Like Linear B, Linear A is found written on clay at Bronze Age Cretan sites. As well as larger tablets, both Linear A and B are often written on ‘sealings’ or ‘nodules’: small lumps of clay that presumably recorded individual transactions, perhaps to be compiled on a tablet later. Handily, these are very suitable for representing in cupcake form. The examples here with two or three signs may be names of people or places who were contributing or receiving goods, while the single signs probably represent commodities (the sign in the middle of the third row, for instance, looks like some kind of tripod to me).
Unfortunately, I can’t tell you a lot more than that about what these nodules actually refer to, because (like Cypro-Minoan) Linear A still isn’t deciphered. A lot of the signs of Linear A and B are very similar, and it’s clear that Linear B was adapted from Linear A in order to use it to write Greek (though exactly who was responsible for this adaptation is debated – were they Mycenaean Greeks from the mainland, or native Cretans – “Minoans” – who had, for one reason or another, learned Greek?).
Either way, a lot of the sound-values of Linear A signs are likely to be similar to the values of the corresponding signs in Linear B. Using Linear B values, for instance, we could ‘read’ many of these cupcakes (e.g. the first two at the top would read SI-KA and DA-KA) but we still don’t know exactly what these words would have been; Linear B spelling conventions mean you can’t always tell exactly how words were actually pronounced, and of course we don’t even know how far these rules would apply to Linear A. Even if we could say for certain that SI-KA was in fact pronounced sika, the fact that it’s probably a name means it’s of limited use for working out what the language of Linear A actually is (the majority of names found in the Linear B tablets are probably not Greek). Most Linear A texts, even on tablets, are pretty short, and the majority are likely to be simply lists of names, so the amount of linguistic information available, compared to the amount that Michael Ventris had when he deciphered Linear B, is tiny.
Even so, comparison of the two scripts has given us some information about ‘Minoan’, as the language of Linear A is conventionally called – for instance, it may well have had a three-vowel system (a, i, u). (For a much more detailed account of the current state of knowledge about Linear A, see this website.) But overall, all we can be pretty confident about is that Linear A isn’t Greek, and there is so far no good evidence to relate it to any known language or even language family. Which is not to say that plenty of people haven’t tried to do just that – and the decipherment attempts listed on that Wikipedia page are only a few of the “more respectable” ones. You can Google the others if you dare…*
*Warning: here be dragons. Or at least some seriously strange theories about the Phaistos Disc.