What Will Software Engineers Do When Chips Are Software?

TL;DR

Chips are the new software and AI tools will automate many junior tasks, shifting engineers toward architecture and orchestration. Engineers will focus on optimizing the whole stack from silicon to AI, using agent copilots to speed design, debugging, and testing, while increasing emphasis on system-level architecture and efficiency.

Transcript

If you narrow yourself, I'm just writing code. You have to rethink your role now and in the future. In the next 3 5 years, we need less human engineers because we assigned everything to an agent engineers. Junior tasks can be replaced by agents. Sassin Gazi is the CEO of Synopsis, the company behind the chips that every modern technology depends on... Read More

Key Insights

• Chips are the new software and optimization must span the entire stack
• Synopsis provides software to synthesize code into chip layouts
• Generative AI can create test benches and RTL code for chips
• AI copilots assist engineers with debugging and documentation
• Agent engineers can replace routine tasks to speed up workflows
• There will be a shift toward system architecture and orchestration roles
• The industry faces a potential shortage of engineers by 2030
• A trillion transistors by 2030 is a referenced industry goal
• Automation will reduce time-to-design from months to weeks or days
• More videos from this channel:
  • How to Become a Software Engineer in 2025-2026
  • How to start coding in 2025 as a beginner
• More from this channel: Sajjaad Khader

Questions & Answers

Q: How does the interview describe the relationship between software and hardware in chip design?

The interview explains that chip design is becoming a software problem where the stack from silicon to AI must be optimized together. Silicon provides compute, software runs on the chip, then a system layer sits above, with AI integrating throughout. This holistic view dictates how engineers should work across levels.

Q: What role do AI copilots play in the Synopsys workflow?

AI copilots are deployed to assist engineers in tasks like debugging, writing documentation, and generating test benches. They help reduce routine overhead, speed up the design cycle, and let engineers focus on higher value work such as architecture and optimization rather than repetitive coding tasks.

Q: What change in engineer roles does the guest anticipate in the next few years?

The guest anticipates that junior tasks will be replaced by agents, leading to fewer pure coding duties for new engineers. Instead, engineers will become architects or system designers who orchestrate AI copilots and agent engineers, focusing on stack-level optimization and strategic design decisions.

Q: How does AI help in designing and testing chips according to the interview?

AI helps by generating parts of the chip design process, including RTL code, test benches, and debugging routines. This enables faster iteration, enables more complex designs to be explored quickly, and helps verify correctness throughout the workflow rather than only at the end.

Q: What does the guest say about the scale of AI chips and the need for optimization?

The guest notes that AI chips operate at very high transistor counts and that moving from billions to trillions of transistors requires optimization across the silicon, software, and system layers. More compute power enables more tasks to be done, driving a feedback loop of improvement.

Q: What workforce shortage is discussed for 2030, and how is it related to the semiconductor industry?

A study cited in the talk suggests a shortage of 1.4 million engineers by 2030, with the semiconductor industry needing about 300,000 more engineers. The required roles involve architects and domain experts who can design and integrate complex systems rather than only writing code.

Q: How does the concept of agent engineers fit into future workflows?

Agent engineers are specialized AI-enabled roles that handle tasks like verification, orchestration, and other repetitive parts of the workflow. They enable humans to focus on higher-level design decisions, while AI agents perform time-consuming tasks, thus increasing overall throughput and reducing design cycles.

Q: What is the proposed outcome for engineers by 2030 in terms of roles and productivity?

By 2030, engineers are expected to be more focused on architecture and system-level optimization, using AI copilots and agent engineers to handle routine work. The outcome is faster design processes, greater efficiency, and a shift toward roles that require deeper understanding of the stack and physics of chips.

Summary & Key Takeaways

• In this interview, the guest explains that chip design is becoming a software problem and must be optimized across the silicon, software, system, and AI stack. The role of engineers shifts toward architecture, orchestration, and leveraging AI copilots. The industry is moving toward faster chip design enabled by AI tooling.

• Synopsis provides software to translate code into physical chip layouts, enabling designers to tame complexity and accelerate production. AI is used to generate test benches, RTL code, and debugging support, reducing cycle times and increasing precision.

• The discussion covers a potential shift in workforce needs, suggesting a move from junior coding tasks to higher-level architectural work, with agent engineers handling automation tasks and humans focusing on design choices and system integration.