[bambax]

> Already, some apps do this by learning patterns in who we left and right swipe on, the same way Netflix makes recommendations from the movies we’ve liked in the past.

Netflix recommendations are utterly bad.

Musk et al. warn us constantly about "AI" and its dangers, but so far AI doesn't seem to work very well. Translating from Italian to English with Google Translate results in complete gibberish; Amazon (the other recommendation engine everyone mentions in that kind of discussion) is only ever able to suggest products one has already bought.

It's probably hard to judge the efficiency of those dating systems since there is no control, and people want them to work. It would be interesting to compare a "sophisticated" dating engine with a more random one and see if there's any difference in outcome.

[mattnewport]

Yeah, I can't think of any service that appears to actually do anything useful to me with all the data they've been collecting. Netflix recommendations are awful (one reason I recently canceled my account), Amazon recommendations are pretty hit and miss, YouTube makes some good suggestions but they are dominated by channels I'm either subscribed to or have watched a lot of recently. OkCupid for all their talk of data analytics seemed to have very questionable match percentages. Facebook fails horribly at putting interesting things in my feed. For all the hype around big data and AI there's a serious dearth of really good applications.

[MrMember]

When Netflix really started to ramp up production of original series I saw a large spike in recommendations that just happened to be Netflix originals. It was pretty clear that their recommendations had nothing to do with what they thought I might like.

[p-funk]

> so far The dangers of AI are theoretical ones that will show up if AI reaches somewhere close to human levels of competence. The fact that it doesn't work well right now should be compared to computers speed today vs computers not being very fast in the 1960s, for example. I'm not totally sure about the "singularity" concept. But I also think today's level of sophistication should not mean any and all concerns about AI should be dismissed outright.

[xirdstl]

Yes, I don't see the "brutally effective" part. It's just clickbait.

[frgtpsswrdlame]

Musk's hype of AI is just marketing.

[chc]

Musk's "hype" of AI? He says we should be scared of it.

[qbrass]

He started a company to research making friendly AI, so he wants you to be scared of AI that isn't his.

[frgtpsswrdlame]

Yes, he says we we should be scared of it because it will be so powerful in the near future - it won't be but the idea that extremely powerful AI is just around the corner is really good for his company.

[p-funk]

Musk's hype of AI is his response to reading Nick Bostrom's Superintelligence.

[basch]

not really on topic, but does https://www.taste.io work for you?

i always like to bring this up when people mention algos only being able to shuffle the past - http://www.bbc.co.uk/blogs/adamcurtis/entries/78691781-c9b7-...

i would think music or photos or writing would be the first place where algos could take WHAT you like about something and find matches. Just liking or upvoting isnt enough in my experience. But with music if I can say I like the lyrics, or beat, or instruments or vocal tones..

[rndmize]

I'll be honest - I think that Amazon's recommendation engine is deliberately bad. There was an article a few years back on Target and their efforts to track/manipulate customers: http://www.nytimes.com/2012/02/19/magazine/shopping-habits.h... . It's fairly long, so I'll quote a few parts:

> One Target employee I spoke to provided a hypothetical example. Take a fictional Target shopper named Jenny Ward, who is 23, lives in Atlanta and in March bought cocoa-butter lotion, a purse large enough to double as a diaper bag, zinc and magnesium supplements and a bright blue rug. There’s, say, an 87 percent chance that she’s pregnant and that her delivery date is sometime in late August. What’s more, because of the data attached to her Guest ID number, Target knows how to trigger Jenny’s habits. They know that if she receives a coupon via e-mail, it will most likely cue her to buy online. They know that if she receives an ad in the mail on Friday, she frequently uses it on a weekend trip to the store. And they know that if they reward her with a printed receipt that entitles her to a free cup of Starbucks coffee, she’ll use it when she comes back again.

<...>

> Pole applied his program to every regular female shopper in Target’s national database and soon had a list of tens of thousands of women who were most likely pregnant. If they could entice those women or their husbands to visit Target and buy baby-related products, the company’s cue-routine-reward calculators could kick in and start pushing them to buy groceries, bathing suits, toys and clothing, as well. When Pole shared his list with the marketers, he said, they were ecstatic. Soon, Pole was getting invited to meetings above his paygrade. Eventually his paygrade went up.

<...>

> About a year after Pole created his pregnancy-prediction model, a man walked into a Target outside Minneapolis and demanded to see the manager. He was clutching coupons that had been sent to his daughter, and he was angry, according to an employee who participated in the conversation.

> “My daughter got this in the mail!” he said. “She’s still in high school, and you’re sending her coupons for baby clothes and cribs? Are you trying to encourage her to get pregnant?”

> The manager didn’t have any idea what the man was talking about. He looked at the mailer. Sure enough, it was addressed to the man’s daughter and contained advertisements for maternity clothing, nursery furniture and pictures of smiling infants. The manager apologized and then called a few days later to apologize again.

> On the phone, though, the father was somewhat abashed. “I had a talk with my daughter,” he said. “It turns out there’s been some activities in my house I haven’t been completely aware of. She’s due in August. I owe you an apology.”

Now my memory isn't great, but I think there was a fair amount of uproar of this and Target scrapped the program. There was a line crossed between "useful" and "creepy" somewhere along the way, and it would be my guess that Amazon seeks to avoid that. I'd fully expect them to have the data and engineering capability to create a very powerful recommendation engine (the Target article is five years old now...) - but the expected value isn't there.

[bambax]

Ha! Never crossed my mind that Amazon's recommendations would be deliberately crippled... It's possible, although, I think, unlikely.

But what about others? What would be creepy about excellent Netflix recommendations for example?