Cohere provides a cloud-agnostic platform for enterprises to leverage large language models (LLMs) for various tasks.
The future of AI interfaces lies in the development of intelligent agents that can carry out complex, multi-step tasks autonomously.
Deep dives
Cohere: Bringing General Purpose Language Models to Industry
Cohere, founded by Aiden Gomez, Nick Frost, and Ivan Zhang, aims to bring large language models (LLMs) to industries by offering access to their models through an API. LLMs have the potential to solve a variety of reading and writing tasks, such as summarization, content creation, and sentiment analysis. In the early days, the market was skeptical about the commercial applications of LLMs. However, with breakthroughs like GPT-3 and the emergence of dialogue models, consumers and businesses have experienced the value and desirability of using LLMs. Now, every board and C-suite is waking up to the immense potential and demand for LLM technology. Cohere positioned itself as a cloud-agnostic platform, providing enterprises with secure and customizable models that can be deployed on-premises or across multiple cloud providers. With marquee customers like LivePerson and Oracle, Cohere is focused on delivering value to customers by continuously improving their models and creating a delightful user experience.
The Evolution and Impact of AI in Tech
AI, particularly large language models (LLMs), is generating significant hype in the tech industry. The progress in LLMs, such as chat GPT, has enabled businesses to explore the value of AI and generate new strategies for its implementation. However, the path from foundation models to tangible business use cases is not always clear. Cohere, founded by Aiden Gomez, Nick Frost, and Ivan Zhang, recognized the transformative potential of LLMs and aimed to facilitate their adoption in industry. Starting with a focus on building the infrastructure to train and fine-tune LLMs, Cohere's mission evolved to provide a general platform for enterprises to leverage LLMs for various tasks. Despite initial skepticism, the market eventually recognized the impact and value of LLMs, leading to increased demand and urgency for Cohere's platform.
The Future of AI: From Transactional to Agent-Based Interfaces
Currently, AI models are often used for transactional purposes, providing quick solutions to specific tasks. However, the future of AI interfaces lies in the development of intelligent agents that can carry out complex, multi-step tasks on behalf of users. This shift towards agent-based interfaces would allow models to plan, execute, and complete tasks autonomously by breaking them down into subtasks and returning a comprehensive solution to the user. While current transactional interfaces like APIs and chat interfaces may still be relevant, the focus will increasingly shift towards long-term planning and agenthood. The challenge lies in improving model accuracy to ensure high success rates and the ability to carry out intricate, multi-step trajectories independently.
The Importance of Independence and Cloud Agnosticism
Cohere distinguishes itself by maintaining independence from large cloud providers. While competitors may have close ties with specific providers, Cohere positions itself as a platform that can operate across multiple clouds or even on-premises. This cloud agnosticism allows enterprises greater flexibility as they are not locked into a single cloud provider, ensuring strategic decision-making and negotiating leverage. Furthermore, this approach facilitates data privacy and security, particularly for highly regulated industries, such as healthcare and government, where concerns about data confidentiality are paramount. Cohere's commitment to independence and cloud agnosticism aligns with the growing market demand for a trusted partner that can offer AI solutions while prioritizing data privacy and security.
It's no secret that the biggest hype in tech is around AI and large language models (LLMs).
Cohere, founded in 2019, builds general purpose language models and provides access to them through an API. Its goal is to help enterprise leverage the power of LLMs to complete a range of useful tasks. With Cohere, companies can bring natural language processing to their applications without spending big on training models or hiring.
Aidan Gomez, co-founder and CEO of Cohere, was credited as a co-author on the influential paper ‘Attention is All You Need’, which established the transformer architecture that drives the world’s most powerful LLMs.
For our first episode, Aidan joined Contrary Research Radio to talk about Cohere, its place in the AI landscape, and his predictions for the future of LLM development.