I believe the GF firmware can also be updated from the cloud, so that could brick the machine if it goes wrong unless it is done in a fail proof manor.
I have implemented many embedded systems that could update the firmware and cope with a power fail during that process, but it amazes me how many products don’t.
True but most updates shouldnt require a client side update where all updates of the muse do.
Yup. Not sure what hardware they have inside, but it’s running a full Linux-ish OS, complete with multiple containers. More complexity = more potential failure points.
[quote=“palmercr, post:21, topic:7900”]I
believe the GF firmware can also be updated from the cloud, so that could brick the machine if it goes wrong unless it is done in a fail proof manor.
[/quote]
Unfortunately this is very true but unlikely.
Lincoln Electric has an update for one of their machines that they specifically tell trained technicians to not use. It WILL brick the machine and the engineers have no intention of fixing it.
BUT, this machine is about 15 years old and not on their priority list.
HOWEVER, I highly doubt that this would be the case with Glowforge since they currently have only two variations of their machines and have shown a high level of respect toward the customer and pride in their work. I’m sure they will test it out a few times before pushing it out to customers.
Amen.
Everyone who joins the firmware team gets this from me:
“There’s really only one thing that absolutely has to work: the updater.”
I don’t believe in ‘fail-proof’ but we built in a lot of redundancy.
If memory serves, it’s a beaglebone computer - but please don’t quote me on that.[quote=“soldiercoleman, post:24, topic:7900”]
I’m sure they will test it out a few times before pushing it out to customers.
[/quote]
We push firmware updates regularly. As with the UI, different machines are on different firmware “channels”.
We’ve had failures, but they automatically roll back to the previous version.
I’m hoping the folks at Glowforge study these unboxing videos for how not to create that “Gee Whiz!” first blush experience. These videos are from a person who cancelled his Glowforge for the Muse.
I’m pretty sure the Glowforge team is shooting for the Apple product experience. It’s little things like providing enough hose and making sure the crumb tray seems new that help. Also, providing an unbroken working product doesn’t hurt.
Um, it should be new, shouldn’t it? 
If it isn’t a demo or beta unit. It seemed new, but had dings. I’d suspect shipments from China.
my PRU came through pristine (now it’s starting to look like Perth-Amboy at low tide…)
Watching his other videos, glad I stuck with Glowforge. There’s just so much other external hardware needed for those other lasers. Not that it’s difficult to set up but it’s very fiddly.
I would depend on the burn-in/testing process. If a GF was tested after complete assembly, it would have some marks on the tray. Or you could mount a scratch tray for the tests.
(My second-most-recent 3D printer arrived with a calibration cube stuck to the build surface, which did inspire at least a little confidence.)
My sewing machine had a test seam in the box.
Firearms often have a few empty shells in the box from a test fire.
Printers often spit out a test sheet when you set them up.
The crumb tray dings were the least of his problems. It hadn’t been used to test his device. The condition was new except for obvious mishandling.
My point? This is 1 of many little things that can impact a first impression. You only get that once.
I really don’t care if there are marks on my crumb tray or not.
That’s what I’d do.
We test every Glowforge with a reference crumb tray so we don’t goop up yours.