Ad
Skip to content

Jonathan Kemper

Jonathan is a freelance tech journalist, author, and media coach. He covers open AI models, Chinese AI labs, and the technical underpinnings that make them tick – from novel training methods and scaling strategies to benchmark design. He's also an avid vibecoder.
Read full article about: Samsung deepens its AI empire with a potential billion-euro stake in Europe's hottest AI startup

Samsung is in talks to invest up to one billion euros in French AI startup Mistral, which would push the company's valuation to around 20 billion euros. That's according to the Financial Times. Less than a year ago, Mistral was valued at 12 billion euros. Samsung already invested in Mistral through its venture arm in 2024, and Swedish investor EQT is reportedly also participating in the new round. Samsung keeps expanding its footprint in the AI industry. The company is one of the world's largest memory chip makers and is also in talks with Anthropic about manufacturing a custom AI chip.

Mistral CEO Arthur Mensch said in a February interview that he expects annual revenue to top one billion dollars by year's end. Just this week, Mistral also expanded its partnership with Microsoft, with the US company committing to billions of dollars in spending on Mistral's compute infrastructure. Mistral is scaling up its GPU capacity with thousands of Nvidia Vera Rubin chips and making its models available through Microsoft's Azure cloud, including in air-gapped environments without internet access for regulated industries like finance and healthcare.

Alibaba's Qwen-Image-3.0 renders full infographic grids and readable ten-pixel text in a single pass

Alibaba’s Qwen team has introduced Qwen-Image-3.0, an image generator that accepts prompts up to 4,500 tokens, renders legible text as small as ten pixels, and supports twelve languages natively. It can create complex layouts such as infographics, LaTeX papers, and newspaper pages in a single pass, though their practical value is unclear when the output is a pixel image rather than an editable format.

Read full article about: Microsoft and Mistral strike multi-billion-dollar deal to build AI infrastructure across Europe

Microsoft and Mistral are expanding their strategic partnership with a multi-billion-dollar deal to build out AI infrastructure in Europe. Mistral is adding thousands of Nvidia Vera Rubin GPUs, and Microsoft will tap that capacity for its cloud and AI services. Mistral's Medium 3.5 and OCR 4 models are now available in Microsoft Foundry. Medium 3.5 is also available in Copilot Studio. Companies can run Mistral models through Azure Local in the cloud, in their own environments, or completely offline. The offering targets regulated industries like finance, healthcare, and manufacturing. Microsoft President Brad Smith said Europe should be able to use powerful AI without giving up control over its data.

Mistral CEO Arthur Mensch sees the partnership as a way to reach companies and public institutions worldwide. Mensch also recently warned that companies risk exposing trade secrets when they use closed AI models. The company, founded in 2023, recently grew its revenue twentyfold to a run rate above $400 million. It invested 1.2 billion euros in data centers in Sweden and, according to Bloomberg, is negotiating a new funding round of about 3 billion euros at a valuation of roughly 20 billion euros. Mistral recently renamed its chatbot Le Chat to Vibe and repositioned it as an AI productivity tool with Work Mode and Code Mode.

Xiaomi-Robotics-1 shows that more data beats bigger models when training robots to move

Xiaomi trained Xiaomi-Robotics-1 on more than 100,000 hours of motion data collected by people using camera-equipped handheld grippers rather than robots. Adding data improved performance far more than increasing model size. The gains haven’t plateaued, though absolute success rates remain low.

Google Deepmind argues video generators already contain the world models computer vision has been missing

Google Deepmind’s GenCeption repurposes a video generator for classic vision tasks such as depth estimation and segmentation, matching state-of-the-art systems with far less training data. The model trained almost entirely on synthetic videos. Its results add to the debate over whether video generators already contain a kind of universal world model.

Read full article about: Zuckerberg's plan to sell excess AI compute could finds its first big customer in Anthropic

Meta is reportedly in talks with Anthropic to rent out compute capacity from its data centers. The deal could be worth up to $10 billion over two years, according to the New York Times. Anthropic pitched the arrangement in June. Meta is still reviewing it, and either side could walk away. For Meta, this would open a new revenue stream. CEO Mark Zuckerberg had already told investors he could sell excess capacity if the company's own AI demand didn't keep pace.

Meta plans to spend up to $145 billion this year, mostly on AI. The company is also building out new revenue sources with paid AI subscriptions like "Meta One" and even investing in space-based solar power for its data centers. Anthropic needs the extra capacity because demand for Claude Code has surged. The company already signed a $45 billion deal with SpaceXAI. Long term, though, Anthropic wants to build its own data centers and has hired former Google executives to lead that effort.

Read full article about: OpenAI wants developers to stop typing commands and start using a joystick to control their AI agents

OpenAI and keyboard manufacturer Work Louder have unveiled the Codex Micro, a compact hardware controller designed for working with AI agents. Instead of constantly switching between chat windows and typing commands, developers can manage their AI agents through a stripped-down keyboard with joysticks and a rotary dial. The top six keys light up with real-time RGB feedback to show each agent's status, whether it's thinking, working, waiting for input, or done. The joystick triggers common workflows like code reviews, debugging, or refactoring. The rotary dial controls the reasoning level, letting users set how much compute the model should spend on a given task.

Modulares OpenAI WorkLouder Codex Micro mit programmierbaren Tasten, Drehregler und verschiedenfarbenen LEDs.
Each color represents a different agent status: thinking, done, waiting for input, or error.

The device plugs directly into ChatGPT Codex, where keys can be remapped. Work Louder's Input software also lets users assign custom shortcuts to every key, dial, and joystick movement across six programmable layers. The box includes 32 swappable icon keycaps so users can match the physical layout to their bindings.

The Codex Micro costs $230 but is currently out of stock. Work Louder says quantities are limited. The device connects via Bluetooth or USB-C and works with both Mac and Windows.