Crystal promises to help you understand how best to talk to any person
wired.com
wired.com
That probably won't stop businesses from wasting truckloads of money on their own automated soothsayer, though.
Edit: Make sure you consult their physical terminal before emailing that big VC! http://imgur.com/UMIlcXz
On the other hand, I think it is a good idea to have a sentiment analysis system in place that advices users on overly negative sentiment in their comments or e-mails. I wish there was one already integrated into YC comments and into Gmail. Potentially can make life quite a bit nicer actually.
P.S. like your wording 'automated soothsayer'. Funny that the next logical step would be wasting money on 'automated soothslayers' ;)
- Give some subjects the advice from this program, and some subjects advice from a knowledgeable co-worker of the person in question (PiQ). (Make sure they don't know which the advice is coming from.)
- Find some way to measure the "success" of the subjects in question -- maybe have them and the PiQ give a subjective rating of the interactions, which should be good enough.
- See if there's a significant difference between the two.
If all such advice is unhelpful, or this service is no better than astrology, you'll see no difference.
Crystal does have a couple cool things going for it, though - for instance, they're in an excellent position to improve user writing quality by presenting standard stylistic advice (don't use passive voice, don't re-use words, etc) as advanced AI clairvoyance.
To make an analogy, imagine finding a room full of un-labeled bottles of chemicals, and an inventory of chemical names. Could you match the chemicals to their names?
Not quite. This would mean that the personality advice is unrelated to the personality of the target. But it could still be good advice! It's quite possible that a machine that always gave you the same advice, no matter who you were writing to, would still improve your results. (What if the advice was "Are you drunk? Are you sure you want to send this?")
If you join a dating site and fail, they lose one customer.
If you join a dating site and are successful, they lose two.
The goal became determining good but not great matches and notify the user regularly, especially if they hadn't logged in for a while. It was basically a game of giving false hope to get them to keep their accounts active and the dollars flowing.
I bailed a couple weeks after I realized this.
Wired writing about new tech products isn't creepy PR infringing on journalism, it's why people read it in the first place.
Crystal gives companies a new and more powerful tool for spreading their messages. Of course the first use is always innocent, but if this tech caught on (I can imagine in some form it will) then there is a lot of potential here for PR firms to interject their messages.
EDIT: I should add, I don't see Crystal as bad. I see it as inevitable. People have been blindly sharing private data on the web for a decade because no one used it? People will start using it.
Thereby freeing us from the drudgery of writing emails to each other. We'll have more free time!
</sarcasm>
well, another, say Quartz, app would read the message and decide to write back through for example the same Crystal. Complete automating of social life of automatons. While they are at it, we can socialize on our own...
Considering the number of bots on Facebook, and the number of humans who communicate with strangers in txspk, it's possible this has already happened.
I can see an inverse business opportunity - a tool to certify that your communication is 100% brain-made and authentically genuine.
Maybe artisanalemail.com?
To be fair, I think it's more aimed at more casual acquaintances whom you haven't gotten to know as well yet.
> EDIT: I should add, I don't see Crystal as bad. I see it as inevitable. People have been blindly sharing private data on the web for a decade because no one used it? People will start using it.
That's nothing - I've been intentionally, publicly, sharing how to effectively communicate with me for a decade - and information about how I communicate.
I can see the ick factor in giving e.g. your local neighborhood spambot this power - but I want to effectively communicate with people. I want to be effectively communicated to. Hell, I even want to be effectively advertizied to - it means they've picked up on how to avoid annoying me and serve my interests (e.g. I click a lot of webcomic ads.)
I have had so many conversations in so many chatrooms where conflicting or just plain ineffective communication styles lead to frustration on one end, or hurt feelings on another. I extol the virtues of reading various articles about effective communication - "How To Ask Questions The Smart Way" [1] and "What have you tried?" [2]. I've even written a few myself to explain my at times abrasive communication style - extoling it's virtues, defending my choice, and just plain explaining my rationale.
It's a great book where someone basically develops Crystal except it can compose messages too. Then it realizes that humans don't have to be in the loop at all..
BeautifulHandwrittenLetters.com
When you email ian cuss like a sailor and spit a lot. Challenge him to a fight in the first three sentences. Tell him star wars > star trek to throw down the gauntlet. As him about his mother.
So if you email me or anyone else who has done that it will I assume incorporate that.
Anyone aware of other stuff like this or want to shamelessly promote something? I'm interested.
It seems a bit miscalibrated though. I tried pasting in the current front-page article from The Guardian (http://www.theguardian.com/world/2015/apr/24/obama-drone-str...), and it becomes almost completely red ("27 out of 37 sentences are very hard to read"). If it categorizes ordinary newspaper prose as "very hard", I wonder what kind of audience it is aimed at.
Yes, the app is apparently quite insane.
It's classic Strunk & White nonsense, the idea that simpler (even to the point of being simplistic) is always, always, always better, founded on the unfounded idea that removing words is the only way to linguistic correctness.
http://chronicle.com/article/50-Years-of-Stupid-Grammar/2549...
The augmentation shows you the personality type of the person you're talking to and suggests the best way to get them to do what you want, cajole/reason/appeal to authority, etc.
These are great. Anyone know of any other similar parody apps?
I'm intrigued by whether the advice it offers focuses on mimicking their own communication style of whether it gets a bit more clever and human, and advises people (e.g.) to phrase things in a very deferential manner to somebody who gives the appearance of pomposity.
Better docs: http://www.ibm.com/smarterplanet/us/en/ibmwatson/developercl...
Free while in beta....
This feels strange but in the movie Avengers, Ultron, an AI algorithm, is born with extraordinary powers that can only be described as Palantir on steroids. Palantir itself is doing amazing work and solving really hard real world problems for Enterprises. Crystal (and others like refresh.io, bright.com, relateIQ) are doing it for the end user. And its pretty evident that something like this has a massive market potential (judging by acquisitions). With the amount of data flowing in the emails, on the social networks, search engines... it must be entirely possible to judge someone with unerring accuracy (the proverbial holy-grail for advertising systems). Mining of information from these signals could help solve a variety of problems-- like helping you find a recruit, a co-founder, a roommate, a life-partner, a friend that shares interests and likes, help create communities and then expand or evolve them, teach you empathy, help you write better, help with thinking... I can see something like this being central to social networking in general. I think Palantir and other companies have demonstrated that these and other similar problems are very much solvable.
Looked interesting in an abstract kind of way, but can't be arsed to play hard to get. Anyone got documentation on these "64 personality types" and the underlying theory?
Spell checking is an indexing and comparison task. Crystal is saying that she's better than I am at social tasks.
Could that be true? Maybe the Turing test passes not by machines becoming more socially intelligent but by humans becoming less so.
I stopped reading shortly after he let the cat out of the bag that that wasn't his actual response to the stated story.
so - worked on me!
Very impressed.
I wonder how it works