I recommend trying out Quiche Browser on iOS, or Orion from Kagi- both will support Kagi as a search option. Spotlight search still goes to Google, but I find that acceptable.
I had similar thoughts about a year ago. My experience is that the results are that good, though sometimes there are changes under the hood which lead to poor or limited results- it is a beta project, after all. I used the free trial, following which I stayed on the monthly sub before paying a lump sum of $100 for the next 10 months. I've enjoyed my correspondence with the CEO as well. YMMV
Happy to see this here; I found it to be a fun language to prototype reasonably performant things in + for learning about certain data structures. I expect to see the comments talking about the lack of docs, language server struggles, and compilation times (the whole stdlib and program is compiled from scratch every time a program is compiled/run, IIRC). The language is slowly growing and I encourage people to stick with its onboarding shortcomings and giving it a shot for their next project.
> I am so bad at iterating over dataframes! It always feels horrible and slow. While doing this though, I discovered that using df.to_dict('records') and then iterating over the resulting dictionary is almost 100x faster than using the pandas built-in iteration tools like df.itertuples() or df.iterrows()!
That's really surprising to hear, any context on why this is? Very fun read BTW, my friends and I have joked about making something similar for our DMs (nicknamed MattGPT) and giving "them" topics to discuss + observing what they come up with.
Depends on your use case and financials. I use it extensively for work as a computer scientist + developer, and have expensed it in the past. I initially found the free plan of ~30 searches / month pretty good for a 'sudo search', when Google and others weren't quite cutting it and I felt that there could be better results.
