In a recent LinkedIn post, Lenny Rachitsky shares key takeaways from a discussion with Alexander Embiricos, OpenAI’s Codex Product Lead, highlighting that the current pace of AI development outstrips our ability to fully leverage its capabilities.
Rachitsky emphasizes that even if AI models ceased improving today, there remains a significant period of product development required to unlock their full potential. He notes that the technology is currently outpacing human proficiency in its optimal application.
“Even if AI models stopped improving tomorrow, there are still years of product work left to unlock their potential. The technology is ahead of our ability to use it optimally.”
Unlocking AI’s Potential Through Strategic Application
A core insight from Rachitsky’s post revolves around how to best extract value from AI tools like OpenAI’s Codex. Contrary to what might be expected, Rachitsky relays Embiricos’s advice that these powerful tools should be directed towards the most challenging problems, not the simplest ones.
As Lenny Rachitsky explains, “The key to getting value from Codex: give it your hardest problems, not your easiest. These tools are built to tackle gnarly bugs and complex tasks, not simple ones. Start with something you’d otherwise spend hours on.” This suggests a paradigm shift in how developers and teams should approach AI assistance, focusing on augmenting human capacity for complex problem-solving.
Evolution of AI Integration and Productivity Bottlenecks
The post also delves into the practical evolution of AI tools and the emerging bottlenecks to productivity. Rachitsky points out that OpenAI’s initial Codex product, which ran asynchronously in the cloud, proved less accessible to newcomers than when it was integrated directly into developers’ code editors.
“OpenAI’s initial Codex product was ‘too far in the future.’ It ran in the cloud asynchronously, which was great for power users but hard for newcomers. Growth exploded when they brought it back to where engineers already work: inside their code editor, on their own computer.”
This shift underscores the importance of user experience and seamless integration for widespread adoption and effective utilization of AI. Furthermore, Rachitsky highlights a surprising limitation: human typing speed and prompt engineering proficiency are becoming the primary constraints on AI-driven productivity.
According to Rachitsky, “The biggest bottleneck to AI productivity isn’t the AI; it’s how fast humans can type. The limiting factors are how fast you can type prompts and how quickly you can review AI-generated work.” He elaborates that until AI can more reliably validate its own output and proactively offer assistance, the full productivity gains remain elusive.
The Shifting Landscape of Software Development and Business Strategy
The implications of AI extend beyond code generation, impacting entire teams and strategic decision-making. Rachitsky notes that designers at OpenAI are now actively writing and shipping their own code, empowered by AI assistance to maintain functional prototypes and even submit them for production.
Looking ahead, Rachitsky suggests that coding might become the universal language through which AI accomplishes tasks. Instead of relying on interfaces or integrations, AI could generate small programs on the fly. This perspective implies that coding proficiency could become a fundamental skill for all AI assistants.
Finally, Rachitsky touches upon the strategic advantage in today’s business environment. He argues that a deep understanding of specific customer needs has become more critical than the ability to build, as the ease of building is rapidly increasing due to AI.
“If you’re starting a company today, deep understanding of a specific customer matters more than being good at building. Building is getting easier. Knowing what to build—and for whom—is the real advantage now.”
This insight frames the future of entrepreneurship around market insight and customer empathy, rather than solely technical execution.
📝 About This Content
This article is based on insights shared by Lenny Rachitsky on LinkedIn.
📅 Originally posted on December 15, 2025 | View original post on LinkedIn →