Why does the universal selector * have a negative impact on browser rendering?

The concept that the universal selector is slow is a myth coined ten years ago when, in fact, there were performance problems, not surprisingly, because of Internet Explorer 6.

Nowadays, in modern browsers the performance impact of the universal dial is insignificant as long as you do not set rules that result in slower effects on it.

Unlike the reference article, personally, I would not define CSS3 effects in the universal selector. Not only because I vehemently doubt that ALL elements would REALLY need that particular effect, but also because I favor reusable generic classes present in CSS frameworks like Bootstrap.

Source: Article on the Telerik Blog via an answer in SOen

    

The main argument against the universal selector was not so much performance, but the impact it had on property inheritance.

In practice, neither loss of performance nor inheritance problems were ever relevant enough to abolish use, because it was much quicker and more practical to use it and make a few specific definitions then including adopting stylesheets specific for the late IE6, which was already obsolete when the use of CSS2 gained strength and table-based layouts began to fall into disuse.

A template I've always adopted, recently upgraded to html5, was:

* {
  margin:0;
  padding:0;
 /* outros estilos com base no projeto */ 
}

div, article, aside, details, figcaption, figure, 
footer, header, hgroup, menu, nav, section {
  display: block;
}

a, a:link, a:visited, a:hover, a:active { 
  text-decoration:none; 
  font:inherit; 
}

li {
  list-style-type:none;
}

Even because, again, in practice, you always have to define the styles according to the project, each one specifically, and in the reset, only the elements that really can get in the way or repeat themselves in creation are inserted.