I only use apt, or compile from source but here is what some guy online wrote:
<span style="color:#323232;">* Creates larger packages from compiling in most/all of the libraries used in development that use up more storage space because you are storing additional copies of libraries already on your system.
</span><span style="color:#323232;">
</span><span style="color:#323232;">* Slower loading because you are loading a separate copy of libraries that may already be loaded and available for use.
</span><span style="color:#323232;">
</span><span style="color:#323232;">* Updates to external libraries that add features, fix bugs, or change functionality are not available to the existing flatpak, which also reduces the testing use base for updates to support libraries.
</span><span style="color:#323232;">
</span><span style="color:#323232;">* Breaks the fundamental concept of *NIX which is to separate out development of different subsystems, meaning that if an external system (library) changes how it does things, your program may no longer work or have features suddenly limited because it is using an older version of the library.
</span><span style="color:#323232;">
</span><span style="color:#323232;">* Depends on the Application Binary Interface (ABI) remaining stable in the Linux kernel, which is preferred but not guaranteed. If an ABI is changed, the flatpak can stop working.
</span><span style="color:#323232;">
</span><span style="color:#323232;">* Can be difficult to adapt to different distributions of Linux. Many Linux distributions have package maintainers that compile programs from source to ensure it works on and is compatible with the particular distribution they are supporting, and flatpaks can struggle to work on a different distribution.
</span><span style="color:#323232;">
</span><span style="color:#323232;">* Sometimes avoids the work necessary to make the source generally compile-able across distributions. It's easier to compile a flatpak on the system it was developed on because all of the testing was done on one particular distribution. However, different distributions may have differences like the files in different locations, different software or libraries installed, different versions, different compilers, and without the program being compiled from source with all these differences, the codebase may never be updated to be take into account the differences in systems. The source code can be very stale, in a sense.
</span><span style="color:#323232;">
</span><span style="color:#323232;">* Not a long-term solution. Flatpaks generally age out over time because they can't keep pace with distribution changes over time, as the open source environment is still very dynamic. Only source code that is legal to share is really sustainable over the long-term.
</span><span style="color:#323232;">
</span><span style="color:#323232;">* Reduce use testing and reporting of bugs. Reported bugs usually get ironed out over time, as most program and package maintainers take pride in their work. With flatpaks, users don't get to test the program against updates to support libraries, so there are less user reports of bugs, and it can mean that bugs don't get identified as quickly or ever.
</span><span style="color:#323232;">
</span><span style="color:#323232;">* Harder to identify bugs. Because flatpaks are monilithic binaries comprising both original and external work, it can be hard to figure out if a given bug is in the original program or supporting library, leading to mis-reporting of bugs, which adds effort to figure out the real source of the bug.
</span>