The definition says it must include data information (“the complete description of all data used for training, including (if used) of unshareable data, disclosing the provenance of the data, its scope and characteristics, how the data was obtained and selected, the labeling procedures, and data processing and filtering methodologies”), as well as code and paramters. Read your link.

The guys at Hugging Face are working on a more open model based on Deepseek as they also claim it is not fully Open Source.

Thank you for stating that “@Hotznplotzn@lemmy.sdf.org is likely a paid actor” being baseless. It indeed is, although your hint is not too friendly.

I respectfully disagree. The analysis provides much more input that Deepseek’s press release claiming its USD 5m budget (and some other points -e.g. of being Open Source while it isn’t, and other points.)

Judge keeps Musk’s DOGE from further digging into US Gov’s spending

Citing potential “irreparable harm,” US Federal Judge Paul A. Engelmayer Saturday blocked Elon Musk’s Department of Government Efficiency (DOGE) from accessing specific records within the Treasury Department, thus acquiescing to a request from New York Attorney General Letitia James and 19 States under Democratic rule.

The plaintiffs contended Musk’s team accessing this data could pose risks to cybersecurity and violate federal law by potentially mishandling or exposing sensitive personal and financial information of millions of Americans.

Engelmayer also ruled that any data already accessed by DOGE must be destroyed immediately. This injunction is in place until at least February 14, 2025, when further arguments involving national security, privacy rights, and political motivations, will be heard.

From an article on an AI summit in Europe with such a title I would have expected that Eurooean LLM projects are at least mentioned.

It’s still censored even if you run it locally.

As an addition:

Modern slavery in China

Since 2018, evidence of forced labour of Uyghur and other Turkic and Muslim majority peoples has emerged in the Xinjiang Uyghur Autonomous Region (Uyghur Region). […] Forced labour imposed by private actors is also reported, in addition to forced marriage and organ trafficking, with vulnerability primarily driven by discriminatory government practices. While China demonstrated some efforts to tackle modern slavery through sustained coordination at the national and regional levels – including by adopting a new national action plan for 2021 to 2030[…] – its overall response is critically undermined by the use of state-imposed forced labour.

No, it’s not open source. Only the model weights are open, the datasets and code used to train the model are not.

… doing things properly and everything is local then Deepseek reportedly has some efficiency advantages that make it worth considering over alternatives

DeepSeek isn’t uncensored if you run it locally (https://lemmy.sdf.org/post/28978937), and this is just one issue among many others.

China, in particular, has developed sophisticated strategies to control narratives and influence public opinion through digital platforms. This phenomenon, often referred to as “networked authoritarianism,” involves state actors using subtle tactics like algorithmic manipulation and strategic content curation to shape narratives on popular social media platforms.

@Onno

No, it’s not entirely open source as the datasets and code used to train the model are not.