[discardorama]

> My life is a failure for not having thought of this.

Actually, I did think of this, many times. But my biggest bottleneck was: how will I scale up the replies? I could handle being a "boyfriend" for, say, 10 women. But any bigger, and I'd need help. I considered MTurk, but thought that quality control would be an issue. (What if the MTurk guy really starts hitting on the woman, they exchange numbers and then he starts stalking her?). Anyways: after considering all the messiness, I gave it a pass.

[enjo]

Anyways: after considering all the messiness, I gave it a pass.

There is a lesson here. "Computer people" have a real aversion to simply scaling up a business by using humans to do things. Sometimes that simple answer (hire a bunch of people to do stuff) is the right answer, provided the economics of the business work.

[vehementi]

What? discardorama was merely wrong - the real service DOES use the approach they considered!!

[dgreensp]

Same here. Actually, I didn't think of it as a "proof" service, I was thinking that bits of positive human interaction ("invisible friend") provide a lot of emotional value, and you could crowd-source a text conversation with a fake friend, modulo quality control, and people would pay for it.

[briholt]

No, this would be so easy to code. Enter 100 basic phrases:

You look so [appearance adjective], my [affection term].

Hey [affection term], I'll see you at the [social location] tonight.

... and so on

Then enter possible variables:

affection term = [baby, honey, sweetheart]

.. and so on

Then buy 100 stock headshot photos and generate a random list of first/last names for each. Set up a FB/email account for each.

Then integrate with Twilio for SMS sending. Write a little cron script to send 1-2 messages to each user per day.

Done. A week's worth of work to launch.