🚀 The Future of Engineering & AI — Insights from CEOs¶
Michael Truell - CEO of Cursor¶
"We're not going straight to a world where AI does everything and engineers disappear. Instead, engineers are shifting roles—from implementers to orchestrators."
Everyone Become Engineering Managers¶
"I think something people don't talk enough about when discussing AI agents and AI engineers doing all this stuff for you… is basically we're all becoming engineering managers."
1. Engineering Manager Analogy:¶
- As AI agents take on more of the implementation work, human engineers increasingly focus on specification, review, and iteration.
- It feels like managing a team of junior engineers that “aren’t that smart yet” — they constantly check in, and you need to inspect everything.
2. Cognitive Load Shifts:¶
- Instead of writing all the code, you spend more time:
- Specifying the task with precision
- Reviewing AI output
- Iterating with tighter feedback loops
3. Two Emerging Workflows:¶
- Macro delegation: Specify everything upfront, let AI execute, then review the whole thing.
- Micro-iteration: Specify a small task, review the AI’s output, iterate step-by-step.
"Most successful users today chop things up. They don’t try to offload giant tasks all at once."
4. The Vibe Coding Trap:¶
- Over-delegating to AI without sufficient understanding can lead to messy, unmaintainable code.
- Cursor aims to empower engineers to stay in the driver’s seat while still leveraging automation.
Source: The rise of Cursor: The $300M ARR AI tool that engineers can’t stop using | Michael Truell.¶
Varun Mohan - CEO of Windsurf (acquired by OpenAI)¶
Why Varun believes 90% of Code Will Be AI-Generated — But Engineering Jobs Will Grow.
1. Code Becomes the Commodity — Taste & Intent Become the Skill¶
- Most boilerplate and repetitive logic will be handled by AI.
- Engineers shift from typing syntax to defining what needs to be built — the higher-value, human work.
- This creates demand for more engineers who can think clearly, not just code.
"AI is lowering the barrier to turn ideas into working software. That doesn’t kill engineering — it multiplies the demand."
2. Software Demand Is Unbounded¶
- As cost of production drops, demand explodes:
- More startups
- More internal tools
- More customization
- Every team will want software tailored to their niche workflows.
"If software gets 10x cheaper to build, you don’t fire engineers — you build 10x more software."
3. Engineers Become Architects and Curators¶
- Engineers will act like project leads, reviewers, product thinkers.
- They’ll manage AI-generated output, ensure quality, handle edge cases, and direct architecture decisions.
- This is analogous to how designers use Figma AI features — but the best still lead the process.
4. AI Generates Code, Not Context¶
- AI lacks full understanding of:
- Business goals
- User experience nuance
- Product strategy
- Human engineers are essential for bridging code to real-world context.
"You still need humans to know what should be built — and why."
5. More Non-Engineers Will Create Software — But Engineers Will Enable It¶
- Low-code and AI tools let PMs, ops teams, and analysts build simple tools.
- But engineers are still needed to:
- Build frameworks
- Ensure reliability
- Handle edge cases
- Maintain standards
“AI is not replacing engineers — it’s leveling them up.”
Source: Building a magical AI code editor used by over 1m developers in 4 months: Inside Windsurf.¶
Scott Wu - CEO of Cognition (maker of Devin)¶
1. Why Engineering Will Shift from “Bricklayers” to “Architects”¶
“One of the ways that we've thought about Devin is really allowing engineers to go from bricklayer to architect.”
-
Traditional Engineering is 90% Implementation
- Much of today’s engineering involves debugging, boilerplate, migrations, CI/CD tasks, and infrastructure work.
- Only ~10% is truly architectural thinking — designing systems, understanding problems deeply, and crafting solutions.
-
Devin Automates the Bricklaying
- By automating mundane and repetitive engineering tasks (e.g., writing boilerplate, fixing bugs, generating tests), Devin frees up engineers.
- Engineers can now focus on problem-solving, high-level system design, and decision-making.
-
Programming Becomes More Important, Not Less
- Even as AI takes over more coding, knowing how to instruct the computer (i.e., understanding abstractions, systems, architecture) becomes even more valuable.
- Engineers evolve into system architects: defining the problem, specifying the desired solution, and orchestrating AI agents to execute it.
2. Skills That Will Grow in Importance:¶
- System design and architecture
- Abstraction thinking and decomposition
- Precise specification and prompt engineering
- Understanding full-stack flows and trade-offs
- Collaborating with AI as a multiplier
3. Skills That Will Diminish:¶
- Low-level implementation
- Routine debugging and plumbing
- Manual CI/CD tasks and migrations
4. Moats and stickiness in AI¶
"People talk about moats in AI — distribution, scale, model performance — but I think the real moat is user stickiness. AI that helps you with your data, your voice, your style, your tools, your business — the more personal it gets, the harder it is to switch."
— Scott Wu