Twitter to Client Developers: Drop Dead
daringfireball.net
daringfireball.net
The part I disagree with is that this will doom Twitter. It will not. They've already achieved a certain level of success. Most people use and will continue to use the Website or the official client and be happy with that.
I do believe that Twitter is doomed to be acquired however. Apple seems the likely frontrunner for this but I think Twitter needs Apple more than Apple needs Twitter at this point.
Twitter is ultimately infrastructure and infrastructure seems doomed to commoditization. Twitter has eyeballs too but social platforms seem fickle at best. There is nothing preventing Twitter from becoming the next Myspace.
There are many reasons I'm glad about Facebook's floundering market debut. This is one of them: it's taking the wind out of the sails of the social hype (IMHO).
Do you think perhaps restricting 3rd party clients is a way to make them more appealing to a potential buyer like Apple?
Twitter on the other hand offers good UX. They have loyal user base of millions of users from different age groups. They are huge in Asian markets. And they have celebrity users, governments, political leaders, olympians etc. It will be very hard to move entire user base from Twitter to some other service.
App.net is an interesting alternative. And they can disrupt real-time information market. I think App.net can have number of use cases like firehose of realtime feeds for devs . But killing Twitter entirely will be very hard!
EDIT 1: Grammar
EDIT 2: Twitter for me is more than infrastructure utility. It's my primary information network.
Devs already have a twitter-like firehose, and it's even open source, identi.ca
Every time I go to twitter in a browser I want to punch whoever designed it in the face. It is by far and away the most annoying site I encounter on a daily basis, which is impressive given that all I want it to do is display plaintext.
I figure out they have 2 designers there, one really good and one that doesn't have the slightest idea of UX and design
Case in point: the way DMs are notified to the user. That's right, in the previous web version it wasn't. And this version is better, but not great
I especially like that I can sync all the tweets I've read between my Windows machine at work, my iPhone, and my Mac at home. I don't know of any other app that works for those 3 platforms and has syncing.
There is a service called Tweetmarker (http://tweetmarker.net) which some clients are now using to flag your last-read location amongst other clients on other platforms (e.g. Osfoora on Mac and Tweetbot on iOS).
You have an interesting point about commoditization, however. It makes me think about the IM space and all the different services there.
As for eyeballs, I think a Twitter competitor can do what media companies do — pay for certain celebrities, authors, speakers to move their microblog to their service — do a few of those, and you'll start to have an audience.
Interesting take. Seems almost exactly the opposite to me. What makes Twitter valuable is not the "eyeballs," or even the infrastructure, but rather the pulse of content. It had what essentially amounted to first-mover advantage in the sphere of mobile content, and that is why none of the Twitter clones have been able to take its place.
That said, this direction Twitter is going toward "more closed" is wrong. Given that it's the market-researcher's goldmine of content that makes Twitter valuable, disallowing developers the ability to build infrastructure on top of that content seems really stupid. Raw data is not useful; the organizing of, packaging and presenting of that data is where the business value can be (profitably) delivered. It can be profitable for both Twitter and developers.
Like others I tire of the dragging out Myspace as that guy who started on weed and ended up an incontinent meth addict in the gutter but the core of this premise is correct, the 'useful' part of Twitter is as an infrastructure.
Small digression, when I was at Sun we had pushed out NFS to anyone and everyone, it was completely 'open' in the sense that we published all the protocols and anyone could build a compatible clients. Since this was web 0.01 the only servers and clients were in the same building generally but still the easy access, the documented protocol, and a thousand flowers bloomed. You could get your implementation 'blessed' as being standard by coming to Connectathon and proving you could interact with all of the other approved implementations there. NFS is everywhere, available for nearly every compute platform. It was infrastructure.
At an important meeting on the future of NFS (and a new proposed product called "ONC plus" which would have per client charges, and strict licensing controls. I argued with Ed Zander (then president of Sunsoft) over the wisdom of changing the NFS model. The business development guys had computed that if everyone that was currently using NFS was paying just $10 per client per year for a license, SunSoft would be the most profitable part of the Sun Microsystems universe. I asked Ed if he knew how many people would run NFS if it were $10/client, he pointed to the bizdev projections, and I told him no, it would be exactly zero. Zero because nobody would pay money to Sun for a technology they were not sure would work (AT&T had tried that with their DFS product in System V), and they certainly wouldn't base a business that needed it if Sun could pull the plug at any time or raise the price. And finally, the whole 'Open' thing only worked when you allowed other people to play. As I recall he reminded me to stick to the technical decisions and leave making a business out of it to people who understood such things.
Twitter is Twitter because it is Twitter. That twisted circular definition captures that something simple and free caught the imagination of millions of people and became something greater than itself. It became an infrastructure. But unlike Cities or other large corporations which have a revenue stream to cover the costs of their infrastructure, Twitter does not.
And so they are in the throes of discovering what is, and what is not, a business opportunity in the Twitter universe. And that discovery process is painful, and prone to missteps. Seeing Myspace wheezing in the gutter I do not think they would make the same exact missteps, while they could end up irrelevant, they have more options. They do need to understand how people value them, and understand how much of that value is "them" and how much is their partners. That is a complex thing. A great example of that process is looking at their on again / off again 'firehose' pricing model.
I think they have a lot going for them, but they have to figure this stuff out, and quickly. Folks like Google and Apple and Facebook aren't going to just sit around and do nothing. Watching them walk through the minefield that is API handling should be instructive to anyone here who hopes to do the same at some point.
> They have an unlimited product placement budget.
Every TV Show, News program, Commercial, Website, Company... places the "Go to facebook.com/ford" or "tweet us @nbcnews".
This results in billions of dollars of advertising and they need to pay nothing. This helped build both twitter and facebook. Google should figure out how to get this for Google Plus. Once they do they can expand as quickly as twitter and facebook.
Spoiler alert - the answer is 'be as popular as Twitter and Facebook'.
T & F didn't become so popular because of product placement. They got the product placement because they were so popular.
Also, proper URLs.
"Add your voice to the conversation! Just go to plus dot google dot com slash 114124849657167573853".
MySpace had that huge advantage of many sticky users, but they stopped improving it, and a competitor (Facebook) made an improved experience.
Twitter is safe for as long as there isn't a sufficiently better competitor - "better" being in terms of the benefit to users, not any engineering quality in itself.
Facebook rolled through phase 1 by being exclusive, then phase 2 by first saturating college students, then they exploded beyond what any social network had ever known by offering a clean, curated and controllable experience that steadily roped in older people by being the place to stay in touch with younger friends and family.
I always laugh when people trot out MySpace as a cautionary tale for Facebook, because it's irrelevant. MySpace was the 500lb gorilla in 2004, but Facebook is the only 5000lb gorilla that ever existed. Whatever bring them down won't be for the same reasons that MySpace failed.
MySpace definitely hit the growth curve early since it was wide open, and Facebook wasn't open to non-college-students until when, mid-2006?
This is what I mean by "figuring out their (Twitter's) value."
Imagine that their API is really a bunch of method calls into the Twitter 'object'. They could, as an example, make their simple API 'free' and as you tried to do more complex things charge for those things. Presumably they would structure it so that folks could make 'useful' twitter clients for free, but they would have to pay some price for making more full featured clients.
Of course the challenge there is that you can't really charge for features that can be built out of the free protocol stubs, because people will just scrape those free ones and get around your code.
This couldn't be further from the truth, and the thousands of copycat networks, not least from Google, are proof that networks can't be commoditized.
The only reason I am still on GitHub, Facebook, hn, reddit and Twitter - despite there being better 'products' and 'infrastructure'- is because everybody else is, and the same applies to them.
Is there anything preventing any social network from becoming the next MySpace? Seems to me it is all in the users, if the users leave you are the next MySpace. I don't see how you could force users not to leave?
I think this is the most important point, one that most people -- Twitter included -- seem to overlook or ignore.
What we need is an open Twitter-like protocol -- but for it to be successful it should be peer-to-peer, not server-based like Twitter. A few year ago I was working on a startup which would build precisely such a protocol, but my partners backed out before we could lift anything off ground.
One of ideas was to make the client double as a Twitter client, in order to attract a wider user base. This would be less viable now, with the new Twitter API restrictions, but I would still love to see someone develop something like that.
Bullshit.
To reinforcer this point: in the early days of Twitter it kept falling over under the load - the one thing infrastructure should never do. It didn't matter.
To further reinforcer this point: OSStatus/Identi.ca has a widely deployed, API compatible implementation of Twitter. That really is infrastructure, and yet it has had approximately zero impact on Twitter's growth or strategy - because Twitter isn't infrastructure.
(Unless you mean it in the way that every single online service is ultimately infrastructure - but I don't think this is what you meant).
Twitter has eyeballs too
You say that like it is a minor point, whereas actually it is 90% of what matters.
It also has the publishers people want to follow, for whatever your particular niche is.
social platforms seem fickle at best
Actually, social platforms aren't fickle: they are usually incredibly sticky (see the huge number of forums that have been running for 10+ years). Most people just look at the "social network" category and see how Friendster/MySpace/Bebo/Hi5/etc all got destroyed by the Facebook juggernaut.
They never talk about the success of LinkedIn/PInterest/Twitter/Reddit/etc.
In every case, those overran competitors of their own (including Facebook in some cases) to dominate their categories.
That's not "fickle", that is platform strategy and product development.
There is nothing preventing Twitter from becoming the next Myspace.
MySpace sold to News Ltd, and then went downhill. Maybe Twitter won't sell itself to a company that is actively internet-hostile. But yes, you are correct: potentially any company could fail.
I don't think that's it. I don't remember the acquisition disrupting their strategy much. Rather I think the quality of their code base and inability to keep up with Facebook's agility is what drove the nails in the coffin.
I don't know to what extent that is fair, but I can only imagine that at the very least highering the best and the brightest to work on a coldfusion project could be.. tough.
I find it very interesting that the rate of growth in MySpace's membership slowed significantly with 6 months of being bought[1].
I think the quality of their code base and inability to keep up with Facebook's agility is what drove the nails in the coffin
I agree 100%. And I think News Ltd was to blame: they didn't (don't) understand the internet, and didn't know how to manage a high-growth property like MySpace.
But then again, it could all be co-incidence.
[1] http://www.petehatesmusic.com/wp-content/uploads/2011/06/mys...
The problem with Twitter is the same problem that plagues innumerable startups - they've built a product without even once thinking about the business model, and are not scrambling for revenue.
Inexorably, everything they do in the pursuit of revenue is going to involve making the product worse for users. Ad injection, loss of 3rd-party apps that support the ecosystem, all of that is just example after example of how a business with no built-in monetization will eventually cannibalize itself in the pursuit of revenue.
Obvious they may be "priced" at a higher valuation - but anyone looking to buy them is going to have to pay up - they are a rare commodity if they are hitting those revenue numbers.
Personally, I've expected this for a while and never understood why anyone would spend time on developing a client - unless you're developing something for a business to manage their social network. I am a bit surprised they don't just put some requirements on what clients have to carry in terms of advertising and that would be it.
Reality of the situation is that most 3rd party clients outside of Twitter probably aren't that big relative to people going to Twitter directly now.
Whenever I get fed up with the complextiy of Facebook or Google+ I'll load up Twitter on an old Apple II, via TTYtter and a serial connection; I'll watch the green text scroll along at 1200 baud and think about how this one simple, geeky text service, pure as a 1980s BBS, somehow made it, worldwide, in 2012.
And now they're hellbent on ruining all of that. Fuck Twitter.
I say when twitter was still this little chick, their approach was "we love all users, we welcome engineers; build amazing tools and surprise us!". I think the reason for that was to speed up the process of spreading the word - a simple fact that geek working on twitter 3rd party is still a human with plenty of friends to spread the word about twitter - so he can be helpful: let him spend his time doing what he likes doing the best - programming and he will become our cheap (free) PR tube.
But now I bet most of a new age civilization knows or uses twitter. So it is time for a reality check: "fuck off of our platform; we don't need you anymore! You got all your friends to know twitter, some even addict to it; now stay away from trying to run your pathetic queries, using our own data stream".
Just my version/2c.
edit: my understanding is that Dorsey still has the most to say in the twitter world. With all its nastiness going on between twitter curtain, I say stay the hell far away from any startup he will do in the future. Sorry, but if he signs up half of the world on his square, what on Earth is stopping him from switching 2.5% to 10% fee?? nothing!! At least the past (present) shows he has the balls to execute moves that average tweeting Joe is not a fan of: shutting down 3rd parties, kicking out linkedin, shutting down instagram access, etc. Bottom line: stay away!
Network effects are the lion's share of the justification behind Square's Starbucks deal.
Twitter got the funding to build that massive infrastructure only because someone believed that they would eventually be able to monetize it. Twitter is not a public utility and everyone knew it.
The conclusion isn't to never use something one particular guy does. The conclusion is this: If you build on someone elses infrastructure, make a contract or make sure you exit before the tide turns.
What is the business advantage of Twitter (or Facebook, or what-have-you) releasing a free, public API to anyone who asks, and how did they plan to monetize it when it got popular? You can't build your business model around "here, use my service for free" and not have a plan how to convert either the users of the 3rd party developers' software or the 3rd party developers themselves into paying customers (or monetize on that somehow, i.e. mining data, selling ads, etc). Maybe I'm just being naive -- I honestly don't have much experience dealing with these sorts of things, so I would love it if someone could break it down.
If a business idea requires free use of other people's services then think of another idea.
However, there's another area that has gotten me wondering, non-app, non-client based websites using the API, in reference to the Display Guidelines..er Rules.
This is the bit specifically.
"Users must have a consistent experience wherever they interact with Tweets, whether on Twitter.com, a mobile client, website, or in an application developed with the Twitter API"
So lets say that I go to GitHub and grab a little jQuery plugin to pull in my tweets on my personal portfolio. Does that also mean I have to make sure I include my own avatar, my username, Tweet actions, and twitter branding, among other things? What if those elements are unnecessary to the design or intention of what I'm doing on the site?
And then there's the fact that all of these jQuery plugins are going to have to start implementing authenticated access (if they weren't already, which many seem to not be.) I don't have access to data on the matter, but I would surmise that there's a significant number of personal and portfolio sites out there pulling in tweets that are either not authenticated, or are modifying the tweet display in some way. All the ones I've interacted with have settings for turning avatar display on or off, or unlinking hash tags or links, etc.
Thoughts?
It needs to be Open Source so anyone can run it and everyone owns their data.
It needs to be compatible with current Twitter apps so all it requires is setting the API root.
It needs to be distributed so anyone can follow anyone anywhere. There is no owner or root, there is no place to shut down.
Proof of concept: https://nstatus.herokuapp.com
Source: https://github.com/maxpow4h/nekomimi
I wrote about the requirements of it here: http://maxpow4h.com/blog/twitter/
edit: you can use any username with any password to sign in to nstatus. It then uses that password for your username. You can even do this from the official iOS Twitter app, just sign in.
In the distributed implementation I would have a chain of trust from the people you follow, where if you report someone as a spammer, the server that represents people who follow you could look at your list and filter spammers based off this. You could also set the depth to traverse. There is a lot you could do with this.
I'm not saying my implementation is right, but someone needs to do it. Tweeting is a type of communication and needs to be distributed.
Maybe even taking Diaspora and taking out everything except 140char status updates AND adding an API compatible with Twitter apps would be a better approach.
edit: clarity to first sentence.
- open-source - open-standards - free - federated
Keep your own data, talk with whoever you wish.
Built on XMPP and actively contributed to by a great group of developers (admission: I am on of these developers).
When you join up you can find me at lloyd@evilprofessor.co.uk
Or jump into the https://beta.buddycloud.org/team@topics.buddycloud.org to ask any questions about dev issues.
As an aside, I feel even better about backing App.net after seeing this news.
We need open protocols and a decentralized social network where we have different implementations with focus on different use cases which are still able to communicate with one another.
A good chunk of the protocols and implementations is already there with OAuth, Activity streams, salmon, RSS and such. And there are already open sourced implementations like status.net and others alike.
The solution is already there. We only need to start using it.
Source: http://techcrunch.com/2012/04/23/zynga-made-up-15-of-faceboo...
Theres no way to stop you from building one anyway, twitter knows that. If you go against their rules, you're a revolutionary, and if you win that revolution, they'll have to deal with you.
You can't expect however, that the incumbent is going to go around encouraging revolutions against themselves.
The only alternative is to encourage everyone to make clients, at which point, they're just a big cloud xmpp server to the world.
Declaring war on your third-party developers is shortsighted at best, fatal at worst.
Professionally I would love to have free access to the Firehose, that's some juicy data. But as a consumer Twitter 'apps' are no more worth my money than email apps.
Is it? Their revenue stream will be coming from paid tweets, so aren't the tweets themselves their main product?
It seems grossly unfair that they punish other 3rd party clients who are doing exactly what Tweetie were doing 2 years ago.
Twitter bootstrap, on the other hand, is something they built for themselves to reduce the friction of creating rich, consistent UIs. Developing it is a sunk cost. Offering it free to the world should cost them nothing as it's hosted on Github [1].
Like others in this thread, I lament that Twitter has become less egalitarian. Moving forward, developers with one idea of usage will have more Twitter clout than developers with "lesser" ideas, but Twitter has bills to pay and investors to answer to. It's far from an ideal world, but I can't offer them a better solution for their problems.
[1] https://github.com/plans/ all plans offer 'unlimited public repos'
Translation: so that we can charge even light API users.
This will surely backfire - some services will switch from API usage to screen-scraping, resulting in an even higher load on twitter's servers.
Most people building products or sharecropping on other people's platform never make meaningful income and yet those platform keep will prefer to announce large sums paid out to developers to encourage you to keep building complements. Apple will claim they paid out $5 billion but spread the numerous app developers it becomes peanut and not enough to pay their bills. They won't tell you that to pay out $5bn they made atleast $2billion based on their 30% cut.
They don't tell you that iOS app success is a "lottery": 60% (or more) of developers don't break even
http://arstechnica.com/apple/2012/05/ios-app-success-is-a-lo...
http://andrewchen.co/2012/08/15/mobile-app-startups-are-fail...
I really hope people will think hard before building their business on the back of Apple, Facebook, Twitter or any such platform. You can use them as as distribution without being dependent on them and that is the way to go.
Be your own bitch and not a Twitter, Apple or Facebook bitch:
http://techcrunch.com/2011/05/23/fred-wilson-be-your-own-bit...
This maybe true for a lot of tweets. But my stream is filled with live updates of interesting stuff from around the world.
Twitter has been instrumental in getting news out of war zones for example.
Surprised they didn't try to make it a bit less obvious.
Seems to me that developers are getting all pissy because they can't have free reign to a platform that isn't even theirs. Perhaps, at the end of the day, Twitter doesn't care - they don't need the developers as early adopters anymore and it must be a fair old strain to keep the api infrastructure supporting them when the resources could be more profitably used building something else inside the company.
app.net suddenly looks more attractive.
The only thing close that I can think of is BitTorrent due to the distrubuted infrastructure, but that's quite impractical to duplicate in this context (how can browsing the content contribute back to hosting and storage automatically?)
In an ideal world I agree, but that's not where we live.
Email and the Internet won because the sector was competitive enough that no one service had enough market clout to not offer e-mail. Once some people got e-mail accounts, other people wanted to e-mail them, and threatened to move off the services that didn't support it in favor of services that did. That scared the online services, so rather than lose, they all started offering e-mail themselves, until they lost anyway.
If you want open social-networking platforms, the solution is to make the market competitive enough that no company can afford not to be open. Usually that'd involve choosing to use the smaller players until they're big enough that no one company is gigantic.
Most of the UKs email over uucp came in via the ukc gateway and IIRC I think I was charged something like 4p/1K for email that was international at that point. From memory my usenet feed came via a US bank in the City and was over a telebit trailblazer modem.
By 1990 I'm pretty sure that cix,MCIMail and Compuserve all at least had gateways to SMTP mail and within a few years it all became more more transparent and free.
We'd need some sort of DNS-for-handles in order to implement something similar for a new protocol, unless you want your handle tied to a domain (in which case you've just remade email with a RESTful API and a length limit)
I don't know of any standards-based activity discussing a twitter-like capability but multiple interoperable implementations, open and closed source, free and for money, provides an 'ecosystem'.
I would contrast an ecosystem to a platform. A platform is owned or controlled centrally whereas a successful ecosystem isn't. A well-managed platform ultimately provides the most benefits to the platform owner whereas a successful ecosystem provides the benefits to the most successful entities within it, a playing field to compete within.
The issues here are open standards, open protocols and the small matter of who pays for the upkeep of the infrastructure necessary to provide the service. Tired, Maoist-style sloganeering is merely making a noise.
This logic assumes, of course, that Twitter's web app and official clients are good enough substitutes for the third-party clients that if the third parties go away people will use the official clients rather than just stop using Twitter altogether. Which remains to be seen.
Not bullshit, these are demonstrably useful products for people who want to utilize new traffic sources in the interest of making money or growing their business.
"But services like Storify and Favstar, which are actually useful and/or fun, those are no good."
ummm...ok? Utility is in the eye of the beholder...Twitter is a B2C product, so they're going to try and limit the number of competing services that are stealing B2C market share from them (why would you expect them to be ok with people using alternative Twitter apps instead of the official one?)
Now, Twitter is not a B2B company, so it makes sense that they would allow those types of services to continue. It's quite possible they're allowing B2B services that are utilizing the twitter platform to continue operating because they plan to acquire a few of them in the future to try and actually make a profit one day.
"Twitter is a B2C product, so they're going to limit B2C"
"Now, Twitter is not a B2B company"
The dangers of sharecropping are well known around these parts, Twitter has squandered their reputation and has brought these things to the forefront of anybody who would consider building a business on top of their platform. It will not end well for them.
1. "consumers" to be analyzed
2. Companies to use the service
Well at least it's rather clear what the new and future values of Twitter are.
Cynically, the more garauntess they can give to the other three quadrants, the more money they can make from them.
More positively, things like requiring reply/retweet/follow buttons with embedded content encourages more interaction with the service.
If a (lazy) journalist makes up an article from collating a few tweets you're now encouraged to reply, retweet and interact with twitter right there rather than using the Facebook Connect comments system at the bottom of the page.
Their value to the other three quadrants is interactions. They're not stupid enough to intentionally reduce that, though one could argue they're not going about it the best way.
But then what does that mean for a Tweetbot user like myself? Less incremental updates? One day the application breaks? Who knows, too early to tell.
I'm mostly surprised Twitter isn't just leveraging the fact they are pumping out that many requests and slipping ads or promoted tweets or promoted tweeters into the API stream and making cash off that. Seems like it'd make sense.
The only other reason I can fathom why they would doing this is they eventually intend to heavily push ads over their network, something that third party apps could interfere with.
Wouldn't an easier solution for that be to tighten the rate limiting for posting?
> The only other reason I can fathom why they would doing this is they eventually intend to heavily push ads over their network
I think that's pretty much it.
Still not sure if Twitter isn't just blogs that include a friends list and a reader (kind of like Tumblr, which seems to be taking off, too...). And the short messages.
http://rstat.us/ http://identi.ca/ http://status.net/
Which is nice, but not really much. Wouldn't it also make sense for, say, Wordpress or phpBB (or any and all other blogging and forum software, these are just examples), to also support OStatus? Everything that has content and updates, basically. Why not build OStatus apps instead of Twitter apps? Am I missing something?
Or are there other, similar standards/attempts? Because to be perfectly honest, I STILL haven't found the tutorial that just "tells me what to do" to go from offering an RSS feed to supporting OStatus, I'm kinda dense when it comes to these things. I need code examples :/ It can be pseudo-code, but it needs to be complete, instead of just a high-level verbal description ala "first you implement subhubbub, oh, and then there is salmon". I digress, but any and all hints would be appreciated. After all, it's for a good cause ^^
As usual what is needed is a decentralized approach, but that always takes time to catch on even if it can be made to work.
LiveJournal too had and has third-party clients. And though hardly popular in the English-speaking world these days, it's still going. But I don't think it has similar guidelines. So what's the comparison?
Would he rather they put one-paragraph labels on each quadrant? Sometimes phrases that sound like MBA buzzwords (and might be MBA buzzwords) are actually useful too.
Anyway, I think the matrix does make it more clear than just saying "we don't want 3rd party twitter clients...and some other stuff".
Social CRM (Customer Relationship Management) is use of social media services, techniques and technology to enable organisations to engage with their customers.
Social CRM is often used as a synonym for Social Media Monitoring, where organisations watch services like Facebook, Twitter and LinkedIn for relevant mentions of their product and brand and react accordingly.
Since when did people consider unlimited access to Twitter's API an unalienable right? It is their platform, and their API. It's also completely free.
Twitter is absolutely free to limit usage of their own API however they wish. If it means they want to change the rate limiting on their servers, I see no problem with that.
The bottom line is that Twitter's communications make it clear that they are grasping at straws for their profitability and that no developer is safe if they think you are capturing too much value. Obviously you always run a risk building on someone else's platform, but Twitter's direction and language around this would make any developer a fool to start anything new on that platform.
Of course if you're cynical you could have called it from the very beginning, but the point is it wasn't wholly unreasonable to believe that Twitter was willing to let developers capture real value in exchange for becoming gargantuan and providing critical infrastructure (ie. the old changing the world thing). Whereas now it's very very clear that Twitter is not willing to share any significant value. They want to capture it all, and they are adjusting the ToS to put them in a position to cut people off at the knees the minute they smell some value escaping from the ecosystem.