Facebook’s Newsfeed Has A Friendship Problem

Blueberry banana overnight oats. Recipe HERE

Via foodfitnessfreshair.com

Last Tuesday night, Shani Hilton — head of U.S. news at BuzzFeed, apparent breakfast lover, and my Facebook friend — posted a simple status: “Has anyone made overnight oats before?&;

Shani’s quest for sensible breakfast advice and its string of replies has remained at the top of my Facebook feed for a week now. Every time I check Facebook for whatever the reason I compulsively check it multiple times a day, the overnight oats post glares back at me.

I never commented on it. I didn’t “like” it. I don’t even like overnight oats. I have no oat advice. This is not a topic I care about.

The intricacies of overnight oats had ground my mental state to the point of madness.

On Sunday, a full five days after Shani had posited the oat query, the oats were still at the top of my Facebook feed. They were even beating out an engagement and a pregnancy announcement — the gold standard of “sticky” Facebook content. I couldn’t take it anymore. The oat talk had taken over my life. Chia seeds, maple syrup, blueberries, whether to use regular or almond milk, the great debate over Greek yogurt. The intricacies of overnight oats had ground my mental state to the point of madness. And honestly, I still don’t really know what overnight oats are (are they different than oatmeal? Do they actually take a whole night?).

You know that thing where you say a word a bunch of times until it sounds really weird and not like a word at all? Oats. OATS. Ooooaaaatttssss. Oats. Fucking crazy, man.

This summer, Facebook announced that it was tweaking its Newsfeed algorithm — the thing that decides what goes to the top of your feed and what gets buried — to prioritize things posted by your actual friends, and to deprioritize brand pages and links to publishers (aka articles on websites like BuzzFeed). The reaction to this was mostly hand-wringing about the potential effect on publishers, who — like BuzzFeed — get a large chunk of their traffic from people sharing their content on Facebook.

Had Facebook toyed with these publishers, hypnotizing them into submission with a giant firehose of traffic, only to cruelly turn the faucet down to a trickle? Did Facebook ever care about publishing anyway? Or did it just care about “media” and “news” only when it was the one thing Twitter was beating them at, and as soon as they crushed their rival (however potentially hamfistedly), they lost interest and moved onto the next shiny thing – video? livestreams? Bots?

Or perhaps it was just going back to basics – people just want to use Facebook to interact with their friends, not to see a bunch of links to articles.

Interesting, but none of this is really relevant to my overnight oats problem.

Shani and I share a lot (97&;) of mutual Facebook friends, mainly coworkers here at BuzzFeed. So some of those replies with oat advice were from mutual friends. That’s an indicator to Facebook that this post must be something I’m interested in — look&033; Your mutual friends are discussing it&033; And because many of those mutual friends are also friends with each other, it kept bumping the post up to them as well, encouraging them to chime in with their favorite mix-ins. The cycle kept going and going, making Facebook more and more convinced that I would care about it.

I asked (on Twitter, of course) if anyone else had noticed lately that old posts had stayed “stuck” at the top of their Facebook feed for longer than normal. Several people had. Indeed, other coworkers and friends of Shani said that they too had had her oats post stuck at the top of their Facebook feed for days.

Facebook’s new news feed is operating based on the idea that you care more about what your friends have to say and their photos and videos than you do about links they post. Which is probably true&033; But even if it’s trying to be more friend-friendly, Newsfeed is still controlled by an algorithm. Based on a years-long friendship and a number of data points, the algorithm has figured out (I’m assuming — it’s a trade secret how the algorithm actually works, but it’s easy to guess) that I tend to be interested in what Shani posts.

Yet there is no way for it to understand that I do not care about overnight oats. That, in fact, I really really don’t care about overnight oats and I really don’t want to keep seeing a bunch of fucking overnight oat tips every time I check fucking Facebook.

What the oats revealed was machine learning&039;s limited understanding of friendship. There’s always been something a little cold and inhuman about the way that Facebook outwardly shows its understanding of how a human uses its service. When it rolled out its first Messenger bot earlier this year, it was a shopping bot to help you pick out clothing and shoes from an online store – the best way I found at the time to describe the experience is “this is something someone who works at Facebook would want to use.” It ignores the whimsy and pleasure of an online clothes shopping experience, where the user can browse for what they like. Instead, it just showed you just a few items in your price range in a generic category like “sneakers”. Messenger bots for weather and news were similarly panned at their launch for being glitchy and unuseful. Tech companies in general seem prone to explaining their services in this weird treacly simulacra of what a real human actually is. For a long time, the demonstration videos for new Apple features always seemed geared toward this imaginary perfect 40-something dad who loves exercising and just wants to share photos of his kids and find a great local sushi spot – I call him “Apple Man”. All the problems in his life can be solved by a slight new improvement in UI.

Facebook is sometimes is like the relative who thinks you still love horses because you were really into horses when you were 14 and keeps sending you birthday cards with horses on them.

But human relationships are messy in ways that technology and social platforms can’t really deal with. They get divorced but feel ashamed to announce it, they have weird passive-aggressive fights with their friends, they repeatedly lurk on the page of their partner’s ex. They have preferences that are not stated in their “Likes”, or they outgrow those Likes after a few years. If you look in your settings at what Facebook is telling advertisers that it knows about you, you will be shocked at both how creepily right it is and how hilariously wrong it is.

Facebook is sometimes is like the relative who thinks you still love horses because you were really into horses when you were 14 and keeps sending you birthday cards with horses on them. This uncanny feeling of someone sort of knowing you but not really is echoed all over the web. Like LinkedIn somehow suggesting you add someone you went on a blind date with years ago to your professional network. Or when you buy that pair of shoes online but the ads from Zappos for them follow you around every new site you visit seemingly forever, like a case of dynamic advertising HPV.

You, too, may have noticed in the last few weeks that certain posts from your friends seem to linger at the top of the Newsfeed for longer than usual – days even. Or that you have to scroll down further to see something that you haven’t already seen. If you’re a Facebook addict who checks multiple times a day and is used to fresh content, this is really annoying. Maybe it keeps showing you the same boring photo or post.

Or perhaps the real problem is that the limit of Facebook’s algorithm is that it’s only given the material to work with of who your friends are. If your friends are boring, you’re stuck with a boring feed.

But assuming you aren’t a monster, you like your friends. And Facebook knows that. But it doesn’t know that sometimes you don’t give a shit about what they post.

Point is, if you have good ideas about how to make overnight oats, please post in the comments below.

Quelle: <a href="Facebook’s Newsfeed Has A Friendship Problem“>BuzzFeed

Mirantis and SUSE: Creating a One-Stop Shop for OpenStack Support on Major Linux Distros

The post Mirantis and SUSE: Creating a One-Stop Shop for OpenStack Support on Major Linux Distros appeared first on Mirantis | The Pure Play OpenStack Company.
This week, at OpenStack Silicon Valley, Mirantis and SUSE announced a partnership which will make Mirantis a &;one-stop shop&; for Mirantis OpenStack supported on all major Linux distributions. Mirantis does not ship a Linux distribution, but rather works with Linux distribution vendors on support of underlying Linux operating systems.This  partnership positions SUSE as Mirantis strategic Enterprise Linux partner providing Mirantis OpenStack customers with enterprise grade SLA’s for SUSE Linux Enterprise Server (SLES), Red Hat Enterprise Linux (RHEL), and CentOS. For ongoing support of Mirantis OpenStack running on Ubuntu, Mirantis and Canonical have had a collaborative relationship for several years, jointly supporting large customers like AT&T.
As a pure-play OpenStack provider, committed to freedom of choice, Mirantis will leverage this partnership to provide more flexibility to customers, reducing lock-in. While &8220;one-stop shop&8221; used to mean that a single technology vendor offered all the components of the solution, services, and support, Mirantis one-stop shop is about being a single source of support, services, and expertise to help customers in their cloud transformation journey using a wide range of certified best of breed technology selections. This approach is hugely valuable to large customers who may be broadly committed to a Linux distribution, but don&;t want to be locked into that choice, or limited in choosing other best-of-breed data center technologies to work with OpenStack..
Mirantis and SUSE will begin engineering collaboration upstream to fine tune Mirantis OpenStack on SUSE enterprise Linux leading to a certified/supported solution for customers. Additional upstream and downstream engineering/support collaboration will accelerate Mirantis taking on front line L1 and L2 support for the entire solution while SUSE provides L3 support for SLES, RHEL and CentOS.
Being free to run OpenStack on a preferred Linux distro is a big deal for enterprises — touching on every aspect of reliability, security, performance, usability, interoperability, and cost. In the past, such freedom has been hard to come by in the OpenStack space, because supporting production OpenStack on multiple spins requires both broad and specialized expertise. In some cases, vendors such as Red Hat have touted the value of Linux and OpenStack being &;co-engineered,&8217; effectively promoting lock-in. Mirantis has historically taken the opposite approach: as a pure-play OpenStack provider, we think of OpenStack as an application that should run on any host OS (or in containers, as our recent announcement about Kubernetes makes clear). This new partnership will help us deliver that kind of freedom of choice and reassurance to OpenStack customers in the real world.
The post Mirantis and SUSE: Creating a One-Stop Shop for OpenStack Support on Major Linux Distros appeared first on Mirantis | The Pure Play OpenStack Company.
Quelle: Mirantis

AWS CloudFormation Adds Support for AWS Certificate Manager, Amazon ECS Service Auto Scaling, and Additional Updates

You can now provision the following AWS resources using CloudFormation.

AWS Certificate Manager: AWS Certificate Manager is a service that lets you easily provision, manage, and deploy Secure Sockets Layer/Transport Layer Security (SSL/TLS) certificates for use with AWS services. Visit our documentation to learn more
Application Auto Scaling: Application Auto Scaling is a general purpose Auto Scaling service for supported elastic AWS resources. Use it to enable Service Auto Scaling in Amazon EC2 Container Service (ECS).

You can now set Service Auto Scaling policies in ECS and determine the target service you want to scale using CloudFormation

Quelle: aws.amazon.com

Preemptible VMs now up to 33% cheaper

Posted by Michael Basilyan, Product Manager

We’re happy to announce that we’ve lowered the price of Preemptible VMs by up to 33%! Since launching Preemptible VMs last year, we’ve tuned our algorithms, improved their efficiency and analyzed usage patterns. Our experience, combined with the growth of Google Cloud Platform, allows us to offer deeper discounts. For example, the price of an n1-standard-1 Preemptible VM instance is now just one cent per hour. That’s 80% cheaper than the equivalent, non-preemptible instance, with no bidding or guesswork. The new pricing is already in effect.

Preemptible VMs are just like any other Google Compute Engine VM, with the caveat that they cannot run for more than 24 hours and that we can preempt (shut down) the VM earlier if we need the capacity for other purposes. This allows us to use our data center capacity more efficiently and share the savings with you.

Over the last year, Google Cloud Platform customers, such as Citadel, have used Preemptible VMs to greatly reduce their compute costs, and have come up with lots of interesting use cases along the way. Our customers are using Preemptible VMs to analyze data, render movies, process satellite imagery, analyze genomic data, transcode media and complete a variety of business and engineering tasks, using thousands of Preemptible VM cores in a single job. We believe that the price reduction for Preemptible VMs will unlock even more computing opportunities and enable you to tackle interesting science and business problems.

Here are some ways you can launch a Preemptible VM right now:

Add just a single flag (–preemptible) in the gcloud compute instances create command, or by using one of our libraries
Check a single box in the Developer Console create instance page
Launch a quick and easy-to-use Spark/Hadoop cluster with Cloud Dataproc
Autoscale Preemptible VMs with managed instance groups
Render a movie with Zync and choose Preemptible VMs, which is now also up to 15% cheaper!

Here are some tips and tricks to help you get the most out of Preemptible VMs:

Resources for Preemptible VMs come out of excess Google Cloud Platform capacity. The load on our Cloud Platform data centers varies with location and time of day, but is generally lowest on nights and weekends — the best time to run large Preemptible VM clusters.
We avoid preempting too many VMs from a single customer and, given the choice, preempt VMs that were launched most recently. This might be a bit frustrating at first, but in the long run, this strategy helps minimize lost work across your cluster. And because we don’t bill for VMs preempted in the first 10 minutes, it saves on costs too.
It’s a good idea to retry once or twice, even if you’ve been preempted early. Combining regular and Preemptible VMs in your clusters will ensure that tasks proceed at an adequate pace.
Manage shutdown and preemption notices with a shutdown script that saves a job’s progress so that it can pick up where it left off, rather than start over from scratch.

For more details on Preemptible VMs, please check out the documentation. For more pricing information, take a look at our Compute Engine pricing page or try out our pricing calculator. If you have questions or feedback, go to the Getting Help page.

We’re excited to see what you build with our products. If you want to share stories and demos of the cool things you have built with Preemptible VMs, send us an email or reach out on Twitter, Facebook, or G+.
Quelle: Google Cloud Platform

Application Insights: Track an Analytics query in a dashboard

Now you can continuously monitor any Analytics query computed from your telemetry in Visual Studio Application Insights, and display the results on an Azure shared dashboard. This means that when you put together a dashboard to help you monitor the performance or usage of your web services, you can include quite complex analysis alongside the other metrics.

As an example, let’s suppose you’d like to track what percentage of requests your website completes within two seconds. You’re only interested in the most popular pages. This is not difficult to write as an Analytics query, but it certainly isn’t offered as one of the standard metrics. Until recently, you could only run the query in the Analytics window, and you’d have to click Go at intervals to keep the chart up to date. But now you can include it in a comprehensive dashboard of your system telemetry.

Dashboards are great for bringing together your most important charts (as you’ve probably already discovered), especially if you want to see metrics from multiple apps in your system at the same time. Furthermore, you can create multiple dashboards. The dev team room can keep an eye on performance of all the front and backend apps, while management can do a weekly review of system usage.

1. Share a dashboard

You can only pin Analytics charts to a shared dashboard, so sharing at least one dashboard is an important preliminary step:

2. Write an Analytics query

Now open the Application Insights resource for your app.

Then click through to Analytics:

Write and test your query. Let’s take the example we mentioned earlier:

3. Pin it to the dashboard

Click the pin icon and choose a dashboard. Only the shared dashboards in your subscription will appear in the list. You might need to upgrade your account to turn on this feature.

4. Adjust the chart on the dashboard

Now go back to the dashboard in the Azure portal. You can adjust the position and title of the chart in the usual way.

The chart is refreshed by re-running the query about every half hour.

Automatic grouping

When you pin a chart to a dashboard, you get a slightly simplified display in some cases.

For example, here’s a chart in Analytics in which we’ve summarized by country:

requests | summarize count_search = count() by client_CountryOrRegion

But when we pin it to the dashboard, it looks like this:

Notice how the long tail has been grouped into “(other)”? If you pin a bar chart of a count of values separated into discrete bins on the Y-axis, then the smaller items are automatically grouped.

What’s next?

This is only the first step in operationalizing your insights by adding them to the Azure dashboard. We still have other capabilities to deliver, including supporting pie chart visualization, editing dashboard charts, and other small settings that will help you set your dashboard parts exactly the way you want them.

As always, feel free to send us your questions or feedback by using one of the following channels:

Suggest ideas and vote on Application Insights ideas
Join the conversation in the Application Insights Community
Try Application Insights Analytics

Quelle: Azure