ObjectBox

Swift developers: build local AI apps with Semantic Index now --> Easily personalize AI experiences with the first on-device vector search for iOS and macOS 🤩- private, fast, offline 💚 https://lnkd.in/dH7qxz-T #iosdev#localAI#offlineAI#vectorsearch#RAG#vectordatabase

The first version of ObjectBox running in the Browser - powered by Flutter, JavaScript and WebAssembly 🎉 #flutter #dartlang #database #vectordatabase #localAI

The first version of ObjectBox running in the Browser - powered by Flutter, JavaScript and WebAssembly 🎉 #flutter #dartlang #database #vectordatabase #localAI

ObjectBox helps developers unlock the power of distributed data directly on the devices where the data is created and used - that's not only superfast, private, and offline-capable, but also more sustainable 💚 #ondevice #edge #local #datamanagement #database #ai

❔ How to make LLM's responses more relevant? Many of us use LLMs like GPT or Gemini on a regular basis. However, their responses are often too general (and sometimes wrong), especially, if you ask for domain-specific information. To make the results more relevant you can provide the context for LLM - e.g. up-to-date information or a set of internal data from the company. Such a technique is called RAG - retrieval augmented generation. In the recent article in Objectbox, we discuss: ❎ What is RAG and how is it connected to vector databases? ❎ Why should you use it? ❎ How does it work? We also compare RAG with long-context windows (another way to enhance LLM's response quality) and discuss future perspectives. 🎀 Bonus: RAG explained in a simple diagram https://lnkd.in/ewYgFpPz

🚀 𝐋𝐚𝐫𝐠𝐞 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐦𝐨𝐝𝐞𝐥𝐬 (𝐋𝐋𝐌𝐬) with their key capabilities - 𝐠𝐞𝐧𝐞𝐫𝐚𝐥𝐢𝐳𝐚𝐭𝐢𝐨𝐧 and 𝐞𝐦𝐞𝐫𝐠𝐢𝐧𝐠 property - led to the creation of a new class of AI models, 𝐟𝐨𝐮𝐧𝐝𝐚𝐭𝐢𝐨𝐧 𝐦𝐨𝐝𝐞𝐥𝐬.  💡 𝐖𝐡𝐚𝐭 𝐢𝐬 𝐞𝐦𝐞𝐫𝐠𝐢𝐧𝐠 𝐚𝐧𝐝 𝐠𝐞𝐧𝐞𝐫𝐚𝐥𝐢𝐳𝐚𝐭𝐢𝐨𝐧 𝐢𝐧 𝐋𝐋𝐌?  𝐆𝐞𝐧𝐞𝐫𝐚𝐥𝐢𝐳𝐚𝐭𝐢𝐨𝐧 refers to the ability of an LLM to apply its learned knowledge to new, unseen data or tasks. 𝐄𝐦𝐞𝐫𝐠𝐢𝐧𝐠 property refers to new behaviors, capabilities, or characteristics that are not explicitly programmed into the model but emerge as a result of its learning process and interactions with data. For more details, see episode 9 of ObjectBox Bites. 💡 How have 𝐋𝐋𝐌𝐬 led to 𝐟𝐨𝐮𝐧𝐝𝐚𝐭𝐢𝐨𝐧 𝐦𝐨𝐝𝐞𝐥𝐬? Due to these features, LLMs can be used not just for specific tasks, but can be adapted to a wide range of downstream tasks. Such models are called 𝐟𝐨𝐮𝐧𝐝𝐚𝐭𝐢𝐨𝐧 𝐦𝐨𝐝𝐞𝐥𝐬. These models are the foundation for diverse AI applications. Instead of training a new model from scratch with massive amounts of labeled data, foundation models adapt their existing knowledge to new tasks, making the process faster and more efficient.  💡 Where to use 𝐟𝐨𝐮𝐧𝐝𝐚𝐭𝐢𝐨𝐧 𝐦𝐨𝐝𝐞𝐥𝐬? Potential tasks for foundation models include language processing, visual comprehension, robotics task specification, code generation, and developer/end-user engagement. 🎬 In the 10𝐭𝐡 𝐞𝐩𝐢𝐬𝐨𝐝𝐞 𝐨𝐟 𝐎𝐛𝐣𝐞𝐜𝐭𝐁𝐨𝐱 𝐁𝐢𝐭𝐞𝐬, we discuss what 𝐟𝐨𝐮𝐧𝐝𝐚𝐭𝐢𝐨𝐧 𝐦𝐨𝐝𝐞𝐥𝐬 are and how they are connected with 𝐋𝐋𝐌𝐬. 👍 Don't forget to hit the like button and share your thoughts in the comments. Produced by and with Dr Anastasia Ragulskaya Video editor: Michael Odafe #VectorDatabases #LargeLanguageModels #ObjectBox #FoundationModels #NaturalLanguageProcessing

Very useful overview of the 𝐃𝐚𝐭𝐚 𝐈𝐧𝐟𝐫𝐚𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞 𝐥𝐚𝐧𝐝𝐬𝐜𝐚𝐩𝐞 𝐢𝐧 𝐄𝐮𝐫𝐨𝐩𝐞 by MMC Ventures - reminded me a bit of the "selling shovels in a gold rush" comic I recently saw, where basically everyone was trying to sell shovels and no one wanted to do the digging.... 🤣 Their advice for data infrastructure founders is to look for opportunities to vertically integrate (section: "What developments in the AI space could help or hurt data infrastructure businesses? ❌") --> How about using their shovels for finding gold? - any thoughts MMC Ventures? Thank you for the recognition of ObjectBox 😊 #datainfrastructure #vectordatabase #edgedatabase #edgeAI #Europe

Finally, the 𝐯𝐞𝐫𝐲 𝐟𝐢𝐫𝐬𝐭 𝐨𝐧-𝐝𝐞𝐯𝐢𝐜𝐞 𝐯𝐞𝐜𝐭𝐨𝐫 𝐝𝐚𝐭𝐚𝐛𝐚𝐬𝐞 𝐟𝐨𝐫 𝐀𝐧𝐝𝐫𝐨𝐢𝐝 is here Some may know it, Android was our first love ❤️ (we did start developing for the Android os before its initial release, though that wasn’t ObjectBox yet) and the Android community is still dear to our hearts, so we’re particularly happy to bring this new tech to Java & Android developers today. Local AI working on Mobile (without Internet, on the actual devices!) opens up a myriad of new use cases in low, no, or intermittent connectivity scenarios as well as any use cases that have QoS requirements, or “just” need - or want - to keep data private. We can’t wait to see what Android developers will do with it 🎨, even when they are faster coding an app than we are with releasing the tech😅 - let us know #android #java #kotlin #coding #ai #vectordatabase #edgedatabase

Python developers 🐍, discover the power of local/on-device RAG on commodity hardware with no need for an Internet connection or data sharing We just released ObjectBox 4.0 for Python. It’s a lightweight and fast vector database you can run on almost any hardware. 💥 And persisting Python object alongside vector data is as convenient as it gets. Don't worry about speed: ObjectBox finds the nearest neighbors within milliseconds, even among millions of documents, on commodity devices. 😮 >> Link in first comment << Stay tuned, more language bindings and benchmarks coming! 👀 #python #coding #ai #vectordatabase #edgedatabase

Understanding the difference between traditional search & vector search is the basis of 𝐮𝐧𝐝𝐞𝐫𝐬𝐭𝐚𝐧𝐝𝐢𝐧𝐠 𝐭𝐡𝐞 𝐦𝐚𝐠𝐢𝐜 𝐨𝐟 𝐚𝐝𝐯𝐚𝐧𝐜𝐞𝐝 𝐀𝐈 𝐚𝐩𝐩𝐬 with vector databases - great job explaining that Dr Anastasia Ragulskaya 💪 #vectordatabase #vectorsearch #AI #databases