-
Software is no longer built for people, but for AI agents, said the CEO of Vercel.
-
“Your customer is the agent who uses the developer or non-developer,” said Guillermo Rauch.
-
AI agents have increased and they could change how apps and software deal with users.
The future of software is not built for people – it is being built for machines, said Vercel’s CEO, Guillermo Rauch.
“Your customer is no longer the developer,” said Rauch about an episode of the Podcast “Sequoia Capital” Tuesday. “Your customer is the agent who uses the developer or non-developer.”
The CEO of the startup of the web infrastructure, last year with a value of $ 3.25 billion, said that the code was not only written for people to read or communicate with it. It is increasingly being written, so that AI agents can understand, use and expand it.
“That is actually a pretty important change,” said Rauch. “Is there anything that I could change about that API that actually promotes the LLM is the, quote unquote, entity or user of this API?”
This new AI-first era means that software tools may have to evolve on the basis of how large language models deal with it.
“The strengths and weaknesses of LLMS will inform the development of runimes, languages, typing roles and frameworks of the future,” said Rauch.
Rauch also said that in the AI era the newer users of Vercel – who may not be developers, but designers, marketers or even AI agents – expect things to work.
Developers were used to dealing with errors and “terrible, negative feedback all day long,” he said. But today’s users have a much shorter fuse if something goes wrong.
Yet he sees that as an “amazing pressure” for product builders. “You want something that works 99.99% of the time,” he added.
Last year Vercel collected $ 250 million in a series E -round led by Accel, with investors including Tiger Global and GV.
Rauch and Vercel did not respond to a request for comments from Business Insider.
Rise of AI agents
2025 is praised as the year of AI agents. They can change how the internet works and how apps and software deal with users.
Bernstein analysts wrote in February that although websites and apps will not disappear, users may no longer communicate with them. Instead, they have access to information, content and widgets via an AI assistant that becomes ‘the aggregator of the aggregators’.
“If it scales and takes place as we think it could be, this changes. Everything. The aggregators are broken down and much of the consumer internet can be structural shorts. Welcome to the Agentic AI era,” the analysts wrote. “There is nowhere to hide.”
But these agents are not perfect. Researchers have warned that agent errors rule and come together with every step they take.
“An error at every step can derail the entire task. The more steps are involved, the higher the chance that something will go wrong by the end,” wrote Patronus AI, a startup that helps companies evaluate and optimize and optimize on his blog.
The startup has built a statistical model that discovered that an agent with an error percentage of 1% per step can be exacerbated to a 63% chance of the 100th step.
Nevertheless, they said that guardrails – such as filters, rules and tools that can be used to identify and remove inaccurate content – can help with reducing error percentages. Small improvements “can cause large reductions in the chance of error,” said Patronus AI.
Read the original article about Business Insider
