open source API puts generative AI within reach of any developer


Almost everyone is trying to get a piece of the generative AI action these days. While most of the focus remains on model vendors like OpenAI, Anthropic, and Cohere, or big companies like Microsoft, Meta, Google, and Amazon, there are, in fact, plenty of startups trying to attack the generative AI problem in a variety of ways. Of methods. is one such startup. While it lacks the brand name recognition of some of these other players, it claims to be the largest open source model API with over 12,000 users, according to the company. This kind of open source traction attracts the attention of investors and the company has raised $25 million to date.

Lin Qiao, co-founder and CEO of Fireworks, explains that his company is not training the foundation model from the beginning, but rather helping improve other models for the specific needs of the business. “It can either be off-the-shelf, open source models or models that we tune or models that our customer can tune themselves. “All three varieties can be served through our inference engine API,” Qiao told TechCrunch.

Being an API, developers can plug it into their applications, get the model of their choice trained on their data, and add generative AI capabilities like asking questions very quickly. Qiao says it is fast, efficient and produces high-quality results.

Another advantage of Firework’s approach is that it allows companies to experiment with multiple models, which is important in a rapidly changing market. He said, “Our philosophy here is that we want to empower users to iterate and experiment with multiple models and have effective tools to put their data into multiple models and test them with one product. “

Perhaps even more importantly, they keep costs low by limiting the model size to 7 billion to 13 billion tokens compared to over 1 trillion tokens in ChatGPT4. Although this limits the universe of words that larger language models can understand, it enables developers to focus on much smaller, focused data sets designed to work with more limited business use cases. .

Qiao is uniquely qualified to build such a system, having previously worked at Meta, and the AI ​​platform development team with the goal of building a fast, scalable development engine to power AI across all of Meta’s products and services. Is leading. She was able to take this knowledge from working at Meta and create an API-based tool that puts that kind of power at the reach of any company without requiring the level of engineering resources of a company of Meta’s size.

The company raised $25 million in 2022 led by Benchmark with participation from Sequoia Capital and angel investors including Databricks and Snowflake. The latter two are particularly interesting strategic investors, as they are both data storage devices, and Fireworks will enable users to put that data to work.

Source link

About Author

Leave a Reply

Your email address will not be published. Required fields are marked *