Peter Steinberger: OpenClaw marks the shift to agentic AI, security risks of system-level access, and the need for AI in customer service | Lex Fridman Podcast

Key takeaways

• OpenClaw is a major advancement in AI, shifting from language processing to agency.
• AI agents with system-level access pose security risks but have transformative potential.
• The age of agentic AI will significantly impact programming and technology.
• Code Rabbit CLI safeguards against errors in AI-generated code.
• Integrating AI into customer service is crucial for efficiency and quality.
• Managing large code bases requires specialized tools for complexity and orchestration.
• Images in prompts enhance the context for AI agents.
• The integration of AI into daily life represents a significant phase shift in user experience.
• Creative problem-solving in AI transcends traditional coding expertise.
• The crypto community can be disruptive and overwhelming, hindering project progress.
• AI-generated content can lead to misinformation if not understood properly.
• The perception of security risks is often exaggerated.
• The security concerns with AI systems differ from traditional systems, necessitating new considerations.
• The evolution of apps into APIs will force companies to rethink their business models.
• AI agents will transform the app market, making many existing apps obsolete.

Guest intro

Peter Steinberger is an Austrian software developer and entrepreneur who created OpenClaw, a free and open-source AI agent framework that achieved viral popularity in early 2026 and amassed over 145,000 GitHub stars. He previously founded PSPDFKit in 2010, a B2B PDF management tool that grew to approximately 70 employees and was valued at around 100 million euros before his exit in 2021. Steinberger has become known for his unconventional development approach, shipping thousands of commits monthly by leveraging AI tools to write and test code at an unprecedented scale.

The evolution of AI capabilities

• OpenClaw represents a shift from language processing to agency in AI. – Peter Steinberger

• OpenClaw… is the AI that actually does things… definitively takes a step forward over the line from language to agency from ideas to actions.

  — Peter Steinberger

• AI agents with system-level access pose significant security risks but have transformative potential. – Peter Steinberger

• A powerful AI agent with system-level access is a security minefield but it also represents the future because when done well and securely it can be extremely useful to each of us humans as a personal assistant.

  — Peter Steinberger

• We are entering the age of agentic AI, which will significantly impact programming and technology. – Peter Steinberger

• We’re living through the OpenClaw moment, the age of the lobster, the start of the agentic AI revolution.

  — Peter Steinberger

• The integration of AI into daily life represents a significant phase shift in user experience. – Peter Steinberger

• It seems like a trivial step but it’s in some sense it’s like a phase shift in the integration of AI into your life and how it feels right.

  — Peter Steinberger

Security and ethical considerations in AI

• The power of OpenClaw comes with significant responsibility regarding data protection. – Peter Steinberger

• OpenClaw represents freedom but with freedom comes responsibility… you also have the responsibility to protect it.

  — Peter Steinberger

• AI agents with system-level access pose security risks but have transformative potential. – Peter Steinberger

• A powerful AI agent with system-level access is a security minefield but it also represents the future because when done well and securely it can be extremely useful to each of us humans as a personal assistant.

  — Peter Steinberger

• The perception of security risks is often exaggerated. – Peter Steinberger

• I think that people turn it into like a much worse light than it is again, you know, like people love attention.

  — Peter Steinberger

• The security concerns with AI systems differ from those of traditional systems, necessitating new considerations. – Peter Steinberger

• The nature of those security concerns are different than the kind of security concerns we had with non-LLM generated systems of the past.

  — Peter Steinberger

AI in customer service and coding

• Code Rabbit CLI acts as a safeguard against errors in AI-generated code. – Peter Steinberger

• That’s where Code Rabbit CLI comes in; it ensures that AI-generated code is production-ready by catching errors at that particular stage of the process.

  — Peter Steinberger

• Integrating AI into customer service is crucial for efficiency and quality. – Peter Steinberger

• I do think for scale, for efficiency, for quality, it’s important to integrate AI into that process and then Fin does a really good job of that.

  — Peter Steinberger

• Managing gigantic code bases requires specialized tooling to handle complexity and orchestration. – Peter Steinberger

• When you’re talking about a gigantic code base that requires a set of tooling that can handle the gigantic context that can handle individual people going in and being able to orchestrate the large-scale refactoring.

  — Peter Steinberger

The role of creativity in AI development

• Creative problem-solving in AI systems is a skill that transcends traditional coding expertise. – Peter Steinberger

• There’s so much world knowledge in there, so much creative problem solving a lot of it doesn’t map from if you get really good at coding that means you have to be really good at general-purpose program solving.

  — Peter Steinberger

• Magic in technology often comes from rearranging existing components in innovative ways. – Peter Steinberger

• Isn’t magic often just like you take a lot of things that are already there but bring them together in new ways.

  — Peter Steinberger

• The journey of building is more important than the end result. – Peter Steinberger

• It’s the journey mhmm my god like those things I don’t think I ever had so much fun building things because I can focus on the hard parts now.

  — Peter Steinberger

Challenges and opportunities in the crypto space

• The crypto community can be disruptive and overwhelming, often hindering project progress. – Peter Steinberger

• They came in and spammed and annoying and on Twitter they would ping me all the time my notification feed was unusable.

  — Peter Steinberger

• The toxicity and greed in the crypto community overshadow the fascinating technology behind crypto. – Peter Steinberger

• There’s so much toxicity there’s so much greed there’s so much trying to get a shortcut to manipulate to steal to snipe to to to to game the system somehow to get money all this kind of stuff.

  — Peter Steinberger

• Changing a project’s name in the crypto space requires a comprehensive and simultaneous update across multiple platforms. – Peter Steinberger

• Changing a name is hard because you have to find everything you know Twitter handle domains npm packages docker registry GitHub stuff and everything has to be you need a set of everything.

  — Peter Steinberger

The potential for misinformation with AI

• AI can create powerful narratives that may lead to misinformation if not understood properly. – Peter Steinberger

• AI is incredibly powerful but it’s not always right… it’s very easy that it just hallucinates something or just comes up with a story.

  — Peter Steinberger

• Society needs to improve its understanding of AI and the critical thinking required to assess its outputs. – Peter Steinberger

• We as a society need some catching up to do in terms of understanding that AI is incredibly powerful… critical thinking is not always in high demand anyhow in our society these days.

  — Peter Steinberger

• AI-generated content lacks the nuances of human writing. – Peter Steinberger

• It missed the nuances to the how I would write it… you can steer it towards your style but it’s not gonna be all your style.

  — Peter Steinberger

The future of apps and APIs

• The evolution of apps into APIs will force companies to rethink their business models. – Peter Steinberger

• I think there’s a space we’re just beginning to even understand what that means… it’s gonna force a lot of companies to have to shift focus and it’s kind of what the internet did right you have to rapidly rethink reconfigure what you’re selling how you’re making money.

  — Peter Steinberger

• AI agents will transform the entire app market, making many existing apps obsolete. – Peter Steinberger

• You think agents will just transform the entire app market… why do you need my fitness pal when the agent already knows where I am.

  — Peter Steinberger

• The rise of intelligent agents could lead to the elimination of up to 80% of existing apps. – Peter Steinberger

• I think that will translate into a whole category of apps that are no longer… I think you said somewhere that it might kill off 80% of apps.

  — Peter Steinberger

The impact of AI on programming and technology

• AI will be a transformative technology for society, but it will also cause significant short-term pain and suffering. – Peter Steinberger

• I’m not saying everything’s good… it’s certainly gonna be a very transformative technology for our society… there’s also… the immeasurable pain and suffering that happens at the short term scale when there’s change of any kind especially large scale transformative change that we’re about to face.

  — Peter Steinberger

• The demand for skilled builders will remain, but the nature of programming will change significantly due to advancements in tokenized intelligence. – Peter Steinberger

• There will still be a lot of demands for people that understand how to build things just that all this tokenized intelligence enables people to do a lot more a lot faster and it will be even more even faster and even more because those things are continuously improving.

  — Peter Steinberger

• The role of programmers is evolving, and they will need to adapt to new expectations and tools. – Peter Steinberger

• I think programmers are generally best equipped at this moment in history to learn the language to empathize with agents… it’s just gonna be called coding again and it’s just gonna be the new normal.

  — Peter Steinberger