More than hundred years after the partial confirmation of the laryngeal theory by the Hittite evidence this aspect of the phonological reconstruction of Proto-Indo-European has remained one of the most controversial ones. While there is widespread agreement about the reconstruction of three “laryngeals”, their phonological interpretation is less clear. Recently, two radically different new interpretations have been proposed: 1) a reconstruction of just one “glottal fricative” (corresponding to *h₂) by Pyysalo [2013] with sharp critique of most assumptions of the received theory, and 2) a reinterpretation of *h₂ and *h₃ as uvular stops by Kloekhorst [2018], mainly based on their continuation as velar stops in Lycian and Carian. While the former proposal suffers from severe methodological weaknesses and is therefore not further discussed here, the latter seems more promising, as it is based on the usual methodology of comparative reconstruction applied to the actual evidence of Anatolian. After reviewing the most crucial general evidence for the reconstruction of “laryngeals”, especially cases of preserved consonants (including hitherto rarely considered evidence from Iranic), the arguments for uvular stops are discussed. Some problematic phonetic details of Kloekhorst's proposal are criticized, with the result that a reinterpretation according to this critique may even strengthen the general idea of uvular stops. Especially the assumption of *h₃ being a labialized fortis is rejected, instead it is argued that the evidence points to a non-labialized lenis obstruent, thus introducing a distinction typical for stops within the IE system. At the end, remaining problems are discussed, and an alternative scenario for the Anatolian development is mentioned (though not extensively discussed). The conclusion is that uvular stops might indeed be the original value of PIE *h₂ and *h₃ but they might have become uvular fricatives already in PIE, or in late Proto-Anatolian, and the stop reflexes of younger Anatolian languages could also be secondary, independent of the original status of the PIE sounds.