Back to the grind.

Long days but a promising amount of momentum.

Feeling like consistency is key…

Topics for this week:

• Android Builds

• Skooling

• Multi-streaming

• Upwork

• GPT-5 Backlash

Android Builds

The Wave app finally works on Android again.

There was a clear fork in the road when I started added Push Notifications a few months ago where I had to choose to focus on one platform over the other.

Finish iOS first, then circling back to Android.

Luckily React Native mostly ensures the app will work on both platforms, so all the code runs for both, but the configuration for both platforms was the issue.

I hadn’t setup Firebase to handle Push Notifications yet so the Android build would fail because I didn’t have a proper google-services.json file created.

But now I do!

With a little digging through Google Console and finding the right Org Policies, I was able to get a google services account created in Firebase to handle push notifications.

As far as I know.

The app builds, but unfortunately the backend service still doesn’t like something about the configuration.

I’ve got some more testing to do there, but after a few months, Wave works on Android again!

And since Android allows for side-loading, I can offer an Android download link to a version of the app before I get it to the Play Store.

If you want to try it out on Android, just let me know!

And hopefully soon it will have push notifications too…

Skooling

Advice from the recent winner of the Skool Games, Nick Saraev, was to just show up everyday.

Not surprising, consistency really is the key to success in anything.

But he did point out an interesting stat.

He averages about 200 activities per day in Skool, which he said seemed to be 15x what other community leaders average.

After a few weeks, I can see why 200 is an outlier because it’s tough to maintain that high of an average.

I’ve been tracking my daily activity in a spreadsheet since July 10, and I’m right at a 15 activity average.

A few 0’s in there certainly aren’t helping, but a long way from 150 or 200 activities per day.

That may start getting easier, though.

Despite low activity count, my community growth seems to be on a good trend.

Up 91 members, now over 100 in about a month.

I’ll take it!

Honestly I don’t think that’s because of activities.

I believe YouTube is doing all the work here.

So some part of my plan is working.

Activity count foundation is important, but the regular traffic is also important.

Either way the principle is the same.

Consistency wins.

The lesson here is more activities AND more YouTube.

Speaking of…

Multi-streaming

For the first week in a long time, I managed to stream 4 days in a row.

Planning for a stream right after writing this too!

But I can start to see the consistency pay off there.

I’ve actually been multi-streaming because there’s really no reason not to in this day and age.

I can log into StreamYard and broadcast to YouTube, Twitch, and X at the same time from a single source.

The game these days is to catch people when they check their apps and a live stream is just another way to do that, and a potentially more engaging way than watching a clip or video since it’s interactive.

But it’s another chance to be top of the feed.

And I noticed an uptick on YouTube with it this week.

Despite not uploading a video for a week, I still had an increase in subscribers.

From what I’ve heard online, YouTube doesn’t prioritize live streams.

It’s easier for them to sell ads through regular videos, but it seems like live streams are still getting some traffic.

And in general, YouTube and other platforms like to see consistent activity, so a few days in a row means it’s more likely to recommend that source.

So regular live streams should still be helpful.

It’s tempting to make live streaming the main focus because it is fun, but it’s not quite sustainable yet.

Need to get some more momentum in a few other places first, but’s not a bad foundation.

Planning to be live Monday through Thursday around 10 am CST so hopefully I can catch you then!

Upwork

Since it will take some time to build an income from streaming, Skooling, and Waving, Upwork is the quickest alternative to a paycheck short of taking a full time position again.

I had some success with my first real push on Upwork this week, getting a response on the first contract proposal I submitted.

Nothing huge, but potentially recurring.

Someone wanting help to complete their n8n automation and wanting to build more automations so certainly an opportunity for continued work.

As if I didn’t have enough stats to track, GPT suggested I submit 6 proposals per day on Upwork to keep the opportunities rolling in, and that’s proving a little more challenging than 20 Skool activities per day.

Right now it seems to take me about 45 minutes to draft a proposal just because I usually record a video and then hand-write a cover letter, but I’m sure after a few weeks, I can get that time way down so I can easily knock out 6 proposals in an hour instead of 2 every hour and a half.

Also there are a lot of contracts posted on Upwork regularly which is good, but the quality varies widely.

Plenty of huge projects that want to pay $20 an hour, which I’m not sure is sustainable.

Hitting 6 proposals per day might not be realistic if there aren’t great options.

But I get the sentiment.

Regular activity leads to results.

Consistency, go figure.

Only takes a few contracts to create solid supplemental income.

Looking forward to more Upwork updates next week.

GPT-5 Backlash

I was really enjoying GPT-5 last week, but it seems I was one of the few.

To be honest, I didn’t really push it to it’s limits.

I used it in Cursor for a while on the day it released and it seemed like it was doing a great job.

I have since noticed a couple updates it made to Wave that seemed like they were working might not actually be working now so maybe it’s not at the level I was thinking.

And the internet (mostly X of course) seems not very happy with it at all.

Definitely some stumbles with the rollout.

The auto-switcher didn’t seem to work like it should, trying to pick the faster or smarter model based on what you asked it but it didn’t always do that correctly.

OpenAI has released a few updates to fix issues, but overall it might still need some tuning.

Direct access to the API version of gpt-5-high is still at the top of the LMArena leaderboard, which used to say gpt-5 so they’ve made a distinction.

That’s not the version you get when you go to ChatGPT.com, but you might get connected to that version if you prompt it correctly and it decides to use that version.

In regular chatting, I haven’t noticed an issue.

The free trial of gpt-5 ended in Cursor though, so I’ll have to start paying to use it now.

I’ll let you know how fast I run through my Cursor budget using gpt-5 tokens and if it’s actually worth it.

But I’d be curious to hear your experience with gpt-5 if you’ve tried it out.

I guess we’ll to wait for AGI with the release of GPT-6 or Grok 5 or Gemini 3 instead 🤞 

And that’s it for this week! Paving the way for consistent progress on Wave, Skool, YouTube, and Upwork.

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!