At a recent Goldman Sachs conference, OpenAI CFO Sarah Friar unveiled the company’s ambitious plans to build an artificial intelligence (AI) agent that can do all the work of software engineers, not just augment their skills. This revolutionary agent, dubbed “A-SWE,” represents the third phase of OpenAI’s development of agentic AI, and its capabilities are set to transform the way software is developed and maintained.
“Not only does it build it, it does all the things that software engineers hate to do” such as quality assurance tests, bug testing, and documentation, said Friar. “So suddenly, you can force multiply your software engineering workforce.”
The concept of A-SWE is not new, with Cognition AI already releasing an AI software engineer called Devin. However, OpenAI’s A-SWE represents a significant leap forward in the development of agentic AI. Unlike Devin, which can plan and execute complex engineering tasks within a gated environment, A-SWE can perform a wide range of tasks, including building, testing, and documenting software.
- Devin, the AI software engineer from Cognition AI, can plan and execute complex engineering tasks, but A-SWE takes it to the next level by performing all the tasks that software engineers dread.
- A-SWE is designed to work independently, using its own judgment and decision-making capabilities to build, test, and document software.
- The development of A-SWE marks the third phase of OpenAI’s agentic AI development, following the release of Deep Research and Operator.
Friar also revealed that OpenAI is working on a new phase of AI agents that will extend human knowledge to create something truly novel and unprecedented in the world. This phase will focus on developing AI models that can create new concepts and discoveries, rather than simply augmenting existing knowledge. “Academics are telling us that AI models are coming up with novel things in their field. They don’t yet know if those novel things are real, because they need to go and test” and see if it’s a new discovery.
Friar also highlighted the importance of emotional quotient or intelligence (EQ) in AI models. OpenAI has been training its GPT-4.5 model to have more EQ, which is expected to make it more human-like and better suited for tasks that require creativity and design.
“We’re actually spending a bunch of time thinking about the more EQ side of models,” Friar said. “We spent a lot more time training (GPT-4.5) to have what Silicon Valley loves to call vibes, but effectively EQ.”
Despite the progress made by OpenAI, human workers remain cautious about the impact of AI on their jobs. A recent PYMNTS Intelligence report found that more than half of survey respondents think AI poses a significant risk of widespread job displacement. The Stargate project aims to solve this problem by providing a comprehensive AI infrastructure provider and applications developer.
OpenAI’s Vertical Growth Path
Friar emphasized that OpenAI is moving beyond being solely a model builder to becoming a comprehensive AI infrastructure provider and applications developer. This shift is expected to yield significant benefits, including improved data center technology and increased IP ownership.
“Today, OpenAI is so much more,” Friar said. “We’re going down into data center technology … and we feel like we’re creating a lot of IP there, and it’s really important for us to own that.”
The Stargate project is expected to provide a solution to the problem of inadequate computing power, which has previously delayed the release of OpenAI’s models. The project aims to yield 10 gigawatts of computing power, which would exceed the entire electrical consumption of Ireland. “We’re not going to be held back by computing limitations anymore,” Friar said.
Friar also highlighted the importance of strategic investments in computing infrastructure. She compared OpenAI’s investment in Stargate to Amazon’s strategic decision to develop AWS rather than outsourcing cloud computing to competitors.
“Think about Amazon, at that moment where they’re rocking it on eCommerce, they see AWS starting to take shape. If at that stage they decided to go outsource to the upstart Google or something, giving all that IP of AWS away, think about how different the company would be today,” she said.
Despite the progress made by OpenAI, Friar acknowledged that computing limitations remain a significant constraint. However, she emphasized that the Stargate project will provide a solution to this problem, enabling the startup to build and train AI models without worrying about inadequate computing power. “We’re going to be able to build and train AI models without worrying about the limitations of our hardware,” Friar said.
Looking ahead, Friar indicated that an IPO is not imminent, given the startup’s development and expansion plans. However, she emphasized that being a public company is good hygiene, and it remains a possibility in the future.
news is a contributor at Soozo. We are committed to providing well-researched, accurate, and valuable content to our readers.
