Selim Uysal

Selim UysalSep 17, 2026

Who wants to do a 30-day writing challenge with me in October? 𝗜 𝗺𝗮𝗱𝗲 𝗺𝘆 𝗳𝗶𝗿𝘀𝘁 𝗺𝗼𝗻𝗲𝘆 𝗳𝗿𝗼𝗺 𝘄𝗿𝗶𝘁𝗶𝗻𝗴 𝗼𝗻𝗹𝗶𝗻𝗲 𝗮𝗳𝘁𝗲𝗿 𝗮 𝟯𝟬-𝗱𝗮𝘆 𝗰𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲. That challenge was Ship 30 for 30. Write and publish every day for 30 days. On day 30 I wasn't a much better writer. But I had 30 pieces out in public, and that was enough for people to see what I could do. Some of them started asking me to write for them. The October cohort starts on October 5th, and this time I can bring people with me. As an AI Writing Skool member, I can give you 𝟯𝟬 𝗱𝗮𝘆𝘀 𝗳𝗿𝗲𝗲 𝗮𝗰𝗰𝗲𝘀𝘀: the October cohort plus everything else inside. It's normally $99 a month, and you can only get the free month through a member's invite. If you've been saying "I'll start writing next month" for a while, next month is October. Enrollment closes Sunday, October 4th.

Selim UysalSep 15, 2026

Most AI demos I see online are something like: “I created a movie with AI.” “I built an entire app in 10 minutes.” “I made 100 pieces of content automatically.” Interesting. But honestly, the AI use cases I enjoy most are usually much more boring :) I recently passed the theory exam for converting my driving licence in Germany, and now I’m preparing for the practical lessons. My driving instructor sent me a PDF with German driving words and their Turkish translations that I need to know. Quizlet already has its own AI and lets you add attachments, so I first thought I could simply upload the PDF there. It didn’t work very well. The PDF was quite low resolution, and there were too many words on each page for it to extract everything properly. So instead, I used ChatGPT to turn the content into a cleaner, structured list first. Then I can move that into Quizlet and study everything as flashcards on my phone. Nothing groundbreaking. No AI agent running a company while I sleep. It just helped me turn a messy PDF into something I can actually use. And I think these are the AI examples I find most useful. Not necessarily the things that look impressive in a demo. But the small things that save time, remove an annoying manual step, or help you do something you were already going to do anyway. Yesterday I was working with LinkedIn data, MCPs and different tools. Today it’s driving vocabulary and Quizlet. Very different use cases. Both useful. I’d actually like to collect more examples like this. What’s one practical or everyday thing you use AI for that other people might find useful too?

Selim UysalSep 14, 2026

Last night I downloaded my LinkedIn data. One of the files is every comment I've left since I joined. That turned into an idea. I've commented on hundreds of people's posts over the years, some of them dozens of times. The feed doesn't care. Someone I talked to every week for a year can disappear from it entirely and LinkedIn will never mention it. So I rebuilt that list myself. I sorted the export into two groups. People I've commented on more than three times, and people I've engaged with once or twice. The first group is a real relationship list. Nobody sold it to me and nobody scraped it. I built it one comment at a time without noticing. One problem: the export gives post IDs, not names. A post ID tells you nothing about who wrote it. So I ran those through the LeadShark MCP to identify and enrich the profiles behind them, then loaded the results into Extrovert as curated lists, so those people get their own feed instead of being buried in the main one. About 200 people are in it now and it's running. 𝗜'𝗺 𝗻𝗼𝘁 𝗹𝗼𝘀𝗶𝗻𝗴 𝘁𝗵𝗲 𝗽𝗲𝗼𝗽𝗹𝗲 𝗜 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗸𝗻𝗼𝘄 𝗮𝗻𝘆𝗺𝗼𝗿𝗲. That's the part of AI on LinkedIn I find worth the effort. Not another generated comment or another few hundred strangers in a sequence. Just using my own history instead of starting from a cold list every time. Happy to walk through how I did it if it's useful to anyone. And please share if you've downloaded your own data and built something with it, I'd like to see what else is hiding in that export.