SuperFocus chooses DeepSeek for cost and privacy

SuperFocus chooses DeepSeek for cost and privacy
SuperFocus chooses DeepSeek for cost and privacy

Steve Hsu, cofounder of AI startup SuperFocus, plans to transition from using closed-source AI models to the open-source DeepSeek system. This decision is driven by three main factors: cost, speed, and privacy. “DeepSeek models are significantly cheaper compared to other proprietary models,” Hsu stated.

The cost savings are particularly impactful in industries like customer service, where the overhead for employing human workers can be high. With the new DeepSeek-V3 model, SuperFocus can reduce costs by up to 30 times compared to their current solutions. SuperFocus frequently deals with applications requiring real-time interactions, such as voice-driven customer service.

Hsu revealed that the models from DeepSeek boast substantial improvements in response times. “The DeepSeek team has implemented clever optimizations to enhance speed, which means the models can generate responses faster,” he explained.

DeepSeek enhances cost and privacy

This reduced latency will make interactions feel more natural for users. Privacy is a significant concern for many of SuperFocus’s clients, particularly those handling sensitive information, such as private equity funds. Hsu noted, “DeepSeek’s open-source nature allows us to run the models on our own hardware or on secure cloud servers, ensuring that client data remains private.” This contrasts with some other closed-source models which require data to be processed through their proprietary systems, posing potential privacy risks.

Hsu also highlighted the broader implications of switching to open-source AI models. He believes it will drive innovation and competition in the field, ultimately benefitting consumers by lowering costs and increasing the quality of AI applications. “The competition between the US and China in AI development is intense,” Hsu said.

See also  Winnti group employs advanced malware in cyber operations

“It’s going to be a close race, but one that pushes the field forward significantly. Consumers and scientists alike will benefit from the advancements and availability of high-quality open-source models.”

SuperFocus’s decision to adopt DeepSeek serves as a strategic move to leverage cost efficiency, speed, and privacy while also positioning the company at the forefront of a rapidly evolving AI landscape.

More Stories