885 karma · joined August 12, 2011
I do think there’s a difference between an Alexa and my employer getting 24 access to my mic. By being in more people’s houses Alexa is much more likely to be detected if doing something nefarious, and the choice to risk it is entirely up to me. Find a new job is much harder than the alternative to voice command
Did a bunch of research, this isn’t covid specific, some huge percentage of people in the icu go through this but it’s not publicized nearly enough. The worst part is it’s a known thing, but there’s zero mental health support coming out of it. Especially since my sister went to rehab, they just weren’t staffed for it but couldn’t get her to people who were until the rehab ran it’s course.
Let's say you see 75% bot traffic. You make $10 per actual user. If youre spending $2 per action with 4 actions, you burn $6(75%) on bots, $2(25%) on a real user($8 total), and get $10 - netting $2. So you're making money, but wasting more.
So, let's say you set your targeting to completely exclude the bots. Great, you spent $2 and made $10, netting $8. The problem though is you ruled out 90% of your traffic - you have false positives on whether its a bot so you exclude legit users(all of this targeting is fuzzy rule based). You can 10x the traffic with bots, making $20, but you cant 10x the perfect targeting, so youre stuck with $8.
But occasionally the “oh f#%^” report comes in...
Opentsdb for instance was built on top of Hbase because implementing one naively in hbase hits tons of performance issues.
It really isnt fair, I don't know employee who actually considers it a benefit. It's like a contract, just write it down so everyone knows what to expect.
But I've subscribed to some food blogs because the recipes are decent and the stories make it interesting - otherwise, a series of recipes in my rss would just be incredibly boring. In this case it's closer to a tv personality you'd watch on tv than a recipe book.
The one that did have qa, the devs were incredibly lazy about testing, relying entirely on someone else to make sure it worked. I’d say the overall product was better with qa but the point at which a developer called it done was way worse.
But a surprising number of people will pay for instant transfer(almost all of that 300m probably comes from this), and the debit cards they get to take in a portion of the merchant transaction fees. They’ll probably be profitable off that, with p2p as a loss leader to drive installs
edit: oh nm i only noticed the ui, not that it totally broke typing
But it was also twelve months of side project time, not full time