Ordo Nexus

🤖 High-Speed Access for AI Labs, Pretraining Companies & Researchers We are testing approaches to provide a high-speed access to raw and processed content for data-hungry entities. STC Library has more than 2T of ready high-quality tokens for pretraining and hundreds of TBs of raw data spanning multiple centuries and languages for your own processing. Reach us at stclibrary@proton.me with the letter describing your needs and conditions and we will find a way to a better future.

Nexus Bots got Updated Safety Features for Uploaders • The Defile Bot, used by our users to upload files, now allows de-watermarking without uploading the file itself, for your safety. Use the /defile command and attach the file plus its DOI or prefix to that message. You can report badly de-watermarked files right there, or use it to clean up your files. • All reports about broken or watermarked files are now forwarded to librarians for review, we have reduced the processing time from weeks to days / hours. If you have an access to academic libraries - start liberating of scholar publications, join our #freetalks group - join the Nexus.

• They know better how to raise your kids. • They know better if you are allowed to gather on streets or not. • They know better what you should read on the Internet. <you are here> • They know better how you should think. <or already here?> What’s the weirdest here: that you’re ordered what should be in your head at all, or that the decision makers here are complete hopeless morons? #имлучшезнать

🤖 AI for Libraries and Researchers Space Frontiers provides MCP and API integration with the entire STC database. You are welcome to test it and integrate the search over a vast of scholar publications and other sources into your own research workflows. Join #freetalks for any questions and discussions, authors of the project is there to answer your questions.

🎉 6 Years of Nexus We have walked this long road, started as a small pet-project search and continuing as a Telegram shadow library and one of the largest suppliers of texts for training AI. None of it would have been possible without hundreds of people who have been uploading papers and books, reading, and building all these years. As usual, wishing you all the best in these restless times.

🎯 Newly Published Research Feed — Relaunched! Space Frontiers now offers a live feed and RSS for newly published papers. Discover the latest research from all publishers on your favorite topics - delivered daily to your preferred RSS reader!

⚠️ Changes in Aaron group TLDR: If you are a regular user, nothing has been changed for you, at all. For a long time, the @nexus_aaron group was a public space where user requests from bots were automatically posted, taken by uploaders, reviewed by librarians, and then accepted into Library STC. The review process has now been moved outside of the public group. Here’s what this means: - Authorized librarians are invited to a private, non-public group. - Files for review are posted in this private group, where librarians vote to accept or reject them. - Non-reviewed files will no longer be publicly available in the Aaron group, so the uploading process is no longer exposed. You may still upload files to Aaron (though it’s recommended to use the Defiler bot) and take user requests from there. After review, accepted files will become available through the bots, just as before. These changes were made to ensure that the file uploading process is hidden from outsiders, enhancing both security and privacy.

Nexus - Pirates of Mind

🌎 Space Frontiers 🪐 Now you can create your own scholarly digest, ask scientific questions, navigate and find related papers, translate articles, and explore an extensive collection of science, books, standards, manuals, and magazines at https://spacefrontiers.org. Simply enter your question and await the response. We have used this platform to produce previous digests, and now you can do the same. Join us in pushing the boundaries of AI into space for our shared future among the stars.

Ordo Nexus pinned a photo

Long ago, this city was home to dragons and their hoarded wealth, but today the wind whistles through the empty halls... Previous #freetalks has been finally banned, so you may find new one in help of bots.

🎉 5 Years of Nexus We've lived through another fruitful year, despite the attacks and growing chaos in the world. Haven't posted updates for a long time, so here it is: 🌍 The IPFS directory with files and a search engine is available at /ipns/libstc.cc (details in the Freetalks group). 🔎 All searches are semantic now; you may pose queries in free form to retrieve more results than before (or use special syntax to use the good old match search, details in help). 📨 Bots are now searching in Telegram channels as well. 👨‍👦‍👦 Our Communities group are growing more and more, becoming an active place to discuss science. 🧐 One publisher has started depositing articles directly in STC and more are arriving. 🤖 Our chatbot Cybrex now uses DeepSeek R1-level LLM and now we are implementing Deep Research capabilities for selected users. During the next month, we are going to publish an updated AI dataset and embeddings dataset for researchers' tasks. Thank you for staying with us all this time.

🏴‍☠️ LibGen issues The provider URDN that seems to be a hoster of LibGen has reported recently: Our installation is currently offline following an act of sabotage by individuals that appear to be from the national police of Ukraine. We are assessing the situation with our legal team. what might be a reason of unavailability of LibGen during last days.

We have now updatable IPFS directories with part of STC content, the update cadence is weekly: https://ipfs.io/ipns/libstc.cc/dois - DOI-indexed IPFS directory with STC files https://ipfs.io/ipns/libstc.cc/repo - Nexus ID indexed IPFS directory with STC files Sites are not openable directly as they are too large and should be accessed using CLI tools. Do not ask how to use it if you do not know what it is; these are for developers and seeders. Also, we are going to release two other datasets: - A new version of the AI dataset soon with many millions of OCRed books and more than 15M OCRed papers + 170M metadata records. - A new embeddings dataset, more than 2.2TB of jinaai/jina-embeddings-v3 built on top of STC for search and ML experiments. Freetalks Chat | LibSTC.cc | SolarMind | Communities Chat

🤖 Conversational Memory and Talking With Books Now you can: - Have longer talks with our LLMs using the "Reply" button - the bot will do its best to maintain the chat context - Reply with questions to files, and the bot will keep file content in context (but the file must be stored with STC). For example: search for shadow libraries in the bot, open the first book, click "Reply" on the message and then ask something like list shadow libraries from the book. More examples may be found in our #freetalks chat

🏴‍☠️ Meta Secretly Trained Its AI on a Notorious Piracy Database Shiny Libgen Meta faces a major setback in a copyright infringement lawsuit as a court unredacted documents revealing the company's use of LibGen for AI training. The plaintiffs claim Meta knowingly used the library's material, while Meta argues its actions are protected under "fair use." The court's decision to unseal these documents may have significant implications for Meta and the AI industry as a whole.

API for STC / Cybrex Looking for people who would like to get access to an API for searching papers and/or LLM models based on STC (similar to /cybrex in bots) for their own projects. Let's discuss ideas in the comments, and some worthwhile ideas may be given beta access to the API.

- The request button is back. - Downloads should also have become faster. - Additionally, Cybrex processes uploaded files immediately after approval in @nexus_aaron. This way, you can upload any paper and then chat with it using the /c command in the bots. For example: /c What is the predicted share of renewables mentioned in 10.1016/j.ijhydene.2024.11.018? Enjoy!

Requesting papers will be re-enabled in a while. We have reworked our internals to make downloads more stable, but it will take some time to bring back all downloads and requesting capabilities. Meanwhile, you can walk through constellation of Nexus projects and our friends: - Library STC - web interface for bots - Nexus Aaron - papers and books uploader - Cassiopeia - AI Digests of recent publications - #communities - talks about medicine, physics, history, and other fields - Solar Mind - feed of publications - Space Frontiers - LLM search over corpus of papers and API for searching