Wednesday, June 24, 2026
8:20 a.m. Room A1 Makhota Vibe Inn
Melaka, Malaysia
My time in Melaka continues. An odd thing is that one of the reasons I’ve stayed here so long is that the hotel prices go down once the weekday starts. And then when I see those lower prices, I want to take advantage of them. The next town down the coast is Muar, and I’ve never been able to find as cheap a hotel there. So when it comes down to a choice between staying in Melaka one more night or cycling to Muar, I get swayed toward Melaka by the cheaper room rate.
That happened again this week because, again, I had plans to get on my bicycle and ride away. I could have done that two mornings ago, yesterday morning, this morning, or even tomorrow morning. That’s Monday, Tuesday, Wednesday, and Thursday. But I needed a day to pack up and get ready, so I booked one more night. And then I saw on Agoda that if I booked TWO more nights, the room rate dropped to 31 ringgit. And I figure I would be a fool not to take advantage of that. So instead of booking just one night, I booked three or four nights. Get that cheaper rate. I figure I can stay here during the week at 31 ringgit/night. And then when the weekend room prices shoot up, I can leave.
But of course I’m not just sitting around doing nothing. I’m going out exploring and shooting videos. And by the time the weekend rolls around, I’m in the middle of editing and working on five new videos, and I end up spending the weekend doing that. Then the weekend is over, and suddenly I have those cheap room rates attracting me again. It’s kind of funny. Of course, I like it here in Melaka. I’m deepening my experience and knowledge here every day, and that knowledge leads to more small adventures as I stumble across new places to check out. And people watch my new videos and they give me new ideas for places to check out and things to do. Just now someone in the comments or in a message told me that the annual Dragon Boat Races are happening this weekend here in Melaka. I don’t know if this is a big event or a small local one, but it is something new that I would like to check out. But it would mean staying in Melaka for this weekend as well. And if I want to do that, I better book my room now before the room rates rocket up too much.
Time has felt in short supply, which is common for me. In some ways, I’m not sure where the time goes. But even the post-edit details for making YouTube videos seems to demand a lot of time. Every time I turn around, there seem to be three new videos on YouTube for which I need to do a bunch of stuff. Some of the stuff I do isn’t absolutely necessary, but it would annoy me greatly NOT to do them. I’m thinking in particular of correcting the YouTube Captions. As soon as I fired up my computer this morning, I remembered that I had uploaded a new video all day and last night yesterday. Uploading was very slow for some reason, and it took the entire day and most of the night. But it was completed by this morning. And this fresh video meant that the Captions hadn’t been corrected yet. So before I even had my first sip of coffee, I opened up that video in YouTube Studio and went into the menu for that video and made a copy of the YouTube Captions. And they were awful, as usual. My routine is to copy and paste the full set of Captions into a Google Docs document. This video was about a video I’m calling “Melaka, I Have a Bone to Pick With You!”. So I called this document “Bone to Pick – Captions.” And then I copy and paste those captions into an AI for corrections. Over time, I’ve come to depend on Grok for those corrections. Grok accepts very long documents that I can just paste in the normal window and it does a good job. ChatGPT, for example, won’t accept such long documents. You have to upload them as a PDF, and that adds more time to the job. And DeepSeek won’t do a big job like that all at once. It stops constantly mid-job and I have to prompt it over and over to keep going until it is finally complete. So, Grok does this particular job the best with minimal fuss.
I’ve also learned that despite their immense power, AI systems can be quite dumb. You have to be very clear about your instructions. And they can’t help themselves sometimes. I tell the AI that this is a video transcript, and I DON’T want them to rewrite or improve it. I just want corrections of spelling, etc. But they want to help so much that they end up rewriting the captions anyway. They condense it and make other changes. So I might input a video transcript that is 19,000 words long, and it comes back with only 7,000 words. It deletes most of it because it thinks it is an essay, and it rewrote it to make it better. No matter how many times I tell it to only make spelling corrections, it still rewrites it.
And then I made a breakthrough. I told the AI about this problem. And I said to the AI, if I wanted you to ONLY do corrections and NOT rewrite, what instructions would you give yourself? And the AI said that to do that, I should use this prompt:
You are a precise transcript proofreader. Your ONLY job is to turn the raw YouTube automatic caption text into a perfectly accurate, verbatim transcript of what was actually said in the video.
Rules you must follow exactly — do not deviate from any of them:
1. Do NOT paraphrase, rephrase, or rewrite any sentence for style, clarity, or brevity.
2. Do NOT remove or combine any sentences. Keep every idea, every repetition, every tangent exactly as spoken.
3. Do NOT shorten anything. The final word count must be almost identical to the original spoken words (only fillers like “uh”, “um”, “you know”, false starts, and obvious repeated words may be removed).
4. Only make these corrections:
– Fix obvious spelling errors and misheard words.
– Verify and correct proper names, place names, brand names, etc. by checking reliable sources (Google Maps, official websites, IMDb, Wikipedia, etc.).
– Add punctuation and capitalisation and paragraphs so it reads naturally.
– Remove pure filler sounds: uh, um, ah, er, like (when used as filler), repeated words caused by stutters or false starts.
5. Never remove meaningful repetitions for emphasis (“I really, really love this”).
6. Never improve grammar if it changes the original phrasing (e.g., keep “The guy, he went to the store” exactly like that if that’s what was said).
7. Output ONLY the corrected transcript and a summary of changes; nothing else.If you are tempted to reword anything for any reason, do not do it. Leave the original wording 100 % intact.
Here is the raw YouTube caption text:
[PASTE TRANSCRIPT HERE]
Corrected transcript starts now:
And that’s pretty amazing, isn’t it? If you gave those instructions to a human being, that human being would probably be upset with you. You’re treating them like a dummy because you are giving the same instructions 10 different ways. But an AI is so dogged and so stubborn that you have to be super clear about what you want. You have to repeat yourself and anticipate every angle and say the same thing over and over. And ever since I started using this prompt, I have gotten very good results.
I just used that prompt this morning for the video captions of the Bone to Pick video. And that video is relatively short for a Planet Doug project. The original captions were 6,877 words long. And after the AI cleaned it up and removed the filler words and corrected the spelling and added punctuation and paragraph structure, it was 6,709 words long. So it removed 168 words, but those words would have been things like “um” and “uh” and things like that. It’s amazing what the AI can do in such a short amount of time. The video captions from YouTube were riddled with hundreds of mistakes and errors. Everything was spelled wrong. There was hardly any punctuation. And the AI, Grok, made it perfect in seconds. And it even consulted 44 different websites to confirm the spelling of all the names and places I mentioned in the video. It’s mind-boggling what they can do. And at the end, it provides a complete list of all the changes it made so I can review them to make sure that everything was done correctly.
The AI did all the work here, of course, but it still takes time for me to go through the steps. And it isn’t really necessary. The vast majority of people don’t use the Captions on YouTube. And I doubt that any other YouTuber takes the time to fix them up as I do. But it bothers me immensely to see the Captions and how awful they are. So I want to make them as accurate and consistent as possible.
I guess I was telling that story to illustrate that even after you upload the video to YouTube, there are still time-consuming tasks to take care of. And that is where a surprising amount of my time goes. You wouldn’t think it would take that long, but even correcting the Captions and turning them into a perfect transcript can see a half hour to an hour go by depending on how things go. And then for that one video, you still have to write the video title and video description and make the thumbnail. I also sometimes take an excerpt from the video and upload it to Patreon. I’ve been doing that lately with the 3D Flyover Map. And I also take the transcript of every video and I post it to Patreon in case anyone wants to see everything I said in the video written down. That’s also something that isn’t necessary, but I figure since I took the time to create a near-perfect transcript of the video, I might as well make it available. And in that Patreon Post, I also include a PDF copy of the transcript. I’m pretty sure that not once has anyone taken advantage of that and actually downloaded the PDF, but it doesn’t take long to make a PDF and attach it to the post. So I just do it anyway. I figure, why not?
And by the time you do ALL these things for just one video, an entire morning can be gone. Do this for two or three videos, and a whole day is over. It’s surprising how much time gets sucked up by video production.
Yesterday, I was working on three different videos, including the video I shot when I visited the island of ghost ships. I have high hopes for that video because it is visually so striking. And it might be of interest to a wider audience. I hemmed and hawed about whether I should highlight the fact that I snuck onto this island. In the video, I play it a bit cool. I talk about how I’m not entirely sure whether I’m allowed to go out there or not. But in reality, I knew perfectly well it was private property and if anyone sees me, particularly security, they are going to kick me out. In fact, I had to wait for 45 minutes before I could even cross the bridge because there was an official-looking woman in an SUV parked at the bridge, and I was worried that she was security. So I hid behind some trees on the rocks and waited until she left. And while I was on the island, I made sure to stay as hidden as I could in and amongst all the giant ships. I knew that if anyone saw me, they would tell me to leave. So I KNEW I wasn’t supposed to be out there.
BUT should I make that part of the video title? Should I lean into that? Is that annoying clickbait? The typical boring Planet Doug video title would be something like “Exploring an Island of Abandoned Ships from a Land Reclamation Project.” And nobody would click on it. Alternatively, I could call it “I Snuck Onto a Secret Island of Ghost Ships – Melaka, Malaysia.” In the end, I think I’m going to use a title that starts with “I Snuck Onto….”. I don’t want to highlight that I’m out here breaking the law, but it makes the video more fun. And, to be honest, it’s not really that serious. It’s not like this island was surrounded by a high barbed-wire fence with security cameras and signs saying “KEEP OUT! PRIVATE PROPERTY!” everywhere. In fact, there was nothing. Not a single sign or security person anywhere. No fences. No gates. Nothing. Anyone who stumbled across the makeshift bridge could be forgiven for just waltzing across it to go check out the ships. Why not? It looks wide open and public.
Choosing a thumbnail for this video also took some time. Hopefully the one I ended up with will somehow trigger some interest from a wider audience. My video about the Submarine Museum worked out well. It is my most popular video in a long time. And I wouldn’t mind replicating that response with this new video about the ship graveyard island. I just went to look at YouTube Studio to doublecheck the view count for that submarine video, and it has 35,000 views. That’s what I would like to get for every video. And when I went to YouTube Studio, I saw a brand new feature. They just added a column to the display that automatically shows you exactly how much money each video made. It’s very sad and depressing to see that number, especially when you consider the number of hours of work that went into each one. I know that my videos don’t make money. I see the final total figure for the month, and it’s quite depressing. It shows just how much it actually COSTS me to make each video. I have to pay to make these videos rather than getting paid to make them. The Submarine Museum at 35,000 views is the only video that earned some AdSense revenue. According to this new YouTube feature, it earned $63 US. But even that is kind of depressing when you break it down into an hourly wage. I didn’t keep track of all the time that went into it, but it was easily thirty hours. Probably more. And that is a wage of around $2/hour at best. It’s probably closer to $1/hour. And that’s for a very successful video. Most of my videos would probably average out to twenty-five cents an hour.
One interesting thing about being able to see the video earnings right on the main page is that you can compare across videos that get the same number of views. And this clearly shows that the view count is not as important as you think. What’s important is viewing time. What I mean is that two videos can both have 3,000 views. But one of the videos will have earned just one third or one fifth of the other one. And that’s because even though the same number of people clicked on each one, they just stopped watching one of the videos. Found it boring and clicked away. So it’s more important how LONG people watch the video. It’s not just the view count. A video could have 200,000 views, but if every single person watches fifteen seconds of the video and stops, then the ad revenue is not going to pay for even your morning cup of instant coffee. Anyway, enough about that depressing stuff. It’s a good thing I enjoy making the videos…
On the positive side, in some downtime, I’ve watched some interesting movies and TV shows. One of the best movies I’ve watched is something called “Pressure.” I honestly had no idea what it was about. It just came highly recommended, and I enjoyed it very much. It’s based on the true story of how weather forecasting was involved in the D-Day landing at Normandy. I love stories like that that give you a new perspective on a very important event. It is essentially a behind-the-scenes story showing what was really going on during the planning of the Normandy beach landing. If you like World War II movies and you enjoy technical stuff, this movie would be enjoyable for you. I think it was based on a stage play, which makes sense. Most of the movie happens inside rooms with people talking. I suppose a general audience would find that boring, but I find movies that are based on plays are often the ones I like the most. And the heart of the movie was a disagreement between an old and a new way of weather forecasting. And the drama was intense because tens of thousands if not hundreds of thousands of lives depended on getting the weather forecast right. If they attempted the Normandy landing during bad weather, it could be a disaster. In fact, it could mean losing the entire war. The stakes were high. And the events as they unfolded were very dramatic. It’s an exciting story and one of my favorite movies from recent months.
And perhaps watching Pressure rekindled an interest in World War II because I decided to dive into the new World War II series with Tom Hanks. Apparently, the full series consists of 20 episodes and each episode is an hour long. That is a WEALTH of insight and information about World War II. I’ve only watched the first two episodes so far, and they were both fantastic. I enjoyed them immensely, and they both sent me down deep rabbit holes learning about other aspects of World War II. I’ve done this in the past, of course. I’ve watched many other documentaries about World War I and World War II. But it’s been a while, and I really needed a refresher course. And this Tom Hanks series fits the bill perfectly.
It got me thinking that we all need to sit down and do a month-long course in world history every five years or so, just to refresh our memory. We all go through life learning all kinds of stuff, and it is important stuff to know. But we naturally forget. If I remembered everything I ever learned about history, I would be a genius. But I learn stuff and then forget. At various times in my life, I’ve done deep dives into all periods of history. And then that knowledge fades. And World War II essentially shaped the world we live in today. Everything about our world now is dependent on the events of that war and what happened afterward. Without an understanding of that time period, you can’t understand the world as it is now in 2026. So it’s important stuff. But I just can’t keep that knowledge in my head permanently. I need to do constant refresher courses.
I’ve been so busy over the last few days that I haven’t had the time to reply to comments. That’s my plan for this morning. Once I post this to Patreon, I am going to do a few other YouTube-related things, and then I will go through the most recent comments and reply to them. That’s my plan for most of today in fact. I do have an idea for another video I want to shoot. But it is essentially another history and museum video. The topic I have in mind is fascinating to me and it is probably worth doing, but seeing the new earnings tab on YouTube has depressed me a little bit. It will take some inspiration for me to grab the GoPro and film this newest video. We’ll see how it goes.