Archive — US — techcrunch
Inception emerges from stealth with a new type of AI model
Inception, a new AI company founded by Stanford professor Stefano Ermon, introduces a novel diffusion-based large language model (DLM) that combines capabilities of traditional LLMs while outperforming them in speed and cost efficiency. It allows for parallel text generation, contrasting with the sequential nature of current LLMs. Early customers include Fortune 100 companies, drawn by the potential for reduced latency and increased performance.
Key points
- Inception is a Palo Alto-based company founded by Stanford professor Stefano Ermon.
- The company has developed a diffusion-based large language model (DLM) which offers faster performance and reduced costs compared to traditional large language models (LLMs).
- Diffusion models can generate text in parallel, enhancing efficiency over the sequential generation method of LLMs.
- Inception has secured funding from the Mayfield Fund and has already attracted several Fortune 100 clients.
- The DLMs can run significantly faster and at a lower cost than existing models, claiming to be 10x quicker and cheaper.
Why it matters
- The emergence of Inception aligns with current trends in AI focusing on efficiency and cost-effectiveness.
- Diffusion technology is gaining prominence in the AI sector, evidenced by its successful application in image and audio generation.
- As AI continues to evolve, the need for faster and cheaper AI solutions is increasing, particularly for enterprise clients.
Inception's innovation in AI through its diffusion-based model could significantly reshape the landscape for language generation, offering promising speed and cost benefits that may redefine user expectations and industry standards.