update: i notice this happens every single time whenever i review the second math card, and conversely it never happens on the first review (after i go back to Decks and start reviewing the same deck again (all my cards in that deck have mathjax))

on review:

preview is more reliable

source:

```
<div>Let <anki-mathjax>\mathbf{C}</anki-mathjax> be a small category and <anki-mathjax>J</anki-mathjax> a Grothendieck topology on C. A presheaf, as before, is simply a functor<br><anki-mathjax block="true">P: \mathbf{C}^{\mathrm{op}} \rightarrow \mathbf{S e t}</anki-mathjax><br>. If <anki-mathjax>P</anki-mathjax> is such a presheaf and the sieve <anki-mathjax>S</anki-mathjax> is a cover of an object <anki-mathjax>C</anki-mathjax> of <anki-mathjax>\mathbf{C}</anki-mathjax>, a <i><font color="#ff0000">matching family</font></i> for <anki-mathjax>S</anki-mathjax> of elements of <anki-mathjax>P</anki-mathjax> is a function which assigns to each element <anki-mathjax>f: D \rightarrow C</anki-mathjax> of <anki-mathjax>S</anki-mathjax> an element <anki-mathjax>x_f \in P(D)</anki-mathjax>, in such a way that<br><anki-mathjax block="true">{{c1::x_f \cdot g=x_{f g} }} \quad \text { for all } g: E \rightarrow D \text { in } \mathbf{C .}</anki-mathjax><br>An <i><font color="#ff0000">amalgamation</font></i> of such a matching family is a single element <anki-mathjax>x \in P(C)</anki-mathjax> with<br><anki-mathjax block="true">{{c2::x \cdot f=x_f}} \quad \text { for all } f \in S \text {. }</anki-mathjax><br>Then <anki-mathjax>P</anki-mathjax> is a sheaf (for <anki-mathjax>J</anki-mathjax> ) precisely then, when {{c3::every matching family for any cover of any object of <anki-mathjax>\mathbf{C}</anki-mathjax> has a unique amalgamation}}.<br><br></div>
```

Anki desktop for mac silicon

Version 2.1.61 (0c1eaf4c)

Python 3.9.15 Qt 6.3.2 PyQt 6.3.1