They increased to 25 to encourage media uploads to train their own models with. They now have collected enough metrics to realize, most valuable content is below 10MB. Now they are optimizing. They won't lose anything valuable to them and the users who are impacted might even buy Nitro now. Win-win for them
I've been a funding member of the Wikimedia Foundation for over a decade. I have looked at their finances several times before and during financing them.
As with a lot of similar non-profits, a considerable amount of donations does not go into "running the servers". You have to judge this by yourself, but they don't embezzle any money and there is a reasonable bottom line. Wikipedia continuously helps tons of people, and the people who run the operation enable that.
You can download a full dump of Wikipedia any day. Compared to other lying companies, they have been true on their promises for some time.
Of all the $1 I could spend in a year, the one I give to Wikipedia is probably the least wrong invested, and that $1 actually already makes a difference
I know that that my opponents just love to bring up how I fucked that one couch only that one time very long ago in the past...
👏 👏 👏 OH YEAH YOU DID YOU HORNY COUCH FUCKING ALIEN ASS MOTHERFUCKER HAHAHAHA 👏 👏 👏
That's the one clapping person he was looking at in this picture.
While that is true and reflected in the article data, it doesn't contradict what I said. Just in case it was meant to 😄
Yes people still buy Russian energy, sometimes through third parties, but total consumption is down.
Because it's clickbait framing. Total imports from Russia are at their lowest point ever. The USA just dropped below Russia in the rankings.
I wasn't actively aware of this for most of my life until I recently visited a clients office. Buying someone a cup of coffee is an entire thing. There's no free coffee. You have to purchase every single cup. And you first have to walk several minutes to the place where they sell the coffee. It blew my mind. I'm used to drinking one cup after the other without even giving it any thought. Coffee machine right next to me or around the corner. There, coffee incurs friction and cost.
So when you invite someone for a cup of free coffee, this can open doors for you. I'm not kidding. People get all excited when you offer them a coffee break on your dime. And there's levels to it too. There's the regular coffee, and there's the premium one. For the premium you have to walk longer and wait in line until the barista serves you.
It's a key component in office politics when coffee access is regulated.
Why anyone would restrict access to legal stimulants in the office is unclear to me though. Put espresso machines on every desk!
I don't necessarily disagree, but I have spent considerable time on this subject and can see merit in decoupling your own error signaling from the HTTP layer.
No matter how you design your API, if you're passing through additional layers, like load balancers and CDNs, you no longer have full control over all responses your clients receive. At this point it may be viable to always signal a successful backend connection with a 200, even if the process resulted in a failure.
Going further, your API may include partial success scenarios, think batch processing, then the result could be a mix of success and failure that doesn't translate to HTTP status.
You could even argue that there is really no reason to couple your API so tightly with a concept of the transport layer it uses.
So you fucked everyone because of a beef you had with AWS. Go fuck yourselves. Moving people off Elastic products is the right move either way. Don't look back.
Respect the Accept header from the client. If they need JSON, send JSON, otherwise don't.
Repeating an HTTP status code in the body is redundant and error prone. Never do it.
Error codes are great. Ensure to prefix yours and keep them unique.
Error messages can be helpful, but often lead developers to just display them in the frontend, breaking i18n. Some people supply error messages in multiple languages, depending on the Accept-Language header.
There are many ways your real IP can leak, even if you are currently using Tor somehow. If I control the DNS infrastructure of a domain, I can create an arbitrary name in that domain. Like artemis.phishinsite.org, nobody in the world will know that this name exists, the DNS service has never seen a query asking for the IP of that name. Now I send you any link including that domain. You click the link and your OS will query that name through it's network stack. If your network stack is not configured to handle DNS anonymously, this query will leak your real IP, or that of your DNS resolver, which might be your ISP.
Going further, don't deliver an A record on that name. Only deliver a AAAA to force the client down an IPv6 path, revealing a potentially local address.
Just some thoughts. Not sure any of this was applicable to the case.
There are many ways to set up something that could lead to information leakage and people are rarely prepared for it.
5 nines imply a downtime of 6 minutes a year, or every 100,000th operation failing. That's not great for a file system. I assume you picked the number arbitrarily, but still think about it.
Classic meme with a sloth
Gotcha. Millions of harmful posts every day. That really does sound like a great place.