<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>Chinmay Deo</title>
  <subtitle>Personal blog of Chinmay Deo — short posts on learning, tech, and life.</subtitle>
  <link href="https://legodud3.github.io/" rel="alternate" />
  <link href="https://legodud3.github.io/feed.xml" rel="self" />
  <id>https://legodud3.github.io/</id>
  <updated>2026-09-28T00:00:00Z</updated>
  <entry>
    <title>Intelligence may not follow division of labour</title>
    <link href="https://legodud3.github.io/p/2026-09-28-intelligence-may-not-follow-division-of-labour.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-09-28-intelligence-may-not-follow-division-of-labour.html</id>
    <published>2026-09-28T00:00:00Z</published>
    <updated>2026-09-28T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>I spent 90 minutes skimming through the deepseek-coder paper. It was released in early 2024. Deepseek (&amp; others) had started giving signals that coding is a great "agentic" use case of foundation models. It's a fascinating paper, even for a non-technical person like me, because of how intuitive, sim</summary>
  </entry>
  <entry>
    <title>BLR8</title>
    <link href="https://legodud3.github.io/p/2026-09-16-blr8.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-09-16-blr8.html</id>
    <published>2026-09-16T00:00:00Z</published>
    <updated>2026-09-16T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>My current hypothesis for generating new ideas is to put myself in new situations. Today, I used Amazon Tours to visit Amazon's Bangalore Fulfillment Center (FC) BLR8. Took me back to ITC days and the start of my career to be in a warehouse. Quite fun! It's a huge space (~0.4M sq ft) for a normal pe</summary>
  </entry>
  <entry>
    <title>Picking myself</title>
    <link href="https://legodud3.github.io/p/2026-09-15-picking-myself.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-09-15-picking-myself.html</id>
    <published>2026-09-15T00:00:00Z</published>
    <updated>2026-09-15T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>I have spent the last few weeks trying to capture my state of mind in this blog post. Until today, the next 'thing to do' has always been assigned to me. While I rejected the obviously bad things to do, my approach was to default try stuff out and reject. I rarely questioned if the next 'thing to do</summary>
  </entry>
  <entry>
    <title>Instinct</title>
    <link href="https://legodud3.github.io/p/2026-09-14-instinct.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-09-14-instinct.html</id>
    <published>2026-09-14T00:00:00Z</published>
    <updated>2026-09-14T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>I got an invite to the viral AI assistant Instinct over the weekend. I have been playing with it for about 48 hours. I have also told everyone I know about it and liberally sent out invites. It's hard to emphasize how rare this is for consumer products. The last time I did this level of recommending</summary>
  </entry>
  <entry>
    <title>Talk less and listen more</title>
    <link href="https://legodud3.github.io/p/2026-08-15-talk-less-and-listen-more.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-08-15-talk-less-and-listen-more.html</id>
    <published>2026-08-15T00:00:00Z</published>
    <updated>2026-08-15T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>Trying to apply "The Mom Test" in every conversation I have. Realizing that it is hard, so hard. What is so hard? To NOT assign labels to the users' conjectures, opinions, or actions in the conversation. Labels are "context-window compaction." To NOT talk about your ideas. My brain tends to subconsc</summary>
  </entry>
  <entry>
    <title>xteink</title>
    <link href="https://legodud3.github.io/p/2026-07-27-xteink.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-07-27-xteink.html</id>
    <published>2026-07-27T00:00:00Z</published>
    <updated>2026-07-27T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>I am tempted to pick up an xteink device. xteink looked at a mature market (e-readers) and had a unique set of insights about that market: Miniatures: Miniature size, not miniature features. This is particularly pronounced in watches, which for a long time have been ridiculously large compared to mo</summary>
  </entry>
  <entry>
    <title>Kindle</title>
    <link href="https://legodud3.github.io/p/2026-07-10-kindle.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-07-10-kindle.html</id>
    <published>2026-07-10T00:00:00Z</published>
    <updated>2026-07-10T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>I enjoy reading on the Kindle. It's a great experience because it does only one thing and nothing else. There is a real commitment from the user when they pick up a Kindle: I am going to read. For the next 10 minutes/30 minutes/3 hours, I am a reader. People enjoy that tag, even if it's temporary, b</summary>
  </entry>
  <entry>
    <title>Thai food &amp; beverages</title>
    <link href="https://legodud3.github.io/p/2026-07-03-thai-food-beverages.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-07-03-thai-food-beverages.html</id>
    <published>2026-07-03T00:00:00Z</published>
    <updated>2026-07-03T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>First, an admission: I have not written a lot over the last 1 month. I can't make many excuses, but for what it's worth, we took Samay on his first international trip to Thailand in mid-June (excellent!). I have had an aversion to Thai food because I had never had Thai food in Thailand. Also, I am a</summary>
  </entry>
  <entry>
    <title>Thoughts on the book "Project Hail Mary"</title>
    <link href="https://legodud3.github.io/p/2026-05-29-thoughts-on-the-book-project-hail-mary.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-05-29-thoughts-on-the-book-project-hail-mary.html</id>
    <published>2026-05-29T00:00:00Z</published>
    <updated>2026-05-29T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>Book spoilers ahead I have been sick nearly all of the last week, so all personal projects are on the back burner. Since I was confined to the house and we will not be entering a movie theatre for at least some time, I picked up a book. Actually, two. I finished "Strange Houses" by Uketsu and then d</summary>
  </entry>
  <entry>
    <title>Don't consume your day with current events</title>
    <link href="https://legodud3.github.io/p/2026-05-27-don-t-consume-your-day-with-current-events.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-05-27-don-t-consume-your-day-with-current-events.html</id>
    <published>2026-05-27T00:00:00Z</published>
    <updated>2026-05-27T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>Let's take an event: "THE LATEST CRISIS." You read, watch, or listen to the take, the counter-take, the social media take, the influencer take, the Reddit community take, your friends' take, and your parents' take. Before you know it, without your knowledge, recognition, or considered consent, you h</summary>
  </entry>
  <entry>
    <title>Prepare for my analysis of the most awaited IPOs of our generation - SpaceX, OpenAI, Anthropic!</title>
    <link href="https://legodud3.github.io/p/2026-05-22-prepare-for-my-analysis-of-the-most-awaited-ipos-of-our-generation-spacex-openai-anthropic.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-05-22-prepare-for-my-analysis-of-the-most-awaited-ipos-of-our-generation-spacex-openai-anthropic.html</id>
    <published>2026-05-22T00:00:00Z</published>
    <updated>2026-05-22T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>Word on the street is that these are dropping soon. I plan to read the S-1s and form my own conclusions with help from AI. I will share them here.</summary>
  </entry>
  <entry>
    <title>To learn something well, start small and follow historical evolution</title>
    <link href="https://legodud3.github.io/p/2026-04-24-to-learn-something-well-start-small-and-follow-historical-evolution.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-04-24-to-learn-something-well-start-small-and-follow-historical-evolution.html</id>
    <published>2026-04-24T00:00:00Z</published>
    <updated>2026-04-24T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>I have been heads-down for the past ~3 weeks understanding agents. I played around with them, listened to podcasts, and explored GitHub repos. H/T to smol-agents (Hugging Face) and pi (Mario Zechner). I am on board with their philosophy of building minimal coding agents and letting the models do the</summary>
  </entry>
  <entry>
    <title>Selling the dream is the most significant driver of market cap</title>
    <link href="https://legodud3.github.io/p/2026-04-10-selling-the-dream-is-the-most-significant-driver-of-market-cap.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-04-10-selling-the-dream-is-the-most-significant-driver-of-market-cap.html</id>
    <published>2026-04-10T00:00:00Z</published>
    <updated>2026-04-10T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>We live in a capitalist world. Consumers, corporations, and the state (i.e., government via laws and regulation) are the most important groups in this world. Consumers have value because they generate demand for various goods and services. Corporations have value because they fulfill these demands.</summary>
  </entry>
  <entry>
    <title>How models approach document processing</title>
    <link href="https://legodud3.github.io/p/2026-03-30-how-models-approach-document-processing.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-03-30-how-models-approach-document-processing.html</id>
    <published>2026-03-30T00:00:00Z</published>
    <updated>2026-03-30T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>I decided last week to shift from (using AI to build vanilla software) --&gt; (using AI to build software with AI features). This is a not-so-subtle shift in my view because I believe that unless a product has real AI-enabled features, (a) it is liable to be subsumed by general-purpose consumer applica</summary>
  </entry>
  <entry>
    <title>Happy birthday to me!</title>
    <link href="https://legodud3.github.io/p/2026-03-27-happy-birthday-to-me.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-03-27-happy-birthday-to-me.html</id>
    <published>2026-03-27T00:00:00Z</published>
    <updated>2026-03-27T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>Today is my 36th birthday. Happy birthday to me! Physically, I am (near) the best shape of my life, though there is a long way to go. Mentally, I feel like I have just turned 30, so a long way to go there too. Samay is here too, to celebrate, which I love. Every morning he wakes up and gives me a wi</summary>
  </entry>
  <entry>
    <title>Strategy formulation should be partly democratic</title>
    <link href="https://legodud3.github.io/p/2026-03-21-strategy-formulation-should-be-partly-democratic.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-03-21-strategy-formulation-should-be-partly-democratic.html</id>
    <published>2026-03-21T00:00:00Z</published>
    <updated>2026-03-21T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>I recently interacted with a few senior folks (think top 10% of the org). Our company had announced a transformative decision—one that will change the trajectory of the company—a couple of months ago. (Execution will tell whether the trajectory change is minimal or meaningful.) What struck me was: H</summary>
  </entry>
  <entry>
    <title>Why I won't use AI to write first drafts</title>
    <link href="https://legodud3.github.io/p/2026-03-21-why-i-won-t-use-ai-to-write-first-drafts.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-03-21-why-i-won-t-use-ai-to-write-first-drafts.html</id>
    <published>2026-03-21T00:00:00Z</published>
    <updated>2026-03-21T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>Unless you have been living under a rock, you know that AI can write better than you (the average person). AI can express ideas in a simpler, clearer, and more compelling fashion...or any fashion you want, really. The question is: Should you let it? It is useful to start by thinking about ghostwrite</summary>
  </entry>
  <entry>
    <title>First trip with Samay!</title>
    <link href="https://legodud3.github.io/p/2026-03-17-first-trip-with-samay.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-03-17-first-trip-with-samay.html</id>
    <published>2026-03-17T00:00:00Z</published>
    <updated>2026-03-17T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>We are currently in the middle of our first trip with Samay, visiting Bombay and Delhi for a few days each. In addition to meeting family and friends, this trip is a great "dry run" for future, longer trips with a newborn. Our parents love spending time with all three of us. Yes, they have expressed</summary>
  </entry>
  <entry>
    <title>What does the future software PDE org look like?</title>
    <link href="https://legodud3.github.io/p/2026-03-12-what-does-the-future-software-pde-org-look-like.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-03-12-what-does-the-future-software-pde-org-look-like.html</id>
    <published>2026-03-12T00:00:00Z</published>
    <updated>2026-03-12T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>I was listening to a wonderful interview with Bret Taylor by John Collison. It got my brain gears spinning on this particular question: What does a software PDE org look like in the next 12–18 months? I think making predictions beyond this year is moot. Specifically in PDE because the rate of change</summary>
  </entry>
  <entry>
    <title>Moats - stronger/weaker</title>
    <link href="https://legodud3.github.io/p/2026-03-06-moats-stronger-weaker.html" rel="alternate" />
    <id>https://legodud3.github.io/p/2026-03-06-moats-stronger-weaker.html</id>
    <published>2026-03-06T00:00:00Z</published>
    <updated>2026-03-06T00:00:00Z</updated>
    <category term="Reflection" />
    <summary>Now that everyone can vibe code (or "agentic code," if you want to be posh) a company of their choice.... Stronger or at least similar as pre-AI Network effects: Where users beget users, or where supply &lt;&gt; demand reinforce each other (e.g., social networks). Data: You can copy the app, algo, UI, mar</summary>
  </entry>
</feed>
