I’ve been longing for this.

Hi there! Running karma here! I’ve been longing for this.

It’s been about two months since I started using ChatGPT for learning English. And for the past couple of weeks, I’ve been playing around with AI video generation — and honestly, I was so blown away by how far AI has come and what it can do, that I decided to start this blog.

At first, this blog was just a test space, a place to practice making blogs. But now, I’m finally kicking it off properly, along with an Japanese version, as part of a plan I’ve had in my head for the past few years. From now on, I want to spend the rest of my life expressing my thoughts, curiosities, and ideas through blogs and YouTube videos — right up to my last breath!

In the past, I used to do a little affiliate blogging for some extra income, but this blog is going to be different. I’m not going to worry much about SEO or chasing trends. Instead, I’ll just share what I love, what I think, and what I wonder about — in my own way.

I’ll leave the video generation site and avatar creation site I used below. You can make a few videos for free each month, so feel free to give it a try!

So, why now?

Because something I’d once given up on — making videos — suddenly became possible thanks to AI!

For example, I can now turn images like this:

into smooth, high-quality (FHD) videos with beautiful camera movements and angles that I love.

And here’s another one:

This one’s in HD (maybe just 720p because it’s from the free version), but my goal is always FHD or higher. I want the quality to hold up on a big 27-inch screen, not just on a phone. A while ago, AI video tools weren’t quite good enough for that. But now, they’re finally getting there. I can even upscale to 4K!

I even tried making a “drone shot” style top image for the blog — starting from a still photo, and using prompts like “walking away from the camera, drone pulling back and rising to show a vast peach orchard and ocean.” Of course, Yamanashi (where I took the original photo) doesn’t have an ocean… but hey, with AI, you can create what you imagine! How cool is that?

I’m not really into weird or surreal images, but super realistic shots like this? They’re amazing.

And while many AI video generators struggle with Asian faces, Kling seems to do a pretty good job!

I once gave up on making videos…

I wanted to make videos because I’ve always dreamed of capturing those beautiful, cinematic shots — where the background is softly blurred, the subject really stands out, and everything moves in slow motion. I was so moved by that kind of visual that I thought, “I want to try making it myself!” So, I started researching.

Turns out, filming in 4K with smooth slow motion (shooting at 120p and converting to 24p or 30p to slow it down by 4–5 times), and keeping the subject perfectly in focus with autofocus — that’s only become realistically possible in the last few years.

And of course, the gear…

If you want that cinematic movement with smooth camera work, you need a gimbal. A drone would be amazing too. Plus, to edit high-quality footage, you need a decent PC.

Then you need a model. Good lighting. Clean audio. Beautiful locations.

I love shooting people, but there’s always the copyright issue. Even if I make a deal with a model, if someone else’s face accidentally shows up in the background, that could become a problem.

And drones… the rules are getting stricter and stricter. You can’t just fly them anywhere.

Even if I drive out far during the perfect season, when the peach blossoms are in full bloom, the weather doesn’t always cooperate…

I really wanted to shoot something like this:

A beautiful woman walking through a city street after the rain, with reflections of car headlights and city lights shimmering on wet roads and puddles. That’s the kind of scene I imagined.

Well, I don’t have any friends who would wear a dress and walk around the city shaking their hips for me, so if I need a model, it’s going to cost money. And if I end up clicking with the model, I’d probably end up paying for things like hotel costs (laughs). Plus, you can’t always count on the weather being good or rain coming at the right time during the shoot.

Trying to do all this by myself… yeah, I realized it’s way too much.

Sure, professional gear is becoming cheaper and more accessible, and I really did think about giving it a serious try. But in the end, it was just impossible. So instead, I ended up helping out when people asked — filming surf contests or skateboarding videos for friends.

But now…

But now, everything is finally in place. With the new Mac Silicon chips like the M1 and M4, I can get a PC with specs that would’ve cost around 500,000 yen five years ago, for just 80,000 yen. (I just checked Mac products and saw the Mac Mini M4 for 80,000 yen, so I ended up buying it…).

Being able to edit 4K videos cheaply on a PC is possible because of this timing.

And really recently, AI video generation has been improving fast, and now I can finally create the visuals I’ve been wanting.

Chat GPT’s Sora was announced in 2024, and the Kling (ver. 1.6) I’m liking now is also from just a few months ago. So, I’m experiencing the cutting edge of AI video generation right now—exciting stuff!

Well, this post is just to announce that I’ve started this blog, and I plan to channel all this energy into it.

I’ve made around 10 videos so far, and video generation is working great. The prompts are key, and it’s better to start by creating images first and then make the video from that, instead of just text. That way, you can create something closer to your vision with fewer mistakes. There are lots of image-generation AIs, but it’s easier and better quality to make them directly within Kling.

It seems like Japanese-only prompts aren’t stable yet, so it’s better to write prompts in English, but I can just use ChatGPT for translations.

Creating beautiful videos is totally doable, and I’ve learned it quickly. Now, I’m challenging myself to create avatars and make them talk!

This is in Japanese, but as a native Japanese speaker, it feels a bit awkward to me. But if I make it speak in English, you wouldn’t be able to tell if it’s AI or not.

Sorry for the quick shift, but while I was researching AI video generation, I found out you can already create your own avatar or a fictional one and make it talk freely. It looks like making your avatar into an influencer might be a trend too. Hmm, sounds tempting… I could make travel videos, review videos, and explainer videos with my own avatar acting like it’s real. Plus, you can even have videos where I talk with myself or travel together with my avatar.

I’m planning to make videos soon where I have conversations with my virtual avatar. Stay tuned for my ‘AI Introduction Diary (Tentative)’ with five episodes where I give it a name and a look. (Haha)

The sites I used this time.

Making Video-Kling

Making Avatar-Heygen

There are plenty of other video generation sites, and most of them let you use them for free a few times a month. If you check out different sites, you can make quite a few videos for free! Run your karma!