Arm, the chip designer behind the processors in most of the world's smartphones, Macs, and even the servers powering AI chatbots, is undergoing a fundamental transformation. Long focused on licensing blueprints, the company now builds its own AGI CPU and is shifting software development toward natural language. Alex Spinelli, Arm’s senior vice president of AI and developer platforms, shared his vision for how this change will reshape engineering. Here are 10 key insights from that conversation.
1. Arm Moves from Designer to Manufacturer
For over four decades, Arm stuck to designing chip architectures and licensing them to partners like Apple and Qualcomm. Now it’s entering the hardware business by manufacturing its own AGI CPU. This chip, already adopted by OpenAI and Meta, lets Arm compete directly with industry giants such as Intel, Nvidia, and Amazon. The move marks a strategic shift from pure intellectual property to integrated hardware-software solutions, giving Arm greater control over performance and optimization for AI workloads.
2. The AGI CPU Powers a New Era
Arm’s AGI CPU is not just another processor—it’s designed specifically to handle the demands of general artificial intelligence. By building its own chip, Arm can ensure that its hardware and software work together seamlessly from day one. This tight integration is crucial for AI applications that require low latency and high efficiency. The AGI CPU will run tasks ranging from inference in chatbots to training smaller models, making it a cornerstone of Arm’s future strategy.
3. Performix Uses AI to Debug and Optimize Code
Arm’s new Performix software suite leverages machine learning to identify problematic code and CPU bottlenecks. It provides engineers with “recipes” and performance insights that speed up debugging and optimization. Instead of manually sifting through logs, developers can rely on AI to pinpoint hotspots and suggest fixes. This tool is expected to reduce development cycles significantly, especially for complex software running on Arm-based hardware.
4. Meet Alex Spinelli, an AI-Native Engineer
Alex Spinelli brings a rare blend of experience: he helped build the TensorFlow stack that powered Google’s Gemini and worked on Amazon’s Alexa team. Now at Arm, he leads software initiatives that bridge AI and chip design. His background in both large-scale machine learning systems and consumer-facing AI products gives him a unique perspective on how engineering must evolve. He represents a new breed of leaders who understand both code and natural language interfaces.
5. English Becomes the New Programming Language
Spinelli argues that computing has steadily moved toward higher abstraction—from punch cards to assembly to high-level languages. The next step is natural language. “English is the highest level language,” he says. Engineers will increasingly write instructions in plain human language, which AI translates into executable code. This doesn’t eliminate programming but changes how we express logic. The skill shifts from syntax mastery to problem decomposition and prompt engineering.
6. Software Engineering Transforms, Not Disappears
With natural language as the new interface, many fear that traditional software engineering roles will vanish. Spinelli disagrees. “Programming doesn’t go away,” he insists. Instead, engineers will focus on architecture, design thinking, and product strategy. The ability to structure complex systems and reason about trade-offs becomes more valuable than ever. AI handles low-level coding, but humans still must think critically about what to build and why.
7. AI Agents Are Where the Real Action Happens
Spinelli emphasizes that the true impact of AI in software isn’t just code generation—it’s agents. AI agents orchestrate multiple models and APIs to perform tasks autonomously. “Agents use a lot of AI and agents are software,” he notes. Building these agents requires a deep understanding of both AI capabilities and software engineering principles. The future of development lies in creating robust, safe, and scalable agent systems.
8. Engineering Blends with Product and Design Thinking
As natural language reduces the friction of writing code, engineers must adopt a broader mindset. Spinelli sees a convergence of technical product management, design thinking, and architecture. The best engineers will be those who can articulate user needs, design experiments, and reason about system behavior—all while leveraging AI for implementation. This blending of roles demands stronger communication and strategic skills alongside technical expertise.
9. Structuring Application Stacks Requires Experience
In the new AI-driven world, the way applications are structured still matters immensely. Spinelli shares that he personally runs an OpenClaw instance with 15 small models for side projects. Deciding where to place each model, how to chain them, and how to manage dependencies requires real-world know-how. Experience in system architecture becomes a critical asset, even as coding becomes more automated.
10. Embracing Change Is the Engineer’s Superpower
The pace of change in AI and hardware is accelerating faster than ever. Spinelli advises engineers to embrace their position in the toolchain and continuously learn. The shift to natural language programming, agent-based architectures, and custom chips like the AGI CPU will disrupt many old habits. Those who adapt—by understanding AI tools, learning to think at a higher level, and staying curious—will thrive in this new landscape.
As Arm redefines its role from chip designer to hardware maker, and as software development moves toward natural language, the engineering profession stands at a crossroads. Spinelli’s insights make one thing clear: the fundamentals of problem-solving and system thinking remain essential, but the tools and techniques are evolving rapidly. The engineers who lean into this change will shape the next decade of computing.