Mark Montgomery’s Post

A good synopsis. Lots of value can be delivered through smaller models focused on specific problems and domains.

For AI Giants, Smaller Is Sometimes Better EXCERPT: The start of the artificial-intelligence arms race was all about going big: Giant models trained on mountains of data, attempting to mimic human-level intelligence. Now, tech giants and startups are thinking smaller as they slim down AI software to make it cheaper, faster and more specialized. This category of AI software—called small or medium language models—is trained on less data and often designed for specific tasks. The largest models, like OpenAI’s GPT-4, cost more than $100 million to develop and use more than one trillion parameters, a measurement of their size. Smaller models are often trained on narrower data sets—just on legal issues, for example—and can cost less than $10 million to train, using fewer than 10 billion parameters. The smaller models also use less computing power, and thus cost less, to respond to each query.... #techgiantsthinksmaller #smallercheaperaimodels #aimodelstentimescheaper

For AI Giants, Smaller Is Sometimes Better - Companies are turning their attention to less powerful models, hoping lower costs and solid performance will win more customers. The start of the artificial-intelligence arms race was all about going big: Giant models trained on mountains of data, attempting to mimic human-level intelligence. Now, tech giants and startups are thinking smaller as they slim down AI software to make it cheaper, faster and more specialized. This category of AI software—called small or medium language models—is trained on less data and often designed for specific tasks.  The largest models, like OpenAI’s GPT-4, cost more than $100 million to develop and use more than one trillion parameters, a measurement of their size. Smaller models are often trained on narrower data sets—just on legal issues, for example—and can cost less than $10 million to train, using fewer than 10 billion parameters. The smaller models also use less computing power, and thus cost less, to respond to each query. Microsoft has played up its family of small models named Phi, which Chief Executive Satya Nadella said are 1/100th the size of the free model behind OpenAI’s ChatGPT and perform many tasks nearly as well. “I think we increasingly believe it’s going to be a world of different models,” said Yusuf Mehdi, Microsoft’s chief commercial officer. Link to Free Full Story here: -https://lnkd.in/gCERWin4 Written By - Tom Dotan and Deepa Seetharaman https://lnkd.in/gCERWin4

LET'S GET SMALL (with #AI): It's obvious in my mind, but I am always focused on operationalization and meeting model / AI #governance requirements; small models are much more congruent with that, particularly since one of my Zoldi-isms is "explainability > predictive power." Perhaps other organizations will back into that belief as a byproduct of adopting smaller AI models; then, perhaps the realization will be that when we no longer need the AI giants, small language models will be within the grasp of internal AI teams. Edited to add: This was one of my #AIpredicitons2024, in fact. 2. Small will be beautiful "Large AI models, including large language models (LLMs) like ChatGPT, Google Bard AI, et. al., are incredibly unwieldy, difficult and expensive to build and operate. Furthermore, they are not easily retrained without cannibalistic entropy. "I have long held an unorthodox belief in “explainability first, predictive power second,” a core tenet of Responsible AI. In 2024, small, highly explainable AI models will be accepted and embraced as being more effective than outsized large models, as organizations focus on practical, reliable, net-positive business outcomes." My full predictions blog is here: https://lnkd.in/gbuVMVfN

For AI Giants, Smaller Is Sometimes Better https://lnkd.in/eUZTt4em

🚀 Exciting developments ahead! The use of #AIAgents is revolutionizing how we approach data and technology. At Observata AB, we're integrating LangChain's cutting-edge AI capabilities with Elastic's robust search engine in our #HYPRSeek Service to deliver next-level search solutions. 🤖✨ This progress brings: 🔍 Smarter Data Retrieval: AI-driven insights paired with powerful search technology. 💡 Enhanced Solutions: More accurate, efficient, and tailored outcomes for our customers. Stay tuned as we continue to innovate and transform industries with these advancements! #AI #Innovation #AIAgents #SearchTechnology #CustomerSuccess #Observata #HYPRServices #HYPRSeek

This article highlights what our Data & AI practice has been applying for a while. We can build more accurate and efficient models. To build more accurate and efficient AI models, incorporating foundational statistical models can be highly beneficial. These models offer a robust framework for understanding and manipulating the underlying data distributions, ensuring that the AI models are grounded in solid statistical principles. LLM benchmarking is key for selecting the right GAI model but incorporating it with foundational methods is a true differentiator. Companies are turning their attention to less powerful models, hoping lower costs and solid performance will win more customers.

Today is a defining moment for Humiris AI, and we’re thrilled to share it with you. We’re 𝗼𝗳𝗳𝗶𝗰𝗶𝗮𝗹𝗹𝘆 𝗹𝗮𝘂𝗻𝗰𝗵𝗶𝗻𝗴 𝗠𝗶𝘅𝘁𝘂𝗿𝗲 𝗼𝗳 𝗔𝗜 (MoAI)— a new way to think about AI infrastructure. Imagine an environment where your models are flexible, scalable, and ethical, perfectly tuned to your needs. That’s what MoAI makes possible: the inner quality your AI deserves. Over the past year, we’ve been on an extraordinary journey: 1. 𝗠𝗮𝘀𝘁𝗲𝗿𝗶𝗻𝗴 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗶𝗻𝗴 𝗮𝗰𝗿𝗼𝘀𝘀 𝟱 𝗽𝗹𝗮𝘁𝗳𝗼𝗿𝗺𝘀 (OpenAI, Mistral, Together AI, Runpod, and Koyeb). Shoutout to Koyeb, a French provider powering the 𝘄𝗼𝗿𝗹𝗱’𝘀 𝗺𝗼𝘀𝘁 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗶𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗽𝗹𝗮𝘁𝗳𝗼𝗿𝗺, for partnering with us to run our homemade model. 2. 𝗚𝗮𝘁𝗵𝗲𝗿𝗶𝗻𝗴 𝟱𝟬𝟬 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 𝗼𝗻 𝗼𝘂𝗿 𝗳𝗶𝗿𝘀𝘁 𝗽𝗿𝗼𝗱𝘂𝗰𝘁, learning from their feedback, and building an infrastructure that’s adaptable and reliable. 3. 𝗥𝗲𝗹𝗲𝗮𝘀𝗶𝗻𝗴 𝗼𝘂𝗿 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗽𝗮𝗽𝗲𝗿, "𝘔𝘪𝘹𝘵𝘶𝘳𝘦-𝘰𝘧-𝘈𝘐: 𝘌𝘧𝘧𝘪𝘤𝘪𝘦𝘯𝘤𝘦 𝘧𝘰𝘳 𝘈𝘐 𝘔𝘰𝘥𝘦𝘭𝘴," showing how mix-models outperform single-models in both efficiency (Humiris Routing) and accuracy (Humiris Reasoning). 💡 Our vision? To 𝗯𝗿𝗶𝗱𝗴𝗲 𝘁𝗵𝗲 𝗶𝗻𝗴𝗲𝗻𝘂𝗶𝘁𝘆 𝗼𝗳 𝗦𝗶𝗹𝗶𝗰𝗼𝗻 𝗩𝗮𝗹𝗹𝗲𝘆 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗲𝘁𝗵𝗶𝗰𝘀 𝗮𝗻𝗱 𝘀𝗼𝘃𝗲𝗿𝗲𝗶𝗴𝗻𝘁𝘆 𝗼𝗳 𝗘𝘂𝗿𝗼𝗽𝗲. We’re building for a future where AI empowers, not overwhelms. A future where 𝗔𝗜 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝘀, 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵𝗲𝗿𝘀, 𝗮𝗻𝗱 𝗯𝘂𝗶𝗹𝗱𝗲𝗿𝘀 𝗹𝗶𝗸𝗲 𝘆𝗼𝘂 have tools that make your lives easier while creating more impactful solutions. 𝗕𝘂𝘁 𝘄𝗲 𝗰𝗮𝗻’𝘁 𝗱𝗼 𝗶𝘁 𝗮𝗹𝗼𝗻𝗲. 🚀 That’s why we’re inviting you to join our 𝗹𝗮𝘂𝗻𝗰𝗵 𝘄𝗲𝗲𝗸: 📆 Monday: Launch of 𝗛𝘂𝗺𝗶𝗿𝗶𝘀 𝗥𝗼𝘂𝘁𝗶𝗻𝗴, a smart routing system dynamically selecting the best LLM for your context. 📆 Tuesday: Explore our 𝗗𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱, your command center for managing and benchmarking multiple models seamlessly. 📆 Wednesday: Discover 𝗛𝘂𝗺𝗶𝗿𝗶𝘀 𝗥𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴, the infrastructure that combines multiple LLMs for 𝗰𝘂𝘁𝘁𝗶𝗻𝗴-𝗲𝗱𝗴𝗲 𝗾𝘂𝗮𝗹𝗶𝘁𝘆 𝗼𝘂𝘁𝗽𝘂𝘁𝘀 and more efficient problem-solving. 📆 Thursday: Advanced features like 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗶𝗻𝗴 𝗮𝗻𝗱 𝗳𝘂𝗻𝗰𝘁𝗶𝗼𝗻 𝗰𝗮𝗹𝗹𝗶𝗻𝗴 for ultimate control over your models. 📆 Friday: Launch of 𝗛𝘂𝗺𝗶𝗖𝗵𝗮𝘁, a new way to interact directly with Humiris Reasoning. 📆 Saturday: Don’t miss our 𝗛𝘂𝗺𝗶𝗿𝗶𝘀 𝗟𝗮𝘂𝗻𝗰𝗵 𝗣𝗮𝗿𝘁𝘆 𝗶𝗻 𝗦𝗮𝗻 𝗙𝗿𝗮𝗻𝗰𝗶𝘀𝗰𝗼 financial district. Come celebrate, network, and imagine the future of AI with us. 🌟 To everyone who’s been part of this journey, thank you. To the dreamers, builders, and thinkers: this is your moment. Thinking about improving the inner quality of your model? Let’s chat: www.humiris.ai #AIInfrastructure #LLMs #MixtureofExperts #AIEngineering #Builders

AI has a data problem. The rule of thumb is that technology will double every 8 months (that's the current metric) But with LLMs this is not happening, we've already started slowing down, even with the addition of GPT-4o (which is great BTW) we're running into the issue of recycled data. LLMs only work well when it has access to new and rich data, but what happens when all of the new data is generated by LLMs? Diminishing returns. So what can you do? Start leveraging your personal expertise with AI. If you have a tech product you're going to need to build that expertise into the non AI parts of your application. What do you think, have you started seeing the data problem in practice?

The Rise of AI Agents: Beyond RAG and Towards Autonomous Reasoning In the rapidly evolving world of AI, one term that has gained significant traction is Agents. Businesses are investing heavily in this technology, yet many struggle to grasp the fundamental differences between standard Retrieval-Augmented Generation (RAG) and Agents. So, what sets them apart? RAG vs. Agents: The Key Difference At its core, RAG enhances Large Language Models (LLMs) by extending their knowledge with external data sources. While LLMs are inherently limited by their training data, RAG enables them to retrieve relevant information from databases, APIs, or documents, making their responses more up-to-date and contextually relevant. However, RAG lacks reasoning capabilities—it merely fetches information and presents it. This is where Agents come in. The Power of AI Agents Agents introduce an additional decision-making layer on top of LLMs. Instead of just retrieving information, Agents analyze, reason, and act based on the retrieved data. They can: ✅ Plan and execute multi-step actions ✅ Interact with external tools, APIs, and databases dynamically ✅ Adapt and refine responses based on the evolving context ✅ Automate complex workflows beyond simple query-answering #Agents#LLM