How I crafted TL;DRs with LLMs and modernized my blog (part 5)
View the series
- See how I crafted story-like tl;drs for my posts with LLMs
- Learn how I generated my llms.txt summary with LLMs
- Grab the "Copy page" button's code, as seen in OpenAI docs
- Check how I optimized images for better blog performance
- See how impressed I was by GPT-4.1's meta descriptions
It's fantastic to work with tools like Lighthouse ↗ that score what you're building, tell you exactly what to change, explain how to do it, and link to articles that show why it matters. With this feedback, you can improve your score and supposedly, your web page too.
The last thing Lighthouse reminded me to fix was missing meta descriptions in my posts:
<meta name="description" content="..." />
Why meta descriptions ↗ matter:
The
<meta name="description">element provides a summary of a page's content that search engines include in search results. A high-quality, unique meta description makes your page appear more relevant and can increase your search traffic.
After I added meta descriptions, my SEO score jumped to 100. That isn't proof they're great, but at least they're no longer missing.
I generated my meta descriptions by passing each post's tl;dr into GPT-4.1 with the following prompt:
I'm writing meta descriptions for my blog posts. Here is the description
of my blog:
This blog is my ongoing, hands-on exploration of AI automation—how to
build, troubleshoot, and refine real-world workflows using LLMs,
automation platforms like Zapier, and both code-based and no-code
tools. Everything here comes from my lived experience learning as I
go, with lots of focus on what's practical, what breaks, and what you
can actually reuse.
I'm Tony Aldon and my blog is served at tonyaldon.com.
Please generate a clear, engaging meta description for this blog post
that is under 160 characters, uses relevant keywords, and speaks to
practical, hands-on problem solving. The meta description should make
sense to someone who hasn't read the post or blog before. Use a
friendly, inviting tone, and encourage readers to click.
Give me 5 alternatives.
Here is the summary I wrote for my blog post:I was so impressed by these meta descriptions that I started questioning my use of tl;drs:
In my posts,
As summaries in my Atom/RSS feed,
As post descriptions in my llms.txt.
I listed the tl;drs alongside their meta description counterparts, shared them with GPT-4.1, and then asked the following:
Between tl;drs and meta descriptions, which should I use for summaries in my Atom feed?
Should I use meta descriptions instead of tl;drs for the descriptions of posts in my llms.txt?
Is it still relevant to keep tl;drs at the top of posts, or should I replace them with meta descriptions?
Part of the reply was:
Meta descriptions are designed as external summaries: search results, feeds, embeds, and LLM guidance. They "sell the post" to potential readers and robots.
TL;DR is an "insider," human-voiced quick abstract.
Keep both!:
Meta description (in
<head>) for SEO, agents, and summaries.TL;DR at the top for human readers.
That helped me decide, once and for all: tl;drs stay in the posts,
while meta descriptions will go in the feed and llms.txt.
Meta descriptions of the first 12 posts
How I explored Google Sheets to Gmail automation through Zapier before building it in Python (part 2): Check how I switched from Zapier automation to a Python polling script, set up Google APIs, and tackled API authentication—just for the sake of learning.
How I explored Google Sheets to Gmail automation through Zapier before building it in Python (part 1): See how I built my first Zapier Gmail alert from Google Sheets updates, plus hands-on tips for troubleshooting Zap errors I triggered on purpose!
How I realized AI automation is all about what you automate: Curious about automating CRM after sales calls? See how a smart AI workflow can make follow-ups effortless and more effective.
How I uncovered Zapier's best AI automation articles from 2025 with LLMs (part 3): See how I processed all Zapier articles from 2025, navigated token limits in OpenAI API, and used Gemini for better AI automation results.
How I uncovered Zapier's best AI automation articles from 2025 with LLMs (part 2): Learn how I scraped Zapier blog articles into JSON and markdown—practical tips, real errors, and workflow automation for your own projects.
How I uncovered Zapier's best AI automation articles from 2025 with LLMs (part 1): See how I used AI to summarize and rank the top 10 Zapier automation articles. Practical tips and hands-on curation for automating smarter!
How I learned the OpenAI Agents SDK by breaking down a Stripe workflow from the OpenAI cookbook (part 4): I cut out the AI agents from a workflow I found on OpenAI Cookbook—explore my practical steps for direct OpenAI API automation.
How I learned the OpenAI Agents SDK by breaking down a Stripe workflow from the OpenAI cookbook (part 3): Dive into AI automation logs—see how logging uncovers agent behaviors, function tools, and handoffs in real-world workflows.
How I learned the OpenAI Agents SDK by breaking down a Stripe workflow from the OpenAI cookbook (part 2): Learn how I experiment with the OpenAI Agents SDK and traces dashboard, tweaking a Cookbook workflow—explore hands-on AI automation with me!
How I learned the OpenAI Agents SDK by breaking down a Stripe workflow from the OpenAI cookbook (part 1): See how I dig into a Stripe workflow from OpenAI Cookbook using OpenAI Agents SDK—real lessons, clear steps, and practical automation tips!
How I implemented real-time file summaries using Python and OpenAI API: See how I built an AI auto-summarizer in Python, tackled OpenAI API issues, and got real-time insights with parallel processing. Dive in!
How I started my journey into AI automation after a spark of curiosity: Inspired by Zapier's AI roles, I'm diving into practical AI automation. Follow along for real insights, tips, and workflow solutions!
That's all I have for today! Talk soon 👋