I'm curious to the sourcing of these experts. Although they must understand design heuristics well, they also should not be associated with the project in order to protect against biases of their own design concepts. Thus, these experts must be designers, but not designers of the particular interface they are critiquing. This seems to cause an issue where most smaller companies will not have the resources (number of designers) to properly carry out heuristic evaluation.

I'm interested in how many participants would be necessary for proper A/B testing to be effective. Basic statistic tests require a sample of at least 30 individuals in each treatment group to reach an acceptable rate of sampling bias, but is this reasonable within the context of interface design? Even if 60 participants were gathered, what if another layer of testing needs to be done? It seems to me that the execution of A/B testing in reality is quite hard to execute.

I was surprised by this statement because the functionality of the sliding toggle switch is so ingrained in my head that its function has become intuitive. However, a person having never used one of these toggles would likely have a much, much harder time discerning its function. I wouldn't have even of thought to even think about usability testing for this feature, as its function is so apparent to me. I would of liked Ko to explain how to design with these specific biases in place, and how a designer can "step out of their own shoes" when it comes to interface functionality.

I agree with this statement and its logic, although I think there's some things to be wary of. I think the perceived cost of building something will typically almost always be less than the realized cost, especially when no prototyping has been done. This is problematic for two reasons. Obviously the cost will be higher, but then at this point you might as well have went through the prototyping stage, as no money was saved in reality.

I'm curious to how these participants are garnered / how many there are in these pilots. In my research methods class, we have talked about all the incredible amounts of biases that can be generated from where the participants are taken from (are they being incentivized, did they respond to an opt-in program, etc.). Although I agree that these pilots are an effective and impactful way to get ride of the questioner's biases, I wonder if it is just shifting the bias to the select individuals in the pilot.

How / where can we strike a balance between differentiation and effective borrowing of successful ideas? Assuming my design was something new to the market, why would a consumer decide to use my product as opposed to a competitors, especially when my design is already based upon theirs's (in addition to the user's loyalty bias). I don't think this is always the case, although I think it could be easy to fall into the loop of TOO MUCH borrowing from competition, especially when it is the easer thing to do as opposed to original thinking.

Although not disagreeing with the use of personas to understand user problems, is it not concerning that through their creation we may be including our own personal biases and backgrounds? I found it hard to truly understand and empathize with the scenario, most likely because my own scenario and lifestyle is very different to Ko's. My positionality as a young, healthy college aged person would impact my ability to define issues with this user persona. I think that the designer's background and situation should be heavily considered before the creation of a user persona or scenario in order to minimize the implications of bias.

Is this really always the case? What about in scenarios where people do not know what they want, or when what they BELIEVE they want is not actually what they would be most satisfied with. I believe that sometimes, the best design could be contrary to what people think they really want or need.

I think this idea is inherently harmful for users. Although a nonuniversal design may not work for some, there is still a net increase in utility in the aggregate. While the hand soap example is disproportionately impacting one particular group, perhaps there are other design decisions that negatively impact other groups. I believe that as long as the benefits to each group are roughly equal in the aggregate of design decisions, nonuniversal designs serve to the betterment of the population.