Release prep continues, but Wave is nearly there.

Plus a website makeover, a new automation tutorial, and a need for more content.

Topics for this week:

• Release Testing

• jaypeters.dev 3.0

• Local Whisper

• All the Content

• AI Updates

Release Testing

One final round of testing this week before finally publishing Wave to the App Store.

Getting dangerously close.

I had one version of the app approved, but I accidentally found a bug yesterday so there’s at least one more round of approvals in the future before Go Live, depending on what the testers also find.

Then the real hard part begins…

Marketing.

The top ideas right now are old-fashioned door-knocking and networking, combined with new-fashioned trend following on social media.

One success probably leads to the other though.

A little social media virality to bring in regular users which creates numbers that investors want to see.

No idea what the timeline on that will be.

Hopefully short but it could take a while.

Anxious to see where Wave’s at in a year, but for today it’s on TestFlight.

jaypeters.dev 3.0

In optimistic preparation for an influx of traffic, I’m re-optimizing a couple websites to be a little more focused on whatever the goal is for that landing page.

After some back and forth with GPT o3 and a brief resubscription to relume.io, I have revamped jaypeters.dev to be more focused on whatever I’m focusing on at the time.

AI Automation is the priority for now since that’s what my YouTube channel is doing, but once Skool is live, the Skool community will be right at the top of jaypeters.dev instead.

I haven’t decided what to do with this newsletter yet since it used to be the first thing on the site.

But fear not, this newsletter isn’t going anywhere!

I will still have a subscription link at the bottom of the new site, still planning to keep this personal newsletter going, but I’m thinking I’ll need a more focused newsletter that’s specifically for AI Automation stuff too.

More to figure out with that later, maybe after the Skool ball is rolling.

Then, divs.design is next on the list for an update next week.

But for now, let me know what you think of the new version of jaypeters.dev!

Local Whisper

It finally happened.

I’ve had a tutorial for running Whisper locally on my to-do list for a couple months at this point so it felt good to finally get that one live.

This is a particularly useful one because this enables downloading and transcribing videos from YouTube fully automatical and free.

I’m a little nervous to put that full tutorial on YouTube since my channel has already been shutdown once, so that full tutorial will have to be a Skool community exclusive hosted on some other video platform.

Looking forward to figuring that out.

As for Whisper, it was surprisingly easy to get running locally.

Install python, run the pip command from https://github.com/openai/whisper and Whisper’s good to go; transcribe to your heart’s desire.

Naturally I way overcomplicated the last time I setup Whisper on my computer back in October because I wrapped it in it’s own service to run in a docker image locally, but I guess I did that with the dream of creating a modular AI pipeline running in the cloud.

Some day…

Running one install command was much faster and easier.

Combined with self-hosted n8n, I had a system for turning long videos into text in no time.

Let me know if you end up giving this Whisper tutorial a try!

All the Content

I had big aspirations this week to produce all the content.

I finally got a video uploaded on Monday, which was a win.

But a few distractions later combined with juggling way too many projects, and the Whisper tutorial was the only social media post for the week.

I chatted with GPT about it a little bit, been having some trouble prioritizing what to do first.

It actually recommended updating the landing pages first but not spending too much time on either - “two day sprint” was how it put it which is a great idea.

Once I have updated links, I can then focus on driving traffic to those links.

Next highest priority is minimum 1 long form video for YouTube, which is easy enough, that can be done in a day.

Then I can use that video to create more posts.

The next most logical piece would be to create short videos out of the long video.

Luckily those tutorials are fairly step-by-step so I’m curious to see if GPT can read the entire transcript from the tutorial (created with Whisper of course) and correctly identify the timestamps to turn the full transcript into a super short tutorial for Shorts, LinkedIn and TikTok.

The tricky part will be the horizontal to vertical video conversion and showing useful clips.

That’s going to need some manual editing, not sure AI can handle that kind of precise cutting yet.

But if it can identify the timestamps for potential clips, it might cut down on the editing time.

After that I can turn transcripts into social media posts pretty easily.

We’ll see, the quick text based post win may take priority while I work out a decent system to create vertical videos efficiently.

The goal for next week will be 2 total social media posts of any kind.

And if it happens to be a short, that would be a huge win.

Expect all the content next week.

AI Updates

A few things I’m looking forward to trying out from this week: Midjourney and Record mode.

Midjourney released their V1 video model and it looks really impressive from what I’ve briefly seen.

Midjourney always had more of a fantasy type feel to their images, so I get that vibe from the videos as well, but they are surprisingly stable in terms of how the physics look with crashing waves and giant dragons floating through clouds.

One advantage to their model is that it is “image-to-video” instead of “text-to-video” like the recently trending Veo 3.

Given an image, the model has much more reference material to work with rather than trying to generate the entire scene from a few words, guessing at all the details it needs to create a full video.

I’ll be curious to see if Midjourney videos end up being as viral as the Veo 3 gorilla that’s been everyone lately.

And one other feature this week that may be super useful, is the GPT app introduced “record mode”.

I’ve used a couple free meeting recording tools, but if GPT can handle the task of transcribing anything that’s happening on my computer, that could be a gamechanger.

No need to take notes any more if AI can write down everything that’s happening.

I’ll give both tools a shot next week and let you know how it goes.

And that’s it for this week! Lots of transcription talk, website updates, and the never ending quest to use more AI.

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!