frauenfelder at June 11th, 2014 20:49 — #1
stephen_schenck at June 11th, 2014 20:59 — #2
I wonder how limited this is by the need to "train" it. Like how the app requires the user to tell it in advance what you're scanning for; it doesn't look like you can just point it at an apple, it will recognize what you're scanning, and provide relevant data - you have to say "I'm scanning fruit now."
And what about wholly unknown compounds? "What's this stain on my sneaker?" Or ones whose spectrum fingerprint hasn't yet been cataloged by the database?
newliminted at June 11th, 2014 21:14 — #3
ryjkyj at June 11th, 2014 21:17 — #4
Never in a million years will this work like they hope it will. Well, ok, maybe in a few hundred.
lloyd_llanahan at June 11th, 2014 21:29 — #5
This is an IR spectrophotometer. That type of technology is not really suitable to identify random chemical unknowns outside of discrimination of a small and previously validated set (ie would be unable to identify a random pill). There are many limitations to this approach that are not addressed by the kickstarter...it certainly isn't a "molecular scanner" in the sense that it will ID random compounds.
chellberty at June 11th, 2014 23:25 — #6
2.5 mil even that "highly illogical" solar road had a million pledged.
Scientist just need to invent more magic like that neverwet water repelling spray.
churba at June 12th, 2014 07:38 — #7
Another one? Well, I really, really hope this isn't a scam like literally every single other one of these devices so far (Remeber Tellspec and it's practically identical claims? Remember Healbe?) but I'm not exactly expecting otherwise, since this throws up more flags than China's National day parade.
For a start, that's not really how Near-infrared Spectroscopy works, you need a IR light on one side of the sample, and a receiver on the other, and the sample must be able to let some light through. Plus, there's been no independent scientific testing of the device, which immediately makes me suspicious.
A quick hunt around reveals I'm not alone on this. There's more than a few people in fields that employ NIR Spectroscopy who find this whole thing incredibly dubious, along with many in the general public.
Amongst the concerns bought up:
They're somehow identifying non-polar molecules with NIRS, the fact that they need a kickstarter because every instrumentation company on the market would buy this technology for billions, or at least license it for hundreds of millions of dollars(if it provably existed)
The fact their aiming at the consumer market when they know as well as we do that a tiny, accurate, low-cost NIR Spectroscopy sensors they could revolutionise chemistry, biology and other labs(and more) around the world practically overnight and make fucking zillions in the process
On more than one occasion relentlessly positive accounts show up in a thread where they're discussed(many of which appear to be blatantly astroturfing)
The incredible amount of variables scan to scan that would be almost impossible to account for - for example, vibration and movement can completely fuck up your reading, yet doesn't seem to be a problem with this device.
The incredibly sudden appearance of the technology, without any verifiable incremental breakthroughs behind it to back it up. This is not a technology that could just pop out of the metaphorical void.
The incredible simplicity of it. Figuring out even a handful of molecules together using high-resolution data and quality fingerprinting is very difficult. Unless they've come up with an incredibly novel way to do this - again, without somehow leaving any trace within the larger scientific community - that in and of itself is enormously suspicious. Let alone doing it purely with an algorithm in the cloud? The needle on my bullshit detector is spinning fast enough to create a plesant breeze.
They're making claims that wouldn't be possible even with regular NIRS equipment(for example, the one about telling how ripe an avocado is through the skin) as well as some suspicous demos, like getting readings through plastic containers, while somehow not showing the composition of the plastic.
And finally, despite more than a few accusations of fraud, they refuse to take even the simplest steps to prove their device isn't a complete scam, such as independent testing, when it would be trivally easy to do so and prove that they've developed what might be one of the most revolutionary devices since the invention of spectroscopy. So, what precisely do they have to hide that's important enough to forgo international fame, incredible fortune, and nigh-on incontrovertible proof of their functionality claims?
In short, I've love more than anything else to be proven wrong, and this is everything - or even most - of what it's claimed to be. But frankly, it doesn't look like this will be the case.
Edit - As a side note, I wonder what Maggie thinks about this? She's bound to have something interesting to say, I'd love to see her dig into this and find out what's up.
bobtato at June 12th, 2014 09:00 — #8
I agree that this is pretty hypeful, but I don't think their claims are obvious lies (especially if you look only at the actual claims, and not the more impressive claims which they imply but don't say outright).
IR spectrometry doesn't in principle have to be transmissive, it could work with reflected light; perhaps their sensor is some kind of Raman microscope, for example? And most non-polar molecules, like hexane, do have IR spectra (you just don't see modes with no dipole change, though they can be detected by Raman). As for identifying individual molecular species out of millions, IR certainly can't do that-- for biological materials, it's nearly impossible even with GC/MS-- but, they don't actually make that claim.
If you scroll down on the kickstarter page, it appears to show some actual spectra, of extremely low wavelength resolution, and implies that their cloud service does some statistical comparison between the whole spectra. It can't say "this lasagne has traces of salamander tears", but it seems believable that it could say "this spectrum is closer to that of chicken lasagne than beef lasagne".
It's like, if you rubbed an avocado on a sample plate and put it in an IR spectrometer, you wouldn't be able to tell what was in the sample, but the spectrum from a ripe avocado might well be recognizable from that of an unripe avocado (assuming that avocados sweat some distinctive compound when they're ripe).
But if that's how it works, it will be very scenario-specific; it might be able to recognize rohypnol in coke, say, but that wouldn't mean it could detect rohypnol in gravy. So it's certainly not the magical tricorder they're hinting at.
churba at June 12th, 2014 09:47 — #9
I'm about to sleep, so I'll have to be brief, but I thought I should say at least something back, since you make some good points.
I'm not trusting their spectra on the kickstarter page. Frankly, I know they're claiming that's how it works, but I've no way of verifying or disproving that in any way, it's not really a testable claim right now. They're saying "Yeah, this is a spectra", but we've no way to distinguish that from, say, a bunch of lines drawn on a square in photoshop, especially with no labels on that group of coloured lines, or a spectrum taken from other equipment.
They're definitely not using Raman Spectroscopy. That would require either a laser(and a decently powerful one to boot), or a light much, much brighter than the device seems to be equipped with - as in, bright enough to be hazardous to the eyes. Though, I might be wrong on that point - I will probably end up saying this many times over this, so pardon me for going all broken record on you - but I'm not a scientist, I'm a journalist, and doing my homework is no replacement for years of education and experience.
They don't make the explicit claim they can identify individual molecular species out of millions(or any other number, for that matter), but the claim is strongly implied, and I would even say necessary for the device to operate. For example, their Tylenol pill - to take an example from the kickstarter page of something it can identify - they have to be able to distinguish corn starch, magnesium stearate, powdered cellulose, sodium starch glycolate and Acetaminophen from each other, and make a decent guess to their amounts to identify the product. Even in a well-equipped lab, that isn't just a done-in-one task, to positively identify all that, you'd need multiple types of spectroscopy(If you didn't want to use some other method, but let's say we don't, considering) and experienced, educated people to be able to filter and interpret the data correctly.
I mean, look at the techcrunch disrupt video. They supposedly differentiate between two different types of cheese, as well as their fat content and caloric content from their specific fingerprint. Now, I'm admittedly not a biochemist, analytical chemist or any other such thing by trade, but every one of them that I consult(Point of order - I've been on the trail of this for about a month now, after the campaign started popping up around the internet, so I've had a little time to work on this, as much as I wish I was magic enough to pull together a crew of scientists at 11:44 pm to look at kickstarter campaigns) notes that this would be an incredibly difficult thing to do with only a single NIRS analysis, and frankly wouldn't currently be possible with the technology we have.
That doesn't rule it out, of course, but it does make it very unlikely - returning again to the point that they would have to have made these incredible leaps in technology without making even a ripple in the metaphorical pond of instrumentation research and development, or any other related field.
alchemist at June 12th, 2014 11:48 — #10
Handheld, low-power Raman spectrometers for material identification already exist; they're just not targeted for the mass market. Thermo, for example, makes several:
SCiO looks like it is basically a low-power Raman spectrometer with the spectrum analysis shifted to the cloud instead of on-device. The device itself would mostly consist of the source (a low-power visible-light laser) and the detector.
churba at June 13th, 2014 01:01 — #11
Which would make it just like the Tellspec, which was an enormous scam.
Also note - all of those handheld spectrometers were made for the identification of narcotics, explosives, and other similar things - a small band of functionality that they can do with reasonable precision. They already have the results from more precise equipment, and all they have to do is see if it roughly matches the result you're expecting. They don't do anything close to what this product claims to do - without context, these devices are almost useless, and having that requirement would, in turn, make the SCiO useless - and they're all easily 10 times the size.
Also note - the cheapest of those units is $20,000 dollars per unit - again, somewhat dubious that they can provide much, much better functionality for under $300, and that they're going to kickstarter instead of instrumentation companies, or even selling to law enforcement and labs themselves. Unless you consider that Labs and Law enforcement tend to require more solid proof that it works as advertised than the public tend to.
As for it being a low-power Raman Spectrometer with the versatility they describe, that requires a 2W laser at the least. Which is a serious eye-hazard, and sucks a lot of juice - go look at the product range of the company Wicked Lasers, who carry a 2W laser - it's easily three times the size of this device, just to support the power requirements(and cooling) required just for the laser diode itself. There have been developments with smaller, lower powered lasers for Raman Spectroscopy, but it's far, far more limited that would be necessary for this product to even remotely approach the claims they're making, and have said they've already demonstrated.
And of course, finally, they're already said over and over that they're using NIR Spectroscopy. While Raman can use Near-infrared light, it's also a different process to NIRS.
jorpho at June 13th, 2014 01:19 — #12
I admire your detailed and articulate posts, and wish I was capable of such despite all the work I did in spectroscopy many years ago.
Well, they could be persuaded to buy dowsing rods.
churba at June 13th, 2014 01:44 — #13
And I wish I could claim as much credit as you're giving me, but I'm just re-phrasing information I've spent the last month siphoning off from far more educated friends, contacts and sources. I'm just a low-rent scribbler with enough practice to look good at it, they're the ones with the real knowledge - and speaking of knowledge, I'm betting you've got what I don't, so feel free to correct me if I get anything wrong. Not only will I not be offended, I invite it.
Oh yeah, the ADE 651. What a fucking mess that was, so much pointless loss of life, all because some fraudster had a particularly convincing pitch and knew his target market. Not sure if I forgot, or was merely trying to, but either way, thanks for reminding me of that.
Not to mention the current furor about Forensics and their use in law enforcement - another possible indication that law enforcement don't tend to look as closely as I was implying at the techniques and devices they use.
bobtato at June 13th, 2014 10:20 — #14
I suppose I was trying to think of ways it could be real just because I like the idea of science-fictional stuff becoming reality-- which, if this is snake oil, makes me their ideal mark.
Scanning tunneling microscopes were unattainable multi-thousand dollar items until it turned out you could build one for $200; and Gracenote can identify CDs with very good accuracy just by looking at track lengths (the digital equivalent of mass spectrometry), even though it can't tell anything about their content. So I don't want to give up hope that disposable sort-of-universal robot noses of this kind are possible. But learning that phoney scientific instruments are a thing (who knew?) has made me more doubtful
churba at June 15th, 2014 10:19 — #15
Never, ever give up hope. Don't let the bastards get you down, just because the last three have been a scam, doesn't mean the next one will be. It can most likely be done, and the question is simply when. The idea is out there, just waiting to be shaped into a tangible reality. And don't get me wrong, I really, really, really want this to be true. I'd be over the fucking moon if it was, imagine the applications that would emerge from everyday use, the data! It would be nothing short of incredible, revolutionary, in fact.
But as much as I hope for it to come true, I'm afraid that I don't think this is the device to do it. I'd be absolutely thrilled to bits to be proven wrong, but so far, that's not looking likely.
And yeah, phoney scientific instruments - admittedly to my utter surprise when I found out - are a thing. I guess wherever there's money to be made, there's someone trying to make it with bullshit and bluster, too.
krigelmanj at June 15th, 2014 13:34 — #16
I don't believe their innovation lies in the improvement of the technology. I think it lies in the the use of crowdsourcing data collection to increase their validated set. I could totally see people using this, even if it doesn't work perfectly at first.
frauenfelder at June 16th, 2014 20:49 — #17
This topic was automatically closed after 5 days. New replies are no longer allowed.