Archive link: https://archive.ph/nrsXa

The President of Venezuela, Nicolás Maduro, ratified this Monday that through technological alliances with the People's Republic of China, the creation of a totally Venezuelan and sovereign artificial intelligence will be promoted.

While specifying that this must integrate the history and characteristics of the country, the head of state stressed that the development of an artificial intelligence representative of the peoples of the Global South will be promoted.

Venezuela has a good alliance with China to build our artificial intelligence, he added, insisting that this is good news for the youth.

“To build our own artificial intelligence, with our history”, he remarked alleging that the new colonialism comes through technologies, communications and artificial intelligence.

By betting on the “balance of artificial intelligence (AI)”, Nicolás Maduro specified that such tool is biased by the interests and vision of the imperialist and colonialist West.

“It is possible,” said the Venezuelan president.

CecilIA is not only the name of a 19th century Cuban novel. In the midst of 2025, it is also the bet of a group of scientists, professors and students of the University of Havana to develop a Cuban artificial intelligence language model, designed from Cuba, with Cuban data and at the service of the country. In a global context dominated by technologies trained with cultural biases, corporate interests and hegemonic languages, CecilIA emerges as a sovereign alternative with identity.

The model, initially trained with Cuban literary texts, national press, political speeches and the Official Gazette, has already begun to show results. Its development has been promoted by the Artificial Intelligence and Data Science Group of the Faculty of Mathematics and Computer Science (Matcom), and its ultimate goal is for AI in Cuba to speak and understand “in Cuban”.

During its second public presentation, this time at the headquarters of the National Union of Jurists of Cuba, technical advances were shared, but above all, something fundamental was insisted on: without Cuban data, there is no truly Cuban AI. For this reason, the developers invite institutions, media, jurists and artists to add documents, scripts, songs, news and legal texts to nurture the model.

A collective project, from Cuba and for Cuba.

The CecilIA model is based on what specialists call “small language models” (SLM), which require fewer resources and are ideal for developing countries. Starting with the Salamandra base model (also in Spanish), the Cuban team carried out continuous training with its own corpus, and is now working on the design of Cuban instructions to fine-tune the interaction and better adapt the model's responses.

Dr. Yudivián Almeida, one of its leaders, announced that they expect to build a corpus with at least 10,000 specific instructions, many of which can be openly proposed by anyone interested. This aims at a collaborative AI, rooted in the knowledge of the people.

But the challenge is not only technical. The team stresses the importance of ethics, prevention of bias, explanation and the urgent need to digitize heritage documents. Libraries, publishers and archives still have valuable information only on paper, which limits access and use for AI training purposes.

Beyond software development, the CecilIA project has a deep cultural dimension: protecting Cuban identity in a world where algorithms often ignore the nuances of the Global South. The possibility of creating AI applications that recognize the country's speech, concepts and cultural references represents a strategic tool for information sovereignty.

During the meeting in Havana, jurists, linguists, sociologists and scientists agreed on the need for all disciplines to contribute to this type of projects. Not only to guarantee technical quality, but also so that the final result is aligned with the social, cultural and ethical values of the nation.

CecilIA is an example of what can be achieved from a public university and a committed scientific system. But its success will depend on society as a whole understanding that digital transformation is not a luxury, but a national necessity. While the work continues, the model is already inspiring other teams across the country. As they said at the close of the presentation: “we grew in everything”.

36Kr learned that the annual D20 Global Design Deans Summit was held in Hangzhou from July 11 to 12. Yang Guang, Vice President of Alibaba International and Chairman of Alibaba Design Committee, said that as a dialogue platform for global design education and industrial practice, the D20 Summit has always been committed to promoting the interdisciplinary integration of design and technology. This time, Alibaba Design, together with ecological partners such as design schools, industry associations, design agencies, and AI technology companies, launched an initiative on "AI Design Talent Outlook" - the "D20 Declaration", covering five major elements including talent standards, curriculum & training system, ecological platform, application transformation, and innovative research, hoping to promote the co-construction of AI design talent standards in various fields.

Pictures of the place

Video of the place

https://archive.ph/q3VzI

The article is a great critique of how what the author refers to as the "Efficiency Lobby" has been pursuing a narrow idea of task oriented intelligence focused on productivity. It's a narrow focus, driven by corporate interests, that necessarily leads to individualistic consumption of AI services, hindering genuine creativity, open-ended exploration, and collection.

A recent paper introduces MemOS with the potential to create a truly collaborative and community driven foundation for AI. The paper introduces a new approach to memory management for LLMs, treating memory as a governable system resource.

It uses the concept of MemCubes that encapsulate both semantic content and critical metadata like provenance and versioning. MemCubes are designed to be composed, migrated, and fused over time, unifying three distinct memory types: plaintext, activation, and parameter memories.

This architecture directly addresses the limitations of stateless LLMs, enabling long-context reasoning, continual personalization, and knowledge consistency. The paper proposes a mem-training paradigm, where knowledge evolves continuously through explicit, controllable memory units, blurring the lines between training and deployment paving the way to extend data parallelism to a distributed intelligence ecosystem.

It would be possible to build a decentralized network where there's a common pool of MemCubes acting as shareable and composable containers of memory, akin to a BitTorrent for knowledge. Users could contribute their own memory artifacts such as structured notes, refined prompts, learned patterns, or even "parameter patches" encoding specialized skills that are encapsulated within MemCubes.

Using a common infrastructure would allow anyone to share, remix, and reuse these building blocks in all kinds of ways. Such an architecture would directly address Morozov's critique of privatized "stonefields" of knowledge, instead creating a truly public digital commons.

This distributed platform could effectively amortize computation across the network, similar to projects like SETI@home. Instead of constantly recomputing information, users could build out a local cache of MemCubes relevant to their context from the shared pool. If a particular piece of knowledge or a specific reasoning pattern has already been encoded and optimized within a MemCube by another user, it can simply be reused, dramatically reducing redundant computation and accelerating inference.

The inherent reusability and composability of MemCubes make it possible to have a collaborative environment where all users contribute to and benefit from each other. Efforts like Petals, which already facilitate distributed inference of large models, could be extended to leverage MemOS to share dynamic and composable memory.

This has the potential to transform AI from a tool for isolated consumption to a medium for collective creation. Users would be free to mess about with readily available knowledge blocks, discovering emergent purposes and stumbling on novel solutions.

https://memos.openmem.net/

The New York Crimes panicking over China's EV dominance and calling for a "Manhattan Program" for EVs. Good luck with that. The US's neoliberal brain worms are dug in too deep. They couldn't do it for the MIC and they certainly won't do it for the auto industry.