markdown on S Anand

Google Meet captions as a local transcript recorder

Sat, 16 May 2026 13:32:06 +0800

I’m a man of simple needs. All I want is: when I’m on Google Meet, I turn on captions. I wanted to click a bookmarklet and save those captions into a local Markdown file. (So that an AI agent can guide me from it.)

Hence, Google Meet Captions. The code is in gmeetcaptions/. Drag the button to your bookmarks bar. Join a Meet. Turn on captions. Click it.

You get a tiny panel with two buttons: Copy and Start Recording.

The bookmarklet writes this kind of Markdown:

# Meeting title

- **Meeting code**: abc-defg-hij
- **Started**: 5/15/2026, 8:00:00 AM
- **Participants**: Alice, Bob, Carol

---

## Alice [0:12]

Good morning everyone.

## Bob [0:18]

Let's get started with the agenda.

That’s it. No server. No extension. No login. No API. Just a bookmarklet page, a script, and local browser APIs.

BUT: Google Meet captions are live and unstable.

A sentence may appear as:

mic, so,

Then a second later become:

mic, So that's a new person. Okay.

Then become:

mic, So that's a new person. Okay. Hey. oh, but,

If I simply append every change, the transcript becomes garbage. So the bookmarklet keeps updating the active speaker turn until it becomes stable. The implementation uses a MutationObserver plus a one-second polling fallback. After four unchanged polls, it treats the turn as final.

The tests are in gmeetcaptions.test.js, using an anonymized fixture at __fixtures__/captions-anonymized.html.

BUT #2: Google Meet’s DOM is not a public API. Class names like .nMcdL, .NWpY1d, and .ygicle can vanish overnight.

So the scraper first tries semantic and structural selectors:

[role="region"][aria-label="Captions"] for the captions region
img[data-iml] and googleusercontent.com avatars to identify caption items
the first <span> as the speaker
the last non-image <div> as the caption text

Only then does it fall back to obfuscated class names. That selector strategy is documented in the README.

Boring, but also the difference between “worked once” and “might work tomorrow.”

The weirdest was Chrome writing to a .md.crswap file while recording. The file appears unfinished until I click Stop Recording. Then Chrome finalizes it.

This is good, actually. It means the browser is safely streaming to a local file via the File System Access API. But it also means: stop the recorder before trusting the file!

I captured these bugs and prompts in prompts.md, because future-me will forget. Future-agent, too.

Why bother? Because transcripts are not the output. They are raw material.

Once a meeting is Markdown, I can ask agents to extract decisions, questions, follow-ups, contradictions, reusable prompts, and blog ideas. I can diff it. Search it. Commit it. Feed it to another workflow.

Meetings now become the “context” in context engineering!

Migrating my blog from WordPress to Hugo

Fri, 02 Jan 2026 12:30:00 +0800

In 2009, I migrated from a self-made Perl static site generator to WordPress because it was slow, WordPress was dynamic and rapidly growing in features, and I wanted to write rather than code. (Also, I had plenty of time in 2009 for such things!)

Over the years, problems crept in. Hosting costs ($200/year) for a slow server. No local writing - Windows Live Writer was dead. I wasn’t using most WordPress features. So it was time to migrate back to a static site generator. (Also, I now have plenty of time for such things!)

I tried in 2024. But the complexity of the migration was higher than my laziness. (I tried with LLMs. Didn’t work.)

Finally, in Dec 2025, coding agents were good enough to get this done. Codex and Claude Code both doubled their limits for the holidays, and I had no meetings. So, over two days, I joyfully migrated to a static site.

Content is written in Markdown via VS Code and pushed to https://github.com/sanand0/blog/
A GitHub Action deploy.yaml publishes to GitHub Pages
It uses a hugo.toml configuration with supporting scripts.

Here’re the useful practices I distilled from my prompts.

Create a `PLAN.md` before complex tasks

Create a plan in PLAN.md to convert my blog into Markdown-based content I can commit on GitHub and is published via a static site generator retaining the same URLs.

My website https://s-anand.net/ hosts a WordPress blog at https://s-anand.net/blog/ and I have exported the content as an XML file using the WordPress export tool into sanand.WordPress.2025-12-28.xml. Read it to understand.

Each post and page must become a Markdown file with front-matter containing all metadata (title, date, tags, categories, slug, author, etc). The Markdown must be stored as

- ./metadata.yml (for site-wide metadata)
- ./posts/yyyy/yyyy-mm-dd-slug.md
- ./pages/slug.md
- ...

You can browse the site https://s-anand.net/blog/ to see the site structure and all the types of pages that are generated.
You can also "ssh sanand" to access the server and `cd www/blog/` to see the WordPress configuration.

Note that some posts / pages have complex content: tables, code, embedded media, shortcodes, JavaScript, etc. Identify all such cases and how to handle them. There may still be edge cases that you don't know how to handle; list them out. This is a major part of the plan - ensuring that you cover all types of content -- either with a plan to handle it easily and elegantly, or explicitly identifying ALL edge cases and listing them.

Include how to handle images and other media. The WordPress export XML file may not contain the actual media files; identify how to get them. Suggest how to store them - I can store them in the Git repo, or in GitHub releases (needs to be fetched during CI process), or in R2 (direct access like a CDN). I would also like to compress media (e.g. to WebP) for better performance.

I plan to use either of these hosting options. Let me know what you prefer:

- GitHub CI to deploy to GitHub Pages (preferred: it's free, easy, and reliable).
- OR, convert locally to static HTML and upload to R2 / CloudFlare pages (priced, more effort, but reliable).

Suggest a static site generator (SSG) to use. I prefer something FAST and simple. I prefer single-binary SSGs or installation-free SSGs (e.g. via npx, uvx) if possible.

Consider any other requirements I may have missed, e.g. plugin features, SEO, redirects, analytics, comments, search, RSS feeds, sitemaps, etc. Going through the site or the WordPress configuration may help identify these. Suggest how to handle them.

It analyzed my blog: blocks and shortcodes used, plugins, URL structure, etc. and came up with a pretty good plan. Here are some notes:

Record unresolved edge cases for manual review. (Wise!)
Prefer raw HTML inside Markdown when conversion would lose fidelity. (Interesting! Also has detailed rules for tables, code, figures, iframes, forms, …)
Frontmatter: Required: title, date, lastmod, slug, author, categories, tags, status, draft, canonical, summary/excerpt, comment_status, ping_status. (Um… too much?)
Use Hugo (single binary, fast, built-in RSS/sitemap/taxonomies, supports raw HTML and shortcodes) and GitHub Pages via GitHub Actions.

There’s no way I would have thought of all of these!

Analyze data while planning

Even while it was planning, another thing struck me. I won’t use GitHub to serve assets. It bloats the repo. Plus, there’s an opportunity to compress better with WebP. So I asked it to analyze my media.

I've rsynced `sanand:www/blog/wp-content/uploads/` to `./uploads/` locally.

Which uploads are unused in my blog posts / pages. Tell me the number, total size, and save them all in unused-uploads.tsv with columns size, filename.

If I convert all used images over 10KB to WebP, how much space will I save? (An estimate is fine; I don't need exact numbers.)

It parsed the WordPress export against my local uploads/, listed unused uploads (1.7K files of 87MB), and found 515 used files of 34MB. It estimated ~11MB savings from WebP conversion of used images >10KB.

Edit `PLAN.md` using the agent, not manually

Editing manually is prone to mistakes, e.g. introducing contradictions.

I am fine with Hugo + GitHub Pages.

/blog must remain the subpath - the URL structure must NOT be disturbed. Prefer publishing into a /blog/ folder.

Import old comments (prefer YAML over JSON for easy reading). The new static site will not have dynamic comments; I may switch to Giscus or Utterances later.

As for media, let's convert and save WebP versions instead of the originals and commit them directly to the Git repo under uploads/. Only used media should be included.

Revise PLAN.md accordingly. Let me know what other information you need, if any.

Useful lesson: Let me know what other information you need, if any. Life is full of unknown unknowns!

It did ask if I was OK changing URLs from .jpg|.png to .webp, which I confirmed in the next prompt.

Implement small steps, run and commit as you go

I know from experience that a single-shot implementation of this size would be too complex today. So I asked for image conversion first.

Now, implement the upload handling. Delete unused uploads (I have backups on the server). Compress JPEGs with 50% quality and PNGs losslessly with 256 colors. Let me know the revised size of assets.

I am OK with image URLs breaking.

Implement this by writing a script (bash, Python, Node JS - anything is fine, whatever is easiest) that will create an `assets/` folder with the converted images from the `uploads/` folder. I will want to re-run this later with some tweaks.

Commit as you go.

Having the agent run the script is the most powerful idea in here. If it makes a mistake, it can figure it out and fix itself.

Commit as you go is useful. I can undo changes later. I also efficiently get a sense of the progress and thinking.

It took a long time (on GPT 5.2 Codex - Extra High Thinking), wrote a scripts/prepare_assets.py, switched from ImageMagick to Pillow for better control (strange!), trouble-shooted PNG transparency bugs, ran it, noted that 41 files in my blog that were missing, and committed everything.

Run post-mortems mid-way

Strangely the JPEG files weren’t converted. So I asked:

Why does the assets/ folder still have some .jpg files, e.g. temperature.jpg? I assumed they'd all be converted to webp.

It said:

Because the script follows your latest instruction: it compresses JPEGs and PNGs in-place and keeps their original formats.

OOPS! My mistake. Anyway, I also needed it to convert MP3 to Opus. So…

Ask for actionable exception reports

Let’s have it do all the conversions.

Convert all JPEG files to WebP, too, with 50% quality. Use cwebp to convert - it is faster and has better quality.

Let's also compress audio files to OPUS with ffmpeg using `-c:a libopus -b:a 12k -ac 1 -application voip -vbr on -compression_level 10`. Modify the script accordingly.

Also modify the script to list missing local uploads/ mentioning the URL on https://s-anand.net/blog/... post or page that references each. Each line should list the missing filename and URL, tab-separated.

But note the last line. It had already identified that some uploads were missing. To action that, I would need a post -> file mapping, which I asked for.

Delegate verification to the agent

In the exception report, there was an amazoncooliri file missing. I couldn’t find it in the blog XML. So I asked:

Double-check the files mentioned in the missing-uploads.tsv. Do these files really exist in the XML? For example, it mentions "amazoncooliri". But there is no such word in the XML, I think, and uploads/amazoncooliris.jpg (which is probably what it refers to) exists. There may be similar mistakes. Make sure nothing required is missing either.

At first, I checked the code to see why amazoncooliris became amazoncooliri. But months of practicing ~~laziness~~ delegation took over and I had it debug itself.

Turned out it wrote \\s instead of \s which fixed it.

Accept that you will make mistakes

Why are there PNG images in assets? I assumed there wouldn't be any and all would be WebP files...

It replied:

Because the script only converts JPEG → WebP. For PNGs, it follows your instruction to “compress PNGs losslessly with 256 colors,” so they stay PNG.

OOPS AGAIN! I forgot to ask it to convert PNGs.

Convert all PNGs to WebP using cwebp maximal effort with 256 color palette, just like the jpegs. Update, re-run, commit.

Verification reports speed up reviews

Now for the big one: converting the blog posts.

Now implement the plan in PLAN.md to convert the WordPress XML export to Markdown files with front-matter. Write a Python script with inline dependencies that `uv run` can execute. Run it and generate all the Markdown files. Verify that everything looks good - especially edge cases.

Make a list of edge cases you could handle that I should verify.
Make a list of any edge cases you couldn't handle.
Share these as links to https://s-anand.net/blog/.... as well as the local relative paths.

Apart from converting the posts, it generated:

reports/edge-cases-unhandled.tsv – an exception report
reports/edge-cases-handled.tsv – a verification report

The verification report was handy. I could see the edge cases (e.g. upload URLs rewritten, iframes/tables/objects/scripts that were retained, WordPress comment blocks removed) and spot check quickly.

Without this, I would have spent a lot more time reviewing. This gave me confidence that it had handled edge cases well.

Generate easy-to-review to review content

I find it productive to have the agent generate content that is easy to review. At this point, I had a bunch of Markdown files that were very easy for me to scan. So I created an extensive list of changes:

Drop redundant or unchanging frontmatter. For example:

- Retain title, date, lastmod, slug, categories
- author: sanand is always the same. Drop
- draft: false and status: publish are redundant. Drop both
- url: is deriable from the slug. Drop. Same for wp_link
- If tags, excerpt, aliases, etc. are missing, drop them
- Drop meta: entirely
- Drop menu_order, ping_status, comment_status
- Drop wip_guid in favor of wp_id (retain wip_id)
- If I missed any other frontmatter, use the same principles.

Insert the featured image as the first element in the post, instead of the featured_image frontmatter.

Rewrite all the upload links (http://www.s-anand.net/blog/wp-content/uploads/xxx) to relative links to uploads/ (e.g. ../../uploads/xxx).

Rerun. Commit as you go.

Finally, write a script to detect all links to s-anand.net/ that are not covered by this approach. For example, I have a bunch of direct assets like https://files.s-anand.net/blog/a/mystic-light.mp3 or other non-WordPress pages on my website. Create a TSV report with the link and the source.

PS: I keep editing prompt.md with my prompts. Keep ignoring it.

The last line was because I was also saving my prompts and it kept getting confused why a file it didn’t create kept changing 🙂.

Give it all your tools

I had it install Hugo and run it. (I could have set it up myself, but why bother?)

I allowed it to ssh to my server to check logs if needed.

In both cases, these are tools I would need to build and test. It makes sense to let the agent use them directly.

Corrections:

- The relative URLs should be ../../assets/ not ../../uploads/ (my mistake)
- No action required on the file URLs and swf links. I'll handle those.

Now, use Hugo (install via `mise use -g hugo`) and generate a static site from the Markdown files.

Ensure that ALL the URLs from https://s-anand.net/blog/... are retained exactly, including URLs posts, pages, categories, tags, author pages, year (or other time period), etc. You can `ssh sanand` and scan logs if that'll help.

This will be deployed via GitHub pages. Create the GitHub action workflow to build and deploy the site on every push to main.

Commit as you go.

Note that I had made a mistake mentioning uploads/ instead of assets/. This keeps happening.

I don’t know any Hugo, so this was a bold step. But the output would be easy to review (it’s a static site), so more reviewability = more confidence.

After a long time, it generated the static site. It managed to self-correct a bunch of stuff. For example:

I removed --minify because inline scripts with <br /> in posts/2011/2011-05-19-eating-more-for-less.md break JS minification; HTML builds cleanly without minify.

Prefer UI reviews over code reviews for ease

In fact, if required, have it build a throw-away tool to help you review.

In my case, the output was functional but ugly. So ugly that I couldn’t review it properly. I suggested a few obvious fixes (like broken links) but the main ask was to pick a theme and make it look like my website.

FYI: I removed mise.toml. Running `mise x hugo -- hugo` should still work, but feel free to reinstall.

Ensure that ALL LINKS are relative. For example, public/blog/index.html links to https://s-anand.net/blog/tamil-ai/ but I'd like to link to tamil-ai/ instead.

Pick and implement a nice, popular, lightweight theme that's suitable for blogs.

Include these features. Where possible, use modern, well-supported & popular plugins or themes rather than custom code.

- Below the title, have a single line showing the date created (date updated shown only if different from date created), categories with links to category pages, tags with links to tag pages (if any) are present.
- Code blocks should be syntax-highlighted.
- At the bottom, include a link to the next and the previous posts (with title).
- Add a footer to all pages that lists
  - All categories with links & post count
  - All year archives with links and post count
  - All pages with links

Run and test.

You can resume if it hangs. Don’t worry about context

This took forever and I think Codex crashed or hung or something. So I killed it, resumed, and asked:

It's been a while... maybe you were stuck? Resume and complete.

It managed to resume. I’m not sure if there was some context loss or confusion, but I’m learning to worry less.

Allow it design flexibility

Internal links were still broken. I didn’t know why, nor enough to fix them. So, rather than make a design decision (e.g. always use full / absolute / relative URLs), I let it decide.

There are several problems due to wrong relative paths. If it will be easier, feel free to switch back to absolute paths for /blog/ to fix them.

- Featured image URLs seem wrong. blog/ai-can-be-held-to-account/ links to "/blog/../assets/pig-court.webp" instead of just "../assets/pig-court.webp"
- When I visit /blog/2003/ the CSS is fine but it breaks in /blog/2003/page/2/ -- and the relative links from that page also break.
- The links from the footers in /blog/2003/ point to /2016/ instead of /blog/2016/ for example

Swap "Next" and "Prev". "Prev" indicates older posts.

Syntax highlighting of code blocks doesn't seem to be working. In blog/openai-tts-cost/ I see a single block like this.

```
<code lang="bash" class="language-bash">curl "https://api.openai.com/v1/organization/usage/audio_speeches?start_time=$(date -d '1 day ago' +%s)&amp;project_ids=$PROJECT_ID&amp;group_by=model" \
  -H "Authorization: Bearer $OPENAI_ADMIN_KEY" \
  -H "Content-Type: application/json"</code>
```

It decided to use absolute paths.

Generate documentation

Apart from adding a few more features / fixes, I had it generate a README.md documenting what I would need to run this in the future.

Add featured images as thumbnails to post listings (e.g. blog home, category pages, tag pages, archive pages). Ensure that there is a placeholder (an elegant blank image) if there is no featured image.

Left-align the footer links.

Now, clean-up and add a README.md that explains the structure of the repo, how to build and deploy, etc.

I also had it fix another design error I made. The placeholders I asked for didn’t look good.

Skip featured images in post listings, post pages, etc. if they don't exist. No need for a placeholder.

Ask for effort estimates

At this point, things were fine. But I was curious if we could refactor a bit.

I’m concerned if it might mess it up, though. Effort is a good proxy for errors. So I asked it:

How easy are these changes? Just tell me, don't implement them.

- Set it up so that the content/ directory is auto-generated from posts/ without needing to be committed.
- Drop the slug: frontmatter and instead, derive it from the filename.
- Embed the comments in the posts as Markdown, clearly distinguished from the content, rather than keep them in comments/

It’s response:

Auto-generate content/ from posts/ is moderate effort…
Dropping slug frontmatter is easy if you accept Hugo’s filename‑based slug rules…
Embedding comments into each post is moderate effort…

Probe for confidence

I asked:

What's Hugo's filename-based slug rules?

I didn’t read the output. My aim was not to learn. It was more to glance at it, see if (based on my considerable experience in this area) if it looked reasonable. In short, I was probing for confidence – its and mine.

Continue the session for the long tail

Normally, I would have created a new session to implement changes and fixes.

But the session was auto-compacting quite well. So rather than lose context, I had it create a build steps and run several minor fixes over the next few days. I didn’t need to specify the context again and again.

Summary

Here’re the lessons I distilled from this migration, tagged by whether it’s new to me.

#	Lesson	New?
1	Create a `PLAN.md` before complex tasks
2	Analyze data while planning	New
3	Edit `PLAN.md` using the agent, not manually	New
4	Implement small steps, run and commit as you go
5	Run post-mortems mid-way	New
6	Ask for actionable exception reports	New
7	Delegate verification to the agent
8	Accept that you will make mistakes
9	Verification reports speed up reviews	New
10	Generate easy-to-review to review content
11	Give it all your tools
12	Prefer UI reviews over code reviews for ease
13	You can resume if it hangs. Don’t worry about context	New
14	Allow it design flexibility	New
15	Generate documentation
16	Ask for effort estimates	New
17	Probe for confidence	New
18	Continue the session for the long tail

Things I Learned - 10 Aug 2025

Sun, 10 Aug 2025 00:00:00 +0000

This week, I learned:

OpenAI supports a tool "type": "custom" that lets it write code as an argument to a tool call. Great for code / SQL generation. Even more powerfully, you can generate output following specific grammars, e.g. STL files, PostgreSQL dialect, Mermaid/PlantUML diagrams, OpenAPI specs, Vega-Lite JSONs, Cron expressions, GraphQL SDLs, Dockerfiles, Terraform HCLs, or any DSL! # #ai-coding
The OpenAI playground has a GPT-5 Prompt Optimizer that can migrate prompts to GPT-5.
Docsify 4.13.1 is 2 years old and uses marked@1.2.9 which is 5 years old. Newer plugins like marked-directive don’t work with it. Though docsify v5.0.0-rc1 is in development, it may be the better option for modern Markdown plugins. Here’s sample code.
CommonMark has a powerful directive syntax proposal that lets you add classes, attributes, and arbitrary plugins to Markdown. For example, :abbr[MD]{#id .class title="Markdown"} for inline directives. Plugins exist for marked, markdown-it and remark.
biomejs and dprint are gaining traction as prettier alternatives. I’m yet to try them but keen to explore.
- Skip biomejs for now. It uses tabs (not spaces) and does not respect .gitignore by default. Handling these is too much work.
⭐ Code generation is more flexible than tool calling. LLMs can’t write a tool-call loop, for example, but they can write code to run an API in a loop. So, I like telling the LLM to “write code using these APIs” than giving it APIs to tool-call. #ai-coding
npx -y ccusage is an easy way of summarizing your Claude Code usage and cost. My cost so far (since 21 July) is about $10. The median session cost is ~50 cents. Most of it ($7) was from a single temporary coding chat that I kept continuing for way too long, building up the context window. #
defuddle can be used in the browser to get the main content from web pages. A replacement for Mozilla Readability. #
Modern Node.js Patterns for 2025 include these 5 features I’m excited by:
- Single-executable bundling. node --experimental-sea-config sea-config.json builds standalone binaries.
- ES Modules. Use node: prefix for built-in imports. import { createServer } from 'node:http';
- Watch mode. Use node --watch file.js auto-reloads when file.js or dependencies change.
- Env file. Use node --env-file=.env loads .env as environment variables.
- node:test is a full-featured test framework with --watch and coverage.
Concise explanations speed up decisions because they’re faster to read and understand (obvious). They’re also easier to combine with other ideas (less obvious). #
I’ve been uncertain about htmx for some time now. This tutorial, HTMX is hard, so let’s get it right, convinced me that it’s too far from my mental model, so I’m unlikely to ever use it.
⭐ Slow, effortful practice (spaced recall, interleaving topics, self-testing) builds lasting knowledge but looks inefficient and doesn’t help with exams. # #beliefs
GitDoc VS Code extension auto-commits and syncs notes. I dropped gitwatch in favor of this.
It’s interesting that Gemini Deep Research cannot access Google Drive while Gemini can. On the other hand, ChatGPT Deep Research can access Google Drive but ChatGPT cannot.
A trend that AI coding will only accelerate: “It is now possible for tiny teams to make principled software that millions of people use, unburdened by investors. … you need far less money and far fewer employees to reach far more customers. That wave is only just beginning.” # #ai-coding
Typed languages are better suited for vibe coding. This will likely lead to the growth of typed languages (TypeScript, Rust, Go) but also of typing in untyped languages (e.g. Python) # #ai-coding
Instead of Celery, Redis, Kafka, etc. as task queues, we could the file system as a message queue. For example, pending/task-01.json moves to wip/task-01.json to done/task-01.json. Folders for state/tags, files for task details.
Foam is a note-taking VS Code extension. The WikiLinks, tags and backlinking features align naturally with Markdown note-taking. Via Steph Ango who uses Obsidian which nudged me to search for WikiLink-ing features in VS Code.
I’m an open data hawk. But here are things I should remind myself of. #
- Privacy incubates creativity. People self-censor when watched. Privacy shields fragile ideas.
- Power assymetry. Big players can leverage openness more, e.g. Cambridge Analytics + Facebook data.
- Context matters. What’s harmless in one setting can be toxic in another.
- One-way door. Data can’t be unshared. Don’t scrap brakes dreaming of perfect roads. Anticipate tyrannical regimes / cultures.
- Not your call. You don’t share your neighbour’s medical records.
One Punch Man is available as manga. I watched the anime first and assumed that came first. Apparently not.
⭐ In “kind” environments (stable rules, rapid and accurate feedback), specialize. In “wicked” environments (rules shift, feedback is noisy/late), generalize. ChatGPT
Models’ ability to orchestrate longer workflows will improve. Factor that into your application design. Claude Code can already handle over 70 tasks in a workflow
What happens when LLMs play Chinese Whispers / the Telephone Game? Here are learnings. ChatGPT
- Drift increases faster than linear with hops.
- Bigger models do better, but constrained prompts (“Copy the text exactly; change nothing.”) have a bigger impact.
- Low temperature improves copying fidelity.
- But even after “forgetting”, LLMs reproduce rare content if they’re trained on it.
“In fact, React Native looks set to become the most engine-agnostic JavaScript runtime around”. The Many, Many, Many, JavaScript Runtimes of the Last Decade
OMDb (simple) and TMDb (comprehensive) are API-friendly alternatives to the IMDb.
copyparty seems one of the most feature-rich file servers out there. Single Python file, runs on any OS, works with any client, and optimized for speed. Video
Quotes I enjoyed from Linus Torvalds’ TED interview
- I want to not have external stimulation. You can kind of see, on the walls are this light green. I’m told that at mental institutions they use that on the walls. It’s like a calming color. … the main thing I worry about in my computer is – it really has to be completely silent. If the cat comes up, it sits in my lap. And I want to hear the cat purring.
- I did not start Linux as a collaborative project. I started it as one in a series of many projects I had done at the time for myself, partly because I needed the end result, but even more because I just enjoyed programming.
- I’m actually not a people person. But I do love other people who comment and get involved in my project.
- The big point for me was not being alone and having 10, maybe 100 people being involved. Going from 100 people to a million people is not a big deal – to me. Well, I mean, maybe it is if you want to sell your result then it’s a huge deal. But if you’re interested in the technology and you’re interested in the project, the big part was getting the community.
- So Git is my second big project, which was only created for me to maintain my first big project. And this is literally how I work.
- Well, I do code for fun – but I want to code for something meaningful so every single project I’ve ever done has been something I needed.
- Apparently, my sister said that my biggest exceptional quality was that I would not let go.
- I can’t do UI to save my life.
- Good taste is about really seeing the big patterns and kind of instinctively knowing what’s the right way to do things.
- Companies like Google and many others have made, arguably, like, billions of dollars out of your software. Does that piss you off? No. No, it doesn’t piss me off for several reasons. And one of them is, I’m doing fine. But the other reason is – I mean, without doing the whole open source and really letting go thing, Linux would never have been what it is.
- I think one reason open source works so well in code (is that …) Code either works or it doesn’t.
The Uses This site has interviewed professionals for decades. From their repo I scraped the top developer apps post 2020:
CloudFlare has an Iceberg data catalog in R2 Data Catalog. Iceberg is like Parquet but supports metadata, time-travel, and schema edits. But I’m yet to find a single publicly accessible Iceberg catalog. Its open-data adoption is not as high as Parquet’s. Apache Iceberg vs Parquet
Observable Notebook 2 is the new notebook format from Mike Bostock. It is vanilla JS and embeddable into other pages. THis would have been a big deal 2 years ago, but with the LLM ecosystem today, I’m not sure if it matters as much.
To add CORS support to CloudFlare pages protected by Zero Trust, add a _headers file to your repo. (This is different from the Zero Trust CORS which allows automated logins.) Sample _headers that lets logged-in users fetch pages via fetch("...", { credentials: "include" }):
```
/*
  Access-Control-Allow-Credentials: true
  Access-Control-Allow-Origin: https://your-site.example.com
  Access-Control-Allow-Methods: GET, HEAD
  Access-Control-Allow-Methods: *
```
As corporates restrict the use of LLMs, I see employees purchasing personal laptops to use LLMs on. An interesting trend!
openai-python has a CLI. You can run uvx openai api chat.completions.create --stream -m gpt-4.1-nano -g developer 'Translate to Chinese' -g user "Hello" for example
Anthropic has an OpenAI compatible API at https://api.anthropic.com/v1/.
Claude Code tips from Things that didn’t work by Armin Rocher #ai-coding
- Speech-to-text. Cannot stress this enough but talking to the machine means you’re more likely to share more about what you want it to do.
- I maintain some basic prompts and context for copy-pasting at the end or the beginning of what I entered.
- I ended up preloading executables on the PATH that override the default ones, steering Claude toward the right tools, e.g. running python asks it to use uv.
- I use the task tool frequently for basic parallelization and context isolation.
- Simply taking time to talk to the machine and give clear instructions outperforms elaborate pre-written prompts.
- Forcing myself to evaluate the automation has another benefit: I’m less likely to just blindly assume it helps me.
Research indicates that we don’t know in advance which prompts will help. Evals beat prompt engineering. Ethan Mollick

How to publish an eBook in 60 minutes

Mon, 31 Mar 2025 12:47:13 +0000

I published an eBook on Amazon. It takes an hour if you have the content ready.

STEP 1 (10 min): Set up a Kindle Direct Publishing account with your address, bank details, and tax info.

STEP 2 (15 min): Export my London 2000 blog archive and convert to Markdown.

STEP 3 (10 min): Reformat the Markdown by writing a script in Cursor. Here’s the prompt:

Write a Python script that reads *.md including the YAML frontmatter, adds the YAML title as H1, date (yyyy-mm-dd) like Sun, 01 Jan 2000 in a new para after the frontmatter and before the content.

STEP 4 (15 min): Convert it to an ePub using pandoc.

pandoc *.md -o book.epub --toc \
  --metadata title="An LBS Exchange Program" \
  --metadata author="Anand S" \
  --metadata language=en \
  --metadata date="31 Mar 2025"

STEP 5 (10 min): Generated a cover page with ChatGPT (5 min) and compressed it into JPEG via Squoosh.

Draw a comic-style book cover page that covers the experiences of an Indian exchange student (picture attached) from IIM Bangalore at London Business School and exploring London. The book title is “An LBS Exchange Program”.

STEP 6 (10 min): Publish the book on KDP. It’s priced at $0.99 / ₹49 because Kindle doesn’t allow free downloads.

That’s it. Here’s the book: https://www.amazon.in/dp/B0F3D55R2Z/

The three things that made publishing in 1 hour possible are:

Amazon’s publishing process is simple.
Open-source tooling (WordPress, Markdown, ePub, pandoc) has built a big part of the infrastructure.
LLMs make the rest (figuring out the steps, generating the cover) very easy.

(An eBook takes 72 hours of review before going live on the Kindle store.)

Voice Chat to Slides: My New AI-Powered Workflow

Thu, 27 Mar 2025 16:34:53 +0000

Here’s my new workflow for creating slide decks:

ChatGPT interviews me and creates Markdown slides.
I use Marp to convert Markdown to slides.
LLMs create supporting images.
I deploy on GitHub Pages.

… and here are 2 decks created this way.

Let’s look at how I built the second example, step by step.

ChatGPT interviews me and creates Markdown slides

While walking 75 minutes from home to IIT Madras to deliver this talk, I had ChatGPT interview me in standard voice mode.

Why an interview? It’s easier when someone asks questions.
Why voice? It’s hard to type while walking. Otherwise, I prefer typing.
Why not advanced voice mode? I want to use a reasoning model like O3 Mini High for better responses, not the GPT-4o-realtime model that advanced voice mode uses.

Here’s the conversation I had with ChatGPT. I began by speaking (not typing):

I want to create an insightful deck in Markdown on how I have been using LLMs in education.The audience will be technologists and educators. The slide contents must have information that is useful and surprising to them.

The slides are formatted in Markdown with each slide title being a level 2 Markdown header and the contents of the slides being crisp bullet points that support the title. The titles are McKinsey style action titles. Just by reading the titles, the audience will understand the message that I am trying to convey.

In this conversation, I’d like you to interview me, asking me questions one by one, and taking my inputs to craft this presentation. I’d also like you to review the inputs and the slide content you create to make sure that it is insightful, useful, non-obvious, and very clear and simple for the audience. Interact with me to improve the deck.

Let’s begin.

(Unsurprisingly, I talk a lot more than I type.)

There were 3 kinds of interactions I had with ChatGPT:

Content. I explained each slide. For example:

“Yeah, let’s move on to the next topic, which is where we had the students learn prompt engineering as part of the course. One of the questions was convincing an LLM to say yes, even though …”

Correction. After ChatGPT read aloud a slide, I corrected it. For example:

“Content-wise, it’s spot-on. Style-wise, it’s almost spot-on. It’s far more verbose. Can you retain the exact same style, but shorten the number of words considerably?”
“These feel generic. I’d like stuff that comes across as insightful, non-obvious, and specific.”

Collation. I had ChatGPT put slides in order. For example:

“Put all the slides together in sequence. Make sure you don’t miss anything.”
“Move the opening questions as the second slide. Move the final takeaways, which is currently the last slide, to just before the final set of questions.”

At the end of the interview, I had all the content for the slides.

Marp converts Markdown to slides

I use Marp, a JavaScript tool that turns Markdown into slides

Why Markdown? It’s natural for programmers and LLMs. ChatGPT renders rich text in Markdown.
Why not RevealJS? Despite a Markdown plugin, RevealJS is built for HTML. Marp is built for Markdown.

I created a bookmarklet that copies text as Markdown. Using this, I converted the ChatGPT slide transcript to Markdown, saving it as README.md.

The Marp for VS Code plugin makes it easy to preview the slides when you adding YAML frontmatter like this:

---
marp: true
title: Visualizing LLM Hallucinations
url: https://sanand0.github.io/llmhallucinations/
theme: gaia
backgroundColor: white
---

LLMs create supporting images

I use ChatGPT or Gemini to create images that support the slides. For example this slide includes an image of a robot psychologist generated by Gemini’s ImageGen 3:

Today, with native image generation in Gemini 2.0 Flash and GPT 4o, I’d likely use those. They have much better character control.

Deploying on GitHub Pages

I use GitHub Actions to render the slides and deploy them on GitHub Pages. Here’s what the key steps look like:

steps:
  - name: Checkout repository
    uses: actions/checkout@v4

  - name: Setup Node.js
    uses: actions/setup-node@v4
    with:
      node-version: "22"

  - name: Generate HTML with Marp
    run: npx -y @marp-team/marp-cli@latest README.md -o index.html

  - name: Setup GitHub Pages
    uses: actions/configure-pages@v4

  - name: Upload artifact
    uses: actions/upload-pages-artifact@v3
    with:
      path: "."

  - name: Deploy to GitHub Pages
    id: deployment
    uses: actions/deploy-pages@v4

What’s next?

I plan to make 3 improvements to this workflow:

Adding images directly with voice prompts.
Adding diagrams (e.g., Mermaid).
Creating a custom GPT that auto-deploys slides on GitHub when I say “Publish these slides.”

If both of these work, I’ll be able to create and publish an entire slide deck just by rambling to ChatGPT for an hour.

A challenge of blog questions

Tue, 04 Mar 2025 11:13:55 +0000

Thejesh tagged me with these questions.

Why did you start blogging in the first place?

I started my website in 1997 on Geocities at https://www.geocities.com/root_node/, mostly talking about me. (A cousin once told me, “Anand’s site is like TN Seshan - talking only about himself.” 🙂)

(As an aside, I didn’t know that searching for Geocities on Google renders the results in Comic Sans!)

I wanted a place to share the interesting links I found. Robot Wisdom by John Barger and Scripting News by Dave Winer were great examples: collection of interesting links updated daily.

In July 1999, as a student at IIMB, I decided to put that into action by creating a custom HTML page updated manually.

What platform are you using to manage your blog and why did you choose it? Have you blogged on other platforms before?

WordPress. Because it was the most fully-featured, mature platform when I migrated to it around 2006.

Before that, I used:

A custom HTML page on Geocities. But it was hard to update multiple links, create individual pages, categories, etc. So…
A Perl-based static-site generator I wrote myself. But as my link count grew, each generation took too long. So …
A CGI-Perl-based blogging engine I hosed on freestarthost.com, handling commenting, etc. But at BCG, I didn’t have time to add many features (linkback, RSS, etc.) So…
Blogger as a parallel blog, briefly. But it didn’t have as many features as (nor the portability of) WordPress. So…
WordPress - moving across a bunch of hosting services, and currently on HostGator.
I also blogged in parallel on InfyBlog, Infosys’ internal blogging platform on LiveJournal.

How do you write your posts? For example, in a local editing tool, or in a panel/dashboard that’s part of your blog?

I started with custom HTML in Emacs (or whatever code editor I kept moving to).

Briefly, I used Windows Live Writer, which was quite a good blogging tool.

Now, I write in Markdown on VS Code and paste it into WordPress’ editor.

When do you feel most inspired to write?

When an idea that’s been bubbling for a while in my mind bursts out.

Do you publish immediately after writing, or do you let it simmer a bit as a draft?

I publish immediately.

What’s your favorite post on your blog?

The next chapter of my life, which I wrote on a one-way flight back from the UK to India to start Gramener.

Any future plans for your blog? Maybe a redesign, a move to another platform, or adding a new feature?

I plan to move it to GitHub Pages with Markdown content and a static site generator. I might write my own SSG again in Deno or use one of the faster ones.

Who’s next?

I’d love to hear from

My learnings as week notes

Wed, 25 Dec 2024 14:27:11 +0000

One of my goals for 2024 is to “Compound long-term goals, daily.” Learning is one of those.

Some people publish their learnings as weekly notes, like Simon Willison, Thejesh GN, Anil Radhakrishna, and Julia Evans. I follow their notes.

I started doing the same, quietly, to see if I could sustain it. It’s been a year and it has sustained.

I’m finally publishing them. My week notes are at til.s-anand.net. Here’s the source code.

Capturing learnings must be frictionless

I learn things when I’m reading, listening to podcasts, listening to people, or thinking. In every case I’m close to my phone or laptop.

If my laptop is open, I add my notes to a few (long) Markdown files like this til.md.

If my phone is easier to access, I type or dictate my notes into Microsoft To Do, which is currently my most convenient note-taking app. It syncs with my laptop. I transfer it (via OCR on Microsoft Power Toys) into the Markdown file.

The Markdown files are synced across my devices using Dropbox, which I find the most convenient and fast way to sync.

The notes have a simple format. Here’s something I quickly wrote down in Microsoft To Do while speaking with a senior at a restaurant:

Government websites like the official press releases cannot be crawled from outside India. Hence the need for server farms in India!

Then I copied that over to the Markdown file as a list item along with the date (which Microsoft To Do captures), like this:

- 15 Dec 2024. Government websites like the official press releases cannot be crawled from outside India. Hence the need for server farms in India!

That’s it. Quick and simple. The most important thing is to capture learnings easily. Even the slightest friction hurts this goal.

Publishing learnings

I run this Deno script which parses the Markdown files, groups them by week, and generates a set of static HTML pages. These are published on GitHub Pages, which is currently my favorite way to publish static files.

It generates an RSS feed as well. I’ve started reading more content using RSS feeds via Feedly, including my own notes. I find browsing through them a useful refresher.

This format is different from my blog. In the 1990s and early 2000s, I published individual links as posts. Then I moved to long form posts. This consolidates multiple links into a single weekly post. But rather than publish via WordPress (which is what my blog is currently based on), I prefer a Markdown-based static site. So it’s separate for the moment.

I intend to continue with these notes (and the format) for the foreseeable future.

Comments

My Year in 2024 - S Anand 30 Dec 2024 10:01 pm (pingback): […] Learning. I took weekly notes of things I learned. […]

How to extend Markdown with custom blocks

Fri, 08 Jan 2021 09:37:48 +0000

One problem I’ve had in Markdown is rendering a content in columns.

On Bootstrap, the markup would look like this:

<div class="row">
  <div class="col">...</div>
  <div class="col">...</div>
</div>

How do we get that into Markdown without writing HTML?

On Python, the attribute lists extension lets you add a class. For example:

This is some content
{: .row}

… renders <p class="row">This is some content</p>.

But I can’t do that to multiple paragraphs. Nor can I next content, i.e. add a .col inside the .row.

Enter markdown-customblocks. It’s a Python module that extends Python Markdown. This lets me write:

::: row
::: col
Content in column 1
::: col
Content in column 2
::: row
::: col
Content in column 1
::: col
Content in column 2

This translates to:

<div class="row">
  <div class="col">Content in column 1</div>
  <div class="col">Content in column 2</div>
</div>

Better yet, we can create our own custom HTML block types. For example, this code:

from markdown import Markdown
from customblocks import CustomBlocksExtension

def audio(ctx, src, type="audio/mp3"):
    return f'''<audio src="{src}" type="{type}">
      {ctx.content}
    </audio>'''

md = Markdown(extensions=[
    CustomBlocksExtension(generators={
        'audio': audio
    }),
])

… lets you convert this piece of Markdown:

md.convert("""
::: audio src="mymusic.ogg" type="audio/ogg"
    Your browser does not support `audio`.
""")

… into this HTML:

<audio src="mymusic.ogg" type="audio/ogg">
  Your browser does not support <code>audio</code>.
</audio>

markdown-customblocks is easily the most useful Python module I discovered last quarter.

Micro-notes

Sun, 29 Nov 2020 03:46:33 +0000

I maintain my (extensive) notes in text files. I’ve explored Evernote, Onenote, Google Keep, Apple Notes, and many other platforms. But text files work. I store them as Markdown and sync them on DropBox.

They used to be relatively large files (50-100KB) each, on broad topics. For example:

todo.txt was a consolidated list of things I had to do
people.txt was a list of everything I knew about people (addresses, birthdays, etc)
towrite.txt was a list of everything I wanted to write about
notes.txt was where I tracked notes about any topics
… and more

This led to a couple of problems.

Searching across files was hard. I wouldn’t remember if I wrote ideas for my next talk in todo.txt or towrite.txt, or if my meeting minutes where in notes.txt todo.txt. I had to open each file and search.
Files were getting too big. Editing them on mobile was harder. Scanning them was harder.

So I changed this system a few years ago into micro-notes. These files became a folder. For example, my notes/ folder looks like this:

time-management.txt has my time management notes
book-never-split-the-difference.txt has book notes on Never Split the Difference
eat-food-sleep-exercise-live-healthy.txt has notes on fitness

The folder has nearly 300 files. Here’s a glimpse of the latest files.

Similarly, my people/ folder has details of my discussions with various people I interact with – friends, colleagues & clients.

What made this change possible is Everything, a fast file search tool on Windows that lets me find files as I type. For example, if I’m looking for my notes on SlideSense, I just type “notes s” and it appears on the list.

I usually sort the files by run count (how often I opened them). That makes it easy to re-open the most used files.

It also makes it easier to edit these notes on mobile. I sync the folder on Dropbox, and use IAWriter to edit them while on walks. Dictation is pretty good, so I’ve been using that to take notes too.

Comments

Saurabh Mittal 3 Dec 2020 1:20 pm: Hey Anand, great to see you back! After a loonngggg hiatus!
S Anand 4 Dec 2020 10:37 am: Thanks, Saurabh!
Ananth 26 Dec 2020 2:32 am: Welcome back Anand. I have been following your “blog” (hope the term is still relevant!) for more than a decade. Your matter-of-fact writing is something that I enjoy and your experiences using various tools/approaches have always been interesting. Hope we can see more of your writing.
Ramamurthy 27 Dec 2020 7:03 pm: Hi Anand, I stumbled into your page accidentally while looking for something about Ilayaraja.I am your chittappa,Lalitha chitti’s husband.We attended your wedding too. Much water has flown under the bridge and I have lost contact with our relatives.Ramnath and Shekar are in touch.How is your Sriram uncle and your dad.I hope you remember me!

Markdress

Sat, 03 Dec 2011 18:05:26 +0000

This year, I’ve converted the bulk of my content into Markdown – a simple way of formatting text files in a way that can be rendered into HTML.

Not out of choice, really. It was the only solution if I wanted to:

Edit files on my iPad / iPhone (I’ve started doing that a lot more recently)
Allow the contents to be viewable as HTML as well as text, and
Allow non techies to edit the file

As a bonus, it’s already the format Github and Bitbucket use for markup.

If you toss Dropbox into the mix, there’s a powerful solution there. You can share files via Dropbox as Markdown, and publish them as web pages. There are already a number of solutions that let you do this. DropPages.com and Pancake.io let you share Dropbox files as web pages. Calepin.co lets you blog using Dropbox.

My needs were a bit simpler, however. I sometimes publish Markdown files on Dropbox that I want to see in a formatted way – without having to create an account. Just to test things, or share temporarily.

Enter Markdress.org. My project for this morning.

Just add any URL after markdress.org to render it as Markdown. For example, to render the file at http://goo.gl/zTG1q, visit http://markdress.org/goo.gl/zTG1q.

To test it out, create any text file in your Dropbox public folder, get the public link and append it to http://markdress.org/ without the http:// prefix.

Comments

Gee 14 Jan 2012 12:38 am: Great news. And all the best in your new ventures. Meanwhile, I made a simple bookmarklet for you: https://gist.github.com/1609547
Thejesh GN 5 Dec 2011 6:47 am: Check this http://markdoc.org I have been using it locally for sometime. Its cool.
S Anand 4 Dec 2011 12:18 pm: James, your suggestion on following URLs is spot-on! I’ll put this in (one of these days, when I get back to this…)
S Anand 6 Dec 2011 10:10 pm: You need to have an empty line before a list. For example:
Test
This is a list: * one * two * three
Michael Lajlev 5 Dec 2011 2:12 pm: I seem to struggle getting list items to work http://markdress.org/dl.dropbox.com/u/623576/test.md Any idea?
Gee 12 Jan 2012 9:14 pm: Markdress is great. I have it running on my server. I would like to render files in a directory adjacent to, or inside the app folder, instead of entering a URL path each time. I would like to enter just filename.txt and it would default back to (effectively) /markdress/filename.txt I looked over index.php and tried a few things but I can’t get relative files to render. Is this an enhancement you’ve considered? Thanks for the great tool.
S Anand 13 Jan 2012 12:51 am: I personally use this sort of a variant, but havent gotten around to making the same changes on markdress. I will do that over the next few days, once I settle down.
The Bootcamp Blog Begins | JFDI–Innov8 2012 Bootcamp 3 Dec 2011 8:16 pm (pingback): […] kind of people who can whip up something like Markdress in a […]
James Manning 4 Dec 2011 9:08 am: Awesome! Nicely done! Not sure if it’s by design or not, but FWIW the example given seems to show that relative url’s aren’t handled correctly - the bottom of this page: http://daringfireball.net/projects/markdown/index.text has this relative url: [tfmenu]: /graphics/markdown/mt_textformat_menu.png which gets faithfully rendered into: That fails, of course, with the attempted get operation returning: http://markdress.org/graphics/markdown/mt_textformat_menu.png Could not fetch http://graphics/markdown/mt_textformat_menu.png FWIW, inserting a base tag in the head seems to work fine, like: However, it would have to be the ‘real’ url in the base - IOW, this doesn’t work: On a related note, IMHO it seems like markdress should handle redirects and ‘pass through’ as a redirect to the markdress version. IOW, looking at this example request from the post: http://markdress.org/goo.gl/zTG1q When markdress requested the page and got back a redirect, instead of following it, it should have instead passed back a redirect to: http://markdress.org/daringfireball.net/projects/markdown/index.text IMHO, that’s a better user experience as it’s more clear what’s getting served up (and it would reflect in the title as well). If/when caching was added (assuming it’s not there now), then you’d want to have the canonical url for targets anyway. :) That behavior isn’t required to fix the relative-url issue, but since you’ll have to know what requests redirect to (the call may be implicitly following it now, so you don’t currently know?), then you’ll have the necessary info to do this piece as well. :) Thanks again for marking markdress! It’s definitely going to be incredibly useful!
Jeff Byrnes 6 Dec 2012 3:14 pm: Markdress appears to have expired, any chance of its return?
S Anand 7 Dec 2012 3:53 am: I didn’t think anyone was using it – so I just let the domain expire. Alternatives are at https://github.com/sanand0/markdress#alternatives. You could also host the code yourself.

markdown on S Anand

Google Meet captions as a local transcript recorder

Migrating my blog from WordPress to Hugo

Create a PLAN.md before complex tasks

Analyze data while planning

Edit PLAN.md using the agent, not manually

Implement small steps, run and commit as you go

Run post-mortems mid-way

Ask for actionable exception reports

Delegate verification to the agent

Accept that you will make mistakes

Verification reports speed up reviews

Generate easy-to-review to review content

Give it all your tools

Prefer UI reviews over code reviews for ease

You can resume if it hangs. Don’t worry about context

Allow it design flexibility

Generate documentation

Ask for effort estimates

Probe for confidence

Continue the session for the long tail

Summary

Things I Learned - 10 Aug 2025

How to publish an eBook in 60 minutes

Voice Chat to Slides: My New AI-Powered Workflow

ChatGPT interviews me and creates Markdown slides

Marp converts Markdown to slides

LLMs create supporting images

Deploying on GitHub Pages

What’s next?

A challenge of blog questions

My learnings as week notes

Capturing learnings must be frictionless

Publishing learnings

Comments

How to extend Markdown with custom blocks

Micro-notes

Comments

Markdress

Comments

Test

Create a `PLAN.md` before complex tasks

Edit `PLAN.md` using the agent, not manually