Accurate Answers, Endless Wonder

Improvements To Pandi’s Thought Terminal

Pandi’s thought terminal shows what Pandi is doing in the background while you wait for an answer. Pandi understands your query, browses the web, reads and comprehends content, composes interactive user interfaces, self-reflects, etc. These steps do take time, but the wait is worthwhile.

I’ve redesigned the UI to look like an old computer terminal. I never used one, but I always thought people who could operate them were some kind of wizards. My first computer was a Pentium II with Windows 95, so it already had a GUI.

The new thought terminal has two new additions:

Navigational Links

Sometimes you just want to navigate to a web page, so now Pandi shows the links and starts browsing straight away. If you want to navigate to a site, you can just click on them, so you don’t have to wait for an answer. The links are displayed as a ranked list, so it feels familiar with classic search engines.

In addition, you can also use the “Go $query” prompt to make it easier to navigate to sites.

Time Saved

While these searches do take time, you’re actually saving a lot of clicking, reading, writing, etc. Pandi essentially does the last mile search for you.

A human can read 200 words per minute. On average, Pandi reads 8,000 words per query, so it saves you 40 minutes. It does all of that in less than 30 seconds. While you wait for your answer, Pandi will tell you how much time you have saved compared to doing the last mile search yourself.

Fast answers are something that users will always want, so I’ll continue making performance improvements.

Leave a comment