Google's Balloons Connect Flood-hit Peru (bbc.com)

An anonymous reader writes: "Tens of thousands" of Peruvians have been getting online using Project Loon, the ambitious connectivity project from Google's parent company, Alphabet. Project Loon uses tennis court-sized balloons carrying a small box of equipment to beam internet access to a wide area below. The team told the BBC they had been testing the system in Peru when serious floods hit in January, and so the technology was opened up to people living in three badly-hit cities. Until now, only small-scale tests of the technology had taken place. Project Loon is in competition with other attempts to provide internet from the skies, including Facebook's Aquila project which is being worked on in the UK. Project Loon recently announced it had figured out how to use artificial intelligence (AI) to "steer" the balloons by raising or lowering them to piggy-back weather streams. It was this discovery that enabled the company to use just a "handful" of balloons to connect people in Lima, Chimbote, and Piura. The balloons were launched from the US territory of Puerto Rico before being guided south.

Comments

It was us...

[SuperKendall]

Project Loon is in competition with other attempts to provide internet from the skies, including Facebook's Aquila project which is being worked on in the UK.

We don't know who struck first, us or them, but we know that it was us that darkened the sky beyond redemption with millions of wireless nodes.

Something about this story

[Anonymous Coward]

sounds like a lot of hot air to me

Loony idea

[fnj]

The concept of freely-floating out-of-control balloons serving any purpose is downright loony.

Re:Loony idea

[swillden]

The concept of freely-floating out-of-control balloons serving any purpose is downright loony.

I saw a presentation by the Project Loon team, and the project leader made the same point. He said that the idea was so loony it had to fail, but they kept failing to find the reason. The Loon team has spent years failing to fail.

That is the Google X methodology, BTW, "fail fast". Find a bizarre idea, think about the reasons it can't work, starting with the most likely to fail, then test to see whether it actually fails for that reason. If it doesn't, move on to the next most-likely reason for failure, and so on. Eventually (so the theory goes) you've either proved the idea doesn't work, or you've exhausted all the reasons for failing... by succeeding.