Longer-form posts about whatever I'm exploring.
I had a bike commute where I happened to join the 2026 May 3rd 5 borough bike tour1. I was curious how much overlap I had. The below is an overlay output from a vibe coded ChatGPT session, extracting the overlay data using gpx from my strava data and gpx from the bike tour that looks like is here2. So in my mind it felt like 5 miles. And looks like it was about 5.92 miles.
Recently, myself and the rest of NYC went through a cold spell and news outlets reported2 the 13 day stretch of sub-zero weather, ending Feb 6th, was not longer than a 16 stretch in 1881. And this was shorter than a 1963 stretch, but tying a 2018-01-13 streak. But this recent winter sure felt extreme. I know there is a recency bias, but I figured, why not also compare the area under the curve too. So I ranked the coldest 14-day stretches, using available data of the past decade. And then tried to visualize the spans of the coldest years too.
Email forwarding mystery solved Feeling several facepalms now that my michal@piekarczyk.xyz email finally forwards to my hey.com email inbox. After several sessions of tweaking, the answer was low tech! Registrar forwarding Originally, I setup mail forwarding on my domain registrar. Only $5 a year, okay why not. After proving to my registrar I owned my hey email, I was all set, but tests yielded silence. I spent several chat sessions with a support engineer at my registrar. He pointed out I was missing SPF and DKIM records for my custom domain and so he added those. Those have nothing to do with delivering mail to my address and only help to authenticate mail sent from my domain, but I figured what the heck let’s try something. That did not work.
Life’s a garden, can you dig it? That is what a colleague of mine had said very often when we worked together, about how you need to put time and effort into stuff. You know, get your hands dirty ! Picking up my shovel Today I finally took a first step to trying out this digital gardening concept on Maggie Appleton’s blog0. I came across back on haha I don’t know why but on 2025-07-04 , I guess it was a kind of independence day in my brain, from the march of chronology.
Came back from a flight from New Delhi , Monday morning, that left New Delhi just after midnight and 16 hours later arriving in New York around 5:30 am. And so far today, Thursday morning, I feel kind of fine, not really jet lagged, in the sense of wanting to wake or sleep at odd hours. The specific strategy I did in attempting to align meal time and sleep time with the destination time zone was to fast the day of the flight until lunch time arrival at my destination. It would have been ideal if I was able to sleep during the red eye though not doing so was also perhaps fine since by the time I crashed on Monday at around 19:00, I had about 16 + 16 + 12 = 44 hours of sleep pressure in the tank. And then when I slept for basically 12 hours, waking at 7 am on Tuesday, and getting the morning sun, I was basically back in my rhythm. And so far, sleep on Tuesday night and Wednesday night was pretty okay.
Fleeting notes and thoughts that might take shape later.
Since I have been using fuzzy find fzf1 lately, switching to fzf for reverse shell search2, I was wondering hey can I use this for switching to a chrome tab? With a super quick trial and error session with ChatGPT, I have an Applescript answer to this question, below. So this works really well for me so far. Will try it out a bit more before adding it to my ~/.bash_profile. But pretty game changing for now.
I was interested in improving my reverse search Control R experience on my shell because many of my previous commands are not easily searchable or they disappear too fast. Initially I just wanted to figure out if you can use fzf for the reverse-search, but ChatGPT made me aware of atuin too. This sounded super fancy and the atuin1 idea of using a local sqlite db for a command cache looked interesting, so I tried to brew install atuin first.
I was worried I would lose the american airlines app notifications but luckily the email notifications persisted. So here is some of my data, extracted from emails from aa.com I should organize this better, but basically I wanted to use some data to tell the story of a recent trip I was taking from NY to Arizona. I left a very engaging group discussion at work because I planned a tight bike ride home to wrap up a quick exit to the airport in time to board my plane.
Trying to see have I wfitten about this before, but weve had data scuence for a while , where given domain knowledge, programming knowledge, and some maths, you can iterate towards data science (just kiddijg thats a blog). You can iterate towards a model that fits your data and generalizes modestly on unseen data. Then we entered ML engineering, where you lean into the formal aspects of rigorous testing and more into the repeatable, so less trial and error perhaps.
Planing on a plane. Used ChatGPT’s DALL-E here. prompt Lets make a fun hyper realistic image with a normal looking scene in an airplane of 737 size , where we see all typical activities, of a plane in flight, but also to make it funny, instead of several rows on one side , is a wood worker, “planing” some wood with typical woodshop vibes. The woodworker , a woman, just happens to be there , no one is concerned, everything is Normal
an interlude to having better grasp of an unwieldly photo situation
This project is a reboot of an earlier project predicting bicycle ride share riders destinations.
This is a simple personal time tracker built with clojure, aws lambda and dynamodb, inspired by Laura Vanderkam’s “168 Hours” book
