Last February, Andrej Karpathy, co-founder of OpenAI and founder of Tesla Autopilot, accidentally invented a now ubiquitous term when he tweeted about ‘vibe coding’. The phrase described a form of AI-assisted coding that lowered the barrier to entry for software engineering and quickly conquered hobbyists and Silicon Valley alike. To celebrate the one-year anniversary of that viral moment, Karpathy has introduced a new, albeit less snappy, phrase: “agentic engineering.”
The difference? While vibe coding was mostly for fun and made possible by early AI coding tools, agentic engineering represents a more advanced phase of software development that is becoming increasingly common in professional environments. The addition of “agentic” into the term makes sense because, for the most part, engineers no longer write code directly. Instead, they direct and supervise agents who do, Karpathy said in a recent post on
Karpathy said he was surprised by the buzz his first vibe-coding post generated, calling it a “throwaway tweet that I just fired off without thinking.” But the researcher’s storied career in the AI industry lends additional credibility to his views. According to Google Trends, online searches for the term exploded last year, with the phenomenon featured in countless news articles and think pieces. It was even named Collins Dictionary’s word of the year for 2025.
A member of OpenAI’s 11-person founding team, Karpathy focused on generative modeling, computer vision and reinforcement learning at the ChatGPT maker before leaving for Tesla in 2017 to lead its Autopilot efforts. Karpathy returned to OpenAI in 2023 and left again a year later to launch Eureka Labs, an AI education company. Outside of AI development, he also has more than 1.2 million followers on YouTube, where he posts educational tutorials.
A lot has changed in a year. While vibe coding originally referred to engineers experimenting with AI during informal weekend projects, Karpathy noted that large language models (LLMs) have improved so much that their use has now become commonplace among professional developers. That’s where agentic engineering comes into play. The approach aims to “claim the impact of using agents, but without any compromise on the quality of the software.”
The rapid evolution of AI-powered coding is reflected not only in new terminology, but also in a wave of investment. Cursor, one of the startups at the center of the vibecoding boom, raised $2.3 billion last November in a funding round that nearly tripled its valuation to $29.3 billion. Stockholm-based Lovable was valued at $6.6 billion the following month after raising $330 million. Replit, another major AI coding company, is reportedly nearing a new $400 million funding round that could boost its valuation to $9 billion. These startups now face increasing competition from established AI developers like Anthropic and OpenAI, both of which have doubled down on their coding roles in recent months.
In addition to his roles at major tech companies and his knack for viral buzzwords, Karpathy is also an active AI investor. According to Crunchbase, he has backed fourteen startups and invested in companies building the autonomous technologies behind trends like agentic engineering. That includes a 2024 funding round for /dev/agents, which is developing an operating system for AI agents, and a 2022 funding round for Adept, which builds assistants that can automate software workflows.
