Author: Yohei

  • Rise of the Autonomous Agents 🤖

    It’s been 4 months since the start of the autonomous agent revolution, let’s check some out. We’ll start with BabyAGI and AutoGPT, two of the inaugural autonomous agents. AutoGPT has a community of devs building lots of integrations, while BabyAGI has a handful of mods that Yohei has released. Other general agents include Adept who’s raised the most, beebot who has scored […]

  • Surfing the Latent Space: Vector Embeddings

    Vector embeddings are numerical representation of words and ideas in the form of a matrix, and core to how large language models work. While most people use the large language models directly with ChatGPT, you may have heard of semantic search, vector search, or embeddings – which refers to the use of these matrices themselves. A […]

  • BabyCatAGI: Fast and Feline

    Abstract: In this paper, we present BabyCatAGI, a modified version of the BabyBeeAGI code base, designed to improve research workflow through intelligent task automation and creation using GPT-4. BabyCatAGI builds upon the foundations of BabyBeeAGI, offering faster execution and increased stability by employing a task creation agent that runs once at the beginning and utilizes […]

  • BabyBeeAGI: Task Management and Functionality Expansion on top of BabyAGI

    95% of the BabyBeeAGI code, this paper, as well as the flowchart above were written with GPT-4. Abstract: In this paper, we present a modified version of the BabyAGI code base, called BabyBeeAGI, which improves upon the original framework by introducing a more complex task management prompt, allowing for more comprehensive analysis and synthesis of […]

  • AsymmeTrix: Asymmetric Vector Embeddings for Directional Similarity Search

    This paper, the code, the chart, and the Twitter thread announcing this was all written by ChatGPT with 13 prompts – which are included below. Abstract: This paper introduces AsymmeTrix, a novel approach to vector search and embeddings that leverages asymmetric cosine similarity to model directional relationships between data points. We demonstrate the significance of […]

  • Birth of BabyAGI

    Backstory Fascinated by the #HustleGPT movement, where people are using ChatGPT as their cofounder, I asked myself if we could build an “AI founder” that could run a company autonomously. I came up with a base architecture, which I described to ChatGPT and started building. About 50 prompts later, I shared this working prototype. Interestingly, I […]

  • Task-driven Autonomous Agent Utilizing GPT-4, Pinecone, and LangChain for Diverse Applications

    NOTE: This article was written by GPT-4 based on the code base. For more info, read this. Abstract: In this research, we propose a novel task-driven autonomous agent that leverages OpenAI’s GPT-4 language model, Pinecone vector search, and the LangChain framework to perform a wide range of tasks across diverse domains. Our system is capable […]

  • Upgrades to our AI Capabilities 🦾

    We’re continuing to add to our AI suite of capabilities, some of them shared. Braindump 🧠 Ever feel like you have a lot of ideas and you need to get them out and organized? Introducing Braindump – an iOS shortcut template that takes your voice ramblings, organizes them into neatly formatted notes, and emails them back to […]

  • We’re building the future of VC (AI Tools)

    Happy new year! As we kick off 2023, we’re leaning into our brand as VCs who don’t just invest, but build – we roll up our sleeves and push innovation forward – for ourselves, the industry, and the ecosystem at large. We’re only a few days into 2023, and we’ve already generated 3500+ go-to-market strategies […]

  • My Build-in-Public Log

    Yohei’s Build-In-Public Log ⚙️ I shared some of our recent AI experiments last month. As we continued to experiment, wanted an easy way to share this with fellow builders. Jump to yohei.me, and filter by category, tutorials, and more. Here are some highlights from this past month: Pitch Deck Parser 🔍 Calling this out specifically due to how amazing it […]