Why AI logos look alike: Founder’s viral X post explains ChatGPT and Sarvam AI comparisons

Founder explains why perfect designs are forgettable.

After social media users began pointing out visual similarities between the logos of OpenAI’s ChatGPT and Indian startup Sarvam AI, the conversation quickly moved beyond comparisons and into a deeper discussion on what actually makes a logo memorable, especially in the crowded AI startup space.

Adding perspective to the debate, Snehil Saluja, co-founder of Overlay AI, shared a detailed post on X explaining that logo design is far more than creating something neat or visually pleasing. According to him, the real challenge lies in balancing familiarity with distinction.

In his post, Saluja said that while he is not a professional logo designer, his background in math and design helped him notice a recurring pattern. Companies, he explained, often compete to claim the simplest and most symmetrical shapes because symmetry feels natural and “correct” to the human brain. However, that very perfection can make logos forgettable.

I'm not a professional logo designer, but I've spent enough time in math and design to notice something about geometry that I think is underappreciated. With trademarks and copyrights, every company is fighting for the most simple, most symmetric, most recognizable figure they… https://t.co/RuWIgpeXBC

— Snehil Saluja (@mesnhl) February 22, 2026

What works as logo, according to Saluja?

He argued that perfectly symmetrical designs may look appealing at first glance but fail to leave a lasting impression. The brain quickly processes them as “right” and moves on. On the other extreme, complete asymmetry grabs attention but feels uncomfortable or broken. What truly works, Saluja noted, is a narrow middle ground, a mostly symmetrical design with a slight, intentional imbalance.

This subtle asymmetry creates tension. The brain senses that something is almost right and instinctively tries to resolve it. That unresolved feeling, he said, is what makes a logo stick in memory. “We don’t remember perfection. We remember what almost was,” Saluja wrote.

To illustrate his point, he shared a personal example: whenhe has accidentally opened the Mercury app instead of ChatGPT and vice versa, multiple times because both icons are highly symmetrical and easy to confuse. That, he said, has never happened with Claude, whose logo stands out more clearly in his mind.

Saluja extended this idea beyond branding, drawing parallels with problem-solving in math and physics. The most engaging problems, he explained, often start with a symmetrical system disrupted by one small irregularity. Finding that disturbance becomes the challengea nd the reason the problem is memorable. He used Saturn as a metaphor where all planets are spherical, it is Saturn’s rings which is single disruption that make it instantly recognisable, even to children.

While designing Overlay AI’s logo last year, Saluja said he consciously applied this principle by starting with a simple geometric form and introducing an uncommon asymmetry. His goal was not perfection, but memorability. “Let’s see if it sticks,” he concluded.