What Is the Future of Personal Computing?

Today, devices like smartwatches, rings, and even glucose-monitoring patches are widely accepted for health tracking. Yet, as AI advances and AR/VR headsets proliferate, we must ask: what comes next? How can future devices fully embody our evolving software capabilities?

In a series of projects, I explored the idea that wearable technology could gain enhanced functionality by integrating directly with our clothing—a medium we’ve used for millennia. For these devices to be both practical and widely adopted, they must blend seamlessly with our bodies while retaining the softness and comfort we expect from everyday garments.

Key Questions

Functionality: Can useful capabilities be provided using only soft, fabric-like materials?
Personalization: Can such functionality be tailored to individual users?
Scalability: Is it possible to produce these devices at scale and at a price similar to today’s smartwatches?

Using Textiles to Enhance Functionality

Drawing on my experience with soft robotics (which uses pneumatics to power motion) and clothing design, I hypothesized that textiles with varying stiffness could direct and control the force transmitted by embedded inflatable pouches.

To achieve a useful stiffness differential, I collaborated with the Self-Assembly Lab at MIT and their knitting specialist. Knitting—a technique with millennia of history—produces materials that are both strong and comfortable, and are manufactured at scale. By experimenting with different types of knots, their arrangements, and various materials, we discovered that simply tuning the knitting parameters could stiffen a fabric by a factor of ten without changing its base material. Moreover, by adding meltable thermoplastic fibers (an idea borrowed from shoemaking), we achieved stiffening factors as high as 400—roughly the difference between soft rubber and rigid aluminum.

We then tested this concept by sandwiching a soft inflatable pouch (25mm in diameter) between a soft layer and a stiff layer. This configuration allowed the pouch to apply tens of Newtons of force, rather than just a few, confirming that our approach could deliver practical functionality.

Advancing Haptic Feedback

Building on this foundation, I explored the use of these inflatable pouches for haptic feedback—the type of tactile alert most people recognize from a vibrating smartwatch. Instead of a single vibration point, imagine multiple contact points across the body that use gentle indentations to convey information. Since the receptors that detect pressure are closer to the surface of our skin (compared to those that sense vibration), this method can provide a denser resolution of tactile information.

To test this idea, we developed a haptic sleeve—a wearable arm cover embedded with an array of inflatable pouches capable of producing various actuation patterns. In user testing, participants interpreted these indentation signals about 25% more accurately than standard vibrations. By rapidly varying the inflation patterns, we also created the sensation of a continuous, human-like stroke along the arm, suggesting exciting possibilities for applications like wearable braille or tactile cues in AR/VR environments.

Details of this work can be found here and were published in Science Robotics.

Personalization

The success of these prototypes demonstrates that it is indeed possible to embed advanced functionality into clothing. However, personalization remains key. Our initial haptic sleeve, which used Velcro straps, was adjustable for about 95% of users but still felt somewhat bulky. Given the low cost and scalability of modern knitting, it is feasible to create custom-fit haptic sleeves based on a 3D scan of an individual’s arm—without significant additional expense.

Designing bespoke garments, however, presents a significant challenge: we don’t fully understand how different knitting parameters—such as pattern choices and material selection—affect the final product. Previous attempts to describe these relationships often missed the mark, focusing more on achieving a natural-looking motion rather than aligning with physical reality.

In this second project, the goal was to develop a simulation model that could translate knit parameters into accurate physical responses, efficiently enough to eventually optimize a garment for

**Step One:** We correlated **experimental** data from knitted samples with a **detailed** **simulation** of individual knots to understand the underlying mechanics.

**Step Two:** With insights from the granular simulation, we developed a simplified, **homogenized** material model that could rapidly predict how entire pieces of fabric would behave without relying on further experimental data.

**Step Three**: Analyze material transitions and **real-world**, heterogeneous fabrics.

By developing our model, we could accurately predict how fabrics with specific knitting parameters behave. However, real garments rarely use a single pattern or material—especially when they must conform perfectly to the body. An additional consideration was how transitions between different textile variations affect the fabric’s overall behavior.

We examined these transitions from various angles and discovered that so-called “heterogeneous knits” can be effectively modeled as an idealized patchwork. In practice, the transitions between different textile sections had little to no effect on the fabric’s performance.

Armed with this insight, we applied our efficient model to optimize garment design. For example, we created a bespoke compression sleeve for someone whose arm diameter varied significantly (such as a very muscular individual). By using a 3D scan of their arm, we were able to optimize the choice of material and pattern in specific areas, ensuring a perfect fit and uniform compression.

Details of this work can be found here and are currently under peer review.

Scalability?

Although there is still work to be done on large-scale manufacturing, leveraging existing production infrastructure offers a promising route to commercialization.