Let’s start with the current state of genetic engineering. Right now we can read the entirety of any genome on Earth. DNA sequencing is cheap and only getting cheaper. Additionally, we have the technology to modify the DNA in almost any way we see fit. Right now this process is tedious, requires expertise and equipment, and is limited in its applications, but with enough resources, we can create any DNA sequence we want. Together these two capabilities theoretically enable us to engineer any genome in any way we please. If we can read and we can write then we can write any novel, code any program, can’t we? Theoretically, there exists a DNA sequence that will give rise to artificial meat, a biological computer, a superior human being. So why can’t we just make these sequences? Because we have absolutely no idea what these sequences are.
The language of DNA is so incredibly complex that we cannot decipher it merely from seeing the sequence. Genomes weren’t designed by engineers with any sense of order. While there are certain patterns and rules in the DNA code they seem to be as frequently broken as they are upheld. The whole system feeds back on itself such that any linear relationship you draw is always a simplification and any change you make will always have a myriad of effects. Right now we can often predict the amino acid sequence of a protein that will be created by a gene, but that doesn’t tell us how that protein will fold, what it will eventually do, wherein the body it will be made, how it will be made, how it will be modified after production and many more unknowns all of which are absolutely critical to understanding what the gene actually does. The current state of most genetic engineering research is using the aforementioned editing tools to change the genetic code in small ways and then observe what happens in the resulting organism. This is a painstaking process with years of work from multiple labs required to get even a rudimentary understanding of a single gene in a single organism.
So, to get to your question, the difficulty of a genetic engineering technology is dependent primarily on how thorough an understanding of the genome it requires. As an example of existing technology, some human diseases such as Cystic Fibrosis are caused by the malfunction of a single gene. Even without an exhaustive understanding of how this gene works, we know all we have to do is repair it in the affected cells. Looking at slightly more advanced theoretical biotech there are several startups trying to use simple organisms like yeast to produce desirable chemicals. The idea here is that since we know the pathways responsible for the production of these compounds we can simply transfer those genes into yeast and they will produce the desired compounds. Of course, it isn’t that simple. You can’t just assume a gene from one organism will function the same way in another. It might, but getting an entire pathway to work is tricky business and relies a lot on persistence and luck rather than an actual understanding of why it is or isn’t working. This is the forefront of genetic engineering right now. We could, of course, attempt more ambitious projects but they would be unlikely to succeed.
It’s difficult to estimate which of your proposed technologies will require a more thorough understanding of the underlying genetics to succeed, but I’ll try to give a few general rules of thumb. Anything that will happen in simple organisms such as bacteria or yeast will likely require less understanding than major projects in mammals or humans. Additionally, projects involving cell cultures of mammals such as lab produced meat will require less understanding than projects involving whole living organisms. Projects involving things we don’t even understand the basics of, let alone the genetics of, such as the human brain or entire ecosystems are likely the furthest reaches.