Current
Summary of historical research sessions
Problem statement
In the process of researching a single question—e.g. “which neighbourhood should I stay in, when I visit Barcelona?”—I might execute a handful of searches, then open and read 10+ results before I come to a conclusion.
The fruits of this labour feel quite ephemeral. I’ll probably “save” the final output of this research by taking a note somewhere—e.g. my trip planning doc. But it’s difficult to “drop back into” a stage of research—e.g. when planning a future trip—without taking a tedious amount of manual notes about the intermediate steps involved.
Ideas
Using https://www.rewind.ai/ and asking it to summarize the outcome of my research session, then saving that in a markdown note somewhere.
Ultra-short USB-C cable
I’ve got an Anker MagSafe battery that is very convenient for on-the-go charging because it magnetically secures itself to the back of my iPhone. The wireless charging itself though is not as good as wired charging. It’s slower, less efficient, and—in a warm environment—results in my phone getting extremely hot, and temporarily pausing charging.
So I bought this 13 cm USB-C cable to use with the magsafe charger. This solves the above issue with wireless charging. This is the shortest such cable I could find, but it’s still unnecessarily long. I’d love to have a USB-C cable that is just 3 cm long, to enable wired charging without a cable sticking out of my pocket.
Archive
Search engine with customizable ranking
It feels like Google search has become progressively worse over the past decade. I often find myself typing a keyboard macro which expands to site:reddit.com or site:news.ycombinator.com to manually restrict results to domains that I trust.
This works okay when I can anticipate the specific domain of results I am interested in. But it doesn’t scale well to multiple domains. Why can’t I instruct Google to prefer results from certain domains, and to avoid from others?
The closest solution I’ve found is setting up a personal Google Programmable Search Engine which indexes specific domains. However, I have been entirely unsuccessful at getting it to slightly up/down-rank specific websites. If I include a more prominent domain—such as The Atlantic, it subsequently dominates more niche domains (e.g. lesswrong.com) in all queries.
Update [2023-10-01]: I came across Kagi, a new search engine which offers exactly this feature.
comments powered by Disqus