Really not surprised to see rauchg [name redacted] steal more ideas for his startup from colleagues.
The entire premise of Zeit is something that I explained to him five years ago in a taxi ride. I was in active communication with Guillermo about developing the idea up until two years ago when he went silent for a month or two and I then I read the announcement for Zeit on Hacker News.
I'm sure he'll sell Zeit to the first high-bidder and completely throw the development team under the bus ( like he did to TJ at his last startup ).
If you don't believe me, you can just look at the Github activity for the Zeit organization. [name redacted] has copy and pastied the same Now project into new projects several times ( which deletes the commit logs of several contributors giving him sole Git credit for the projects ).
Not when you have to actually make money haha. Can't really execute without money, that's the entire point of startups, starve out the little guy with your free offerings until you can sell the team and hype.
Well, as somebody who tried building a bootstrapped and self funded Node.js PaaS competing against Heroku, Dotcloud, and Marak back in 2010[1][2] (the good ole'days) I am certainly sympathetic to the struggle competing against well funded companies. Ultimately though, I prefer lifestyle businesses these days. Funded companies tend to slow down, carry ridiculous baggage, politics, and usually result in less money in the founders pockets. Being self-funded and thus moving fast and agile is an advantage.
Im sympathetic to your money-disadvantaged position in this pursuit, but its just what it is. You got out-monied in this and have to continue looking for a feature that wont be copied down easily. It's not a defensible position if a competitor can copy it immediately.
Much what i was going to say, until he mentioned the continuous conversation over the ensuing 3 years. That's still no slam dunk. Semi-regular emails fleshing out an idea are still far from actual "execution", but we don't have more info and the nature of the collaboration, so i withhold judgment.
I'll say this as a more general comment on this sort of thing: I work more within the realm of academia, and of course in that realm there is a very strong requirement and culture of citing all sources, influences etc. In that context, this issue might not appear, or not as much. Giving credit as a tertiary contributor is both easy and "free". That's harder in the startup world because boundaries if activties are more blurred, and giving any sort of formal recognition of influence or minor collaboration can cost a lot in terms of equity etc. At least, that's my impression, from the academia end of things.
We aren't talking about what Heroku does here. I understand the confusion since the space is similar, but it's like comparing oranges to orange juice.
I won't be going into details here on HN because it's a somewhat pointless waste of time. There are valid reasons why TJ and I are upset with rauchg. I've said my piece.
There is a big difference between emulating or being inspired by another's idea versus actively collaborating with them and then simply stop working with them once you feel you don't need them anymore.
Without people like TJ and myself, Zeit wouldn't exist. Buyer beware to whoever decides to acquire Zeit.
Rauch talks as if he invented the one-command deploy, Heroku did this shit yearsssss ago, not to mention dozens of other platforms and tools. All they're doing is making a really crappy version of Heroku with 1/20 the features, poor performance characteristics, and putting a free price tag on it to keep people like me from competing.
It's really not an innovative idea. Trade secrets? We can all see a bunch of the code, and while I found it to be edifying in some respects, it's nothing special. Features provided to users cannot be trade secret unless you're somehow planning to hide them from your users.
While I don't want to sound condescending or like a promoter of fad philosophies, I've found that meditation and reading Stoic philosophy and really helps to build a thicker skin.
You sound as if resentment / envy is deeply poisoning you, and yet in the scope of business schemes it is really nothing.
Also, based upon your attitude here, it's not surprising that collaboration ended. Most people are insecure and envious, but you can't found companies with people like that as they will kill the company.
This was clearly rushed to get ready to compete with Up, it pretty much copies my blog post exactly in terms of implementation. Expect to see the other ideas from the repo in there soon.
As I said on Twitter, it's a CLI, literally anyone can build this. Startups have their place but in this case it's such a cowardly way to run a business.
Funny thing is it doesn't even follow best practices, they clearly don't have much experience in the space, which is evident in that they built almost nothing in two years, this is basically a hobby project.
I can't compete with free though, so congrats, best of luck with your joke of a "company".
All of their resource management for DNS, certs, domains, etc is done using sub-commands which are not reviewable.
Sure they can have an audit trail of events, but Terraform/CloudFormation are industry standard now, as changes can be easily previewed and reviewed by peers.
Single-command deploys are cute, don't get me wrong, but it's ultimately a gimmick. It's great for solo or small teams, but for larger teams, they'll ALWAYS use CI, and they'll ALWAYS want to restrict damage that a single team member can do to the infrastructure.
There's nothing novel about single-command deploys. Hell at Segment we had "/deploy blog" in Slack for the simple cases that we weren't worried about, tons of companies do this stuff. Most at very least have git -> CI deploys which is a more correct way to do things for real teams.
Yeah I have an issue open for it, but there's no point supporting others until AWS is very robust and feature-complete. Now's can't even tail logs by the looks of it haha, extremely premature announcement.
So you're implying that Zeit has been working on this for less than that time? The idea of a universal deployment tool seems like common sense once your business already deals with deployments, and behind the scenes uses multiple providers.
Why the automatic assumption this idea was "stolen"?
Learnboost. They were responsible for creating some of the most popular node.js packages back in the day. But from what I recall, TJ was the most prolific developer there.
Anyone care to share a bit more context? If you ignore the tone for a bit I feel there might be some valuable lesson I'm missing. I mean, TJ seems to imply that while he has to work for money, Reich somehow gets it for free because of something unfair. What am I missing?
What's the mature way to handle startups ruining opportunities for people who actually want to earn a real living, instead of hiding behind VCs like cowards?
Sorry if my ranting comes off as immature, you're probably not wrong haha but I'm perfectly happy to discuss about how ridiculous Silicon Valley is, founders aren't special, they're not visionaries, just con artists.
I'm not saying life is supposed to be fair, startups are designed to keep people like me from competing, it's just a huge joke to me.
I just think you need to flesh out your case more. What exactly is your beef? That they are operating something that doesn't have value, but burning VC cash? Perhaps the niche you're operating in simply isn't profitable?
Seems like you're arguing that VC are funding something worthless or not worth much... but that you should be getting paid for building the same product?
I'm a nobody in this field but know enough about you to get that you're extremely productive and talented. Why not go this route:
Plenty of so called "lifestyle" businesses that make plenty of money. Of course it's not that easy, but find a niche and build a great product. Pure tech plays like cloud deployment infrastructure seems to be operating at such a hyper-competitive level, that there's no way to avoid either big VC money, or Google/Amazon themselves.
I am an indie developer already. It would be possible to avoid VC, if startups weren't able to offer everything for free for long periods of time. Competing at a the level of skill instead would be nice, their product certainly does not require a team or funding TBH.
Hell there's even a timeseries startup, been around for ~5 years, they had no clue how to write databases, got paid for 3 years to finally realize columnar makes sense for analytics... no offence to them but you see this all the time, re-loaders with some vague trivial "vision", bunch of clueless VCs and buzzwords, and boom they have millions. It's such a sad way to compete, at least have the skills first. It's like no one wants to actually earn things these days.
Marketing your idea to the public is skill. As you might recall, Steve Jobs couldn't code.
Also, picking an original idea is skill.
If the idea is not original or unusually compelling, then marketing is even more important.
There's more to business than code. Happy to offer some tips on competitive strategy and moat building if you'd like, altho I'm just an hobbyist investor with an econ degree from a former life.
> Most cloud providers have created different proprietary APIs to expose lightweight services to the cloud.
True. However, in the case of Google Cloud Platform, the Protocol Buffer definitions for their services have been open-sourced here: https://github.com/googleapis/googleapis
There are two immediate benefits/possibilities: you can generate your own gRPC client libraries for GCP services and you can re-implement GCP services using their open-sourced interface definition. One example would be the Google Cloud Functions Emulator [1], which implements the service defined in the Cloud Functions service' Protocol Buffer [2]. You could deploy that Emulator somewhere for a sort-of "dev" version of the production Cloud Functions service, and the Gcloud SDK could talk to it.
This is really cool. We're a one-man shop (me) that uses Heroku though. Even though I can get on AWS or GCP really quickly with this (but probably not because we have a Rails back-end and React/Express front-end), I wouldn't then know how to best leverage/manage it. So this is a bit of a non-starter for me.
So far I'm seeing this as useful for side projects at most. The layer isn't opaque enough where I can troubleshoot serious issues and its too new to know there won't be any. More layers add more complexity in the early stages, over time though this could get quite compelling.
Yea their every cloud is pretty limited. The world isn't limited to AWS/GCP. For VM hosting, there's Vultr, DigitalOcean, Linode. And for app hosting you have Heroku, Jelastic, et. al.
It's difficult to create something that works universally since the APIs can be radically different. Terraform is a nice attempt to plug in different hosting solutions and configuration layers (puppet, anisble, chef), but your configuration still varies pretty heavily per provider and I've often struggled with it enough that I gave up and just wrote scripts to call my providers API directly.
"Aside from "Function as a Service" providers, we also are working hard on bringing the Now experience to container management solutions such as Kubernetes. These systems provide a lot of power to their administrators, but typically don't offer a great "last mile" experience for product developers and designers."
I'm running a one-man shop too. And I deployed many websites to Digital Ocean/Vultures with the help of Docker machine, 5 minutes from starting the VPS to the website is ready
I've also done this with Dokku specifically and it was really cool. The project has probably matured a lot since then (8+ months ago) but...I do have a pretty high level of trust in Heroku that's been hard to erode.
(Waiting for all the comments saying I shouldn't have that trust in t-minus...)
I'm a little confused, is this basically a competing service with zeit's already-abstract 'now' deployment service? Or is this completely separate from that now service that doesn't even require a zeit account/plan?
The docs are very light on technical details. It appears for aws they are using lambda and API gateway. If so, how do you connect to a sql db that requires connection pooling?
It looks promising, but it's unclear how to use it for a real app. Perhaps the idea is you maintain a separate set of servers which talk to your db?
I saw this at the reactriot hackathon. Since it works with multiple clouds I think I will try to use this from now on. If it allows me to migrate easily there are distinct advantages especially such as lock-in.
[1]: https://medium.com/@tjholowaychuk/blueprints-for-up-1-5f8197... [2]: https://github.com/apex/up