By and large, the recent Mac malware scare has been grossly overblown.
These malware had a total infection of 600K. In less than 3 weeks, it has gone down to 30k.
(source: Kapersky (and corroborated with Symantec)
Kaspersky: Number of Macs infected by Flashback drops to 30,000 - Update - The H Security: News and Features)
That is a very small infection. The duration of this problem is very small relative to other infections.
On the other hand, Microsoft had a zero day exploit that went un-patched for 6 months. The Duqu was far nefarious and can be exploited with a true-type font. It was a bootkit. A bootkit is far the most dangerous because it is resident in memory and upon boot, the NT kernel cannot detect it.
Microsoft pushes out emergency fix to block Duqu zero-day exploit
I work with about 300 computers. 60 macs and the rest are PCs. PCs tend to get infected easily. This is just a wake-up call for mac but by and large, I'd say macs can be very safe if you harden them.
I can run zenmap/wireshark/nessus and scan the network and find vulnerabilities in various open ports on Windows in minutes.
You can run a Python script and nuke a harden 64 bit Win7 machine from a rooted Android phone.
Yes, a phone can DDOS a Winbox in seconds. There are buffer-overflows attacks that when a Window user launches Window Media player, can easily root the entire computer in minutes. Again, exploit you can do from a phone/tablet.
for this very reason we have wifi on completely separate firewall networks. And machines that need to be on our wifi networks need certificates which excludes most android devices.
I'm more scared of exploits that can be launched without user-intervention. The kind that finds open ports and spreads like wildfire.
E.G. Conficker.
In 2005-2006 , I remembered it clearly. It spread to 3 million computers in 24 hours. You didn't need to open an email or go to a website. It scanned the network, found port flaws and injected itself. That was the worst nightmare in career. Every client, every vendor, every business partner we knew were infected.
Social engineering attacks like facebook likes, javascript injections, flash/pdf hijacks can be locked down.
None of our macs were infected because we run a proxy server, run IDS (snort), users don't have admin access to install Flash and Java was disabled.
Seems like most people here are wise to not open mysterious emails or go to bad websites.
But what do you do if you get hijacked just by being on a certain computer network? Now this is what I am more scared of.
If I had to compare something like Mac Flashback to Windows' Duqu, I take the Mac trojan any day of the week. Bootkits scare the living daylights out of me. The Flashback trojan can be fixed with a command line delete. Some of the Windows malware I dealt with were much, much harder - boot in safe mode, delete registry keys, scan for boot-level memory inprints. You needed actual physical hardware access in most cases. On the mac, you just SSH into it and 2 lines and you are done.
Linux can just be as dangerous. I work with engineers and they all run super-users. They don't understand permissions vis-a-vis groups so they just chmod 777 everything. I can just SSH in as a "regular" non-root user, go to their /etc directory and start playing with their host files, firewall rules,etc as a non-root user.
I also tutor high-school and college student web developers who run linux. Whenever I have access to their laptops/workstation to check something out, all the permissions are wide-open. Who the hell 777 their entire var folder? I've seen it way too many times.
And I've been into data centers where even harden linux servers have SELINUX disabled just because having it on has been problematic w/ certain apps.