I started off following along pretty well…until finally a few posts back I gave up and fell back on just liking everything haha.
Being able to hack a glowforge…would be pretty amusing at times, gotta say.
makes me think of “unprotected wifi-printer graffiti” where one drives around with a laptop looking for open printers, and then send art to them. It’s neat when it’s a single page of interesting or funny art (“what is this? where did it come from? This is great!”). It’s malicious when it is 500 pages of pure black (“why is the new $250 toner cartridge empty already? where is all of our paper? This is horrible!”)
3 Likes
If only everyone thought like the funny art person and not the pure black pages person!
2 Likes
Cant say I havent sent a few butt stallions over the airwaves before >_>
1 Like
Re: Spoofing… it actually happens more often than you think, but it’d depend on how untrusting GF intends to be. You could easily pull off a SSL man-in-the-middle using the RPi, if the GF doesn’t strictly check the entire SSL certificate chain all the way up to the Root CA. Some wireless routers, proxy servers and antivirus already do something similar, which sucks if there’s not some way to change how trusting a device should be.
1 Like
You’re getting at the edge of my knowledge. Understood what you are saying but have little real world experience other than setting which protocols are to be used. I’m assuming that how the communication protocols are implemented is set in H/W. However I would guess which protocols are used is more likely in the firmware. That’s one of the reasons why I hesitated to say the firmware wouldn’t have to be changed. But that change might be minimal.
This is an interesting discussion with lots of tangents. Maybe, someone already made this observation and conclusion??
Im thinking that it would be an opportunity for a small team to create a replica of the GF cloud service in a small server with WIFI that is in the same room as the gf . Essentially, transplanting what the cloud service is doing into a local PC server. It might be slower, but it could be the easiest route instead of reverse engineering, and possibly bricking 10k GFs. Maybe, if the time comes for GF to move on we could invest some time and funds to keep our GFs going, by licensing/buying out the code and installation procedure (if needed). Of course one would need access to spare parts too. OK, my coffee is settling down. Thanks for an interesting read this morning. Have a great weekend!
I realize some people take a lot of issue with GF going to the cloud, but replicating the GF ‘web experience’ isn’t what we’re discussing here. People should not get their hopes up for that.
It’s important to recognize that the work the GF team is doing in the cloud is the lion’s share of providing a simple, easy to use experience for users. In effect it’s “the application” people traditionally use on their PCs.
What we are discussing is akin to a “print server” aspect of the Glowforge, with the physical 'forge itself being “the printer”.
The best outcome you could expect would be that “transplanting what the cloud service (does) into a local PC” requires some other application to create the print data. The entire user experience would be entirely different, as you’d need to use some other application to generate data for printing.
Any such an application would not have any idea how to use the on-board cameras, for instance. You’d not be able to raster-lase penned objects using the camera. Or identify objects for a database of materials. Or in general, visual anything.
1 Like
Exactly. A print server, not an application. That’s what I was trying to describe in my non-technical explanation. I understand that all of the “magic” is server side as explained earlier in this thread, which I find both enlightening and interesting.
I understand that there is a web interface part which is very important to the user side of the equation.
It would need to be there too for many of the existing users. It would be very great if that part would work with a dedicated “laser print server” too.
I was not advocating creating an app, sorry for the confusion. Just copying the whole thing over to a dedicated server that runs whatever it needs to provide both user interaction and whatever the GF would need.
Gosh, it sounds like we are planning a funeral for something that isn’t even born yet, feels a little weird, although understand the concern due to newness and all.
One note: while we’re not supporting any particular aftermarket changes, we actively designed with the notion in mind that people might flash the firmware. By contrast, for obvious security reasons, our design prevents a malicious party from intercepting your Glowforge’s communications with the Glowforge server - which would likely, as a side effect, prevent the raspberry pi method suggested above (unless you flashed the firmware). You really don’t want a home wifi-connected piece of hardware that can be spoofed to talk to an alternate server!
–dan
3 Likes
@dan In regards to that, can you comment whether the firmware is going to be built onto the controller, or if it’s something removable like an SD card? It’d be nice to be able to swap between two SD cards rather than going through an entire reflash process…
I would think that at 90ish days before 10,000 units are planned to be en-route, that that ship has sailed. It would be hard to fathom all hardware decisions not having been cast in stone at this point if the delivery date scheduling is still considered viable.
It’s flashed via wifi (like everything else).
On the issue of preventing cloudmageddon bricking … as has been said the GF magic is in the cloud, and they would be business-foolish to open source all of that. So as customers we are all taking a risk, along with Dan, that this company is sustainable.
I don’t think open source firmware on the dumbest layer of the stack is much insurance for all the reasons discussed above … so my question for GlowForge is: is an escrow being set up for the cloud source code that will release it to the community in the event of a service shutdown?
(Don’t get me wrong! I’m heartened to see the commodity end of the GlowForge stack open sourced. It should mean the community can optimize for scenarios that emerge as important to only a subset of users. Though I worry that choosing GPL as a license will prevent the emergence of a commercial ecosystem serving such users.)
Excellent point hans. Yes, that is the correct question to ask!
That happened with a software vendor we had (when I was pre retirement). We were using and happy with software vendor A which was subsequently purchased by vendor B basically to obtain the userbase. B then gave notice that software A was being phased out and all users would be required to migrate (at a high price) to software B. The userbase A group stood up and said “hey B, you are obviously unaware of our escrow agreement with A that gives us the code in the event the software A ceases to be supported”. Userbase A knew it was coming and literally stood up and made the claim at the public announcement by Vendor B.
Needless to say, B had to continue with the A software because of the escrow agreement. They didn’t want to lose the user base they purchased nor access to the related code. Everybody eventually migrated or in a few instances moved on, but it wasn’t with a gun to the head.
It might seem to be the dumbest layer for the simplicity it conceals, but in reality it can be quite complex.
The application software could have other potential uses than the Glowforge. Bre Pettis is a key investor, involved in 3D printing. They’re not going to stop with laser cutters. It wouldn’t be much of a pivot to add camera support in it to do 3D extrusion (edit: or CNC milling) as well.
If they haven’t announced an escrow plan by now, I wouldn’t expect it to happen, ever.
As I’ve said before, we have no plans to release our cloud software.
If you purchased under the impression that it was, or might be, or was in escrow to be, or any other such misunderstanding: my apologies, and we’ll get you a refund processed ASAP.
No need! I went into this with a Kickstarter-like attitude. I know I’m taking a risk.
Your sale to me is not under discussion, only your NPS. 
It is troubling that you would confound escrow with “releasing [your] cloud software”. I don’t suggest that you share the crown jewels. I only think insurance against my Glowforge ending up sitting in a box with my BERG Little Printer would make good business sense.
Without a significant company history or a user-protecting exit plan, Glowforge is a wonderful, exciting experiment but not recommendable as a consumer product.
(Which is fine for me, most of your buyers, and I expect for investors at this stage! I’m only commenting because you opened the discussion by framing the open-source firmware as Skydog-prevention. Because you all have been so open does this feel like a misunderstanding, but it otherwise would smell of subterfuge.)
2 Likes
I think the point would go along the lines that an escrow agreement would be nice, but it would represent a huge impediment to any potential sale of Glowforge (the laser cutter business) to anyone interested in acquiring their cloud technology, but not interested in supporting a large userbase (e.g. the 10,000+ current GF owners) that is not generating any income stream from the “no cost” cloud service.
IMO, that potential, the sale of Glowforge (the business) at some point down the road could represent the single largest risk to the GF laser cutter/etcher owners at this point. I guess along the same lines, the second most significant risk could be that Glowforge (the business) is a one-trick-pony that saturated its space right out of the gate. Overheads are now getting higher and the cloud service that costs money to operate doesn’t generate any (under the current model anyway). In any event, the answer is pretty simple, anyone not willing to accept that risk only need pull the plug on their purchase via refund.
3 Likes
@dan - if access to the Glowforge servers disappears for whatever reason, what is Glowforge’s vision or plan for how owners will still be able to use our GF laser cutter/engravers? Presumably we will be able to download the GPL-licensed firmware that was mentioned in the original topic, but what else will a buyer need to do in order to continue using our Glowforges with the as-shipped features? As a follow-on, exactly what does the firmware do for us? I’m kind of confused on that point.
Sorry if these questions are trivial in nature, but my coding experience is far too old for me to reliably guess the answers to the above questions.