TL;DR
Get business pricing on networking and server gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
Linear co-founder Karri Saarinen argues that software reflects the judgment and experience of its makers, and warns that AI agents may encourage teams to produce disconnected changes. His essay is an argument about product design, not a report of a new product launch or measured industry finding.
Linear co-founder Karri Saarinen argues that software products take their character from the people making them, and warns that AI agents could make products less coherent if teams focus on producing changes rather than shaping a unified experience. His essay is a statement of perspective, not evidence that software quality across the industry has measurably declined.
Saarinen writes that two teams can speak with similar customers and use similar technology yet create different products. He attributes that difference to how makers interpret what they learn, what they choose to prioritize and the way they work. A product’s character, in his view, appears in decisions about what to build, what to leave out and what to keep refining after a feature already works.
He also explains what Linear means by valuing intuition over process or data. Saarinen says this is not supposed to mean choosing based on unexamined preference. He describes intuition as judgment built through experience: working on a problem, speaking with customers, making mistakes and learning which details matter. He says customer conversations before Linear launched helped the team test and improve its own perspective.
The concern in the essay is that software work organized around AI agents could make it easier for individuals to produce changes independently. Each change may appear reasonable on its own, Saarinen writes, while the combined product lacks a shared point of view. He compares the result to “Frankenstein’s monster” rather than a deliberately shaped whole, and says products can accumulate capabilities without improving their central purpose.
Why Product Coherence Is at Stake
Saarinen’s argument matters as more software teams explore using AI agents to produce code and changes. If those tools increase the volume or speed of implementation, teams still have to decide which changes belong in a product and how they fit together. The essay frames that selection and editing work as a human responsibility, rather than something that follows automatically from generating more features.
For users, the distinction is between a collection of functioning capabilities and a product that feels consistent. Saarinen argues that a feature can work correctly while making the overall experience worse. That is a design judgment, not a claim that agent-produced software has already caused widespread harm. The source offers no comparative data on product quality, development speed or customer outcomes.
The piece also challenges a common assumption that more output necessarily means better software. Saarinen’s position is that easier production should give makers more room to understand problems and refine solutions. Whether AI tools produce that result will depend on how teams organize review, customer feedback and decisions about the product’s overall direction.
As an affiliate, we earn on qualifying purchases.
How Saarinen Defines Product Judgment
Saarinen grounds his argument in his experience at Linear, the software company he co-founded. He says the company has valued intuition, but defines it as a perspective developed through work and contact with users—not as a substitute for learning. In his account, experience helps people decide what evidence to notice and how to act on it.
That account connects the essay’s two main points. First, makers shape software through many small decisions, including choices not to add a feature. Second, production methods that let people work separately may make it harder to sustain a common understanding of what the product is for. Saarinen does not say teams should reject data, process or automation; instead, he argues that those methods cannot by themselves supply a product’s point of view.
He describes a pattern he perceives in products from successful companies: changes appear more focused on what technology can add than on improving the product’s form or central problem. This is his observation, not a documented survey or an assessment naming specific companies. The essay provides no examples or metrics to establish how widespread the pattern is.
“I think good products carry something of the people who make them.”
— Karri Saarinen, Linear co-founder
As an affiliate, we earn on qualifying purchases.
What the Essay Does Not Establish
The source is a personal argument by Saarinen, not an independent study. It does not quantify how often products have become less coherent, compare outcomes from agent-assisted and conventional development, or provide examples showing that AI agents caused the pattern he describes. It is also unclear how broadly his observations apply beyond his own experience and the products he has encountered.
Saarinen does not specify which agent workflows he has in mind, how teams should divide responsibility between people and automated tools, or what measures would show that a product’s character has been preserved. The essay does not announce a change to Linear’s product or policies. Its claims about messy products and additive development should be read as his assessment, not as verified findings about the software industry.
product design collaboration software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
How Teams Put the Argument to Work
The essay sets out questions software teams can test in their own work: whether individual changes support the product’s central purpose, whether someone is accountable for the combined experience, and whether decisions remain informed by customer feedback. Saarinen does not offer a formal framework or a promised next step, so these are implications of his argument rather than announced plans.
No follow-up research, product release or company initiative is identified in the source material. The broader issue will become clearer as teams publish evidence about how agent-assisted development affects the consistency, usability and quality of software over time. For now, Saarinen’s contribution is a warning—and a hope—that making software easier will leave more room for deliberate product judgment.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is Karri Saarinen arguing?
He argues that a product’s character comes from its makers’ experience and judgment, and that AI-assisted development could weaken coherence if changes are made independently without a shared direction.
Does Saarinen say AI agents make software worse?
Not as a proven or measured conclusion. He raises a concern that agent-based workflows may encourage disconnected additions; the essay provides no data showing that agents have caused a general decline in software quality.
What does Linear mean by valuing intuition?
Saarinen describes intuition as judgment developed through experience, customer conversations and mistakes. He says it is not simply personal preference without thought or evidence.
Did the essay announce a change at Linear?
No. The source presents Saarinen’s views on software design and AI-assisted work; it does not announce a product update, policy change or company initiative.
Source: hn
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
