I bought a free website for $50 billion after my wife left me.
I was forced to buy a free website for billions after I posted cringe during an ambien binge
and now I'm in hock to some saudi guys for a return on investment
$100 a month and he'll throw in a set of steak knives and a hair trimming kit. Please pay me, please, please, please, please. I'm a fucking idiot please give me money please the Saudi's are riding me for a return on their investment or they'll snap my thigh bones please please i'll be your friend please i'll dance like a little racist raccoon please pay $100 a month please please please everyone hates me please I cry myself to sleep holding a replica Blade Runner pistol I bought on etsy for $19,000 I was told it was an original prop but it was made from toilet roll tubes and PlayDoh please please my children don't return my calls please oh god please.
Saudi’s are riding me for a return on their investment or they’ll snap my thigh bones
They have a bone saw for that.
reminder that Twitter has never had a software development life cycle, one of the most basic industry best practices imaginable, despite it being legally required of them since 2011
Wait, not even like, waterfall? What exactly do they do, then?
relevant section from the twitter whistleblower release:
- Software Development Life Cycle (SDLC): An SDLC is a uniform process to develop and test software, and a basic best practice for engineering development at commercial companies. Twitter's need to implement an SDLC was more than a best practice, it had been required since the 2011 FTC Consent Order and reported regularly to the Board of Directors.’ In or around May 2021, Mudge instructed that the Board Risk Committee receive accurate data showing that the company only had a template for the SDLC, not even a functioning process, and by Q2 2021 that template had only been rolled out for roughly 8 to 12% of projects.
Why does the government require development cycles for private companies?
it was part of a consent decree caused by Twitter's previous violations of FTC regulations, so like in lieu of prosecution they agree to certain shit
Lmao. “You violated the law. The punishment is that you have to actually do work”
looks like they also had to pay $150m but I don't know if that ended up sticking
just to follow-up, here is the actual consent decree with more background info:
https://www.ftc.gov/news-events/news/press-releases/2011/03/ftc-accepts-final-settlement-twitter-failure-safeguard-personal-information-0
and more digestible form, the DoJ's press release about it:
https://www.justice.gov/opa/pr/twitter-agrees-doj-and-ftc-pay-150-million-civil-penalty-and-implement-comprehensive
I know Twitter had profits prior to Musk, probably engulfed by the debt servicing subsequently but we'll see. I've worked for/consulted at a surprising number of massive companies whose SDLC was about as well-established Twitter's template + barely any process whatsoever.
If he removes the API, bots are just going to scrape instead which will use more bandwidth and cost more money.
Tweets are delicious candy. When you eat candies it comes with packaging and a wrapper, like the twitter website interface. API allows software to access the candy without the wrapper or packaging. If Elon removes direct candy access for the bots and other software, the bots will instead take packaged candy. When the bot removes the wrapper from the candy, we call this scraping. The candy packaging is useless to the bot because it just wants the candy. The extra candy packaging is going to cost twitter more money.
API is a way for software to access website data in a block of data. When a human visits a website, there is all sorts of html, css, javascript that gets interpreted by the web browser and makes the website data visually readable for humans. A computer program just wants a block of data, ex: the twitter handle, the tweet text, the tweet time stamp, etc. If a bot can't access an API, it will resort to web scraping, that is loading the full page and extracting whatever data they want from it. Full page loads use much more data than API requests and will cost most bandwidth fees to twitter.
https://en.wikipedia.org/wiki/API
https://en.wikipedia.org/wiki/Web_scrapingAPIs are access points to twitter that twitter has given you access to. You can get whatever twitter decides to reveal to you. Public APIs are available for anyone while private ones are limited.
Data scraping (or screen scraping) is a computer accessing a web page like a human does. It loads it in and then reads the data and then does this over and over again.
APIs are optimized for low resource use (usually) while scraping uses all the resources a human accessing a web page does and the scraper is doing it as fast as possible, possibly thousands of times a second. If you've gotten a CAPTCHA just to open a webpage that CAPTCHA is probably there to prevent data scraping.
API lets you ask for a specific part of the website, scraping means you load the entire thing.
HTTP APIs can be queried, eg you can ask for tweets from a specific user (probably) and get just formatted text with the contents of their tweets, and maybe some key stats about the tweet if you want.
Scraping works like a normal web browser, so any images or other stuff like tweets from random other people which you'd see in your browser as a normal user gets sent to a theoretical bot. It basically is forced to be as resource intensive as a normal user, or it would be identifiable and they can stop what they can identify. I'd imagine people will probably just scrape by emulating the mobile app though, that's usually easier when a site/service has it.
The tweet is about API, not AI. API is the interface that software uses to receive data from twitter. API basically sends an application a block of data instead of loading a full webpage. This makes it easier for software to interpret the data and reduces the amount of bandwidth that the site uses.
If Elon makes the API private, it's going to break all the open source twitter apps and twitter frontends like nitter. Also the supposed malicious bots will just switch to scraping, which will use more bandwidth and cost twitter more money. I think Elon just wants to break frontends like nitter.
Oh for sure, he sees things like nitter. Bots are more nebulous, but he definitely can't stand free-loading. He'd take the whole site pay-per-view but even he's not that stupid yet.
Ah, yes, because the wealthy would never abuse an API 🤡
This is also true.
scammers and opinion manipulators, famously MORE unable to handle a $100 upfront cost than the lucky duck wage worker
That $100/month is assuming you're only making 500 requests a month, which is fucking insane because even small applications which use the Twitter API can average thousands of requests a day. Total fucking lunacy,
5 API calls per dollar feels like it definitely has to be fake
Lmao this will kill off so many apps that rely on that access for all kinds of things. The extended network of tools that a lot of Twitter runs on will be wiped out overnight.
And it still won't stop bots.
Yes elon, break every third party service in the Twitter ecosystem. I'm sure it will go well.
Also I think nitter is safe from this, right? IIRC they do not use that API
Very Smart Techbro Demagogue discovers that overwhelming majority of internet traffic is bots, proceeds to publicly suggest committing website-suicide in response.
That’s so expensive holy shit
Dude if you charge a nickel a month it will have the same effect
I am pretty sure someone designing a botnet could spin up some headless chrome and manipulate your shitty website just fine, even with your checks and balances.
API, being abused... what a croque of shit.
Captchas aren't even remotely useful nowadays either, unless you use some unbelievably frustrating ones with ~1min timers for an answer. There's a couple that machine learning can only reliably crack that quickly 1/1000 times, where humans get it 1/20 during their first exposure to it and 1/2 the time after repeated exposure.
He's going to absolutely kill every cool bot that posts architecture or cats because he forgot that API posts are already identified as "Automated" so it's harder to use them for scamming people
