I searched for "parsec". Kagi gave me these results, including a value that’s incorrect. A parsec is ~3.08 x 1016 m, not 5.38552341 × 1014 m.
I asked Claude what that value might be. Its response:
That number isn't a mislabeled different named constant — it's a computation bug. It's off from the real parsec (3.0857 × 10¹⁶ m) by almost exactly a factor of 57.2958, which is 180/π (the radians-to-degrees conversion factor).
Concretely, the value Kagi shows equals 1 AU × 3600 meters (149,597,870,700 m × 3600 = 5.3855233 × 10¹⁴ m — matches to 7 significant figures).
The real definition of a parsec is: the distance at which 1 AU subtends an angle of 1 arcsecond, where that arcsecond has to be converted to radians as (1/3600) × (π/180) ≈ 4.8481 × 10⁻⁶ rad, giving parsec = AU / 4.8481×10⁻⁶ ≈ AU × 206,265.
Whatever produced Kagi's number appears to have used 1/3600 directly as the angle in radians — i.e., it converted arcseconds to degrees (dividing by 3600) but never converted degrees to radians (the ×π/180 step got dropped). That single missing factor is exactly the 180/π discrepancy, and it lands you on AU×3600 instead of AU×206265.
So it's not secretly the correct value of some other real constant (light-minute, light-day, AU-related quantities, etc. don't match) — it's a parsec calculation with the degree→radian conversion silently omitted.
This happens in Safari 26.6.2 on macOS 26.6.2.
If it's going to give me a value like that, I expect it to be the correct value.
