starknet ($STRK )vs arbitrum ($ARB )
Starknet without of doubt is a very good project. However, their technology is just as good as any other layer 2 solutions (e.g. polygon).
The only benefit of their system is that their snark algorithm relying a more standard assumption (i.e. collision resistance) than the assumption used in snark for polygon. Hence, for theoreticians, starknet can be considered more secure + quantum resistant than other L2 solutions.
However, to drive price up, one needs to rely on onchain activities + tvl. Currently, arb has 3b tvl, and strk only has 100m. Market cap of arb is at 2b, and strk is at 1.4b. I strongly believe that strk is overvalued.