Remember Tomorrow (hereafter called RT) has many of the same strengths that 3:16 does - it's easy to learn and remember, distributes narrative power across the table, reduces the GMing prep burden to nil. In fact, you could argue that those strengths are actually amplified in RT. It's better at what 3:16 did well than 3:16 itself was. In our first session of RT, we told a wild, unpredictable story, and only once had to look anything up in a rulebook. Bravo!
However, RT has many of 3:16s flaws as well - limited differentiation between characters, mechanics that are so grainy you can't apply any modifiers, the occasional pacing and structural issues caused by unrestricted narrative rights, rules that encourage Player-Vs-Player conflict but don't really reward it enough to make anyone comfortable getting aggressive - and these flaws too are amplified beyond their presence in 3:16. RT is definitely drawn from the same well as 3:16, but then adapted to a less-structured genre. It's everything you loved or hated about 3:16, dialed up to 11.
The lack of setting can be a source of confusion, such as when two different people disagree about what it means to be a "rogue AI", and how they can interact with the world. The lack of solid GMing advice or definitive guidelines, the competitive elements, and the scene-by-scene narration rights rotation can combine to make for a very uneven pacing and sometimes a lack of fairness. The conflicts caused by setting vagueness are not insurmountable, especially if you're just using it for a one-shot or short-run mini-campaign. Just the same, they render it probably a better game for use with friends you have an established rapport with, than for dumping on a group of strangers at a convention.
We encountered these lack-of-guidance stumbling blocks in our group when one person kept starting all his scenes with one of the other PCs (the rogue AI mentioned above) yet again trapped inside another unfamiliar super-computer. Eventually, those sorts of issues would sort themselves out, but it was definitely frustrating in the short term. We approached the game from a point of view of generic cyberpunk future, and so some of us narrated things as gritty and dark, while others went for sufficiently advanced tech and a more lighthearted cinematic approach. In hindsight, I think the game would play much better if the group agreed up front to use a setting they are all familiar with, such as CP2020, Max Headroom, a specific series of novels, or The Matrix. That way everyone's on the same page about what you're after thematically, and you have a reference point for what the tech in the setting can do. The other half of it, learning how to play fair in a game with no GM, is just a trial-and-error issue. (And though it bugged me to see the same "trap" again and again, what else can you do to challenge a rogue AI?)
A few other minor issues were caused by the non-traditional structure. Most notably, there was a particular type of scene that was conspicuously absent from RT. When one PC wants to talk to another PC, that's generally trivially easy to arrange in any RPG. Unless it's the middle of a fight scene, or one PC has been taken captive, the players have it entirely in their power to start up a conversational scene between their characters, often even in campaigns where GMs keep a tight reign on the narration rights around you. In RT, surprisingly, this paradigm is inverted. Though the rules really don't draw attention to it, it is in fact nigh impossible for two PCs to even communicate, let alone work together in RT. I can easily start a scene where one PC attacks another, or where a Faction attacks a PC, or where a Faction contacts a PC and strikes a deal with them, but I am implicitly forbidden to stage a scene where my PC contacts another PC and tries to help or team up with them. There are color narration scenes, but as written those must begin with the intention of turning into conflict, and only become peaceful scenes by means of a happy accident when no one escalates. Which means that only those players most inclined to abuse rules-systems or metagame are even capable of teaming up. Of course, if the other player misunderstands you attempt to subvert the system, you could easily end up in a combat between your PCs, which makes people not want to risk trying to interact with the others.
As a result, one of the players at our group compared the game to multiple solitaire, and that wasn't an entirely unfair comparison. Indeed, the momentum always lies with the Factions within the game, not the PCs, which is a bit awkward. PCs feel less like protagonists and heroes, and more like the poor dumb schlob that everything happens to. Even when the PCs win the die rolls, they still have a somewhat passive role in the game, since their scene goals are usually something along the lines of "don't get shot". Worse yet, the structure feels terribly counter-intuitive, as though you aren't actually playing the character you're holding, instead you're just waiting for the next threat to rear its head. The rules as written render the PCs incapable of actually pursuing anything, and they instead just have to roll with the punches. That seems like a house-rule might be in order:
A Very Simple House Rule: This rule adds an additional scene type: Dialog Scenes. You may choose your PC and any other PCs or Factions, played by the appropriate other players. You play this scene out as a conversation and light interactions. Unlike normal Deal or Face-Off scenes, Dialog scenes give no mechanical rewards or penalties, and involve no die rolls. They exist solely to facilitate the interactions between PCs, to allow for a bit of flavorful characterization, to illustrate (but not advance upon) your goals, or to function as a "pass" if you're stuck for a better idea when it's your turn to frame the scene. Color | No Conflict scenes may still be initiated in the normal way, by having a Face-Off scene peter out. Face-Off scenes that result in no conflict still get their usual rewards, as the risk of conflict was present.
That simple house rule (and a clearer discussion up front about the setting) would solve most of my objections to the game, and allow it to run a lot more smoothly and intuitively. It would then function as the quick-to-pick-up, easily improvised alternative to CP2020 or Shadowrun, a role at which it would excel. If I wanted to run CP2020 next week, I'd have a lot of reading and planning to do between now and then, but if I wanted to run RT in the next 5 minutes I'd have a one page summary to read and the other 3 minutes of my prep would be breaking open a pack of blank index cards. In this hectic world where we don't always have as much time for gaming as we'd like, light-weight games like RT are a real boon.