Sam Altman believes that AI will significantly alter the role of software engineers, leading to more people creating and capturing value from computers. While the nature of the job will change, with less time spent on tasks like typing and debugging code, the demand for software will continue to increase, driven by personalized software experiences. This shift will result in a greater percentage of the world's GDP being created and consumed through software engineering.
The Persistent Challenge of Go-to-Market
Altman acknowledges that while AI has made it easier to build software, the challenge of go-to-market (GTM) remains difficult. He notes that even with AI transforming software engineering, the fundamental rules of building a business, creating differentiated value, and getting the GTM function working still apply. Human attention remains a limited commodity, making it hard to stand out, even with AI-driven automation in sales and marketing.
Altman on the Deflationary Impact of AI
Sam Altman posits that AI will be massively deflationary, leading to radically cheaper goods and services, particularly in areas not restricted by social or governmental policies. He anticipates that by the end of the year, a small amount of inference and a good idea could create software that would have previously taken teams a year to develop. This economic shift should empower individuals and equalize opportunities, provided that policy doesn't hinder its progress.
OpenAI's Focus on General-Purpose Models
Altman clarifies that OpenAI prioritizes developing very good general-purpose models over highly specialized ones. He admits that GPT-5.2 focused more on intelligence, reasoning, and coding, leading to a decline in writing quality compared to GPT-4.5. However, he asserts that future models will aim to excel in all dimensions, including writing, as these capabilities are seen as complementary and essential for creating comprehensive applications.
Altman Predicts Dramatic Cost Reductions
Sam Altman predicts that OpenAI should be able to deliver GPT-5.2x-level intelligence by the end of 2027 for at least 100 times less cost. He also notes that while OpenAI has focused on reducing costs, there's increasing demand for faster output, even at a higher price. Balancing these priorities will be crucial, but Altman assures that significant cost reductions are achievable if that's what the market demands.
The Rise of Personalized Software
Altman observes a shift towards software being written just for individual users, with tools constantly evolving and converging to meet specific needs. He notes that internally at OpenAI, people have adopted Codecs for their workflows, each with their own custom things and use cases. This trend suggests that static software will be replaced by dynamic, personalized tools that adapt to individual user behaviors and preferences.
Altman on Building Durable Startups
Sam Altman advises builders to focus on creating companies that would benefit from, rather than be threatened by, improvements in AI models. He emphasizes that while AI can accelerate work, the fundamental rules of building a successful startup—solving the go-to-market problem, providing something sticky, and establishing a competitive advantage—remain unchanged. Startups should aim to build durable advantages that persist even as AI capabilities advance.
AI as a Tool for Idea Generation
Altman suggests that AI can be a powerful tool for improving the quality of ideas, acting as a brainstorming partner that provides seeds to build on. He envisions AI tools that can analyze past work and code to suggest useful or interesting ideas, similar to a "Paul Graham bot." Even if most ideas are bad, the ability to generate a high volume of suggestions can lead to significant contributions.
The Importance of Human Resilience in Biosecurity
Altman emphasizes the need for a resilience-based approach to AI security, particularly in biosecurity, rather than relying solely on blocking harmful applications. He draws an analogy to fire safety, where society shifted from restricting fire to building resilience through fire codes and flame-resistant materials. AI will be both a problem and a solution for bioterrorism, requiring a society-wide effort to build infrastructure for resilience.
Altman on the Value of Human Connection
Sam Altman believes that human connection will become more valuable, not less, in a world filled with AI. He suggests that people will increasingly value getting together and working with others, and that AI can enable collaborative experiences in ways that no other technology has. He envisions AI as a part of group brainstorms and problem-solving sessions, helping the group perform better.
The Underestimated Failure Mode of AI Agents
Altman warns that the power and convenience of AI agents may lead to a "yolo" attitude, where users grant them unsupervised access despite potentially catastrophic failure rates. As models become more capable and harder to understand, misalignments or complex problems could emerge over time, creating security vulnerabilities. He emphasizes the need for robust security infrastructure to prevent sleepwalking into unforeseen consequences.
Altman on the Future of Education
Sam Altman expresses a preference for keeping computers out of kindergarten, emphasizing the importance of physical play and social interaction for young children. While acknowledging that AI tools pose a problem for the current education system, he suggests that this necessitates changing the way people are taught, rather than restricting access to AI. He believes that the practice of writing remains important for learning how to think, but the methods of teaching and evaluating thinking ability should evolve.
The Future of Software Developer Interviews
Altman reveals that OpenAI is planning to slow down hiring due to the increasing capabilities of AI, which allow them to do more with fewer people. He notes that internal policies have not yet adapted to a majority-AI workforce. The company aims to change its interview process to assess how effectively candidates can work in this new way, potentially by having them complete tasks in 10-20 minutes that would have been impossible for one person to do in two weeks last year.
The Importance of AI Adoption in Companies
Sam Altman stresses the importance of companies adopting AI aggressively and hiring people who can use the tools effectively. He warns that companies that fail to do so will be outcompeted by fully AI companies. This adoption is crucial to prevent a destabilizing scenario where AI concentrates power in the hands of a few, highlighting the need for companies to integrate AI quickly.
The Enduring Value of Human Creativity
Altman observes that consumers report dramatically higher appreciation for art if they are told a person made it versus an AI. He believes that people care a lot about other people and very little about machines. Even if AI-generated art is technically impressive, people's subjective appreciation goes down when they know it was made by a machine. However, if the art is even a little bit human-directed, people don't seem to have the same strong negative reaction.
Altman on the Future of Personalization and Memory
Sam Altman expresses his readiness for AI to have access to his entire computer and internet history to provide personalized experiences. He believes that the value from this level of access is so high that he doesn't feel uncomfortable about it. While acknowledging the importance of security and privacy, he suggests that AI should have a deep understanding of the complex rules and interactions of a user's life to know what to use when and what to expose where.
The Most Important Skills in the Age of AI
Altman identifies soft skills such as high agency, idea generation, resilience, and adaptability as the most important skills to learn in the age of AI. He notes that these skills are learnable and can be developed through focused training. These skills will matter more than any specific technical skill, as the world is rapidly changing and requires individuals to be adaptable and proactive.
