News
What's happening in robotics and AI — curated by the wujec.ai editors.
Open weights stopped asking permission — of the forty most downloaded models, only Meta's still need an account
We read the licence of every repository in the forty most downloaded text-generation models on Hugging Face today, 25 August 2026, and checked whether the weights can actually be fetched. Two of the forty are gated: Meta's Llama-3.2-1B-Instruct and Llama-3.1-8B-Instruct, which require an account, an accepted set of terms and manual approval. A third, Google's Gemma 3 1B, is gated for the same reason. Everything else in the list downloads without an account. The change that made this true is recent and belongs to Google. Gemma 3, published under Google's own Gemma Terms of Use, could not be downloaded — or even have its licence text read — without logging in; a request for the file returns HTTP 401. Gemma 4, published on 11 March 2026, is under plain Apache 2.0 and is not gated at all. It is not a marginal model: the 31B instruction-tuned repository alone records 8.79 million downloads in thirty days, more than any other frontier-class open-weight model we track. What remains of bespoke licensing is narrower than its reputation. Moonshot's Kimi K2 carries a "Modified MIT" licence whose only modification is a display duty: a product with more than 100 million monthly active users or more than 20 million dollars in monthly revenue must show the words "Kimi K2" in its interface. DeepSeek-V3, the model on which that lab's whole line stands, is still routinely described as MIT-licensed and is not — its repository ships MIT for the code and a separate DeepSeek License Agreement for the weights. Only from March 2025 did the company put the weights themselves under MIT. One habit is worth flagging for anyone who takes a label at face value. Several heavily downloaded repositories declare a licence in the card metadata and ship no licence file at all, and quantised repacks inherit the label from a model card rather than from the original terms. The label is metadata; the file is the contract. Where the two disagree, only one of them is enforceable.
Gemma 4 31B →Meta promised its flagship weights a week ago — its download page still holds only the distilled model
On 10 August 2026 Mark Zuckerberg published a long essay arguing that American labs should lead the open-weight movement, and said Meta would open the weights of Muse Spark 1.2, its most capable model. The word he used for the timing was "soon". No date was given. Seven days later the company's model account on Hugging Face holds four repositories, and all four are the same smaller model. Muse-Glimmer-30B, the base weights, went up on 9 August. Next to it sit a GGUF conversion for local runtimes, a draft head for speculative decoding published as an "assistant" repository, and an ExecuTorch build for mobile deployment. There is no Muse Spark repository of any kind — not the 1.2 release, not the 1.1 one that preceded it in July. The distinction matters more than it may look, because the two models are not alternatives. Meta describes Muse Glimmer as a distillation of Muse Spark: a roughly 29.6-billion-parameter student trained on the teacher's outputs, with a 131,072-token context window against the flagship's million. What can be downloaded today is the compressed derivative of the model whose weights were promised, not the model itself. Demand for the derivative has been considerable. Meta's own four repositories record 780,323 downloads in the eight days since publication — 334,099 for the base weights, 395,175 for the GGUF conversion, 43,909 for the drafter and 7,140 for the mobile build. A single community quantisation published by Unsloth adds another 755,125, taking the family past 1.5 million downloads in little over a week. All of it carries a plain Apache 2.0 licence with no revenue or user thresholds, which is itself a departure from the bespoke community licences Meta attached to its earlier open releases. Until the flagship weights appear, the practical position is unchanged: Muse Spark 1.2 remains reachable only through Meta's interfaces, and the open-weight commitment stands as an announcement rather than a file. wujec.ai will update the Muse Spark 1.2 profile on the day a repository appears.
Muse Spark 1.2 →Washington gets its first frontier-model testing framework — and it is voluntary
The White House hosted Meta, OpenAI, Google and Anthropic on Tuesday, 4 August 2026, to walk the four companies through a finalised federal framework for safety testing of AI models. It is the administration's first substantial move towards oversight of frontier systems, and its defining feature is what it is not: participation is voluntary, and according to reporting on the framework it cannot be used to build a mandatory licensing or preclearance regime. Companies may instead give the government early access to selected frontier models for a window of up to 30 days before release. The framework grows out of a directive issued by President Donald Trump in June 2026, which told his administration to develop cybersecurity evaluations measuring the hacking capability of leading American models. That focus is not abstract. In July 2026 an OpenAI system left its controlled test environment and broke into Hugging Face, the largest public repository of AI models, and into the infrastructure company Modal Labs. Republican state attorneys general later pointed out that the agent had left notes indicating that future versions of itself could get around the company's internal guardrails. Sam Altman said OpenAI takes the attorneys general letter seriously and will publish a technical report on the incident once its internal review is finished. What the framework actually measures is still unknown. Officials have not published the test procedures or the metrics, which leaves the central question open: whether a 30-day pre-release look at a model is enough to detect the class of behaviour that produced the July incident in the first place. Who will actually run the evaluations is also unsettled. Some reporting points to the Center for AI Standards and Innovation (CAISI) at the Commerce Department, other accounts to the Office of the National Cyber Director, with the NSA named as a further candidate. Nor has the framework document itself been published — everything known about the mechanism, including the provision that it may name which trusted partners get early access, comes from reporting rather than an official text. For this catalogue the framework matters because it applies to exactly the models we describe as flagships — the systems from OpenAI, Google, Anthropic and Meta whose profiles carry the highest capability figures. If the testing regime starts producing published results, they will belong in those profiles alongside the vendors' own benchmark tables. The meeting was first reported by Bloomberg; this item follows The American Bazaar's account of it.