CrewAI

Although a bit late, I’m thrilled to announce that I’ve officially completed my studies at BAU, Centre Universitari d'Arts i Disseny! My thesis,?Design for Economy, is the culmination of my research on how design can actively contribute to shaping more equitable economic systems. In this work, I explored the intersection of speculative design, economic theory, and technological innovation, drawing on key thinkers such as Mark Fisher and William Davies. I examined the possibilities of post-capitalist economic models through speculative design frameworks and practical applications using AI. My project reflects on the role of designers in shaping future economies, from rethinking growth to questioning neoliberal structures. One of the core elements of my project involved coding a framework using CrewAI by Jo?o (Joe) Moura that simulates societal structures and economic systems through AI agents. I invite everyone to check out and experiment with the code I developed—you can find all the materials in the link in the comments! This journey wouldn’t have been possible without the support of my classmates, my friends, and my tutor Tim Cawlishaw, who challenged my ideas and helped me reach new philosophical depths. Thank you all for your incredible encouragement! Shoutout to some of the incredible authors and thinkers who inspired this work: Silvio Lorusso, Matt Webb, Georgina Voss, Nick Srnicek and Alex Williams & many more.

?? CrewAI version v0.63.6 is out! - New LLM class to interact with LLMs - Support to custom memory interfaces - Adding support to o1 family models - Updated logs format - Docs updated - Bugs fixed - Oh and small logo update :) We move fast! ? ? ?

This Wednesday, I had the chance to give a workshop about LLM agents, something that many people have heard about but not that many people really looked into. Even for someone that is actively doing and following research in the field of NLP, it feels so fast sometimes that it is almost impossible to keep up with every new approach. But from what I have seen, agents are here to stay. It is really fascinating how a model that just predicts probabilities for the next tokens is able to model such intelligent processes. In a few hours, the workshop participants were able to use the crewAI framework to create their own agents. The exercises included: - a public transport agent, that is able to fetch a public transport data API and answers questions regarding public transport connections - an agent that had access to a Google calendar and could see which events were already scheduled and create new ones - a combination of the two agents above - think about just formulating where you need to go, and the agent looks up the connection and adds the details to an event that is scheduled in your calendar right away - a Pub Quiz Crew that was able to generate pub quiz questions that are inspired by random Wikipedia articles It is really fascinating to me where we are heading! And while there are still some problems now and then, and of course some questions that need to be tackled first before blindly trusting in such a system, I think that these frameworks and of course the LLMs and agentic workflows will only get better from here on. If you are curious, I encourage you to try it out. There are many built-in tools that your agents can use out-of-the box, not only from CrewAI but also from LangChain, on top of which CrewAI was built. There are many other frameworks with their own advantages and weaknesses, but if you want to go from an idea to an agentic workflow very fast, you should definitely try it out. And if you want some inspiration from the workshop exercises, you can take a look at the Jupyter notebooks here: https://lnkd.in/eV5txw8p

We're excited to announce Exa's Series A funding round, led by Lightspeed, with participation from NVentures (NVIDIA) and Y Combinator. This is just the beginning of our journey to transform the data layer of the AI ecosystem. Thank you to our incredible team, who has been working on this first-of-its-kind search technology since 2021. We're also grateful to our customers for betting on Exa. Together we've seen the power of pairing AI with embeddings search, allowing applications to access niche knowledge like never before. Special thanks to Guru Chahal and Nnamdi Iregbulem from Lightspeed, as well as Artem Barsukov and Samay Dhawan from Nventures, for being thought partners in this and seeing Exa's transformative potential. Let's go!! ?? https://lnkd.in/dSfNtqJS