They mentioned the reason they weren't able to know was because they only kept the logs for a short amount of time.

It's just ridiculous that they had it open for 3 years, didn't have proper metrics so don't know exactly how many were breached, but there estimate was 500,000 and they didn't disclose???

C'mon Google, why'd you have to try and hide this...

Just to add my experience, the demo worked perfectly straight away (Chrome 70 on macOS). Really cool to see this tech, hope it's easy to integrate.

thanks a lot for the feedback :)

Take a look at this standalone demo if you like https://github.com/Picovoice/Porcupine/tree/master/demo/js

it should be a matter of tweaking a few parameters.

Web demo on your main page worked great for me at first, but stopped working after about half a dozen commands.

That's odd. If there are errors in the console. Please open an issue in the GitHub repo is possible: https://github.com/Picovoice/Porcupine

Otherwise, I am thinking it could be something related to the microphone?

That setting will only appear if you have a 15" with a dedicated GPU. 13" only have "Energy Saving" integrated graphics and don't need the setting

Android 6.0 was the first version to implement permission control. Looks like this guy was running 5.1, so we at least know how Facebook got access to that info.

Doesn't justify it though.

Android always had permissions:

https://developer.android.com/guide/topics/manifest/permissi... (see "introduced in")

Many apps request the phone permission to change their behavior on incoming calls, or during calls. For example, they would stop playing audio.

Access to text messages is sometimes requested to automate text message-based one time pad flows for the user.

By permission control they mean: https://developer.android.com/training/permissions/requestin...

>The infiltration likely lasted more than a year.

Wow, that's pretty significant... How was it only now they caught on?

There are reports telling that the attackers were observed for some time and potentially got honeypots with insnsitige data served.

However there are no formal reports published, so mostly rumors and speculation. The parliament's security council should receive a (non-public) report soon, only after full analysis there might be some public information released ...

Infosec on federal and state level is generally regarded as very poor. Hardly unique to Germany, though.

I would not say it is uncommon though. AFAIK intrusions into major companies like Google also happen to stay unnoticed for months.

I'd argue however that Spotify's song recommendation engine is one of its differentiating factors. I'm always hearing friends rave about how much they love each weeks discovery weekly playlist.

AFAIK all of the major players have this feature and even developed it a while ago :

Google Music have been displaying some kind of 'I feel lucky' button for ages (it is not random but based on your listens). They also had artists maps years ago.

Deezer has its Flow

Apple Music has 'for you'

Etc ..

There are many variations in how these recommandations are surfaced to the user though.

I would be hard pressed to recommend one service over another as far as their quality. It is just all over the place depending on your tastes, listening styles and how much new tracks you want to listen to.

IMO Spotify has mostly found a sweet spot in how to present these results to the user.

Spotify has years of my listening data and preferences, other services don't. So if I compare Spotify's recommendations to others, it will always be better, and I'd have to put in a lot of hours into another service for it to catch up.

The Apple Music 'for you' section was something I never really bothered to check. I found the recommendations were poor. I love Spotify's auto-generated playlists. They have a slight problem where if you decide to listen to a new genre for a couple days it suddenly assumes that's all you want to hear, but that's a minor issue

Spotify's is based on machine learning working on the actual audio tracks.

It seems to me that Google Play Music's recommendations are based on what artists you like, based on machine learning, but not using the actual music, just pattern matching with other people.

I use Google Play Music, and am actually considering trying Spotify to get better recommendations.

Do you have sources on this ?

Always interested to read more about how these recommendations work.

This is the article I read that talked about it. Sources are a the bottoms, and they seem legitimate.

https://hackernoon.com/spotifys-discover-weekly-how-machine-...