- jaypeters.dev
- Posts
- Grammar, An Anniversary, and Reporting
The Weekly Variable
More Wave testing this week.
Turns out apps are pretty complicated.
Also started a new YouTube project, considering a new AI helper, and I’m developing a new appreciation for reporting.
Topics for this week:
Certificates and Grammar
As Wave inches closer to being published on the App Store, last minute testing has revealed a few issues I missed.
Turns out users will probably need a way to reset their passwords, and unfortunately I can’t do it for them so I’ll need to setup a basic web portal if nothing else to allow someone to enter their new password if they forget theirs.
Also found that I used the wrong license for Push Notifications and missed it because I haven’t tested phone to phone, I was only testing sending messages the iOS simulator on my MacBook to my phone.
Since the iOS simulator can’t get Push Notifications, I was basically only testing one way, and forgot that my local services running on my computer were using the old certificate which worked.
The service in AWS was using that same old certificate which no longer worked for this version of the app so Push Notifications would not send due to an invalid Key.
Luckily that’s a quick fix and we were able to confirm that’s good to go now.
Also missed some simple phrasing issues like the notification for a promoter that read “Jay is looking for Club near you!” when I submitted a new Plan that I was looking to find a Club nearby.
Turned out this message template wasn’t going to work:
<user> is looking for <activity> near you!
Not exactly grammatically correct for most options.
According to Gemini Pro, the better option for now is each option has it’s own custom phrase instead of coming up with complex grammar logic to decide between “a Bar” vs “Late Night Eats options”.
Decided “to Pregame” was the correct phrasing.
Given some of these minor issues, it makes sense to do one more thorough round of testing before go-live just to catch anything else like this while there aren’t a ton of user to disrupt with numerous app updates.
Hopefully by next week, Wave will be all shored up.
YouTube Build Series
Last week I was talking about creating “all the content”, and so far this week, I’ve made some progress.
2 complete social media posts, including a 49 minute n8n tutorial, and a YouTube Community post to supplement it.
I was researching a couple n8n tutorial ideas, when I decided the thing that I really need is something that turns long form video clips into short form video clips.
There are tools out there that try to do this, but I haven’t been super impressed with them.
Since I’ve been playing around with that idea for a while myself, I figured why not turn that concept into a video itself.
So I had a goal in mind but thought I’d show the whole process of trying to build a system that could create shorts from full YouTube videos.
This will most likely be a 3 part series, but we’ll see how the rest of the process goes.
The first part is fairly straight forward:
get the transcript from the video (including timestamps) with Whisper
have GPT try to find relevant clip timestamps
generate an ffmpeg command to pull those specific clips out
It turned out pretty well given only a few attempts, but still ended up being a 50 minute process.
The next, and probably hardest part, will be reformatting horizontal video into vertical video, and figuring out how to use GPT to systematically find the proper portions of the video to use in mobile format.
Hoping I can come up with something decent, but we’ll see.
The last part is easy enough though, just need to stitch all the vertical clips together into one vertical video.
It was fun to create the first part of the series so I’m looking forward to getting to the next part soon, maybe this week if I’m feeling particularly inspired.
Let me know if you check out the tutorial and have any feedback!
Divs Anniversary
After more than a year of divs.design being in service, the business has officially doubled!
I took on a new client and got their website up and running last week: https://www.thenorthyardllc.com/
And the first divs client just renewed their yearly contract this week so I successfully kept a client for one full year, which is weird to think.
I didn’t end up getting around to updating the Divs Design website yet, but that’s still high on the list.
Wave will be a primary feature with the new design, along with a new pricing model soon.
Just in time to celebrate one full year of Divs, with more years to come!
Gemini CLI
No major AI model upsets this week, the streak of new releases has quieted down at least for the moment.
An interesting, and maybe not completely surprising release was Google’s Gemini CLI, which is very similar to Claude Code and OpenAI’s Codex.
All of these command-line tools are on my list to try out, I just haven’t taken the time to set them up and get them running.
Cost has also been a factor.
I tend to stick with the browser only because model usage is free in browser, it costs money to run AI APIs.
But it looks like Gemini CLI is free with a personal account so I might have to give that a shot.
I see a ton of value in autonomous agents quietly churning away at bugs and small features in any project really, but I don’t want to run up too much cost on projects that aren’t funding themselves yet.
Depending on how Wave development continues, I may give Gemini CLI a shot at being my background backup so I can better manage a couple projects at the same time.
Very cool to see multiple options popping up for these automated coding agents, though.
I’ll be anxious get one of them going.
Proper Reporting
I’ll admit, as a full-time engineer, I didn’t always see the value in “reporting.”
Reports and reporting was often a high priority for the business people I would interact with, and I knew reports mattered because it was the only way they would have insight into what apps and projects were doing, but I don’t think I had enough context to really appreciate that insight.
And to be fair, I was dealing with my own reporting issues of making sure Sprint and engineering metrics were satisfying enough to keep teams out of trouble.
But once Wave is on the App Store and people I will never talk to will be using it, I am understanding the need for metrics within the app.
New user sign-ups, how long users stay on the app, what time do they usually use the app, what features do users click on the most, what bugs or errors are popping up, and plenty of other things will become very important to track everyday.
I’ll be starting with just the basics around active users but I can definitely see why platforms like Grafana become so popular, because they give you a way to visualize all this data about users and app performance in a way that’s specific to your app.
Once Wave’s been tested and is finally ready for release, it will be time to implement a proper reporting dashboard.
And that’s it for this week! Wave testing, a new YouTube Series, and a full year of Divs.
Those are the links that stuck with me throughout the week and a glimpse into what I personally worked on.
If you want to start a newsletter like this on beehiiv and support me in the process, here’s my referral link: https://www.beehiiv.com/?via=jay-peters. Otherwise, let me know what you think at @jaypetersdotdev or email [email protected], I’d love to hear your feedback. Thanks for reading!