{22} Drawing an Eye, Semantic Search & Mental Health
{22} Drawing an Eye, Semantic Search & Mental Health
Featuring Sktchy, Chris Staudinger, AI Breakfast, Roshan Patel, and Nick Bryant
Design: Drawing an Eye
Sktchy
Tiktok failed to load.Enable 3rd party cookies or use another browser
I found this advice to be surprising. I assumed the teacher would talk about scaffolding a drawing with basic geometries then progressively add detail. But this artist is far more concerned with light and “bypassing what your brain sees to focus on what your eye sees”.
This is interesting to me because your brain tends to work with abstractions. For example, It’s not useful information to think about the gap between the top of the eyelid and the start of the eye lashes. So that bit of visual information is lost in your mental model of an eye. The sum of all these little abstractions results in an eye shape that feels unnatural when sketched.
Focusing more on what your eye sees and less on your mental model helps you develop the skill to reconstruct the details to make your sketch feel more realistic.
Frontend: em vs rem
Chris Staudinger
I know em and rem units can be confusing. I’ve often avoided em units because being relative to the parent can lead to unexpected behavior. As you iterate on a design, it’s often necessary to make adjustments to the markup. Sometimes this means adding wrapper elements. If those wrapper elements have a different font-size than the previous parent, you might randomly notice a shift in rendered font sizes and not know why.
Rem units are far more predictable. You can have the stability of pixel units with the ability to shift all of the text on the page up or down with a single line of code. That is particularly useful when using media queries for mobile.
Backend: Semantic Search
AI Breakfast
Semantic search has long been a pipe dream for indie makers. We don’t have the resources to build and maintain Google-like infrastructure. The alternative is to do the best we can with simpler keyword-based solution.
That is all changing. With ChatGPT, OpenAI, and access to vector databases such as Pinecone, people are building incredible things without the need to make huge upfront investments.
Filechat is turning PDF documents into chatbots. Imagine if the PDF is a dense research paper. I’d much rather talk to a chatbot about a complicated document than invest a significant amount of time consuming something I’m only mildly interested in.
Marketing: Networking
Roshan Patel
Unsolicited deck pic. 😂
I can empathize with solicitation fatigue. Living in SF for a decade made me dread pitches. But why the hell are you going to a networking event if you are just going to pull this jackass card to block people from pitching?
This could be an opportunity to rethink the concept of pitching. What if you go to a networking event and as you walk in the door a mobile app shows a map+list of everyone at the event along with their pitches (1 paragraph max). Then you gravitate to the people with pitches that resonate and skip some of the painful pitching formality.
Team: Mental Health
Nick Bryant
I’m a member Nick’s Discord. Having a space to talk about tech and mental health is a nice change of pace.
Startups should optimize more for mental health. It’s not uncommon to put people in daily stressful situations and not deal with the fallout. When people burnout, they are replaced by the next human “resource” in line and the revenue engine keeps chugging along.
People should be happy and feel valued. Happy people think less about anxiety and more about the problem at hand. Revenue is factored into the equation. But I find it to be more balanced and sustainable to also factor in wellbeing.
Re: Startup mental health. We've been framing the startup life as training to be an athlete - you can train hard and push, but rest and recovery are a natural part of the cycle and you can't get the results you want without making the time to rest.