Maisa AI

Introducing Vinci KPU: the latest evolution of our AI Operating System, built to tackle multi-action resolutions. Execute Digital Work with Agentic Cloud Functions. Our Digital Workers solve problems, make decisions and execute workflows autonomously with full traceability. The KPU OS serves a reasoning and execution engine. With a model agnostic design, it offers full traceability, mitigation of hallucinations, autonomous problem solving, lower latency, and predictably lower cost. We are pleased to share the results of our latest benchmarks, matching and even surpassing leading LLMs, such as Claude Sonnet 3.5 and OpenAI’s o1, on challenging benchmarks like GPQA Diamond, MATH, HumanEval, and ProcBench. Sign up at kpu.maisa.ai and put it to the test – challenge it with your toughest problems or workflows. For more technical details about the system visit our technical report: maisa.ai/research

We're #hiring a new Technical Product Marketing Leader in Spain. Apply today or share this post with your network.

"Es la primera vez que la IA se vuelve transparente y auditable." This article published today in Expansión by Javier García Fernández does a great job in highlighting the unique benefits maisa's agentic AI solution rpovides to our Fortune 500 clients ... finally AI, or actually Digital Workers, that companies can rely on and hold accountable! ... Digital Workers that their human collegaues can oboard like any other new colleague and that self-improve through continuous learning, leveraging existing and creating new institutional know how. #ai #agents #digitalworkers #agenticai https://lnkd.in/dnCtgpHq

The Agent Company - already feasible or still more of a headline story?! ... Researchers at Carnegie Mellon university tried to find out and created a virtual software company where AI agents have to do the real jobs ... 175 different tasks like managing team projects, dealing with HR stuff, handling company finances or writing & fixing code. Agents had acess to common work tools. The final results were surprising and surfaced things that LLM agents still struggle like common sense, social skills and navigating complex websites XD For a quick summary I recommend to listen to our latest podcast episode, or take a look at the thread on X, for getting into the weeds I'll also post the link to the full paper ... all in the comments. #ai #agents #agenticAI #digitalworkers

Forbes: "Artificial intelligence has a trust issue. For all the hype about generative AI and large language models, many enterprises are reluctant to really embrace the technology. They worry about the well-documented hallucinations LLMs are prone to generating and they dislike the opacity of black-box systems where they can’t check how the model arrived at its results. Spanish start-up Maisa promises it can deal with these problems." https://lnkd.in/dTV2waa6 #ai #agenticai #kpu #chainofwork #proofofwork #proofofsource #trust

What a week it’s been! 🚀 Since launching the KPU Studio last week at Maisa AI, we’ve been blown away by all the messages showing how you're using it and how much you love it! The variety of Agentic AI use cases we’re seeing is incredible. It's inspiring to see KPU tackle the real-world problems you've set. We're truly grateful for all your messages and feedback – it's teaching us a lot, and we're already working on the next updates. Can’t wait for what’s ahead 🧠✨

Our latest podcast episode on educating us all on agentic AI systems will help us to better understand how LLMs learn to reason. https://lnkd.in/d7Xsmvyk #ai #agenticai #llm

In preparation of the public launch of the new version of our KPU, a frontier AI reasoning and execution engine, we're publishing a new episode of our podcast: Beyond Knowledge: Why LLMs Struggle with Multi-Step Instructions In this conversation the hosts go deep into the key elements necessary for reasoning and limitations of existing benchmarks to evaluate reasoning capabilities of current AI systems. https://lnkd.in/e57GbiBe #reasoning #o1 #openai #agents #ai #kpu

Our next podcast episode is out "Agents Thinking Fast and Slow", a deep dive conversation about Google Deepmind's new paper on Talker-Reasoner Architecture ... we think of it and call it internally "Intelligent Interpreter" (multimodla) and "First Principle Thinker" https://lnkd.in/dJWjYMG7 #ai #agi #architecture #firstprinciple #agents #agenticai #aisystems #futureofwork

In March we introduced the KPU: The first leap in AI reasoning ... soon we will launch access to our next release ... maisa's Knowledge Processing Unit is a Reasoning and Execution System for Agents (Talkers (LLMs),NPCs, Services, Workers ...) that leverages the intelligence and reasoning power of LLMs and solves many of their intrinsic limitations like lack of reliability (hallucinations) and lack of traceability and explainability. If you're interested to learn more about our beginning, listen to our first episode of our podcast or read the paper. https://lnkd.in/dNvUyGTi https://maisa.ai/blog/kpu #ai #reasoning #knowledge #context #management #agi