PieFed allows you to follow pretty much anything - posts that aren't yours, individual comments, users, communities, etc., and it can even allow you to unfollow things too, none of which Lemmy allows.
Lemmy really does seem pretty far behind whenever I start thinking about the feature comparisons between the two. To be fair Lemmy does have a far better search feature and is a tiny bit more polished in some aspects (PieFed can get buggy in places, like there is a Preview option for both posts and comments but somehow it got disabled for top-level comments), but given its current rate of development I'd be surprised if PieFed didn't surpass both aspects in about 3 months time, just as it has already surpassed Lemmy in so many other areas.
How is it on api availability for third party apps, and further adoption by third party apps?
That seems to be a big driver of Lemmy support, the fact there are lots of apps that support its api. That's what killed kbin, I realize this is a post about its fork, mbin, but similarly does not have much app support.
Most of the top apps now already support PieFed - I don't know where to find a list but a few I can recall of the top of my head are Voyager, Thunder (that one might not yet be in the fully public version rather than nightly beta), and Interstellar (that one also supports Mbin too!), but I know there are several others that I simply am not recalling the name of since I always just use a web browser.
That said, the PieFed API itself is very brand-new, and does not support all the features that the webpage interface does. e.g. you could definitely set up keyword filters on the website and then I think those would still function even if your choice of app did not support the actual setting of them and instead merely delivered the feed as calculated by the PieFed back-end, but I do not think that some of the most popular features such as the initial sign-up wizard, the categories of communities (which are user-customizable and shareable), the collection together of comments across all cross-posts that helps deal with the increasing fragmentation effect present on the modern Threadiverse, etc. are available via apps.
So the answer is a bit more complex than simply yes or no - as in technically yes apps work, but you do not get the full experience with them. It still seems a step up from Lemmy to me tbh, as in no worse and even noticeably better, even if lacking a lot of the "wow!" factor offered by the webpage UX.
That was not the only thing that killed Kbin though - Ernst simply tried to do too much, and could not handle being both the sole dev of the software (many changes were offered to him, languishing for months on end that he never bothered to so much as look at or comment about) as well as the sole admin for the primary instance kbin.social, the latter of which started (I am not joking about this) sending spam messages to the entire Threadiverse, forcing other instances to defederate from it or least those specific communities that were left unmoderated as people began to abandon it - the latter not always on purpose btw but they simply could not gain reliable access to the website itself, instead always getting an error whenever they tried. Ernst had some personal problems in his life with his partner, and some medical issues, but ultimately it was his decision to not accept any help at all (for either the software development or the routine instance admin duties) that lead to the platform becoming unstable and then finally succumbing to the trials that he faced through little to no fault of his own.
Something similar happened to another Lemmy alternative btw, Sublinks, where the sole dev had a baby and then never released it, with it now having lost traction (though hopefully still being developed in secret! the dev is quite friendly and seems quite capable).
I say all that not for the purposes of denigrating Ernst or the situation, but to point out that situation with Rimu - the chief dev of PieFed - is the absolute polar opposite. He listens to feedback, admits when he is wrong, has an entire team of people working both on the primary instance (piefed.social) as well as the software engineering, and he is constantly accepting whole entire feature developments offered by many helpful people. As you learn more about PieFed, you will be absolutely IMPRESSED, if not by every tiny little detail, then by the overall way that things are being handled. This project looks to be going places!
And if the absolute worst-case should happen, it is FOSS and can be continued, much like Kbin was - although many people (myself included) don't like the look of K/Mbin or how it combines the threaded (Reddit-like via Lemmy federation) and non-threaded (Twitter-like, via Mastodon federation) aspects of the Fediverse together, and in such confusing ways (there are up-votes, and on top of that also Boosts, and there are down-votes, but on top of that also Reduces - except just kidding, although actually wait no, there are those, but you cannot see them for some reason, even though you can see all people who offered Boosts; oh and communities are now magazines, and then there is also microblogs and... what were we talking about again?!:-P). I just liked kbin.social for the reason that it was not developed by authoritarian tankies with questionable morality, not because of its tie-in to Mastodon microblogging - plus have you noticed how its entire interface seems geared towards the microblogging first, but then the Threadiverse integration seems like an after-thought? (e.g. images are displayed with a lot of wasteful space, much better attuned to the style shared via microblogging than via community-based threading). Which is another strong reason that PieFed is liked by so many, but unlike K/Mbin it seems eclipsed by how fantastically well-received the features are by so many people (and also unlike K/Mbin, the community-based threaded discussion format is its primary focus, not an after-thought).
Now this is a meme template I've not seen in a while...
PieFed allows you to follow pretty much anything - posts that aren't yours, individual comments, users, communities, etc., and it can even allow you to unfollow things too, none of which Lemmy allows.
Lemmy really does seem pretty far behind whenever I start thinking about the feature comparisons between the two. To be fair Lemmy does have a far better search feature and is a tiny bit more polished in some aspects (PieFed can get buggy in places, like there is a Preview option for both posts and comments but somehow it got disabled for top-level comments), but given its current rate of development I'd be surprised if PieFed didn't surpass both aspects in about 3 months time, just as it has already surpassed Lemmy in so many other areas.
How is it on api availability for third party apps, and further adoption by third party apps?
That seems to be a big driver of Lemmy support, the fact there are lots of apps that support its api. That's what killed kbin, I realize this is a post about its fork, mbin, but similarly does not have much app support.
Most of the top apps now already support PieFed - I don't know where to find a list but a few I can recall of the top of my head are Voyager, Thunder (that one might not yet be in the fully public version rather than nightly beta), and Interstellar (that one also supports Mbin too!), but I know there are several others that I simply am not recalling the name of since I always just use a web browser.
That said, the PieFed API itself is very brand-new, and does not support all the features that the webpage interface does. e.g. you could definitely set up keyword filters on the website and then I think those would still function even if your choice of app did not support the actual setting of them and instead merely delivered the feed as calculated by the PieFed back-end, but I do not think that some of the most popular features such as the initial sign-up wizard, the categories of communities (which are user-customizable and shareable), the collection together of comments across all cross-posts that helps deal with the increasing fragmentation effect present on the modern Threadiverse, etc. are available via apps.
So the answer is a bit more complex than simply yes or no - as in technically yes apps work, but you do not get the full experience with them. It still seems a step up from Lemmy to me tbh, as in no worse and even noticeably better, even if lacking a lot of the "wow!" factor offered by the webpage UX.
That was not the only thing that killed Kbin though - Ernst simply tried to do too much, and could not handle being both the sole dev of the software (many changes were offered to him, languishing for months on end that he never bothered to so much as look at or comment about) as well as the sole admin for the primary instance kbin.social, the latter of which started (I am not joking about this) sending spam messages to the entire Threadiverse, forcing other instances to defederate from it or least those specific communities that were left unmoderated as people began to abandon it - the latter not always on purpose btw but they simply could not gain reliable access to the website itself, instead always getting an error whenever they tried. Ernst had some personal problems in his life with his partner, and some medical issues, but ultimately it was his decision to not accept any help at all (for either the software development or the routine instance admin duties) that lead to the platform becoming unstable and then finally succumbing to the trials that he faced through little to no fault of his own.
Something similar happened to another Lemmy alternative btw, Sublinks, where the sole dev had a baby and then never released it, with it now having lost traction (though hopefully still being developed in secret! the dev is quite friendly and seems quite capable).
I say all that not for the purposes of denigrating Ernst or the situation, but to point out that situation with Rimu - the chief dev of PieFed - is the absolute polar opposite. He listens to feedback, admits when he is wrong, has an entire team of people working both on the primary instance (piefed.social) as well as the software engineering, and he is constantly accepting whole entire feature developments offered by many helpful people. As you learn more about PieFed, you will be absolutely IMPRESSED, if not by every tiny little detail, then by the overall way that things are being handled. This project looks to be going places!
And if the absolute worst-case should happen, it is FOSS and can be continued, much like Kbin was - although many people (myself included) don't like the look of K/Mbin or how it combines the threaded (Reddit-like via Lemmy federation) and non-threaded (Twitter-like, via Mastodon federation) aspects of the Fediverse together, and in such confusing ways (there are up-votes, and on top of that also Boosts, and there are down-votes, but on top of that also Reduces - except just kidding, although actually wait no, there are those, but you cannot see them for some reason, even though you can see all people who offered Boosts; oh and communities are now magazines, and then there is also microblogs and... what were we talking about again?!:-P). I just liked kbin.social for the reason that it was not developed by authoritarian tankies with questionable morality, not because of its tie-in to Mastodon microblogging - plus have you noticed how its entire interface seems geared towards the microblogging first, but then the Threadiverse integration seems like an after-thought? (e.g. images are displayed with a lot of wasteful space, much better attuned to the style shared via microblogging than via community-based threading). Which is another strong reason that PieFed is liked by so many, but unlike K/Mbin it seems eclipsed by how fantastically well-received the features are by so many people (and also unlike K/Mbin, the community-based threaded discussion format is its primary focus, not an after-thought).