Principles of Making

by Dineth Mendis

The Japanese do socks right

In Principles of Making, I explore what it is like to work in a technology companies of all sizes. My insights here reflect how I navigate my need to be a maker despite the many issues in working for an enterprise.

Dineth

  • , ,

    Writing software at an enterprise require different architecture tradeoffs

    This post delves into balancing autonomy and consistency, why duplication isn’t always bad, and how simplicity can outshine forced uniformity. It offers practical principles for creating systems that adapt and thrive in complex environments.

  • Bluesky seems cool

    I gave up on most things “social media” years ago, resorting to just messaging friends.

    However, I was fascinated by the idea of Bluesky. AT Protocol seems very interesting. So does its idea of applying composable moderation and selecting an algorithm to augment your personal experience instead of deciding a one-size-first-all approach.

    Perhaps the best thing about it is that it works like the open web. Really cool. I love this.

    Follow me: https://bsky.app/profile/dinethmendis.com

  • I moved to WordPress

    I’ve been writing a blog since I’ve been in high-school. At first, it was fun to make websites and be part of the Internet (this was pre-Facebook). I was in love with my monochromatic colour palettes. The posts… well, they were the voice of my teenage brain — scattered and overly sassy. I continued to…

  • , ,

    Allow projects to have exploration time

    We’ve been working on some terrific discovery stage projects in my team. We are all super pumped about it. Team has come up with some fresh ideas and a vision to execute on. However, we are very much seasoned in working within fixed structured, mature projects. Usually, we either work on small pieces or large…

  • How neural networks and LLMs work?

    If you are looking to dive into machine learning, how neural networks and LLMs work underneath, this is a really good YouTube course to watch.

    Really good chapters list:

    1. But what is a neural network?
    2. Gradient descent, how neural networks learn
    3. What is backpropagation really doing?
    4. Backpropagation calculus
    5. How large language models work, a visual intro to transformers
    6. Attention in transformers, visually explained
    7. How might LLMs store facts

    https://www.youtube.com/playlist?list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi

  • ,

    Difficult job of focusing in a large organisation

    It is time to work through our division operating plan for 2025. We’ve had many hours of discussions around which customers to focus on, volatility in the market, where we should build trust with customers and what kind of initiatives will amount to strong growth. I find it energising to spend time thinking about these.…

  • Peru

    After a 4 year hiatus, we returned to doing slightly longer holidays. This time, we were in South America for the first time! Peru was a great adventure. We spent time in 3 regions of Peru that had very different climates and ecosystems.Open the post to see the pictures from our trip.

  • Investing to catchup vs leapfrog

    It helps to find your own path to success. Good read about investing on catchup vs investing to leapfrog.

    …if your competitor’s boat gets ahead of you at the starting line, the instinct is to chase it. To tack where it tacks on the same path. But if you do, you will always be behind, because you will be sailing on the same path, subject to the same wind. And in fact for much of the time, you will be sailing in its ‘dirty air’ — i.e., it gets the full wind in its sails, which breaks the wind diminishing its effect on your sails. Your only hope is for the leader to make a terrible blunder. Otherwise, you will race for hours and hours, tacking dutifully behind the leader and crossing the finish line behind.

    https://rogermartin.medium.com/the-tragic-futility-of-investing-to-catch-up-aaf4b5c90e0f

  • Can AI make art?

    Fantastic article on The New Yorker on “Why AI Isn’t Going to Make Art“. Thought provoking and plenty of good threads to reflect on.

    Article dives into the need to make so many little decisions in order to create something of value. With a prompt, the amount of decisions made are minuscule.

    Even with code generators like Github CoPilot, this has been the issue. LLMs can generate basic code really quickly. It is certainly a great tool for picking up programming or learning a new language. Perhaps it will make a mid-level engineer efficient by 2-3%. However, for a competent coder, it offers little in the way to help solve actual problems. Code’s just an expression of a solution. Solving problems in code require a lot of decision making.

    It also talks about how real connections matter:

    Some individuals have defended large language models by saying that most of what human beings say or write isn’t particularly original. That is true, but it’s also irrelevant. When someone says “I’m sorry” to you, it doesn’t matter that other people have said sorry in the past; it doesn’t matter that “I’m sorry” is a string of text that is statistically unremarkable. If someone is being sincere, their apology is valuable and meaningful, even though apologies have previously been uttered.

    Perhaps my favourite part:

    We are entering an era where someone might use a large language model to generate a document out of a bulleted list, and send it to a person who will use a large language model to condense that document into a bulleted list. Can anyone seriously argue that this is an improvement?

    No, it isn’t.

    https://www.newyorker.com/culture/the-weekend-essay/why-ai-isnt-going-to-make-art

  • ,

    Strategy and focus

    Great quote from “Good Strategy/Bad Strategy: The difference and why it matters” by Richard Rumelt: “If the business is really successful, then there is usually a good strategic logic behind that success, be it hidden or not. But the truth is that many companies, especially large complex companies, don’t really have strategies. At the core,…

  • What kind of principal engineer are you?

    What kind of principal / staff engineer are you? This is a really good guide around what each type of role entails. One of the reasons why a single levelling guideline for an IC is really tough.

    https://staffeng.com/guides/staff-archetypes

    The four common archetypes of Staff-plus roles I encountered are:

    ❶ The Tech Lead guides the approach and execution of a particular team. They partner closely with a single manager, but sometimes they partner with two or three managers within a focused area. Some companies also have a Tech Lead Manager role, which is similar to the Tech Lead archetype but exists on the engineering manager ladder and includes people management responsibilities.

    ❷ The Architect is responsible for the direction, quality, and approach within a critical area. They combine in-depth knowledge of technical constraints, user needs, and organization level leadership.

    ❸ The Solver digs deep into arbitrarily complex problems and finds an appropriate path forward. Some focus on a given area for long periods. Others bounce from hotspot to hotspot as guided by organizational leadership.

    ❹ The Right Hand extends an executive’s attention, borrowing their scope and authority to operate particularly complex organizations. They provide additional leadership bandwidth to leaders of large-scale organizations.

  • Great read: the man who killed Google Search

    …Google’s response to the Code Yellow was to roll back changes that were made to maintain the quality of search results.

    Yikes. It is yet another story about corporate greed destroying a great products. Passionate write up, well worth the read: https://www.wheresyoured.at/the-men-who-killed-google/

    Thanks @Rob for the share. Coverage of Boeing troubles by John Olivier
    was another hot take on good products going astray under new management.

  • A week in Phuket

    Recently, we spent a week in Phuket (Thailand) to get away from our busy work schedules to just chill by the beach. Phuket is an island at the south of Thailand. It is mostly a tourist island with stunning beaches, bright blue waters, great resorts and delicious local food. Here are some pix!

  • ,

    Future of the web

    I recently had the opportunity to attend WordCamp Asia in Taipei. It was my first time in a WordPress community event. One of the key takeaways for me was that WordPress adoption was in decline. Not only that, behind closed doors, I got to hear that number of websites created on the web overall is…

  • Euclidean distance vs Cosine Similarity for text searches

    Good article on the topic: https://cmry.github.io/notes/euclidean-v-cosine

    Cosine similarity is the angle between two points, while euclidean distance is the actual distance between two points. Cosine doesn’t care as much about magnitude so does well for different lengths of data (think how far into a specific direction something is as long as they are in that direction).

  • Watch: Pantheon

    Great show about living forever. Best sci-fi show I’ve found this year. Check it out!! IMDB

Popular Reads

Maker Series