Cringe. I promised myself that I would not get involved in any more security discussions. However, since this is a holiday...
The real problem with Wi-Fi security is the shared key. All wireless clients on your network use the same shared key. If the key is compromised, so is the entire network. There are complex ways to sniff the traffic and recover the WEP/WPA key, but it's much easier to simply borrow a laptop on the network, and recover a hashed key from the registry:
In other words, the very concept of a shared key is lacking.
What's needed is a one time key, which does not need to be remembered. This is accomplished with WPA-RADIUS. The user is presented with a unique per-user login and password. The RADIUS server then delivers a one-time, per session, and unique key. You could sniff the key, but it would only be good for that session. Few home networks offer this level of key management, although it's common in corporate networks.
Chuckle. I've been tempted to offer a prize to anyone that can demonstrate a streaming wireless connection that will do 300Mbits. I know that it's been done in the lab (controlled environment) and with dual band channel bonding, but I seriously doubt it can be done in the presence of interference and uncontrolled reflections. The only reason manufacturers offer gigabit ethernet ports is that they would look rather foolish offering 100Mbits/sec ports on a router theoretically capable of 300Mbits/sec wireless.
As for wired being more secure, I beg to differ. I have a small collection of ethernet taps, that I use to sniff traffic for network troubleshooting. If I wanted to sniff your network, I would install one between your broadband connection and router. Taping a single ethernet LAN port won't work because it will only see traffic on that port and broadcast traffic.
I've done about 700Mbits/sec. I forgot the exact hardware but I do recall that I had to tune both the client and server computers IP stack to get decent performance. Out of the box, I think it was about
300Mbits/sec. For testing, I use iPerf and JPerf.I've never seen an easy run on a rework job. There's always some complication involved. The easier it looks, the more complicated it will become.