Seems like an intriguing product, though certainly with the potential to be "creepy" if executed in the wrong way. I was struck by an implementation paragraph though:
> Every major component of Refresh is a fuzzy system. When it sees “Lunch with Tom” – without exaggeration 100 bots go out and vote who Tom is. One says “You have a Facebook friend named Tom”. One says “You met with a Tom person 3 months ago at the same location.” One says “It’s not your cousin Tom — you hate him.” Some are heuristic and some are statistical.
Putting my grad-student hat on, from a Bayesian perspective, they seem to be effectively "marginalizing" the distribution on "facts" over the hidden identity-of-Tom variable given the event and your current network:
The app's output seems to be, effectively, samples from that "facts" distribution, which can be generated by taking samples from the identity distribution and accepting/rejecting based on likelihood. It's a more formal way to look at fuzzy logic systems like this one. The cool thing is that you can eventually plug other things into the system, such as taking into account the distribution over types of facts that the current user likes to look at, in a mathematically sound way. But it's rare (imho) to come across systems like this outside of academia.
It pings me about 5-10 minutes before I meet someone with a dossier on them. It includes highlights from social media, things we have in common, and our last interactions. Very helpful in catching back up where you left off.
One feature I like is that once the meeting is over, the app asks if you want to remember anything about them and gives you notes field to edit. This shows up the next time you meet them.
It's gotten REALLY good at highlighting interesting things about over the last few months.
All in all it's very impressive technology and I look forward to the push messages. It's hard to say that about many apps.
Every once in a while you meet someone who's really good at remembering not just your name but the little details of your life, your interests, what you last talked about, etc. It's either a gift or the person is disciplined enough to keep notes and review them before you next meet. Either way it's impressive.
Now, with an app like this, it will cease to be anything remarkable. Yeah, someone you met two months ago seems to remember everything like it was yesterday, eh big deal, he's just using Refresh.
I don't really know how I feel about it, but it seems like it's cheating somehow.
I am one of those people. It's neither a gift, nor a product of discipline; I just remember trivial shit about lots of people that I have ever met (but certainly not everyone I have ever met), all the way back to the 4th grade.
This app will help the forgetful types, but it won't win over people who have an innate ability to remember some utterly trivial facts. IIRC, the majority of successful CEO's share this trait...try to pull Refresh on them, and you'll find out how much closer you got to dog house than if you just kept your trap shut.
> I don't really know how I feel about it, but it seems like it's cheating somehow.
I used to feel the same way about phone numbers: before cell phones arrived, I prided myself in knowing the phone numbers of all my friends by heart, but these days, I can barely remember one more phone number besides my own.
I felt a bit bad over this for a while, until I realized that I preferred to use my memory for things that are more meaningful.
Way too creepy for me. I'm (approximately) Cringely's age, and it doesn't seem to bother him, so maybe it's not an age related thing, maybe an introvert vs extrovert thing.
What do the 20-something hipsters on HN think of this? Are you OK with computers becoming more and more adept at mining your personal information and relationships? Of course Google, Facebook, Twitter, et al. already do this. But they (pretty much) keep the results to themselves. They don't shove the mining back in my face.
Seems very wrong. For some reason, especially the mention of "meeting" the girl at the bar, and of course, "befriending" people on this basis seems wrong overall. I don't think these are really friendships. Seems to be a business world thing.
I find it so creepy, I might go out of my way to post hilarious fake information about myself. Like how I baked gummy-bear pie for my niece, and how my stocks in BLAH something were faring. This way I would know who was doing this to me. I would disdain them... If i were "important" that is =)
Likely it doesn't bother Cringely because he's such a public figure. Like most celebs, I bet he gave up trying to hide the vast majority of his life anyway.
A scan of his Wikipedia page [1] gives me more information than I know about the vast majority of people I interact with on a daily basis.
Based on my use so far, it only displays information that the user has access to, so I don't see anything creepy about it.
What's creepy (in a good way) is how they manage to pick up the things that are actually interesting and relevant, which I often miss since my friends' walls are so busy already.
My view is, if you don't want your data out there... don't put it out there. If it's out there, don't be surprised to see it mined and used :) Especially for people that use consistent handles across the internet.
> Every major component of Refresh is a fuzzy system. When it sees “Lunch with Tom” – without exaggeration 100 bots go out and vote who Tom is. One says “You have a Facebook friend named Tom”. One says “You met with a Tom person 3 months ago at the same location.” One says “It’s not your cousin Tom — you hate him.” Some are heuristic and some are statistical.
Putting my grad-student hat on, from a Bayesian perspective, they seem to be effectively "marginalizing" the distribution on "facts" over the hidden identity-of-Tom variable given the event and your current network:
$ p(facts | network, event) = \int p(facts | identity) p(identity | network, event) d identity $
The app's output seems to be, effectively, samples from that "facts" distribution, which can be generated by taking samples from the identity distribution and accepting/rejecting based on likelihood. It's a more formal way to look at fuzzy logic systems like this one. The cool thing is that you can eventually plug other things into the system, such as taking into account the distribution over types of facts that the current user likes to look at, in a mathematically sound way. But it's rare (imho) to come across systems like this outside of academia.