On June 29, Adam Cochran shared his thoughts on the current state of AI development, specifically comparing OpenAI’s ChatGPT and Anthropic’s Claude.
Founded in 2015, OpenAI began as a non-profit research company with the mission to ensure that artificial general intelligence (AGI) benefits all of humanity. The organization has since transitioned to a “capped-profit” model to attract further investment while maintaining its commitment to its original mission.
OpenAI has made significant contributions to the field of AI, including the development of GPT (Generative Pre-trained Transformer) models. Their most famous creation, ChatGPT, took the world by storm in late 2022, showcasing the potential of large language models in generating human-like text and assisting with various tasks.
The company is known for its policy of gradually releasing its AI models to the public, aiming to promote safe and ethical AI development. OpenAI has also been vocal about the potential risks of AI and the need for responsible development practices.
Anthropic, a more recent entrant to the field, was founded in 2021 by former OpenAI researchers. The company places a strong emphasis on developing safe and ethical AI systems, with a particular focus on what they call “constitutional AI” – AI systems designed with built-in safeguards and values.
Anthropic’s flagship AI model, Claude, is designed to be helpful, honest, and harmless. The company aims to create AI systems that are not only powerful but also aligned with human values and interests.
Unlike OpenAI’s more public-facing approach, Anthropic has generally maintained a lower profile, focusing on research and development rather than consumer products. However, their work on AI alignment and safety has garnered attention within the AI research community.
According to Cochran, “Anthropic in one release cycle really clobbered OpenAI.” He notes that while Claude 3.5 Sonnet isn’t significantly more intelligent than GPT-4 overall, and both are mid-sized models, Anthropic’s approach has given them an edge.
Cochran points out that Anthropic focused on tailoring their AI to customer needs, while OpenAI concentrated on developing real-time voice capabilities, which he describes as a “novelty.” He states that Claude Sonnet has a “really locked in instruction set” and “prioritizes completeness” compared to ChatGPT-4. Additionally, Claude prioritizes the use of artifacts as a tool.
These factors, Cochran argues, allow Claude to cover “90% of commercial/corporate usecases with way lower compute and therefore less API cost.” In contrast, he suggests that the “OpenAI GPT market is mostly unused,” and that OpenAI is “spending 100x the inference cost going after voice which won’t be a big commercial API seller.”
Cochran concludes with a key observation about product development in the AI space: “Goes to show you don’t need a 10x leap in tech to build a 10x product. You just need to tailor the experience to the customer.” This statement encapsulates his view on why Anthropic’s approach with Claude has been successful in challenging OpenAI’s position in the market.
Featured Image via Pixabay