tuned version of Qwen3-8B for automated creation of simulation models in ocean engineering. Pass@1 rose from 0% for the base model to 51.43% after fine-tuning and RL optimization.
Snapchat has introduced the Easy Lens CLI tool, which allows developers to create AR Lenses using the AI assistants Claude Code, Cursor, and Codex without writing code. It works on Mac, Windows, and Linux.
The vibe-coding platform Lovable says it plans to offer business automations and daily AI summaries for small business owners next year. The company states that 80% of users come with a financial motivation, but only 35% of them are currently making money from their project.
Amazon SageMaker HyperPod Inference now supports the open-source Dynamo framework from NVIDIA for distributed, multi-node inference of large language models with separation of prefill and decode phases.
GitHub has released GitHub Copilot Evals into public preview — a tool for repeatable testing of AI agents, skills, and MCP tools in Copilot, with the Vally evaluation engine, available for Individual, Business, and Enterprise plans.
OpenAI has introduced background completion windows in the Responses API in public beta: the two-hour window is 30% cheaper and the eight-hour window is 40% cheaper than the standard tariff, in exchange for higher response latency.
AWS described a five-phase methodology for rewriting mainframe Assembler code into Java Spring Boot using Amazon Bedrock, Kiro IDE, and AWS Transform, with COBOL as an intermediate step and human validation of business rules.
Claude models now support the HTML-in-Canvas feature, which embeds interactive web elements (text, buttons) into canvas graphics while preserving accessibility. It works in Chrome/Edge 157+ and Claude desktop 5.11492+; Safari and Firefox are still missing.
Author Sebastian Raschka published a technical tutorial on his blog Ahead of AI about positional embeddings in LLMs – from classic absolute and relative methods to RoPE, with math, visuals, and a PyTorch implementation.
The Microsoft C++ Language Server for GitHub Copilot CLI now generates the configuration for MSBuild projects in C++ itself. Manual preparation of compile_commands.json and running a separate configuration skill are no longer needed.
Hugging Face has introduced the concept of decision models: an LLM that, instead of generating text, returns probabilities for predefined options. It uses the "decision" pipeline in the transformers library and the GEV-26B-Decide model.
Simon Willison had the Claude Opus 5.5 Medium model port his 3D web application from the Three.js library to Vanilla JavaScript with WebGL. According to him it succeeded, the result is functional, although the details differ from the original.
According to its own statement, the company Cult.fit has deployed OpenAI's Codex tool so that teams across departments can build internal tools themselves and speed up work. The source does not provide details or figures.
GitHub has expanded the Copilot usage metrics API with separate tracking of BYOK (custom keys/local models) and custom enterprise models, so they are not merged with Copilot's hosted models.
According to Bloomberg, Goldman Sachs has restricted access to Claude Code and GitHub Copilot to engineers only. Traders and bankers who had used them for months to experiment lost access, for productivity and risk-management reasons.
With the help of the Claude Opus 5.5 model, Drew Breunig built an experimental GitHub project called nesformer — an NES processor emulator. According to the author, the model independently found and used a 15-year-old processor conformance test suite and created a CLI for running ROM files.
Epsilon built a shared service for testing web interfaces for its product teams. Amazon Nova Act and Amazon Bedrock AgentCore Browser Tool allow steps to be specified in natural language and complement the existing Playwright regression tests.
AWS Machine Learning Blog has published a guide to migrating applications from PHP to Java using Kiro. Shared skill files define an eight-phase process with approvals; a parity harness is intended to verify behavioral equivalence.
A guide from AWS demonstrates distributed fine-tuning of Qwen2.5-1.5B using Ray Train DLC, KubeRay and EFA on Amazon EKS. The example uses eight NVIDIA L4 GPUs and, according to AWS, does not require building a custom container image.
AI Radar monitors Czech and international sources every day, looking for changes that truly deserve attention.
MonitorsOfficial AI company blogs, specialist media, and research sources.
Selects and combinesFilters out information noise and combines articles about the same change into a single event.
Summarizes and explainsExplains significant events in English: what happened, why it matters and where the information comes from.
The result is a quick overview of what has actually changed in the AI world, rather than another stream of articles.
Use the CS/EN switch to read the same Radar in Czech or English. English content is published after its translation has been checked, so new and older items may appear later.
Everything you need to navigate the AI world
Today’s briefingAn overview of news from the past 24 hours and a main story when available. Beside them is Live · AI Flash, followed by research and links to other Radar sections. On mobile, the blocks appear one below another.
AI FlashAn ongoing feed of brief updates with an evidence status. Links lead to a Radar detail page when one is ready, otherwise to the original source. You can also find reset and outage histories here.
Practical applicationsWhat new tools and features can do, what you can try and what their actual impact could be.
Model selectionModel comparison by type of work, capabilities, price and speed.
Research and archiveA separate research overview, topic search and older events by date.
One event, everything that matters
Each row represents one event — not one article. At a glance, you can see its significance, credibility and main point.
Illustrative example, not a current news item.
Importance: ▮▮▮ majorOpenAIModels✓ 6
Agent mode is available to all paying users
Until now, the mode was available only on the highest plan; it is now available on all paid tiers without a waitlist.
▮▮▮ major · ▮▮ important · ▮ we're tracking = how significant the change is✓ 6 = six independent publishers, not the number of articles or feeds✓ official = a clear release, law or incident is substantiated by the relevant authority1 source = no independent confirmation yetbold = who is behind the changegray text = a brief summary of what happened
The detail page contains a fuller summary, its significance and original sources. Practical impact appears in the detail and the For individuals and For businesses views. An AI Flash item reaches the main selection only after it has been expanded and meets the publication rules.
The same news, two practical uses
We first summarize each event in the same way for everyone. Based on those same facts, we then explain what the change means for your own use and what it could mean for how a company operates.
For individualsWhat you can use or try, how the change can help you at work and what to watch out for.
For businessesWhat impact the change could have on processes, costs, risks and other business decisions.
Today’s briefing is the same for everyone. Pages
For individuals and For businesses
can be found in the main navigation — they select only events relevant to the given use case.