2026.05.13 - Implementing Updated Interface, Finalizing Logo, and Launching Social Media Presence
I spent the past two weeks in China, the abstract behemoth that is challenging the great United States of America in every area imaginable, not least, AI. But this post isn’t about that. I mention that context just to explain why I haven’t been doing more work on Scaffold.
Did you know Claude, the AI powering Scaffold, is not available in China? That’s my excuse! (Also, in case you’re curious, while most people in the U.S. just can’t stop talking about Claude Cowork, Code, etc., people in China are largely unfamiliar with the company. They are better versed in ChatGPT.)
Yesterday, though (Tuesday), Brendan and I finally got together to recalibrate our workstreams.
While I’ve been traipsing around China, Brendan has been working hard on Scaffold. He implemented the design overhaul that we completed with significant help from Sophia. In case you’re curious, here’s what the interface looks like now:
We started the meeting talking about Brendan’s progress building Scaffold along three different workstreams: (1) Scaffold for desktop; (2) Scaffold for mobile; and (3) The informational website introducing Scaffold. All the screenshots above are for Scaffold for desktop. Brendan affirmed that he has made significant progress on that workstream; less so on the others.
We started by talking through that progress. To begin, Brendan asked me, “Does everything look okay with the page?” nervously smirking a bit. I was confused by the question–everything looked okay to me, pretty consistent with the Canva mock-ups that we had previously developed. “Yes…” I responded hesitantly. Why?
“Well,” he replied, “I may have used vibecoding a bit.” I was a bit surprised but also slightly amused by this answer. As I’ve documented in previous meeting recaps, Brendan has had a long-standing aversion to vibecoding. I personally never had significant issues with vibecoding, given that (1) I wasn’t the one doing that technical work anyway and (2) I was drawn to the efficiency gains.
I asked Brendan why he changed his mind, and he shared that he had come to the realization that, well, most people don’t care. Recently, he’s come across a number of websites/applications that have gained a lot of traction despite, in his view, being obviously vibecoded. Seeing this, he realized that, if vibecoded work was socially acceptable and widely recognized as good, then there wasn’t a significant reason to reject the tool.
That being said, Brendan emphasized that his concerns about security remained, and so before launching Scaffold more publicly, he would personally go through the code to address any risks.
Brendan’s explanation made sense. But I was still a bit confused by the question, because I don’t know how I would’ve been able to recognize that the Scaffold interface was vibecoded. That’s because the design was not vibecoded–Sophia, Brendan, and I developed the outline together, and then I took a late night to more fully build it out in Canva. The (vibecoded) website merely implemented those designs, which were drawn out by humans.
Anyway. We proceeded to review the desktop view more comprehensively and made a list of relatively small adjustments we’d like to make. Among them: increasing the size of the top bar, replacing the “back” formatting bar with an eraser, adding borders, and creating a button to access our prompt ideation feature.
We then moved on to briefly talking about the mobile view. Brendan disclosed that he didn’t feel as confident designing for mobile, a statement that surprised me–as a non-software-engineer, I’m not very familiar with the processes for designing on desktop/mobile and was unaware that they required different skills. He also mentioned that some of our features/designs would be difficult to implement for mobile because the screen simply is not large enough. In reply, I suggested that we could remove those features for mobile entirely and keep them exclusive to desktop. I’m familiar with other applications that do that (e.g., Substack). I didn’t mind having reduced functionality in mobile, especially because people probably wouldn’t be playing around with as many features on mobile, anyway.
I wanted to prioritize mobile development because, from my experience overseeing testing, it’s easier for people to test on mobile–it’s a device that’s on their person, whereas most people do not always carry around a laptop. However, after discussing with Brendan, we agreed that we could put mobile on hold for now. As an immediate next step, we wanted the mobile view to be good enough, where people could comfortably see and access all available features, but we didn’t need to optimize it just yet.
Next, we turned to talking about the informational website. Brendan said that he had not focused on this workstream. I had initially wanted to prioritize this as well because I thought that this page would be important to show to people unfamiliar with our product. However, Brendan indicated that he was more interested in something else: figuring out how to build a database, so that we could give users the option of saving and accessing historical chats.
Through this conversation, it became clear that Brendan and I were each thinking about progress somewhat differently. I was thinking about dipping our toes in a variety of different workstreams: desktop, mobile, informational website, etc. But Brendan seemed to be leaning towards really deepening our work in one particular workstream–in this case, desktop. By building out the database, the functionality of our desktop view would then fully reflect what we are currently imagining.
Ultimately, Scaffold is a passion project. So, as long as Brendan was doing what he was interested in and moving the project forward in some capacity, I had no problems with continuing to work on the desktop view.
After talking about these implementation workstreams, we moved to talking about a new set of logos that we had received from Sarah. Like the previous two sets, they were absolutely amazing! But while truly every iteration has been awesome–sleek, minimalist, chic–the set that we received this week was definitely the most aligned with our vision, as well as compatible with the look of our updated interface. Sarah and I had spent some time looking at the updated interface, including a post-volleyball meeting where we thought about alternative designs (but ultimately settled on what we started with). All that is to say, Sarah has been intimately involved with this process, and the most recent designs seem to reflect that understanding of what Brendan and I are trying to build.
The logo that we decided to go with is shown below:
It makes me excited just looking at it! I can imagine that image appearing as an application icon.
After talking about the logo, we then turned to talking about an entirely new workstream: social media! Whoohoo…
I started a TikTok and Instagram account, both under the handle @musingsbyhill. I will share more reflections on this process in due time… Let’s just say that it’s been really humbling, but in a good way. The content that I’m creating right now is so far in quality from the content that I would normally consume. But every day, I am seeing improvement. And, the engagement with my content has, in the words of my friend, been accelerating on a “crazy curve.” (My first video received 3 views; my second received 153; and my third received 833, with 69 likes! These stats are as of May 14, 2026, at 2:44 pm…)
Currently, I am focusing on completing a daily “critical thinking challenge,” where I use Scaffold to critically think about a topic of my choice: whether it be something personal, something in the news, or something I’ve noticed about the world around me.
I decided on this approach following the advice of creators whom I’ve consulted, among them Sophia and Magan. You have to get into the habit of posting, they told me. As long as you commit to it, you will see growth. So, that’s what I’m trying to do.
Talking with Brendan, however, it didn’t seem like he completely agreed with my approach. He didn’t think that we needed to post videos daily for Scaffold. And, for himself, he wouldn’t want to commit to posting videos daily. Further, he didn’t want to develop a reputation as a public figure independent of the work he’s doing with Scaffold.
All incredibly fair points. At the start of this process, Brendan and I had very similar, tepid views on social media. Maybe I’ve changed, though. Maybe it’s because I have more time. But I want to give this media a chance. I don’t want to be a creator, I told Brendan (hopefully, convincingly?) Nonetheless, I believe that social media is a critical channel for distribution, and I want to create the framework for it to grow.
So, despite our different perspectives on the issue, we were able to arrive at a consensus. I will continue posting on @musingsbyhill independently, doing my little challenge. I will also separately create a Scaffold account, where I can cross-post content and where Brendan will exclusively post his content. Me taking the time to post every day is an investment–but hopefully it pans out.
We concluded the meeting as we always do, talking about the list of people we want to test Scaffold (which we still haven’t started making progress on), and discussing next steps.
Lots to do!
Food for thought:
Is it “better” to have made little bits of progress in different workstreams or prioritize making more meaningful progress in one particular workstream?
What is the best way to spotlight Sarah’s amazing work on the logos and overall website design?
How should I approach creating content if I don’t want to be a career content creator? How much time/effort should I invest in learning the craft? How much of content creating is putting my head down to learn, versus just showing up everyday and learning through experimentation?
Next steps:
Make discussed immediate edits to desktop interface
Make mobile interface “good enough”
Finish mocking up the homepage for the informational website
Research/begin implementing database
Plan and record a video introducing Scaffold
Reach out to friends for testing
Review guide from Anthropic on developing programmatic and model assessments to refine the Scaffold user prompt