Category: Work

  • Technical Writing in Anthropic

    Tom Johnson over at I’d Rather Be Writing has a nice podcast up with the technical writer at Anthropic, Sarah Deaton.

    Sarah explains how she “orchestrates rather than writes” the documentation. She reviews the documentation in GitHub so that all of her reviews are tracked and can be fed back.

    Some suggestions that she had for technical writers:

    • Learning how to a steer a module through skills, plugins, or customization layers.
    • Build your own style guide and then try to get the model to follow that guide.
    • Start tinkering.

    I get surprised sometimes when I hear that a company that I follow for various reason has tech writers. And then, it would make total sense to me that as a technology company, they would have tech writers – if not in India then in some other part of the world.

  • World TechComm Day 2026 in Pune

    The Society for Technical Communications (STC) India chapter held the World TechComm Day 2026 celebrations simultaneously in Pune, Bengaluru, Hyderabad, and Mumbai on 6 June 2026.

    I went to the venue with Rakesh ettan in his car, Hatchi. We reached on time and met up with some former colleagues. One of my former colleagues was also presenting.

    Building a Strategic AI Roadmap

    The first session that I liked was on Building a Strategic Artificial Intelligence (AI) Roadmap for Technical Communication by Snehal Yerawadekar of Philips. The talk was about implementing AI in a technical communications team within a highly regulated sector – medical devices.

    Snehal began by talking about why AI matters in technical writing. She says that product and content complexity, faster release cycles, localization demands, and higher customer expectations. AI enables TechComm teams to meet this demand.

    She continues that an AI strategy must provide business value (align the AI strategy with the business unit strategy), focus on the key aspects (turn the priorities in the business unit into measurable outcomes), keeping with the trend (anticipate and adapt to the latest trends and review organizational policies and change existing business models to advance the TechComm team), and Collaboration.

    She then spoke about having an AI ready team (with an AI mindset), manage processes with Management, and leveraging technology by using solutions that align with the strategy. She then went ahead about how she used these principles to implement an AI strategy in Philips and the future projects that they are working on.

    Shifting Style Guides to AI Skills

    The second talk I want to talk about is Editing at AI Speed: Shifting Style Guides to AI Skills by Amruta Hungund at Salesforce.

    She begins by saying that human is at the center at Salesforce. It is AI + human. She says that all the style guides that technical writers so adore must be rethought of as a rules engine for AI.

    She says there are two types of checks one does against a style guide. One is deterministic checks (which are yes/no rules) like spell check, grammar rules, punctuation rules, mandatory format requirements etc., AI can fix these easily without needing a human in the loop. For probabilistic checks, like context and meaning, nuanced tone adjustments, logic flow and arguments, etc., it is better to have humans in the loop.

    She says that the Rules Engine does not need conversational essays on brand philosophy, basic grammar remediation, sometimes rules etc. She suggests that rules must have a list of negative constraints, versioning, product names, voice and tone, sentence simplification, etc.

    The style guide must then be optimized for the use as an AI rules engine. This means having conditions like “max_words_per_sentence”: 25 instead of “Avoid overly long sentences to ensure readability and clear flow.” After optimizing the style guide, she suggests using Google Gems (or your stand-in for skills). She defines a Gem as an addition of a prompt to a rules engine.

    She also suggests using a Gem to collect data from the style guide to optimize the Gem.

    She asked what was the job of human in the loop – governance, conflict resolution, and understanding the larger context.

    She suggested a podcast towards the end of her talk – The Culture of Writing podcast. Episode 30, in particular. But, it is a Salesforce channel podcast.

    Quiz

    I was part of a team that won a quiz. Mugdha (who took my interview at my present workplace), Rakesh, and I were in the team.

    We wrote answers in the first round on a piece of a paper as a team. After we submitted the paper, we ended up as one of four teams that was part of an oral round which we topped after breaking a tie-breaker.

    Overall

    We loved the tea and snacks at the venue. This was my first networking event as a technical writer and felt that it was quite well organized and I enjoyed the quality of people and conversations we had. I didn’t take any photographs. Will add them if I get them with due credits. 🙂

    Snehal wrote this nice appreciation post on LinkedIn.

  • Managing your Energy

    Photo by Vlad Cheu021ban on Pexels.com

    The above image represents what I think when people say you need to be energetic at work. Since I could not do this, I just gave up.

    That was when I saw a video that YouTuber and productivity savior1, Matt D’Avella. The video is titled, “Why you’re so bad at managing time?” Do not be carried away by the title. It is an energy management video in the guise of a time management video.

    YouTube video version of this article

    Time Management was introduced by factories to make sure that workers worked like clockwork in the factory. Although many people work in factories today, many of us today are knowledge workers.

    As I have said before in this blog, I do knowledge work. So, this is something that I want to try.

    Matt says energy management works better than time management. He gives a few suggestions on how to work better as per your energy.

    Find your peak performance period:

    Matt says that everyone has a peak performance period. A time when they have the most energy. He suggests using a notebook or an app to track your energy daily. One can then identify their peak performance period. This time can change or stay the same.

    Find how much energy your task requires:

    This is also an estimation. Your estimation. How much energy do you need to do something. Split your tasks into high energy and low energy tasks.

    Refill your cup:

    Matt says, “rest is productive”. We should take rest before we get tired and not after we get tired.

    Make flexible to-do lists:

    The to-do list is where this energy management comes into action. Do tasks that take most energy during your high energy period. Do the low energy tasks when you feel low.

    Don’t be hard on yourself when things don’t work out. Remember that you are not perfect.

    Matt reminds us, “We may all have 24 hours in a day but they are not the same 24 hours.” We are asked rhetorically how we compare to achievers like Jeff Bezos and Elon Musk who have the same 24 hours we do. But, the difference is that they probably do not have to do their own laundry. They probably outsource these tasks unlike us.

    1 Matt does a lovey ad for Squarespace where he creates a website for a cult of his own called Productivity Savior. He creates a website for the same using Squarespace called productivitysavior.com.  I appreciate the amount of work he puts into his videos. Whois shows that someone has registered the URL.

    Tumblr version of this post.

  • Knowledge Work

    Ryan Holiday had the above podcast episode with Cal Newport for the Daily Stoic podcast. While both practitioners shared their ideas of how they practiced their crafts, I especially liked the part where they spoke about knowledge work.

    I consider what I do as knowledge work.

    Cal shares that Peter Drucker defined knowledge work in the 1950s. Since then, he says not much has changed. Workers are given a task and are then left alone to do their task. The organisation does not interfere in the performance of this task.

    He suggests that probably knowledge work may improve with the involvement of the organisation. He suggests that NBA and NFL teams in the US do this much better. They know what they want to get out of a player in terms of his personal performance, his team’s performance in the game and in the league.

    On improving our own knowledge work, Cal suggests that we do lifestyle-centric career planning. This involves first understanding the type of lifestyle we would like to lead and then select a career accordingly for best results. For the first two years of one’s career, he suggests not letting anything fall through the cracks. This makes sure that managers stay out of the way and gives the knowledge worker more autonomy.

    Another snippet that I liked in their conversation is how Cal approaches a new idea. He says that knowledge work involves a lot of thinking and the present day has no slack for this very important part of a knowledge worker’s work. He builds the idea by thinking about it, talking to others and writing about it. This progression places more and more burden to be structured and relevant levels of research.

    Tumblr version