I was initially attracted to Savage Worlds not becuse of its uniquely different approach, but because of its similar approach. From what I'd read, it seemed to share many of the same approaches and goals as I'd attempted to address when writing PowerFrame - a generic, simplified and streamlined distillation of 90's gaming, with an emphasis on low prep time. As I read the rules, my first impression was that it was more successful in its goals of low prep, simplicity and streamlining than I had managed. Everything in Savage Worlds was pared down to the minimum; by comparison, PowerFrame had begun to suffer from rules bloat.
Savage Worlds' resolution system appears to be much quicker - a single roll (two dice, pick the highest) against a static Target Number (usually 4, unless you're fighting in close combat or inflicting wounds) - this compared to PowerFrame, where in a contest both sides roll one exploding dice and add their Ability Level, with the defending roll setting the difficulty for the active character (or using static Target Numbers when there's nobody in direct opposition). Both systems apply modifiers, often count successes based on the margin, have exploding dice (although PowerFrame dice only explode once), and include a way to reroll the dice.
Another of Savage Worlds' "draw-cards" is its initiative system, where cards are dealt from a standard playing card deck to each player or group of NPCs. With a method for splitting ties between suits, there's no chance two people will end up on the same initiative - which admittedly, can take a little while to deal with in PowerFrame. However, while Savage Worlds deals initiative each turn, everyone in PowerFrame rolls once and then cycles around the list, so comparing the two systems is swings and roundabouts. Savage Worlds does have a slight advantage in that it can be a waste of time to set up an initiative list in PowerFrame only to have combat last just a couple of turns.
One thing both games have in common is a number of detailed and defined combat options. Savage Worlds even manages to keep non-combat-specced characters involved with Taunts and Tricks, which is something PowerFrame does not mechanically define - it's up to the GM to adjudicate any such attempts.
I haven't had a good look through the Powers chapter, although Rohin said it seemed to have some fairly large gaps. I guess the corebook just has a stripped down list, and you'd find more comprehensive Power sets in setting books or supplements like the Fantasy Companion; I haven't investigated those options, though.
I did find it a little difficult to gauge the power of NPCs; there's no real guidelines apart from "don't use the character system, just make them up!" which is all well and good if you're experienced enough to eyeball things. With just the corebook, though, there were precious few example NPCs. I guess the cross-genre nature makes it hard to come up with definitive examples, but even a rough idea of what a cannon fodder enemy, challenging enemy, and boss enemy looked like might have helped. It's also tricky to judge how the creatures will fare; I ran a small test with +Paul, whose monster-hunting priest nearly got taken out by two zombies because he couldn't pump out enough damage to keep them down! A second test against mostly d4-rated Nazis went much better.
One thing Savage Worlds has that PowerFrame doesn't is its lists of Edges and Hindrances - a system of advantages and disadvantages that is fairly integral to the characters. In fact, it's assumed that all characters will take their full complement of Hindrances, thus allowing them to buy advantages including unique Edges. Much like the Stunts in Spirit of the Century, these represent a large amount of up-front information that needs to be digested by each player before they can make fully informed choices about their character. The only comparable thing in PowerFrame is the Ability list, which is admittedly quite long, and which does contain Abilities with specific game effects... Maybe it's just familiarity that makes my groups comfortable with PowerFrame, but resistant to learning large amounts of stuff for a new system? Once the effort's been made to learn one system in detail, it can seem like unnecessary work to repeat that process with another system.
I was curious to see how Savage Worlds performed in actual play compared to PowerFrame, since the systems shared so much in common yet used different approaches. During Easter 2012, I picked up a hardcopy of the Interface Zero setting, and Rohin fell in love with it. He began to plan a campaign (Interface Zero: Las Vegas), and I suggested I run a brief test game (Scavengers of the Sunken City) so the group could test out character creation and get used to the basics first.