Cohere CEO Aidan Gomez sees AI’s pathway to profitability

Cohere's Approach to Enterprise AI and the Future of Generative AI

Cohere, a Toronto-based company with close ties to Google, is bringing generative AI to businesses. In an interview with Cohere’s President and COO, Martin Kon, he discussed how its machine learning models are being used within enterprise companies. Cohere’s focus is on making AI accessible to all developers in a manner that is not prohibitively expensive, and is not hard as the barrier to entry is low.

Cohere has two types of LLM (large language model): generation and representation. The former is what ChatGPT does, the latter is for understanding language (for example, to do sentiment analysis). Each type comes in different sizes: small, medium, large, and xlarge. There are various trade-offs between the size of the model and the speed it can work at. Cohere’s base model has 52 billion parameters, based on the Stanford HELM rankings (Holistic Evaluation of Language Models).

Cohere has a strategic relationship with Google and is the largest consumer of TPUs (Tensor Processing Unit) outside of Google itself. This gives Cohere access to enormous compute resources that are needed to pre-train the models.

Cohere’s models are showing higher accuracy than a number of well-known models which are 3x larger, including GPT-3, Jurassic-1 Jumbo, and BLOOM. Cohere’s latest model, Command, gets re-tuned every week, which means that every week, the performance of command is expected to improve.

Cohere is working with customers to create custom LLMs based on their proprietary data. Cohere is also focused on data protection, deploying its models inside enterprises’ secure data environment. This makes Cohere an attractive option for businesses looking for AI solutions that are cloud-agnostic, accessible, customizable, and data-secure.

Cohere is experiencing high demand from major enterprises and is excited to launch enterprise-grade generative AI solutions for its clients. The company aims to help organizations integrate generative AI into their operations, redefine business processes, train and upskill workforces, and use this emerging technology to tackle some of the toughest current challenges.

Cohere, a Toronto-based startup, is gaining attention in the race to build the most capable large language models (LLMs). Founded by Aidan Gomez, Ivan Zhang, and Nick Frosst, Cohere aims to make AI technology accessible to all developers in a cost-effective and efficient manner. Unlike OpenAI, which offers consumer chatbots, Cohere is focusing solely on large corporate customers, providing them with AI models that can be fine-tuned with their proprietary data for specific use cases.

Cohere's approach is to build models that are efficient and scalable, with a focus on business-critical capabilities. Their latest model, Command R+, is tailored for real-world business applications and workflows. It offers retrieval-augmented generation (RAG) with citations to reduce hallucinations, multilingual coverage in 10 key languages, and tool use to automate sophisticated business processes.

Cohere's mission is to make their technology accessible to developers in a manner that is not prohibitively expensive and is not hard to use. They operate on multiple clouds, including AWS, Google Cloud, Oracle Cloud Infrastructure, and Microsoft Azure, making their models available to a wider range of customers.

As the AI industry continues to grow, Cohere faces competition from Big Tech companies like Google and Meta, as well as other startups like Anthropic and Mistral. However, Cohere's focus on the enterprise market and its partnerships with major cloud providers make it a strong contender in the AI race.