Fastly acquires Glitch
fastly.com
fastly.com
1: I know it’s been years in the making, but I’m still slinging Docker containers in a single AWS region.
https://www.glassdoor.com/Reviews/Idera-Inc-Reviews-E317065....
This one made me laugh:
> - Idera Inc has an anagram: Acridine. Here are some facts about Acridine:
> ○ It's a colorless (and probably soulless) solid obtained by dousing a foul smelling compound (coal tar) in a destructive liquid (sulfuric acid)
> ○ Both it and its homologues are weakly basic
> ○ Acridine is a skin irritant
In other Joel track record exploits, he's also our co-founder over at HASH.ai [1] and a driving force behind the Block Protocol [2][3]
[1] https://www.joelonsoftware.com/2020/06/18/hash-a-free-online... [2] https://blockprotocol.org/ [3] https://www.joelonsoftware.com/2022/01/27/making-the-web-bet...
A long-ranging conversation with Joel on, among other things, confluence of Catholic theology and the Talmud plus the memorable phrase “Shouldn’t you apply your skills to something which isn’t totally bullshit?” caused me to have a sharp reassessment of life and career goals, but for which it is unlikely I would have made a serious go of my consultancy, successfully launched my following few companies, or continued writing at anything like pace observed over the interval.
Those who were there would be better at telling the story, but up until the Atlassian acquisition and Joel deciding to spend most of his time on a collection of other projects[1], my sense of FC was "software for developers." The original product they made for project planning (FogBugz) and its SCM companion (Kiln) continued on for a while, but seemed to have faded as Glitch ramped up.
The other one that comes to mind is Copilot, whose development was chronicled in the Aardvark'd movie - I think Tyler took over the project and is still running it on the side, last I heard.
Kind of bittersweet to see these fade away, but Glitch seems to have had more critical success outside of their traditional user base so they're running with it.
[1]: https://www.joelonsoftware.com/2019/12/05/so-hows-that-retir...
edit: Maybe you shouldn't download random PDF's from the internet; here is the text and links:
--
Copilot is Closing
Hi,
I'm Tyler, owner and founding member of Copilot.
Copilot was originally built by a team of interns in 2005 at Fog Creek Software in New York City. I was one of those interns. In 2014, Copilot became its own company and relocated, with me, to Denver, Colorado. For the past 8 years, I have been running it independently, providing people all over the world with simple, effective screen sharing.
However, this past year, it has become clear that Copilot is no longer providing the level of service that I would like it to. After much thought, I have decided it is time to shut Copilot down. The reasons are simple: Technology is changing rapidly, and as an independent software vendor, I am no longer able to keep up with those changes, while maintaining compatibility with the older systems that so frequently need support.
The Good News
Because user experience is a key factor in closing Copilot down, shutting it down without an alternative was never an option. So I set about trying to find another company who offers similar services to partner with, to continue giving Copilot users a great screen sharing service.
To that end, Copilot has chosen HelpDesk by RemotePC¹ to provide future screen sharing services. I've spoken with the team at RemotePC and trialed their software extensively. I'm confident it will meet and exceed your needs.
As a bonus, they're offering Copilot users 95% off the first year, meaning that you will get HelpDesk for a full year for just $24.98, with unlimited users and computers. Or, if you just want to kick the tires, you can try it free for 7 days. Sign up here.¹
How It Will Happen
Copilot's servers will remain operational until May 1, 2022, and your account will continue to be active until that time. On May 1, all services will cease and all remaining subscriptions will be ended. For those who choose to move to HelpDesk by RemotePC, your account will be canceled automatically when you sign up for a paid account. As always, you can cancel your Copilot account at any time on your account page² or you can email support@copilot.com with any other questions or concerns.
Thank You
I want to thank everyone who has used Copilot over the years. It's been my privilege to be a part of Copilot from the beginning and to hear how many people it has helped over the years. Truly, thank you.
All the best,
Tyler
It's been kind of wild to see how FogBugz went from their main, almost exclusive priority, to something of barely any significance to the company, and now I'm guessing it's a forgotten product in a subsidiary of a giant CDN. Who knows how long until it's shut down.
It's sad; I still love the whole evidence-based scheduling part of it, and have never seen anything else like it.
I think a few forces came together to diminish the relative importance of EBS in project management:
- Rapid shipping got easier; rather than uploading executables (or minting CDs!) we shifted to the SaaS model and with that, continuous delivery, etc. In this world, coordinating a "big release" became more of a marketing/communication topic than an engineering one. In the FogBugz customer base, it was the game development companies that held on to EBS the longest.
- Developer tools in general got easier and faster to use, and along with that all of our tolerance for managing timers and estimates went down. Estimation and work tracking I think are still hugely valuable, but there's an ever-higher UX bar to hit to actually have people use the software, and we want the computer to be smart enough to figure it out on its own. EBS never achieved that fluency of UX and it really needs diligent users for it to perform well.
For the last ~10 years of Fog Creek, we were largely structured so that we had core groups of developers focused on our mainstay products, like FogBugz, which were happily profitable revenue sources and could fund all of our assorted bits of inventiveness. After pushing on FogBugz and Kiln in an innovative way for a few years, we came up against an adoption wall of sorts--- changing those products to increase their user base was harder than inventing entirely new products. Trello, in many ways, represented our next stab at productivity and software development tools, and making FogBugz more Trello-like was never going to be as compelling as Trello already was. This pattern kept repeating, and so we did our best to stabilize FogBugz while inventing other sorts of things that would show us a more compelling path to growth.
Glitch was the biggest next invention and its interest and adoption so greatly outpaced what growth we could achieve in FogBugz that it make sense to reorient around it. But of course FogBugz paid the bills and Glitch wasn't doing so yet, so that lead to a VC raise for Glitch and, ultimately, a sell-off of FogBugz and Kiln.
I still love FogBugz, and all of the users of FogBugz were ultimately the seed funders of inventions like Kiln, Stack Overflow, Trello, Glitch, HASH, CoPilot, and a dozen others that we never let past internal testing. Thanks, FogBugz :-)
When I interviewed at Fog Creek, they had a DVD copy of Aardvark'd in a care package in my hotel room. I watched it that night, and for my interview day in the morning it felt like everyone I interviewed with was a movie star. Sneaky plan, Joel. Well executed.
- HyperDev [ First public branding / early launch ]
- GoMix [ First major relaunch, now with Anil as CEO ]
- Glitch [ After acquiring the sweet glitch.com domain from the kind folks at Slack ]
Source: I helped work on it
It's in "Alpha" and works pretty well.
Enjoy!
Many union contracts also have an explicit "successor clause" too, although I don't know about the Glitch union contract.
[1] https://www.nlrb.gov/about-nlrb/rights-we-protect/the-law/ba...
Things have been good pre-acquisition, so let’s dissolve the union immediately before we get new management. Fellow techies I beg you: stop being so gullible!
As they bet the future on wasm-at-the-edge saving them from the crushing downward pressure of bandwidth pricing, having a compelling product for the "entry level"* developer market will be crucial to getting the mind and marketshare needed to hit revenue targets (eventually, not this year).
Put simpler: fastly made itself overly exclusive and difficult to use, glitch is the polar opposite. Fastly's customer base is 45 year old vp's of infra/devops/sre, glitch's is 27 year old app/front-end dev/designers. There's more money in a deal with the former, there's more of the future in the latter.
* I want to be clear, with ~20 years exp when I sit down to figure out a vue front end to my latest flask thingy I'm entry-level af in js/front-end land. Lowering barriers to entry, much like handicap accessibility, is usually a win for everyone.
Adopting and experimenting with all of that stuff will require developers fully embrace the CDN (oops, "edge") as part of their app architecture. A lot of places still treat the CDN as an ops/infra "config and forget" tool.
More heroku for rust at the edge than varnish as a service.
- Compute@Edge is a far better product than CF Workers. I think CF is catching up, but right now Fastly is ahead.
- Video. Fastly's streaming video offering is so much better than CF.
- Large file support (ex: +200gb game downloads), unless it was added recently CF is more limited on file size that can be cached effectively.
- Morals. Fastly has morals, CF is severely lacking. I don't know if this one could be "acquired" though.They also just announced a couple of crazily useful features that Fastly does not have [1] [2]. In fact, you can bet that most new things Cloudflare releases are complete innovations. They are changing the landscape with nearly every new offering.
On top of that, Cloudflare workers are faster [3]. Not sure how many people need 200+GB downloads, maybe that's a niche Fastly does cover, though given their history of dubious claims I'd want to see that working before trusting their declarations.
Then, they are open sourcing the workers codebase [4], which is absolutely huge, along with forming a working group for standardizing this type of interface [5]. It's everything you could dream of because they're leading the way before you know you need it.
Seriously no idea what you are talking about with morals. Like, in regards to which websites they allow to be on their system? I'm really tired of this argument. We can't just close off core internet services to people with whom we disagree. That only increases tension in society. If anything, that speaks to Cloudflare's higher moral fiber.
[1] https://blog.cloudflare.com/service-bindings-ga/
[2] https://blog.cloudflare.com/workers-for-platforms/
[3] https://blog.cloudflare.com/network-performance-update-platf...
[4] https://blog.cloudflare.com/workers-open-source-announcement...