Trends in Tech and Corresponding Complexity Levels

by Leo

It wasn’t too long ago that electronics were something that only large corporations did - the components and materials necessary to dabble in electronics were simply unaccessible to the average consumer, and the learning curve was steep for a large portion of the machinery.

With the advent of Arduino, however, interest in hobby electronics saw a steep rise. Coupled with the push for accessibility, efficient delivery systems, and a focus on consumer satisfaction, electronics were bought to the masses. In the same way that some amateur artists burst into the music scene from humble beginnings, we saw incredible levels of innovation coming from electronic hobbyists with no more marketing than a simple phone video. Electronics became fun to play with and easy to assemble, and as we can see with the so-called useless boxes that only function to turn themselves off, they became more amusing and design-oriented. Rather than simply viewing electronics as a solution-provider, more and more people were using electronics to create simple for the sake of creating.

Today, we’re seeing it return to its roots. Hobby electronics have turned serious and more complex - with innovation came more difficult methods and a drop in accessibility due to the rise in the level of complexity. As electronics become more impressive, the creation of such electronics begin to require bigger teams and more funding, thus returning electronics back to its corporate origins. This is a cycle that will likely repeat for some time, and at this time, I believe thar we can look for the creation of new methods that will reduce the learning slope of today’s complex methodologies.

We can see computer vision and robotics follow the same cycle; at this time, they have recently become more consumer-accessible. In 2019, and likely for the next few years, we’ll see two things: a rise in machine-vision-focused projects, and a drop in electronic ones.

Innovation in the Age After Deep Learning

by Evelyn

In a recent article written by Karen Hao, the MIT Technology Review downloaded some 16,000 abstracts from arXiv in an effort to break down the trends in deep learning through time. What they found was that despite deep learning being at the forefront of AI research in the past decade, it seems likely to fizzle out in the next few years like all other technological trends.

Figure 1 - Changes in language used in deep learning articles, from Hao

Figure 1 - Changes in language used in deep learning articles, from Hao

Trends in deep learning have shifted from its limitations to its potential, as seen in Figure 1. With the incredible leaps forward that machine learning has made in the past five years or so, it’s no surprise that this advancement could eventually sputter and slow down, although it likely won’t stop. If we choose to believe that the age of deep learning could come to an end, then no matter the reason, it would make sense for us to begin reconsidering older methods in a new light, and begin searching for new approaches as well.

Article, by Karen Hao: https://www.technologyreview.com/s/612768/we-analyzed-16625-papers-to-figure-out-where-ai-is-headed-next/

Watertight Design Guidelines

by Daniel Righetti

Figure 1 - The tube

Figure 1 - The tube

Recently, a client came to us with a design which required a unique container to hold liquid. Naturally, it needed to be watertight, but the main challenge was in the materials – glass, acrylic and 3D printed components. Some quick context on how the container would function: liquid would be poured in from the top, flow through both glass tubes simultaneously and be stored in the base as it waited to be used. When it was completely full, the liquid would fill the glass tubes and half of the top compartment. The clear acrylic casing simply acted as structural support and protection for the glass, and didn’t come in contact with any liquid. The transparency of the materials allowed the user to know when the container was running low by looking at the liquid level in the glass tubes. The complexity of this project was focused on the unique aesthetics the client was trying to achieve.

We began with trial and error when designing with food-safe, PETG, 3D-printed parts – testing the hole sizes, spacing, depth and tolerances for the glass piece to fit comfortably and securely. After a few iterations, we began working towards creating a strong, watertight join between the glass and PETG.

We chose to apply a Super Clear Sealant to the joining areas that was transparent, food-safe, waterproof, and strong. It was used to fill the gaps in the holes and around, and block anything from escaping through the top (as indicated by yellow arrows in fig. 2).

Figure 2 - Plastic and glass connection point

Figure 2 - Plastic and glass connection point

Initial tests showed that the sealant helped to hold the glass tubes in place while we added the outer acrylic tube and provided a watertight seal. This same method was applied to a similar piece on the other end of the acrylic tube, creating an airtight seal in the area surrounding the glass.

Unfortunately, we were unable to test the liquid with the expected amount of pressure, as it was not completely assembled at this stage. This eventually led to leakages when the sealant was later tested – because it was silicone based, it didn’t form a strong bond with the glass or the plastic, and so it started to give way under pressure and flood around the glass (see fig. 3 – the dark blue area should not contain water).

Figure 3 - Water flow and flooding

Figure 3 - Water flow and flooding

Another issue later discovered was that when the plastic components themselves were put under the same pressure and left overnight, the liquid would seep out through tiny, random holes in the 3D-printed body.

Our first solution was to insert additional 3D-printed parts around the connection point of the glass and plastic, which we sealed around the edges (see yellow indicators), to help stop the liquid from leaking out (see fig. 3). Although we knew this would slightly affect the flow rate, it wouldn’t make too much of an impact and would leave us with a decent, quick solution to our leaking problem.

Figure 4 - Sealant locations for new part

Figure 4 - Sealant locations for new part

Although it was theoretically a good counter-measure, the additional parts were not 100% effective. After some planning and thought, we decided to drill holes in the acrylic tube that encased our parts (which also provided an air hole to avoid cracking under increased pressure during high-altitude flights) and used a thin tube and syringe to inject resin in to the area surrounding the glass (see fig. 5). We used an Easy Cast clear epoxy resin to get the job done, which took time to test when allowing for the resin to cure but ended up being extremely effective and aesthetically pleasing. To avoid the resin flowing into the plastic parts and clogging, we waited for 5-10 minutes after mixing the resin to allow it to set slightly. This level of viscosity allowed us to fill the desired area and any gaps from our previous attempts, without the worry of filling the plastic chamber below.

Figure 5 - Adding the resin

Figure 5 - Adding the resin

With this resin method, we can bypass our previous, less successful solutions and speed up the build process of each new container we make. Although the previous attempts were made redundant with the resin, exploring our options was necessary to fully understand the problem and the viability of various solutions. This kind of research and development has made us far more confident in our ability to make these custom containers water tight, and we now have a tested solution that we can implement in future projects should a relevant situation arise.

AI on Creativity

CloudPainter, an AI robotic arm that is programmed to try to make as many independent aesthetic decisions as possible. Developed by Pindar Van Arman for the past 10 years, making this one of many multidisciplinary programmers, roboticists and artists to explore the AI's limits on the topic of generative creativity. 

The results of the generated arts are a hit and miss, which is not the main take-away point. This attempts to recreate and explore the meta-process of creativity would bring us closer to understanding the human creativity itself. 

More information on the latest development can be found on this link: http://www.cloudpainter.com/

3D Printing for Home Users

3D Printing never took off as predicted: 3D printers in everybody's home quickly turned from a vision into a trend that is losing its steam.

One of the reasons is that the process of designing, setting up, maintaining and post-processing are just too cumbersome for the majority of the adopters.

The last component to make the whole equation works - the AR mixed reality aspects to design 3D objects in 3D environment - never took off in time to fuel to the hype. Hopefully in the future when Augmented Mixed Reality has become more mainsteam and more accessible to the majority users, 3D printers in every homes might become a thing again. However probably it will take another 5-10 years for the tech to go its full cycle and arrive at everybody's homes.

It's a different story for the early adopters and the 3D printer enthusiasts, however. I for one happened to jump right in front of the early adopters of these 3D printers, to the point of becoming an enthusiastic hacker when the 3D printing became available to the pro-consumers. This Replicator Original, one of the first 3D printer kit made by MakerBot, is one of my trophies. Thanks Dave for donating this beautiful machine to me! The wooden body quickly gives it an antique even if it is just about 5 years old.

The Original Replicator!

The Original Replicator!

Given my background in 3D CAD design, knacks for tweaking gadgets and tons of curiosity, I quickly got over the complexity barrier and got comfortable with adopting 3D printing as one of my tools in the shed.

Long story short, it finally come to the tipping point where I found 3D printing a bunch of pulleys are more convenient, cheaper, and faster than purchasing it either from the local hardware store or online!

3D Printed Pulley
Assembled 3D Printed Pulley
Final Work!

The pulley were in the end installed on a weighted boxes and secured 5-meters-long tensioned nylon rope for an art exhibition. 

Product Design 101

Preparing to give a presentation on product design for a hackathon was the best opportunity to gain fresh perspectives on my basics.

Classic example on S.C.A.M.P.E.R. Diagram

Classic example on S.C.A.M.P.E.R. Diagram

I decided to open the presentation with a classic example of SCAMPER diagram of the design of a spoon. Design is an iterative process of idea generation and elimination. If brainstorming does the idea generation randomly by association, SCAMPER technique does it systematically. I proceeded to let the participants to get a taste of both before they dive into the hacking actions.

Stepping out of the freezing hacking space, I asked the participants to take notice that everything around us are designed. Some are designed well, some are not so. 

Stepping out of the freezing hacking space, I asked the participants to take notice that everything around us are designed. Some are designed well, some are not so. 

It helped that the hacikng space is surrounded by whiteboards. Both the participants and the mentors were happy to fill them with ideas, good points, critics, and feedback notes. I love post-its so much!

It helped that the hacikng space is surrounded by whiteboards. Both the participants and the mentors were happy to fill them with ideas, good points, critics, and feedback notes. I love post-its so much!

This was my view when we are just starting. Empty white boards, sleepy people just waking up in the morning. Fresh starts!

This was my view when we are just starting. Empty white boards, sleepy people just waking up in the morning. Fresh starts!

We the mentors did a little bit of data analytics on the topics and the stages of the groups. if we didn't stop ourselves, probably we would have ourselves came up with an app to optimize this process. 

We the mentors did a little bit of data analytics on the topics and the stages of the groups. if we didn't stop ourselves, probably we would have ourselves came up with an app to optimize this process. 

Some of the creative works we did in the middle of our mentoring. 

Some of the creative works we did in the middle of our mentoring. 

Design Requirements

Designing, like a fluid, doesn’t flow in vacuum. Fluid behaves like how it is on earth due to a combination of gravitational force, pressure differences, temperature conditions and the shape of the earth rocks. Likewise, design needs a set of limits and rules to define its universe. Change the rules even a little, and the balance will be thrown off and we might not even live through to tell the difference.

If each design is analogized as creating a new universe, gathering the requirements is like defining its law of nature. There is no point in painting planet’s landscape if an added semicolon might wipe off the entire galaxy. I do not start designing until the requirement is set in stone, to emphasize how important the requirements are to a design.

My mentor once said, that understanding of the problem is one skill that differentiate good designers and the great ones. In my experience, a problem is unsolvable only because it is not understood properly. A completely deep and thorough understanding is like a fully defined river banks, which enables the design to flow so effortlessly to find the solutions at the end of the process.

Gathering the requirement is an art on its own, as the designer usually is not part of the universe that is to be created for the clients and their users. This is where the designer needs to have a far-reaching emphatical heart, paired with observant eyes to gather what would work for the clients and users. This is the core of user-centric design philosophy popularized by IDEO’s Tom Kelley. Personally, I think the understanding of the users should be done even before designing begins.

Requirement is also useful to separate the designed universe from the designer’s own preferences and styles, to ensure that the designer can create what is truly best regardless of his/her own personal choices. From this perspective, the act of designing is very similar to that of acting, where one needs to kill his own ego to fully take on the role of a hero, an assassin, a villain or even a demon. A very selfless profession albeit the highlight and glory given to the few celebrated ones.