Archive for the ‘Algorithms’ Category

Rigging Amazon Flex’s gig economy algorithm

Monday, February 10th, 2020

Using a device you can buy on Amazon! Amazon Flex gig takers get crap money to do vicarious work for the 2nd biggest company in the world. To get more gigs they figured out how to game the gig distribution algorithm.


Hacking the Google Maps algorithm

Sunday, February 2nd, 2020

It’s quite simple. Pull a handcart with 99 phones with Google maps loaded on them through a street. Google Maps now thinks 99 cars are slowly moving in that area and will direct people around to another less congested area.

Google Maps

Manipulating the algorithm to get hired. Faking it with robot recruiters.

Wednesday, January 29th, 2020

South Korea of course is using A.I. to recruit people and get rid of the recruiter. Cram schools are now giving lessons on what to say and what facial expressions to use in order to fool a system designed to replace humans who hire humans and judge these humans on their natural expressions. Confused? The same way we change our accent and voice stressors to have Alexa and Google Home understand us, will we start changing out expressions to keep the robots happy?

One AI video system reviewed by Reuters asks candidates to introduce themselves, during which it spots and counts facial expressions including ‘fear’ and ‘joy’ and analyses word choices. It then asks questions that can be tough: “You are on a business trip with your boss and you spot him using the company (credit) card to buy himself a gift. What will you say?”

A.I. predicts lightning strikes. What could go wrong?

Tuesday, January 28th, 2020

A.I. predicts time and place for lightning strikes based on weather data. Which means lots of scientists are going to be creating devices to try and store the energy for these strikes. That won’t go massively wrong at all. It’s aliiiiiiive!

Rigging the Christmas number 1 – Alexa, play Christmas songs

Wednesday, January 1st, 2020

This reminds me of when Apple put a U2 album on everyone’s phone.

“Alexa, play Christmas songs”

Amazon put an Amazon exclusive song from Ellie Goulding on all their Christmas playlists. Those streams counted in the official charts plus it helped sales of 78,000 singles and it got to number 1. Loads more detail here. But that makes me think, could this be done again for everyone with a smart speaker?

Amazon rigs the music charts

Songs are getting shorter because of the Spotify payment algorithm

Tuesday, December 31st, 2019

Using the algorithm to make more money. via Quartz

Payments from music streaming services like Spotify and Apple Music made up 75% all US music revenues in 2018 … Streaming services pay music rights holders per play … Every song gets paid the same. Kanye West’s 2010 five-minute opus “All Of the Lights” gets the same payment as West’s two-minute long 2018 hit “I Love it”.

The amount of filler or interludes on some albums is crazy. But yeah if they are paying per song played and the money is shite, then why not have an album with 20 tracks instead of 10 and get paid double per album play?

Gaming Netflix’s recommendation engine

Monday, December 30th, 2019

via Hollywood Reporter

Norwegian Viking Comedy but done with 2018 dialog. It’s good too.
Netflix not really into spending money on marketing for acquisitions, how do you make it a hit?

The key to landing on Netflix’s radar, he knew, would be to hack its recommendation engine: get enough people interested in the show early. Then, hopefully, Netflix’s mysterious algorithm would do its thing.

with additional pushes in Minnesota, Wisconsin and South Dakota, three states with large ethnic Norwegian populations.

Marketed the show to specific groups of people on Facebook ahead of the launch. Tested which ads worked. Spent $15k on the ads. Built momentum. It worked.

Netflix now wanted it to be an original and invested in pushing it. Kind of like force multiplication.

I’m reminded of a movie Danny Devito directed and was in called The Ratings Game where they rigged TV ratings to create hit TV shows.