I was telling a friend recently about how there was this "golden age" when you could access all sorts of free APIs, and how I still long for this time.<p>I remember the public Netflix API, Twitter APIs and Flickr API with particular fondness. My personal site was a big mashup of all of my data.<p>I also abused the hell out of Yahoo Pipes - I would run RSS feeds through like 15 different languages with Babelfish before back to English, just for kicks.<p>My friend seemed very skeptical such a time ever existed.
As the founder of one of the screen-scraping tools he alluded to in the video (Selenium), I just want to say the video has one of the best explanations for the difference between automating a process through a user interface vs an API. In the end, entropy always gets you, but you can push it off a little bit longer if there's an API.
I'm reminded of this little gem: <a href="https://hookrace.net/time.gif" rel="nofollow">https://hookrace.net/time.gif</a><p>(Relevant post <a href="https://news.ycombinator.com/item?id=14996715" rel="nofollow">https://news.ycombinator.com/item?id=14996715</a>)<p>But on the topic: is there actually a dearth of APIs "these days" vs peak Web-2.0, or have the major players just restricted theirs due to abuse, and thus it seems like the whole world of possibilities have veen restricted? One can easily find lists of public apis (e.g <a href="https://github.com/n0shake/Public-APIs" rel="nofollow">https://github.com/n0shake/Public-APIs</a>), but perhaps the video was more about the facilitators, like Yahoo Pipes.
I really liked this video, it highlights that a lot of software we write is ephemeral and will one day either be retired or stop working.<p>The frequently updating title thing is cute, it'll be interesting to see what dies first - YouTube pulling the "title update API" or Tom's script running wherever he's put it
<i>If it's actually spot on, it's a miracle.</i><p>I got the counter and the title both showing 3,690,744 when I first opened the link - so how unlikely is this actually? Probably not really too unlikely. Or I got really lucky.<p>EDIT: Thinking about it, as YouTube probably updates the view count only every couple of seconds or minutes it might actually be spot on most of the time if the title gets updated at about the same frequency.
This "open APIs" feeling where you can build and mash up all kinds of services together to build cool things is how writing apps on Ethereum feels right now. All of the data and functions of other people's contracts are on chain and available to you for use in whatever way you want to use it. It's very powerful and makes developing fun again for me.<p>As an example there is a project called Maker which produces a stablecoin called Dai which is pegged to $1. Another project called Compound took Dai and used it without asking anyone at Maker to create automatic loans where you can put in money and get interest automatically. A third project, Pool Together, started using Compound, again without asking, to pool everyone's funds together for a month and give the interest earned to one winner as a "no-loss lottery". I bet in a few months something will be built on top of Pool Together as well.<p>None of these teams needed to work together or ask permission. They just built cool things. An added bonus is that these projects can't be turned off by anyone which means Pool Together can trust that their app will work next year just fine, which isn't really something you can rely on in Web 2.0. It's a very exciting time for composability and neat experiments and I'm looking forward to what else will be built.
Reminds me of <a href="https://news.ycombinator.com/item?id=3742902" rel="nofollow">https://news.ycombinator.com/item?id=3742902</a> ("Show HN: This up votes itself") which I came across right after joining HN and really set the tone for what HN is really all about, for me. Thanks olalonde :)
The voice, oratorical flourishes, and narrative style really remind me of James Burke's Connections. "And that's why I chose to film this here..." Delightful!
>If it's actually spot on, it's a miracle.<p>Not that it matters, but I think Tom may have gotten this wrong. If his code is invoked many times faster than google updates it's video count then the odds of seeing an exact match in the total is proportional to that difference.<p>Which, ironically, means it's using even more cycles than necessary to do his intentionally silly trick, further proving his point.
Despite that the video starts with<p><i>"The title of this video won't be exactly right. [..] If it's actually a 100% spot on it's a miracle"</i><p>the title <i>was</i> exactly right when I saw it the first time. I even screenshotted it.<p>I also wondered if it would work on HN. Is there a limit on the number of times you can edit a title on HN? Obviously there isn't on Youtube, which I find quite surprising.
Reminds me of the CGP Grey video that shows how much the video made in ads (<a href="https://www.youtube.com/watch?v=KW0eUrUiyxo" rel="nofollow">https://www.youtube.com/watch?v=KW0eUrUiyxo</a>)
Speaking of the YouTube view count, Tom Scott also did a great explanation on distributed computing and eventual consistency: <a href="https://www.youtube.com/watch?v=RY_2gElt3SA" rel="nofollow">https://www.youtube.com/watch?v=RY_2gElt3SA</a> "Why Computers Can't Count Sometimes"<p>And here's a video from Computerphile about overflow: <a href="https://www.youtube.com/watch?v=vA0Rl6Ne5C8" rel="nofollow">https://www.youtube.com/watch?v=vA0Rl6Ne5C8</a> "How Gangnam Style Broke YouTube"
My twitter bot, @fiveobot[0], lives on, within a `screen` session on my VPS. Its time zone and geographical data are both at least 2years out of date, but I'd have to adapt new tools, or hook into APIs that will eventually fail to access up-to-date data. I made it for an audience of one, and I'm still amused by it, today.<p>[0] <a href="https://github.com/jachee/fiveobot" rel="nofollow">https://github.com/jachee/fiveobot</a>
If anyone is curious, I took a crack at building out a bot that does this with Node.js and the YouTube Data API! (<a href="https://github.com/stursby/this-video-has-x-views" rel="nofollow">https://github.com/stursby/this-video-has-x-views</a>)
This would be noteworthy if the counter was included <i>in the video</i> (like the similar video that shows its own URL [0]). But as it is, it's just the title that changes to match the number of views, so... not much to see here.<p>[0] <a href="https://news.ycombinator.com/item?id=20452013" rel="nofollow">https://news.ycombinator.com/item?id=20452013</a>