Silicon Valley invests $120 Mill. in $400 juicer that works as well as hands

[Muxr]

Don't turn it on!  Take it apart!

https://blog.bolt.io/heres-why-juicero-s-press-is-so-expensive-6add74594e50

Tim

Nice.. the cost of the product itself for what it is really isn't overpriced.. the problem is that it's a Rube Goldberg machine.

For the price and force I was actually expecting hydraulics.

...and using ball bearings for supporting the sliding press plate is somehow innovative? My desk drawers all have that...

(That URL reminded me of AvE, someone who I'd probably enjoy seeing taking something like this apart. "Keep your dick in a vise Juicero" )

I mean yeah, it's not anything special nor is it particularly affordable.. we don't know the quantity it was made in.. if it's low quantity it could be reasonable just from the pure manufacturing stand point. This whole other issue aside.

[Kilrah]

You don't make low quantities of something that needs to support large continued sales of consumables...

[AndyC_772]

Beautiful parts and construction indeed - but looks like it's overkill even for the forces involved

I like overkill. Overkill is reliable. I like reliable.

I don't think anyone has suggested that the machine is actually bad at juicing, nor can I believe anyone would have even noticed this machine if it weren't for the comedy IoT features.

Expensive, but nevertheless attractive and well built kitchen gadgets are nothing new or remarkable. This one doesn't need much in the way of modification to make it a perfectly acceptable product IMHO:

- remove network connection, QR scanner, and the whole idIoT way of thinking
- design and produce refillable juice bags (a simple roll top, with a clip in the machine to hold it closed?)

[MarkS]

- remove network connection, QR scanner, and the whole idIoT way of thinking
- design and produce refillable juice bags (a simple roll top, with a clip in the machine to hold it closed?)

It appears that their business plan revolves around a strong DRM mechanism and mandatory juice packet subscriptions. I would guess that the machine itself is priced at or just slightly above cost. Remove the DRM and subscriptions and they will be out of business in a month. Not that I'd complain though...

[Kilrah]

I don't think anyone has suggested that the machine is actually bad at juicing

No, but not particularly good either.

Yup. If they removed the IOT stuff and made reusable bags then it would be no more than a conventional juicer, and they'd sell 0 of their machine they'd now have to price at $2000 because the subscription model they'd make their actual income with falls apart.
Even a super premium design conventional juicer wouldn't sell at those prices.

The fact the price of the machine has gone down from $700 to $400 indeed shows they're now likely only making a tiny margin on it if any at all.

[AndyC_772]

Remove the DRM and subscriptions and they will be out of business in a month. Not that I'd complain though...

Sure, I've no problem with that either. In fact I'd be surprised if someone doesn't start a little business of their own, selling very simple control PCBs which completely replace the electronics in the Juicero, along with a few refillable bags.

[RGB255_0_0]

Doesn't look as though it'd be that difficult to modify with a switch you have to use to press the juice, and add some protection so people don't crush themselves in it with some sensors.

I don't see the surprise in the fact people would but this: Juice+ and Slim Fast have been around for a long time, particularly the latter and its psychological DRM (points).

Preying on people's naivety and anxieties will never truly fail.

[MarkS]

Sure, I've no problem with that either. In fact I'd be surprised if someone doesn't start a little business of their own, selling very simple control PCBs which completely replace the electronics in the Juicero, along with a few refillable bags.

I am highly tempted to do that myself, but lack the resources.

[Kilrah]

In fact I'd be surprised if someone doesn't start a little business of their own, selling very simple control PCBs which completely replace the electronics in the Juicero, along with a few refillable bags.

That would only work for a short while if the company folds and you find machines at on clearance at $100 and people who weren't interested in the original concept pick them up.

Otherwise you'd be trying to sell something to people who precisely don't want what you offer. People who will be seduced by the juicero offer and will buy this thing are those who don't want the time/mess to prepare fruit and cleanup, they are precisely after the "quick and no-mess" disposable bag aspect, so a refillable bag is the opposite of what they want.

[TheAmmoniacal]

I wish someone would take the 2 hours necessary to replicate this whole machine with an arduino, ESP8266 and linear actuator. Just to point out the absurdity of it.

[AndyC_772]

That would only work for a short while if the company folds and you find machines at on clearance at $100 and people who weren't interested in the original concept pick them up.

I wouldn't have invested $118.5 million, though.

If the company folds and its servers go down, *every* buyer of the machine would need refillable bags.

It's annoying that the (required) power supply is part of the same PCB. I wonder how hard is to flash new firmware via that internal USB connection? (Let the arms race begin!)

[Kilrah]

If the company folds and its servers go down, *every* buyer of the machine would need refillable bags.

Nah, they'll just trash the thing and all its beautiful engineering without second thought. I doubt any significant fraction of owners would go through the hassle of preparing their bags just to keep using it. Given the initial price and running costs the owners aren't in a category where they'll care about the $400 more than the hassle or their time that made them buy it.

I wonder how hard is to flash new firmware via that internal USB connection? (Let the arms race begin!)

I read it was an STM32, so it's likely just a connection to the built-in DFU bootloader. Piece of cake.

[T3sl4co1l]

I wonder if the press mechanism would be useful for machine shop applications.  Forming powders and plastics, embossing sheet metal, maybe even some small hydraulic-press applications (replace the door and pusher plate with a small piston and slide).

Very likely, everything is under MCU control, so that once hacks are developed, it could be fully repurposed as a Wifi connected CNC device!

Tim

[MarkS]

I'd just like to pick one up to see if I can hack it. There has got to be a way to trick the QR code reader. Problem is that I have to spend $400 to get my hands on a juice bag to get the QR code to run tests. :/ I doubt their data scheme is 100% secure, but I'm not going to waste $400 to find out.

[Kilrah]

Guess you'll need to develop a fake server as well.

[MarkS]

Guess you'll need to develop a fake server as well.

It depends entirely on the contents of the QR code. You are most likely correct though.

[Kilrah]

The initial outcry was precisely over the fact that you could not use the machine at all without a connection. Suggests the QR is just a unique ID that gets sent to the server which checks whether it matches a bag that went out of the factory, that expiry date isn't passed, returns the type etc, then marks it as consumed in the database so it can't be reused again.

[MarkS]

The initial outcry was precisely over the fact that you could not use the machine at all without a connection. Suggests the QR is just a unique ID that gets sent to the server which checks whether it matches a bag that went out of the factory, that expiry date isn't passed, returns the type etc, then marks it as consumed in the database so it can't be reused again.

That is most likely the case, but not guaranteed. It depends whether they thought it would be cost effective to make the data contained in the QR code secure. It isn't something the average person will know how to access or change. It is also likely that it says "Orange Juice". That's why I want to take a look and precisely why I will not spend $400 to find out.

[timb]

It seems pretty basic for a juicer, in the cutaway view you can see the motors, wifi board (WHY! JUST WHY!), switches, and a power supply. I think the wifi board and the custom power supply, along with the casing amount for the majority of the cost. I don't see where the other $150 is coming from though.

If you're interested, I found this miscellaneous control system on amazon. Supposedly for rc cars.

- https://www.amazon.com/DROK-Controller-H-bridge-Regulator-Current/dp/B017FZDVJY/ref=sr_1_10?ie=UTF8&qid=1493089180&sr=8-10&keywords=wifi+control+board

Read the article. All that custom machined aluminum is really expensive to produce. It has quite a few parts required custom machining. And not just one type of machining either! A single part have to pass through several different CNC setups before it's finished. That's slow and expensive to produce and prices don't go down much as quantity increases.

The electronics are most likely the cheapest part of the thing. If they're using the CC3200MOD, it would cost $10-$15 in 1000 quantities, however the CC3200 by itself would be closer to $5 (plus another $2-$3 of support components that would be included in the module version).

Actually, it's a bit weird they're using the CC3200 *and* and STM32 on the same board. They should shave a couple of bucks off by using the CC3100 instead. (CC3200 has a user accessible ARM-CM4 that can be used as a primary MCU in the system, whereas the CC3100 doesn't.)

[Kilrah]

99% sure it isn't.
Would be too easy to bypass even maybe just by cutting/pasting the code from another bag, and as we all know regardless of how "secure" the code is there would still be a possibility for someone to crack it and short-circuit their business by selling 3rd-party bags, too much risk.

The server-based approach is the tightest, and as they've gone for tigntness all the way and haven't exactly spared on engineering costs it would actually be surprising if they used any other solution.

[Kilrah]

Actually, it's a bit weird they're using the CC3200 *and* and STM32 on the same board. They should shave a couple of bucks off by using the CC3100 instead.

You answered yourself in the first part of your post. Given how overboard they went with everything they couldn't care less about shaving a buck or 2 on the STM.

[MarkS]

99% sure it isn't.
Would be too easy to bypass even maybe just by cutting/pasting the code from another bag, and as we all know regardless of how "secure" the code is there would still be a possibility for someone to crack it and short-circuit their business by selling 3rd-party bags, too much risk.

The server-based approach is the tightest, and as they've gone for tigntness all the way and haven't exactly spared on engineering costs it would actually be surprising if they used any other solution.

Understand that I'm not arguing with you and I believe you to be correct. Still, this is all supposition until someone actually checks the contents of the QR code. I want to look at the hacking of it from all angles and the QR code is one of those angles. Maybe a dead end, but still something to explore.

[Kilrah]

Sure, understand the interest of looking at all aspects.

I actually just read that Juicero has put forward the ability to block packs from being used if there was a recall on a batch of product.
That forces the remote authentication scenario.

[timb]

Actually, it's a bit weird they're using the CC3200 *and* and STM32 on the same board. They should shave a couple of bucks off by using the CC3100 instead.

You answered yourself in the first part of your post. Given how overboard they went with everything they couldn't care less about shaving a buck or 2 on the STM.

It still doesn't make sense. Lowering the BOM cost means more money in their pocket. It's not like designing in the CC3100 would be any harder than the CC3200. The power and RF design is *identical* in pretty much every way between the two parts, the CC3200 simply has additional I/O for the MCU functions.

In fact, it would have been *easier* to design in the CC3100, since you could just talk to the STM32 via SPI right out of the box. The CC3200 would need to be programmed with some custom firmware to allow it to send data back and forth between the STM32.

[AndyC_772]

I actually just read that Juicero has put forward the ability to block packs from being used if there was a recall on a batch of product.
That forces the remote authentication scenario.

...or if the pack has passed its expiry date, or if it's already been pressed, or even partially pressed but then aborted.

Seems that refusing to press packs is a part of the business model too. There's no "oh, ffs, just press the damn thing" button.