TJ is Blogging!

Coding is cheaper? In the current day and age, since the start of AI coding agents, the new software products are getting high rocketed. The official GitHub stats are proving this already.  There're news that GitHub is barely holding against new traffic with AI agents and some companies like Anthropic are trying to switch to new version controlling systems that can scale with the rate of AI agent traffic. This means coding is a software product is getting cheaper and cheaper everyday compared to coding with humans as it used to be. This means new software products are getting created and people are starting to build new products even without any developing or engineering background who are vibe coders.  Now, it seems a unique code or a feature is not going to be a flagship product to a certain company anymore. This is getting proven everyday with replicating products of flagship brands that used to dominate the market. We're seeing replicas of Office suites, image editing, vid...

Planning your vacation abroad is not a simple task when you do it on your own. It's easy when you go through a travel agent but it's always much more cheaper if you do it yourself. This time I decided to do it on my own. However it's a hassle and much more research as always, but this time I had ChatGPT for help. Which made my entire trip planning much more convenient.  Destination First thing's first. I created a project in ChatGPT for my vacation 2026 and opened chat. I provided my requirements. Per person budget that I'm willing to spend Dates of travel With whom I'm traveling. My family in this case. What we like to do and what we don't like to do Our meal preferences Shopping expectations Must included and optional things A list of my personal preferred destinations With this information ChatGPT gave me a few options. 1. Malaysia 2. Thailand 3. Singapore + Malaysia With some back and forth discussions with the family and shuffling designations we decide...

 Recently I was travelling abroad with my new Pixel 10 pro and I noticed the GPS accuracy was way off. It hardly correctly pin pointed my exact location. It was a disaster to hail a taxi using an app like Grab or Uber. I was wondering if that's an issue with my phone or something else.  Actually it was both. It turns out weather, sky scrapers in a dense city could impact GPS accuracy significantly. So, it is expected to have a messy GPS when you're in a new city. I found two things I needed to follow from my end. 1. First mistake I did was battery saver ! I had battery saver on for most of my time coz I was on the road and wasn't sure when to charge. It  turns out the battery saver takes away some good part of the GPS accuracy. So, whenever you hail a taxi, make sure to turn off battery saver on your phone. 2. Improved GPS accuracy settings.  Android GPS Settings Go to Settings > Location > Location Services > Google Location Accuracy, and turn on Improve Lo...

  I've been using ChatGPT for my everyday work for quite some time now, including some project planning, generating code snippets, documentations, proof reading, and various tasks. One of the features I like is Projects in ChatGPT. By creating a project, you could consolidated your related chat conversations into a single space. You can start multiple threads with multiple topics in the same project.  The issue The issue is the context window and the memory. When you're working on a large project which could drag for several days, and say you wanna generate a document based on the brainstorming or any research you did with ChatGPT. When longer you go through the chat conversations, it tends to forget your previous conversations.  If you're using the built in canvas for the documentations within ChatGPT, it helps to create and edit short documents while conversing. However, if it's a longer project, the chances are when you ask to update it later, it might have messed up...

Using AI for text generation is the new normalcy, and we have to accept that it's really efficient and helps us to save a lot of time. Yet, I personally feel that it's kind of polluting the originality of the data on the World Wide Web. I mean, the LLMs are trained on the original data created by humans in the first place; now the same data is being generated by AI, and before long there will be a (I think we're already there) time when it won't be able to distinguish human-created data from AI-generated data.  By data, I don't mean only text, it could very well be images, audio, and video of all formats.  Can LLMs change our language, culture as we know it? Let's try to imagine the bigger picture here. The first generations of LLMs were trained on human-generated data, no doubt. Then the next few iterations might have some polluted data generated from the first few generations of AIs coz they're constantly getting trained on the public Internet. Which means...

 Use the following commands to install MariaDB 10 in Amazon Linux sudo yum install mariadb105-server.x86_64 sudo s ystemctl enable mariadb sudo systemctl start mariadb sudo systemctl status mariadb

 I wanted to create a Java List with a single element. Yet, I wanted to add more elements later. So, I was looking for a couple of ways to do this. So far there are multiple elegant ways to create a list with a single element with a one-liner. Not so much for a modifiable list though. Here's what I gathered so far. Followings are a few ways of creating a list with strictly a single entry. Can't add more elements. 1. Collections.singletonList() This returns an immutable list that cannot be modified or add more elements. // An immutable list containing only the specified object. List<String> oneEntryList = Collections. singletonList ( "one" ) ; oneEntryList.add( "two" ) ; // throws UnsupportedOperationException 2.  Arrays.asList() This returns a fixed-size list consisting of the number of elements provided as arguments. The number of elements provided would be an array hence the size is fixed to the length of the array. // Returns a fixed-size list List...

 I wanted to check the connectivity to an email service from a server that belongs to a client. I had restricted access to this server so I couldn't install new packages like Telnet.  If I had telnet it's much easier, but fortunately, the CURL package was already there. Curl is usually available to any user in a typical Linux system. In this case, CURL comes to your rescue. Did you know that Curl supports telnet commands as well? Curl is known as a tool to test HTTP layer functions. Whereas Telnet supports lower TCP level connections. With Curl supporting telnet, it can also be used for testing the TCP connections. eg:  curl -v telnet://www.google.com:80 -v switch is being used to spill verbose output from the request. In this example, I've used port 80, but since it is capable of using TCP, it can use any port.

In Linux, "find" command is a powerful command that we can use not only to find files. I've used it for different requirements, so thought of posting some of the convenient usages of find commands in a series of posts. For some application issue, I've got one of my directories in Ubuntu system swarmed with empty files with 0 bytes. Some of the files got data in them but some of them were empty. I just wanted to delete the empty files and it wasn't possible to identify them by file names or dates. So find command comes to rescue. find . -size 0 -type f  If I needed to isolate the files by file extension, we can use something like this. find . -size 0 -type f -name "*.tar" You can get the file count by piping it to wc command. find . -size 0 -type f -name "*.tar" | wc -l The awesome thing about the find command is you can use it for deleting files just by adding the -delete flag to the command. find . -size 0 -type f -name "*.tar" -dele...

 If you get to troubleshoot some unknown Linux server, running these few commands will be useful. 1. Check who's logged in w This is a cool one-letter command to see who else logged into the server. It will list the usernames and even you can find the IP and the connection type. In the following example, I've logged in as the "user". If you see, some other users, watch out for unauthorized access. 2. Identify the OS There are different commands to identify what kind of Linux distribution that you are dealing with. Knowing this is crucial to decide what commands to use later. cat /etc/os-release Usually, in Linux distributions, you have a file /etc/os-release . Basically, you can see everything you need to know about the OS in this file.  3. See running processes top This basically lists the processes running with the CPU and memory consumption. With this, we can get an idea about what kind of apps running on the server and if any of them uses too much CPU or memory. 4...

 If you write a single-page app where the routing is handled on its own, it works perfectly on your local host with the dev server. For an instance, if there is a route in your app for login such as "/login", if you try this in the localhost it will work without any issues.  If you host this with an Nginx server as a static web app, using the direct route in the URL will end up with 404. There is an easy fix for this updating the Nginx configurations.  Nginx default static site configuration is located in " /etc/nginx/conf.d/default.conf ". This may differ based on your installation and OS. However, the configurations are pretty much the same. You can add the following line to the configuration  try_files $uri $uri/ /index.html?q= $uri&$args ; This makes the Nginx look for files in the file path, and if found then serve it. Or else, route the request to the index.html.  In our example, it will look for a file called login in the static directory, since we...

This has been one of the common interview questions for senior engineers for the Java Spring framework. I was pondering this question myself for a while about the pros and cons of using the constructor injections. Spring framework provides the means for dependency injections using 3 ways using annotations (@Autowired). Constructor injection Field injection Setter injection Out of these 3, the Field injections and Setter, injections are somewhat similar, but field injections are more commonly used than the Setter injections. I believe that is due to the fact that it is more convenient to use and implementing setters are kind of automated using libraries such as Lombok. I can think of a couple of reasons to use Constructor injections over Field injections. 1. Immutability Using constructor injections, it buys us immutable objects of the dependent class. Whereas if the Fields or Setters were used, they can be modified during the runtime which would make it not exactly thread-safe. Having ...

There are two main commands in Ubuntu for creating users. I've found them confusing and I always forget which to use when. So, thought to write them down clearly as a quick reference.

Replace the server name and the port. Usually the port will be 25.  openssl s_client -connect <hostname>:<port> -starttls smtp eg:  openssl s_client -connect smtp.office365.com:587 -starttls smtp Copy the output section of the certificate that looks like below. -----BEGIN CERTIFICATE----- ###bunch of encoded text### -----END CERTIFICATE----- You can copy it and save it to a file with ".crt" or ".cert" extension and that's the mail server's certificate. Make sure to include the -----BEGIN CERTIFICATE----- and -----END CERTIFICATE----- lines. Note: OpenSSL is required. Ubuntu:  sudo apt install openssl Other OSs:  https://wiki.openssl.org/index.php/Binaries

Issue The code deploy events keep hanging in the pending status without starting.  Possible Reasons 1. CodeDeploy agent not installed? Make sure to install the CodeDeploy agent in the EC2 instance (If you are using EC2 instances to deploy the application) Refer to the documentation for the CodeDeploy agent installation . 2. Is the code deploy service role properly created? Make sure to check the proper permissions for the service role that you have used for the code deploy service role.  https://docs.aws.amazon.com/codedeploy/latest/userguide/getting-started-create-service-role.html 3. Make sure the IAM instance profile is properly created https://docs.aws.amazon.com/codedeploy/latest/userguide/getting-started-create-iam-instance-profile.html 4. Is the code deploy endpoint command endpoint reachable?  For each region, there is a CodeDeploy command endpoint. Make sure that you can reach the command endpoint from your EC2 instance.  eg: for us-east-1 ping codedeploy.us...