This Video Has X Views
youtube.comI 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.
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.
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.
My friend seemed very skeptical such a time ever existed.
It reminds me distantly of an earlier time when just the web was sort of owned ... by the web folk.
In the sense that even corporate sites if you found it would have a little corner where the 'webmaster' had a page that mentioned the server, or his cat, or some silly pic. Some sort of character or tidbit before any of the branding drones were really aware of the web. All just because the 'webmaster' was the only one really in charge / who understood the site was even there and they wanted to share.
I suspect to some extent the APIs were the same. Someone who really didn't mind was all "Yeah sure if someone wants to see what I did.. awesome."
>> a little corner where the 'webmaster' had a page that mentioned the server, or his cat, or some silly pic.
Or a “links” page. I haven’t seen that in a long while. No affiliate garbage or anything just a page linking to other sites that the webmaster liked or whatever. It’s hard to remember when that fell out of fashion but it did seem to add a personal touch as well.
I used to beg people I met online to host pages for me, give me shell access, etc. I was a child in the very early 90s with basically no money.
The same, in fact my life would be very different if a random guy from (I think Kansas, given I'm from the UK) on a forum bought me a domain and gave me a slice of his dedicated hosting when I was mid-teens.
I've always wanted to pay it forward in the same way, but lots of things on the web seem overly complicated now that'd make that hard to do, and I've lost attachment to most communities.
It's not very common, but I had people begging my friends for this kind of services.
raspberrry pi in home-network DMZ, with fail2ban, unattended-upgrades, and a free dynamic DNS service gets you most of this. (plus, you're root, so you can run things more complicated than web-services)
> It reminds me distantly of an earlier time when just the web was sort of owned ... by the web folk.
I remember those times, I miss them. I had a modem back in the late 1990s and used to buy .net magazine (in the UK, back before there was a framework of the same name) on my way home from school and it had the number of people estimated to be on the Internet printed on the spine. It all seemed too good to be true, we were worried it might get shut down by governments. There was the TV program "the net" (https://en.wikipedia.org/wiki/The_Net_(British_TV_series)) that used to give you an "info dump" at the end that you were meant to record on your VCR and play back frame by frame.
In Jan 2000 I got my first proper job at a web hosting company and used to read Wired magazine before it became (as far as I recall) fascinated by the stock market.
I miss the optimism and simplicity of those times.
> I miss the optimism and simplicity of those times.
It seems like every generation (minus those who came of age in WWI/II) think this is true for them.
Is it really true, or are we just all reflecting back to when we were younger and the world was simpler because we understood less.
To be clear I was speaking specifically about the Internet/Web and not the world in general. I would say that there was a lot of optimism for the web back then, and it was definitely simpler.
Back when I had three hundred bookmarks instead of, like, six. God I'd kill to still have that list.
I remember these days fondly. Surfing the web on 486 at 28.8kbps. it felt slow, but worth it. Most information to be found existed because someone thought it was worth sharing. A golden age to be sure.
Thats the web we love.
> 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.
Yahoo Pipes was one of the greatest services I used, just when I started getting into programming. Maybe it was so cool because I was naive, but I really miss being able to pipe services together in the same way. Anyone know of any similar attempts that is open source + offers a hosted version with paid plans?
Not sure about open source and paid hosting, but there is node-red, https://nodered.org/, which is open source and easy to get started with. Combining stuff is something Zapier excels in, https://zapier.com/, and the free tier can suffice for some tasks (not open source). You could also try my attempt at a spiritual pipes successor, https://www.pipes.digital/ (but it's also not open source). If there is something missing there to reproduce how you used Yahoo Pipes I'd definitely be interested in hearing from you, so I can restore it :)
So, the requirements of open-source + have entity with paid hosting are both equally important. First one to ensure I can continue using whatever I setup on the hosting, in case the entity behind it cannot. And the paid hosting is important because it gives better odds towards the service actually sticking around.
What I likes with Yahoo Pipes compared to NodeRED (at least as far as I looked at NodeRED, I might be wrong) is that Yahoo Pipes worked out-of-the-box with services out of the box. I seem to remember that you could use the Google Search API for example, with Yahoo Pipes and pipe that into other things. That's what Zapier does as well, but with less flexibility than NodeRED.
So I guess my dream would be something like the integrations provided by Zapier but with the UI and flexibility of NodeRED.
Haven't seen pipes.digital before, I'll take a look as it looks interesting, but for anything serious, open source is a hard requirement (gotta learn from the Yahoo Pipes history :) )
I don’t think those requirements will be met until people like us who miss it put our money where our mouth is ;)
I think people like us (at least me) miss a lot of things, some more important than others, and we have to carefully choose what we spend out time on :)
I'm about to open source it now. Which means I just did, but haven't announced it yet. https://github.com/pipes-digital/pipes now contains a very new version you could run locally.
r-w is right though: More users paying for pipes would allow me to invest more time into it and have more requirements covered.
I haven't used it but https://n8n.io/
It reminds me of zapier, which reminds me of Yahoo pipes.
From a quick search, it looks like someone is recreating Yahoo Pipes here: https://www.pipes.digital/
Unfortunately, it looks like it's not open-source, their Github repo is only for bug reports: https://github.com/pipes-digital/pipes
Yeah as I mentioned https://news.ycombinator.com/item?id=21641615
A Yahoo pipes! clone really needs some open source love. Too bad the maintainer is more interested in making it a one man SaaS.
The video is interesting because I thought he was talking specifically to me about the http://2fb.me product. But anyway I think this is a reminder that even though API’s have gotten more restrictive that more creative ways of using them bubble to the surface. :)
> A Yahoo pipes! clone really needs some open source love. Too bad the maintainer is more interested in making it a one man SaaS.
I guess that's fair.
I want(ed) to run Pipes as an SaaS because I think that model would be highly advantageous to me, and that would be very advantageous for Pipes. But that with the predecessors history as background this approach is a critical concern is something I do understand. I will now try to have my cake and eat it too: https://github.com/pipes-digital/pipes contains a new FOSS version that is meant to exist in parallel with pipes.digital, as long as that model works out. So worst case now is this FOSS approach fails, and then only the SaaS would die, not the software.
Check out my project, riko. https://github.com/nerevu/riko Open source, but no hosted plan.
For a self hosted Yahoo Pipes alternative I’d recommend checking out Huginn - it’s been featured on HN a couple of times and the creator is tectonic@HN
I don’t, but Node Red is giving me the same fun feeling I had with Yahoo Pipes, but for Home Automation.
Dark is similar. I think they're still in private beta, but I got my invite pretty quickly.
I believe https://luna-lang.org has plans like this, though they're kinda definitely not hurrying things...
I remember when you could just stream in every tweet and reddit comment ever posted with a few lines of code. There used to be dedicated API methods for doing so. Now, there's all sorts of namespacing, rate limiting, pagination, and upper bounds on data access that make this impossible, or at least infeasible.
Twitter wants to capitalize on the fact that tweets are massively public, and they also want to capitalize on the fact that they are not.
Researchers are passing around lists of tweet IDs ("dehydrated", they call them) that can be "rehydrated" (that is, turned back into full tweets) if you have the right permission from twitter to do so.
The whole setup is really shameful.
It would be de-facto illegal to build a "Google for Twitter" today. I settled on doing it for ActivityPub/Mastodon because it's less likely I'll get sued into oblivion for creating a search engine that way.
Hi sneak,
I am working on some fediverse aggregator projects. Let me know if you want to talk.
You have no standard contact methods available on any of your sites that do not require registration. Consider publishing your email address; I am happy to chat. My email is in my profile and on my website.
Sent you an email
This reminds me of an endpoint that Google used to run. It was a never ending stream of RSS feeds. Anyone could subscribe (it was just a never ending HTTP GET, if I remember right), and be told which blogs had updated in near real time.
Nowadays you'd need to sign up for an API key, probably pay some amount of money, and provide twenty different forms of contact to use something like that. That's assuming it was allowed to exist in the first place, since it might take views away from google.com.
> 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.
For a while it was common to find SEO-spam sites composed entirely of posts generated this way. They would translate from a source language back to it in a roundabout way and end up with an article that was "different enough" to count as unique content to Google.
A British MP (I cannot for the life of me remember who) came under fire for owning a company made its money doing this - take existing content, shuffle the language round a bit and publish it with loads of ads.
Likely to be Grant Shapps
> In 2012, Google blacklisted 19 of the Shapps's business websites for violating rules on copyright infringement related to the web scraping-based TrafficPayMaster software sold by them
https://en.wikipedia.org/wiki/Grant_Shapps#Business_ventures
Yes! That's him. I only knew about that one, but this other one mentioned afterwards is hella shady too:
> It cost $497 and promised customers earnings of $20,000 in 20 days. Upon purchase,
> the "toolkit" was revealed to be an ebook, advising the user to create their own
> toolkit and recruit 100 "Joint Venture Partners" to resell it for a share of the
> profits
Sort of like a pyramid scheme, really. Incredibly this appears to have scarcely affected him, he is still an MP :-O
I also remember vividly those optimistic times. You should show this video to your friend as proof: https://www.youtube.com/watch?v=6gmP4nk0EOE
There was a lot of interesting promise in Web 2.0 that was completely wasted and then died. I think a big part of the problem was even things like Yahoo Pipes were too complicated for a lot of people and anyone trying to use third party APIs for commercial purposes ran afoul of EULAs or just plain old rent seeking. Once privacy invading advertising became the norm APIs were further restricted or discontinued because the user wasn't running a bunch of client side code tracking and scraping all their behavior on a web page.
It didn't help that a lot of Web 2.0 darlings sold out to on-the-way-out Web 1.0 companies \cough\Yahoo\cough\. Yahoo's management couldn't even monetize money, let alone Web 2.0 properties. So instead we got social media silos. You can put stuff in but good luck ever getting it out. You can share it with whomever so long as they also join the same silo.
There are 3 big problems with open APIs
1. they enable people to do things that other people think shouldn't be done
2. people get upset at companies when (1) happens
3. people get upset at companies for removing or restricting apis when, or in fear of, both (1) and (2).
I think you are missing what is probably the biggest factor: they can be very expensive to run if successful.
If you create a popular API, people are going to find creative uses for it, and because they can, by definition, be automated, you can get rapid growth in traffic with not that many users.
There is a bit of a 'tragedy of the commons' that goes on, because the people writing apps that consume the API have no incentive to moderate their usage, or try to be efficient.
Since the company that is providing this API is paying for the resources to run it, they can quickly get very expensive. Unless there is a clear financial benefit for allowing it to continue, most companies will shut them down eventually.
Is it actually expensive, or is it just a result of everyone moving to cloud providers that charge an arm and a leg for performance equivalent to a cheap laptop (when you account for "CPU credits" and all that) and try to nickel & dime you on everything, including bandwidth (despite bare-metal providers somehow being able to stay in business by offering unmetered 1Gbps bandwidth)?
Also the API forms a committed interface you are stuck with. Whatever way you want to "improve" your offering you have to make sure it's compatible with the old API, at least for a while. Especially as you can't reach your consumers.
Couldn't this all be solved with a p2p network that hosts restful services?
The framing of open APIs disappearing because of bad actors doesn't ring true to me. In my view, the golden age of APIs disappeared one-by-one as tech companies realized they won their respective markets and consolidated their power.
I think what you said, yes, but also bad actors. Basically they disappeared because what business purpose did they serve, they were a cost center not a profit center - bad actors just increase the cost.
Those times were absolutely great. It's only human to look at the past with a feeling of melancholy and forget how great is the present and the long way we came. Today you can actually build your own APIs with very little knowledge and effort: easy to get started frameworks, free or inexpensive hosting in the cloud, lots of freely available data. That's pretty awesome too.
Heck yeah! Years ago I had a nice little feed of my latest Tweets and Flickr photos on my personal website! It was fun incorporating them cleanly/seamlessly into my custom design. One day I'll get my site going again and have fun stuff like that -- if the 3rd party APIs even allow it these days :P hehe
Even Google Search had this. I read an old O'Reilly book back then that was centered around API mashup projects (remember that term?) using the many Google APIs, which have mostly been shut down.
It surprises me that we don’t have crypto-based API micro-transactions yet.
Even if you had to pay 1 satoshi for 10 queries, it would go a long way towards making APIs viable long-term.
What's the transaction rate for bitcoin these days? Is it still a handful of transactions per hour? What are the transaction costs?
There's a layer 2 technology called Lightning Network that lets you transfer Bitcoin almost for free and in seconds, no transaction number limit. The biggest issue is liquidity, it's not easy to move large sums in a single transaction, but that shouldn't matter much for micro-payments.
There's even ongoing work on making paid API's work seamlessly by adding some middleware, using the HTTP 402 (payment required) status code, no usernames/e-mails/passwords involved: https://lightning.engineering/posts/2020-03-30-lsat/
(And yes, LN is open source, non-custodial and moves real Bitcoin, not IOU's or other tokens. It's actually an open spec with at least 3 major implementations in 3 different languages.)
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 got my first job in software writing selenium code, your library literally changed my and my family's life.
Your work has had a huge impact in my career and life.
Introducing Selenium to testing teams has been a complete game-changers for a lot of organisations I work with. Often traditional orgs will assume you can't automate front-end testing because it requires a user to go through all of these processes in different front-end SaaS apps. It's literally saved hundreds of man-hours across the organisations I've introduced it to and allowed the project teams to focus on features and reporting instead of testing.
Thank you!
Selenium is awesome, but the browser GUI for it in Firefox seemed to... Regress heavily from where it was a few years ago? It used to allow you to export a script as any type of code and now that's no longer possible. The newer gui based tools have maybe a fraction of the features that the selenium GUI used to have
What happened?
I love and use Selenium all the time. Thank you!
Thank you for your work sir. I love Selenium.
It's a team effort! I'm more of a fan and advocate for the project these days.
Amazing work!
Thank you for Selenium. I recently used it to parse some hotel data for Covid.
I'm reminded of this little gem: https://hookrace.net/time.gif
(Relevant post https://news.ycombinator.com/item?id=14996715)
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 https://github.com/n0shake/Public-APIs), but perhaps the video was more about the facilitators, like Yahoo Pipes.
What the GitHub list doesn't show is the amount of APIs that require you to enter a relationship with the API provider.
APIs I can just point my script at to get data are fun and useful for mashups. APIs where I have to sign a contract with someone are only worth it in rare cases, or if I need them for a business.
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.
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 can go find NES games and Gameboy games in my parents basement. If the NES worked I could play it or find a knock off on Amazon. Pretty sure I have a Doom floppy.
Kids these days will be lucky to remember their favorite mobile games. Let alone be able to play them.
Will they ever have an iOS or Android equivalent of ROMs?
> Kids these days will be lucky to remember their favorite mobile games. Let alone be able to play them.
I'm already feeling the pain of this. There was a game for iOS called GeoDefense (and a sister game, GeoDefense Swarm). To this day these were my favorite games on the phone. But iOS 10 ended support for one of them, then a later OS update bricked the other one.
The developer hasn't updated these games to work in new iOS, so they're lost to time.
If I had a time machine, I would go back and warn past-Me to reserve a 4S for just playing these games.
Sorry for self-reply , but hopefully someone out there can point me to some similar tower-defense games? Everything I see on the app store now is bloated with graphics or IAP and shitty gameplay. If you're familiar with the GeoDefense games[0], you know what I'm talking about.
It's not exactly tower defence, although i think it's close, but i am currently recovering from a pretty serious Bad North obsession:
You defend procedurally generated islands from viking hordes, but rather than towers, you use up to four squads of soldiers from an army you grow over dozens of islands. The art and sound design are lovely, and the gameplay is engrossing. Each island takes a few minutes to play, and a campaign lasts a few hours.
It's something like 5 dollars on iOS and Android, and has no ads or IAP.
Mindustry and Infinitode are both lots of fun
+1 for Mindustry. Factorio with more tower defense emphasis. And FOSS!
GeoDefense! Oh hell yes. I played so much of that. I still have a 4 and a 4s somewhere. Might try it out!
I still miss Aurora Feint from the dawn of iOS.
I have a huge folder of iOS apps, but.. unfortunately in I think iOS 9 or so, Apple made it so backing up your iPhone stopped backing up the applications, and I believe made it impossible to install applications from your hard drive backups (that may have happened in a later iOS version - I forget). Meaning... iOS apps are now simply unrecoverable once they aren't on App Store anymore. Good thing I have a couple old iPhones which actually allow installing these old backed up applications. Even then, the entire ecosystem is completely reliant on internet authentication to even install the OS, so if your phone needs to be restored -- good luck. Very sad times, indeed.
Yeah the personal effect of this for me is, I can't show people the song I had in Tap Tap Revenge anymore. I do have an old iPhone that still has it installed, but it's not like I carry that around with me. The app is no longer on App Store and I thus could never install it on my newer iPhone, nor restore my backed up copy to it (and even if I could restore it, it wouldn't run anyways because of the breaking of backwards compatibility in recent iOS updates). Overall, a whole ton of user-hostile product choices in the iOS ecosystem, sadly. Not the Apple I grew up with, where I could modify anything to my heart's content (ResEdit, anyone?)...
Some of those NES and Gameboy games might not work as originally intended. Lots of them had internal batteries in order to save information and those batteries are starting to die. Floppies from that era and early CDs are also at the age in which they are likely to degrade. So while some of this old tech might have a longer shelf life than tech today, as the video says, entropy will get us all in the end.
Well you can still download the rom and the machine is fully emulated, forever. Not sure there is a way to play geometry wars in any fashion. Besides batteries can be replaced and even then you'll miss your high scores but you'll still be able to play.
I've still got Geometry Wars on my 32-bit iPad. It's my go-to airplane game—if there's a 120v outlet to plug into.
No multiplayer, though.
Could you reinstall it if it was uninstalled?(real question)
I don't think so. It's part of why I haven't.
Most of my floppies are dead, but most NES games did not have batteries, and it's really easy to replace the batteries in the ones that did (it's usually just a CR2032).
Replacing the battery isn’t hard. There are plenty of youtube instructional videos on it.
This is what I experienced with Tap Tap Revenge on iOS. What used to be one of the highlights of iPhone and iPod touch gaming was eventually bought out by Disney, ported to Android (poorly), and then phased out. Luckily, users had dumped some of the song packs and put then online, and with some reverse engineering and hooking, you could swap the game’s server endpoint with a custom one and play it again. This was my main side project in high school junior/senior year and a highlight in my college application.
Unfortunately, it also became victim to Apple’s discontinuation of 32bit apps.
Forget kids, I'm an adult that's missing some early adulthood games.
Flash, man.
iOS and Android have been dominating for a decade - i bet there will absolutely be ROMs out there.
The flash games I played as a teenager at the turn of the century are about to be completely gone, like tears in the rain.
* https://www.kongregate.com/games/scarybug/chronotron * https://www.acno.tv/acno/
You might be happy to know about Flashpoint, the Flash games conservation project: https://bluemaxima.org/flashpoint/
The developer of the game has the responsibility for it. It's very simple: He/she can just release the code and game assets. That will make sure that the game lives on forever. Anyone interested in it will pick it up and port it over to new platforms.
Even if the developer does not do this, it's still possible. Some people will get the game assets and reverse engineer the game. There are countless examples for this. Unfortunately sharing the game assets will be illegal (at least for a while, until the copyright ends), so this will make it harder for the game to survive.
Also, there will be emulators at some later point, where you can play the original game binaries.
I have a humble bundle set of android APKs I think.
The will once emulating an iPhone is something that people can feasibly do. While there's been commercial interest in this area, it's not widely available yet.
Will they ever have an iOS or Android equivalent of ROMs?
Yes. Torrent sites are littered with archives full of apk files already.
It's pretty easy to backup and install Android apps to/from a file.
You see, that’s not in Apple’s business plan. If they can’t extract money from you while pushing down developers, they don’t care to do it. We don’t own our devices like we used to. The terms of service alone are egregious.
All the software we write is ephemeral and will one day either be retired or stop working.
You can say that about anything we do.
However, at my previous job I integrated an LFBGS[1] implementation into our production code.
That was written in the early 80s in Fortran '77.
The code outlived all hardware that it ran on when it was first written (we ran it on an OpenMP cluster for a scientific-computing problem related to lithographic mask optimization), if not its authors.
It will continue to exist, and run, for a very long time.
Sure, all software is ephemeral. But as you say so, you probably used SciPy/NumPy. Deep inside, there's an implementation of LAPACK/BLAS doing the heavy lifting for you[2]. It started in 1979, and is still kicking.
(And it's still in FORTRAN)
[1]https://en.wikipedia.org/wiki/Limited-memory_BFGS
[2]https://en.wikipedia.org/wiki/Basic_Linear_Algebra_Subprogra...
Wow, I'm actually using some of the C++ code based off of MINPACK for one of our current projects at work. It is quite humbling to realise this dates back so far.
Unless you're a game developer! Lots of people are still playing Super Mario Brothers daily! They're running it in an emulator now, most probably, but all of the original code is doing the work.
Obviously it won't outlast the heat death of the universe or anything, but I'd say it's has the potential to live as long as any other human creation.
That's a grand assumption :) Who knows whos GitHub code will be included in the "Big MasterScript" that is written into the consciousness of the universe in the future, making it forever.
I know, far out there and in theory, I agree with you, everything is temporary and nothing is forever. But who knows, maybe in the future, things will no longer be ephemeral.
Probably left-pad
Like ourselves and the rest of the universe.
I think that thinking this way is a coping mechanism in response to the massive loss of control we experience being subject to the whims of giant corporate platforms deciding what we can and cannot run on our own devices.
The alternative is to face how insanely unfair and belittling it is to have bought an app and a device and due to circumstances out of your control arbitrarily no longer be permitted to run “your” app on “your” device.
You think that being aware of our own inevitable deaths, and the eventual end of the universe, is a coping mechanism to avoid the unbearable reality of closed software platforms? Am I talking to Richard Stallman?
You never know.
If humans discover a way to defeat entropy, even at a limited scale, whatever device is created to preserve information into the era of a universe solely made of black holes and hawking radiation will be capable of running doom.
I fully expect the last computer to ever be built by humans to still be able to load original .wads.
> lot of software we write is ephemeral and will one day either be retired or stop working.
You're giving me a lot of credit by implying the software I write works to begin with! Can we hurry up and get to the day with software like Facebook and Twitter stop working?
If it's actually spot on, it's a miracle.
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.
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.
Same here. However I think that the way he is communicating information in this video is meant to capture your attention not just now, but when this video is visited months or years from now.
It will be very cool for viewers to stumble across this video when it doesn't work, effectively proving his point.
My guess is it's more likely some YouTube employees hacked around this as a show of "support" for YouTubers?
More likely they have a cache for the views with a longer expiration time set than the update interval of Scott's script.
That's unnecessary. The slower the view count updates (and it's very slow), then the more likely the title is to be correct.
That's what I thought too. He probably got a rate limit lifted.
YouTube's view counts update at a notoriously slow rate. It's not hard to find new videos with more likes/dislikes than views.
> new videos with more likes/dislikes than views
That doesn't mean that the view counter is slow. Views are only counted when you watch the video for X seconds (or percentage complete, can't remember) while you can like/dislike the video by going to the page, clicking the like/dislike then close the page before the view would even count.
Same here, they matched.
I think his point was more that it won't increase every time you refresh the page.
But then neither will the actual view counter.
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.
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.
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.
I guess one big difference is who is paying the fee. In an ethereum "app", the code stays dormant until a user interacts with it and pays a transaction fee.
Yeah, interaction with Ethereum APIs costs a little bit instead of being completely free, but that’s what makes it sustainable, I think.
As a result Ethereum apps/platforms don’t need to be centrally owned or become ad-supported and won’t die when its maintainers vanish. This also serves to stop abuse like spam, which would become too expensive to perpetrate.
Free apps and APIs were a good way to bootstrap wide internet adoption, but I think users might now be comfortable paying fair nominal fees for interactions instead of dealing with free ad-filled, privacy-invading services.
Pool Together is sponsored by Maker.
Reminds me of https://news.ycombinator.com/item?id=3742902 ("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 :)
> really set the tone for what HN is really all about, for me
Can't tell if you're referring to clever hacks or upvote farming, but I agree!
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!
You’ll enjoy the rest of his videos then. He’s one of the few Youtubers whose videos I watch regularly. In addition to being interesting, I always find them well-researched, well-produced and exactly as long as they need to be without any fluff or clickbait.
You should watch his other videos. He often comes up as recommended for me. I think it's the accent and the cadence of his speech that is so appealing.
I don't know if it's accidental or learned but he's really nailed the speech style of typical BBC documentary programming. Either way, his videos are generally very good and entertaining
>If it's actually spot on, it's a miracle.
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.
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
"The title of this video won't be exactly right. [..] If it's actually a 100% spot on it's a miracle"
the title was exactly right when I saw it the first time. I even screenshotted it.
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.
It's apparently very common to see it be exactly right. My guess is that YouTube is doing enough batch/cache/snapshot magic to view counts, and applying the scheme equally to the web UI and API, such that it's not actually necessary for the script to poll super-frequently.
Or, someone at youtube saw the video and thought it was cool.
Like how this video will always have 301 views: https://www.youtube.com/watch?v=oIkhgagvrjI
The view count is definitely cached for some second at a time. Part of me wonders if that statement is just part of the viral-magic of this video.
No way! He said it would be a miracle!
Pretty sure only hn admins can edit a title.
Users can too, at least I can, but probably there is a karma threshold.
In the opening snapshot of his code when the video starts, Tom's referencing a URL to YouTube which takes you to, you guessed it, a Rick Roll. :)
This was brilliant, and being in the trending youtube list, a majority of viewers will never have found that.
I can't take credit, I found this same comment on Reddit. :)
Reminds me of the CGP Grey video that shows how much the video made in ads (https://www.youtube.com/watch?v=KW0eUrUiyxo)
Speaking of the YouTube view count, Tom Scott also did a great explanation on distributed computing and eventual consistency: https://www.youtube.com/watch?v=RY_2gElt3SA "Why Computers Can't Count Sometimes"
And here's a video from Computerphile about overflow: https://www.youtube.com/watch?v=vA0Rl6Ne5C8 "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.
If anyone is curious, I took a crack at building out a bot that does this with Node.js and the YouTube Data API! (https://github.com/stursby/this-video-has-x-views)
Does anyone know of some reading material on "entropy will get us all" perspective? What does entropy mean in this sense?
If you're into sci-fi short stories, Isaac Asimov's "The Last Question" [1] is more or less the story about this, although it's heavier on the twist ending aspect and lighter on the solving-the-problem-in-reality aspect.
heat death of the universe.
I love Tom Scott's videos. He does a fantastic job of explaining complex things in a simple way.
I want to make an Instagram that will only ever have 42 followers ;)
Instagram has one of the least interesting APIs of all popular services, with basically just 2 GET endpoints or something. So if you do want to build this, you're gonna have to do a lot of reverse-engineering of the smartphone application. Probably a fun project on just it's own.
You could automate something in their m.instagram.com site with selenium :)
Guessing they are redirecting to instagram.com based on user-agent or something, because I end up on the lame read-only desktop website. Does m.instagram.com allow you to post content and everything the mobile app allows you to do?
Sorry for the late answer. I can post content from m.instagram.com if I change the user-agent as you mention, but this is quite easy to do with the browser default dev-tools, just press F12 and change the device to any mobile. I think the only thing it does not work is the stories.
Web 2.0 was a big heart of love.
:(
The title was exactly correct
This would be noteworthy if the counter was included in the video (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.
You seem to have missed that this is a science-communication video aimed to help lay-people understand APIs, standards, online abuse, privacy issues, and how automation is playing into the issues we have in society stemming from social media.
I don't think that changes the fact that the gimmick, that gives the name to the video and the submission in HN, is pretty unimpressive.
Who said it has to be impressive? It's simply something fun that relates to the deeper dive he takes in the video.
I'd go further and say that it needs to be simple above all else, to help communicate the message. Confusing that with the ideas of video editing/compositing and uploading would make the message less effective.
> This would be noteworthy if the counter was included in the video
Pretty sure that would be impossible, unless I am missing something.
I was thinking it would be a video counting from 0 to 1 million or whatever. Combined with a service to redirect you to the video at the correct timestamp using YouTube's "t=1m23s" parameter.
But 10 hours of video with 1s timestamp parameter resolution would only get you up to 36000
Similar things were done by generating a live video. Heck, you could just show the page for the video in the live video.
The problem would be in keeping the transmission live for a long time.
It would start out working fine at zero.
I agree, but when someone does something we consider to be impossible, that's interesting and noteworthy. This, not so much.
Yeah unfortunately Youtube doesn't let you edit videos, which sucks for creators who want to fix minor mistakes after a video has gone viral.
You can, however, add and edit "cards" on top of the video.
On another note, it would be noteworthy if the title on Hacker News also included the counter.
It’s sucks for them (and their viewers). It doesn’t suck for viewers who want to see the thing that went viral - not some new video with product placement etc.
Like Reddit comments with a thousand edits after they become popular.
If such a feature existed, it would be used for abuse 99% of the time
Vimeo does have such a feature. The ability to do such could be based on karma on YouTube.
Uh, why the downvote? Can HN institute a policy that a reply is required if you downvote?
You're just farming more downvotes by clearly being affected by them and then complaining about them.
If this is how you respond to pointless vote counts, I would avoid revisiting comments after you leave them for the sake of your mental health.
> Please don't comment about the voting on comments. It never does any good, and it makes boring reading.
> Please don't submit comments saying that HN is turning into Reddit. It's a semi-noob illusion, as old as the hills.
I didn't leave a downvote, but that sounds like a horrible policy. You'd get a ton of short comments on the least-valuable (as judged by voters) comments, bloating the comment tree and making it way less readable.
It is fortunate, not unfortunate, that YouTube prevents editing of the video itself after a video has been uploaded, viewed, interacted with, shared, commented on, and started to accumulate karma of various sorts.
That would be spooky, we would hire an engineer like that on the spot due to the cross-domain knowledge required.
I imagine it's possible to update the counter as closed captions for the video....