TL;DR:
- Growing code generation trend: A substantial portion (40%) of code is now AI-assisted.
- Startups rising: Companies emerging to prevent AI-augmented code errors.
- Funding success: Digma secures $6 million, Kolena raises $15 million, Braintrust unveils $3 million.
- Braintrust’s mission: Operating system for engineers to ensure AI code reliability.
- Impressive backers: Renowned investors like Elad Gil, Adam D’Angelo, Clem Delangue, and Greg Brockman support Braintrust.
- Founder’s background: Ankur Goyal’s journey from biology to AI software reliability.
- Unique offering: Braintrust operates within partner companies’ cloud environments.
- Competitive landscape: Deepchecks, an Israeli startup, is also securing seed funding.
Main AI News:
The landscape of code generation is undergoing a significant transformation, with an increasing portion of code labeled “AI-assisted.” According to Scott Guthrie, Microsoft’s Executive Vice President of Cloud and AI, approximately 40% of code uploaded to the AI developer tool GitHub Copilot is now both “AI-generated and unmodified.” This trend has ignited the emergence of startups dedicated to preventing AI-augmented code from causing disruptions, and investors are taking note.
Recently, Digma, an Israel-based startup, secured $6 million in seed funding for its continuous feedback platform. This platform operates locally on developers’ machines, offering in-depth code analysis, including generative AI-created code, to pinpoint potential issues. In parallel, a San Francisco-based testing platform named Kolena secured $15 million to develop tools aimed at testing, benchmarking, and validating the performance of AI models.
Today, Braintrust, a four-person startup from the Bay Area, unveils a fresh funding round totaling $3 million. Braintrust, as described by co-founder and CEO Ankur Goyal, functions as an “operating system for engineers building AI software.” It equips developers with the tools necessary to ensure the accuracy of AI models and prevent them from producing erroneous results. For example, developers working on customer support chatbots can use Braintrust’s technology to ensure that their chatbots provide accurate responses, avoiding the dissemination of false information.
Braintrust boasts a roster of savvy backers, including renowned angel investor Elad Gil, who played a pivotal role in incubating Braintrust’s initial product. Notable investors like Adam D’Angelo of Quora, Clem Delangue of the AI powerhouse HuggingFace, and OpenAI co-founder Greg Brockman have also lent their support to the startup.
The question remains whether this impressive investor syndicate can propel Braintrust to the forefront of the AI software reliability landscape. In the interim, Goyal is steadfast in his commitment to addressing the challenge of safeguarding against AI code disruptions in company workflows.
Goyal’s journey into the world of AI software reliability was not a traditional one. Raised in Pittsburgh in a family of doctors, he initially aspired to follow in their footsteps. However, a high school linear algebra class that introduced him to Google’s PageRank algorithm changed the trajectory of his career. He shifted from biology to computer science at Carnegie Mellon University, where his passion for technology led him to drop out during his junior year. His endeavor to build a relational database system at MemSQL, an early Y Combinator alumnus, ultimately paved the way for his co-founding of Impira. When Impira was acquired by Figma late last year, Goyal assumed leadership of its machine learning platform.
Goyal’s experience at Figma provided valuable insights into the challenges of building high-quality software in the era of AI. His realization that AI code is inherently non-deterministic and requires extensive real-world testing prompted him to launch Braintrust. The startup, although yet to commercialize its product, collaborates with data-rich companies like Zapier and Coda, who are currently beta testing Braintrust’s innovative solutions.
One of Braintrust’s unique offerings is its ability to operate within the cloud environments of its partner companies, addressing compliance concerns that could otherwise hinder adoption within enterprise settings.
While it’s still early days, competition in this space is expected to intensify in the coming months and years. Deepchecks, an Israeli startup specializing in “continuous validation for AI,” is another player that recently secured seed funding.
Goyal views Braintrust as the solution he wished he had during his tenure at Figma, a product that, until recently, was absent from the AI landscape. He emphasizes that while continuous integration methodologies have evolved over the past decade in software development, a similar structured approach for AI software was lacking until the advent of Braintrust.
Conclusion:
The growing investment in startups dedicated to safeguarding AI-generated code underscores the increasing importance of code integrity in AI development. Braintrust, with its unique approach and prominent backing, aims to lead this market by providing essential solutions for developers to ensure the reliability of their AI software, addressing a critical need in the industry. As competition in this space intensifies, we can expect further innovation and growth in AI code quality assurance solutions.
