OZZZER · AI NEWS

Latest AI news.
Editorial picks.AI News

Selected signals with a direct route back to each original source.

475 stories

Latest stories

Unclassified

Supporting independent journalism in Ukraine

Publisher preview

Working with WAN-IFRA and AIRPPU to protect and rebuild independent journalism through conflict in Ukraine. The following was issued as a joint press release on Monday 7th September 2026 from WAN-IFRA, AIRPPU, and OpenAI. The Newsroom AI Masterclass Series and Newsroom AI Catalyst support Ukraine’s independent news sector with a focus on AI adoption, innovation, and resilience. OpenAI, WAN-IFRA, and the Association of Independent Regional Press Publishers of Ukraine (AIRPPU) have announced a new initiative to support and strengthen Ukrainian news publishers by building practical artificial intelligence capabilities and accelerating…

OpenAI · 7 Sep 2026 · 02:00 CESTRead story →Join free ↗

AI NEWSSource illustration

Unclassified

An Alien Mind

Publisher preview

In mid-2023, within the “RLSlow” research project, we saw the first results that gave us confidence that we will be able to scale the training of reasoning models, unlocking the capability of pretrained models to form their own chains of thought. Szymon and I spent that night at the office, thinking not about the incredible benchmark numbers, products, or scientific results that this technology will deliver - but rather, trying to process the sobering fact we will actually see machines meaningfully smarter than ourselves in our lifetime, and we already…

OpenAI · 6 Sep 2026 · 11:00 CESTRead story →Join free ↗

AI NEWSSource illustration

Research

Research acceleration: The view inside OpenAI

Publisher preview

For AGI to benefit all of humanity, we believe it must be democratically governed. This can only happen through an informed public debate about the capabilities, risks and safeguards of highly capable AI systems. People everywhere need to understand the likely future trajectory of frontier AI, so they can have a meaningful voice in how it develops. Transparency about specific risks, incidents and safeguards is necessary, but not sufficient. We believe the public also needs to understand how the most capable systems are developing, and how they are driving research…

OpenAI · 6 Sep 2026 · 10:00 CESTRead story →Join free ↗

Models & tools

Introducing WeatherNext 3, our most advanced and accurate global weather AI model

Publisher preview

Our flagship AI weather forecasting model now includes real-time satellite data, hourly refreshes, higher resolution, precise precipitation forecasting, and clean energy variables. It’s now integrated across Search, Gemini, Maps, Google Maps Platform, and Cloud. Google has launched WeatherNext 3, an advanced AI model that provides more accurate, high-resolution weather forecasts by using real-time satellite data instead of traditional physics simulations. This update delivers hourly, localized predictions that are five times sharper than previous versions, offering significant improvements for agriculture, renewable energy, and daily planning. You can now access these enhanced…

Google DeepMind · 3 Sep 2026 · 17:02 CESTRead story →Join free ↗

AI NEWSSource illustration

Unclassified

Daybreak for Frontline Defenders: $1B to protect essential services

Publisher preview

OpenAI is committing $1 billion in subsidized Daybreak access, training, technical support, and partnerships to help frontline defenders protect essential services. Today OpenAI is introducing Daybreak for Frontline Defenders, a new global initiative to help frontline defenders use frontier AI cyber capabilities to protect essential services in the United States and around the world. The initiative includes: A $1 billion global commitment to expand subsidized access to Daybreak cyber models and products, training, technical support, and partnerships in the United States and internationally. Daybreak for America, bringing together all of…

OpenAI · 3 Sep 2026 · 15:15 CESTRead story →Join free ↗

Unclassified

NeoMME: an efficient Multimodal-native and Multilingual Encoder

Publisher preview

We introduce NeoMME, a family of 260M and 800M multilingual multimodal encoders. Unlike many generative visual language models, NeoMME does not use a separate pretrained vision tower or a causal language model. A single bidirectional Transformer processes both text tokens and raw image patches, and we train the entire model from scratch with a masked discrete-diffusion objective. We fine-tuned NeoMME for visual document retrieval using ColPali's page-image approach. NeoMME-Retriever returns dense and late-interaction embeddings in one forward pass. Both model sizes lie on the ViDoRe v3 Pareto frontier for nDCG@10…

Hugging Face · 3 Sep 2026 · 15:13 CESTRead story →Join free ↗

AI NEWSSource illustration

Unclassified

Legora reviewed 41 documents in minutes with GPT-6 Astra

Publisher preview

Legora uses Astra to improve accuracy, completeness, and reliability, while keeping final judgment with legal professionals. Legora is an agentic operating system for legal and professional work, used by more than 100,000 professionals across more than 1,800 in-house legal departments and law firms in over 50 markets. Its legal engineers work directly with customers to understand how they operate and adapt Legora to their end-to-end workflows, from contract and agreement review to legal research. One of the more tedious workflows is financial-statement tie-out: checking every figure in draft accounts against…

OpenAI · 3 Sep 2026 · 14:00 CESTRead story →Join free ↗

AI NEWSSource illustration

Unclassified

Playco cut manual fixes 50% prototyping games with GPT-6 Astra

Publisher preview

Using GPT‑6 Astra, Playco built three themed game prototypes from one grey box foundation, with most working on the first take. Playco is using GPT‑6 Astra in building Playbot, an AI-powered IDE for professional game developers. It connects directly to engines such as Unity and Godot so AI models can edit scenes, play and test games, validate changes, and work in parallel inside the tools developers already use. In game development, a model needs to do more than write code. It must reason about space, visual references, responsive interfaces, game…

OpenAI · 3 Sep 2026 · 14:00 CESTRead story →Join free ↗

Unclassified

GPT-6 Astra: A new generation of intelligence

Publisher preview

Introducing GPT-6 Astra, our most intelligent and aligned model yet, with state-of-the-art capabilities across computer use, coding, cybersecurity, and science.

OpenAI · 3 Sep 2026 · 13:00 CESTRead story →Join free ↗

Coding & Development

Training a coding model to paint watercolours with TRL and OpenEnv

Publisher preview

On 23 August, Surya Narreddi posted a beautiful video of watercolours painted by a language model. The model writes JavaScript through p5.brush, a library that "adds natural drawing tools to p5.js". The video went viral fast, over 1.5M views at the time of writing. The video came with a blog post explaining the training behind an earlier and narrower stage of the project, close-up flowers rather than the full compositions in the video, sadly without open artifacts yet. His site says a full technical report is coming, so ensure you…

Hugging Face · 3 Sep 2026 · 02:00 CESTRead story →Join free ↗

Coding & Development

Give Your Coding Agents a Memory You Own

Publisher preview

Earlier this year, Software Forgets: Agent Traces Are the Memory made the case that coding agents already produce the record we keep losing. As they search a codebase, try approaches, hit errors, read documentation, and change direction, they leave behind a dense account of not just what changed, but why. While the diagnosis is correct, traces are only potential memory. The session logs of an agent are still just an archive. You cannot grep your way to “why did we move off the streaming parser?” across ten thousand turns. For…

Hugging Face · 3 Sep 2026 · 02:00 CESTRead story →Join free ↗

Models & tools

Fine-tuning a 350M Model for Better Structured Outputs in 100 GRPO Steps

Publisher preview

Structured output is one of the most common real-world tasks for LLMs, yet most benchmarks fold it into broader reasoning or extraction scores rather than measuring it on its own. Whether a model reliably returns valid, parseable output in the requested format and shape — schema compliance — is often what decides whether it can be wired into a downstream system at all. Note that the training pipeline described here is not the one used to train the RL model described in the IFStruct blog. This notebook doesn't aim to…

Hugging Face · 3 Sep 2026 · 02:00 CESTRead story →Join free ↗