A new AI startup is on the move to compete with AI behemoths, OpenAI and Anthropic.
The San Francisco-based startup is called Writer, a new AI model in the game with a more 'cost-effective' approach, compared to other giants.
As reported by CNBC, it only spent $700,000 to cover all expenses, including data, GPUs, and manpower. Developing similar AI models typically costs millions. Currently, Writer is in the process of raising up to $200 million from investors, which will bring its valuation to $1.9 billion. In 2023, it has a value of over $500 million after raising $100 million.
Now, what made the AI startup cost-effective is its use of synthetic data, which is AI-generated information designed to imitate real-world data while maintaining user privacy. With concerns of confidentiality swirling in the industry and the risk of running out of publicly available training data in the coming years, it has become a popular way to train AI models instead.
Amazon uses synthetic data for Alexa, while Meta uses it for training its Llama models. Recently, OpenAI has also caught a whiff of the trend and has begun using it for its models.
Is Synthetic Data Better?
For users, companies turning to using synthetic data is better as it can protect their privacy. However, VCPost mentioned that experts believe it will lead to lower model performance and worsen biases among AI models.
For Writer, their approach is to refine and refine their synthetic data. CTO and co-founder, Alshikh clarified that they do not train their AI models on fake or random data. They take up real and factual information and convert it into synthetic data to make it "cleaner" for model training.
Writer's generative AI, therefore, will allow users to use its LLMs for human-like texts for several purposes. Besides generating texts, it can analyze, summarize, and develop custom AI apps. Currently, Writer has over 250 enterprise customers who enjoy their approach more.
Join the Conversation