Awesome TED Talks (and similar enlightening lectures)

[Nappi]

In typical light detection applications, stray light problems can be eliminated, to an extent, by using a lock-in amplification scheme. They are probably already using a similar technique to synchronize the pulsed laser and the detection unit (they have to know which pulse the detected photons originated from), but considering how weak the twice-scattered signal must be, I would guess that moderate stray light would still completely ruin the experiment. Besides, a camera as sensitive as that is bound to be saturated in broad daylight conditions.

I should probably read the article before I start guessing though.

The car example is complete bullshit for so many reasons. The next generation cardioscope at least makes some sense.

[Twig]

You guys know there are different frequencies of light, right. Like, infrared vs ultraviolet, etc.

I don't know if the tech will ever reach that kind of level - and kind of doubt it - but the idea isn't completely out of the question, as far as I know.

[Nappi]

You guys know there are different frequencies of light, right. Like, infrared vs ultraviolet, etc.

Yes. I study optical engineering. There would be plenty of stray light at all the wavelengths in the more outrageous application examples that he proposed, so I don't see how a wavelength resolved (or filtered) measurement would solve the problem.

Besides, there are plenty of other serious problems with the principle.

Still, awesome tech, ridiculous examples.

[Twig]

Computers can be pretty good at filtering out noise from signal.

[Nappi]

If you are referring to bandpass filtering in the optical frequency range, then no, computers cannot do that. As for detecting modulated signal in noisy background, the signal that has been scattered three times (once from the target car and two times from the conveniently placed wall) is bound to be way way below the noisy background of the measurement. The natural fluctuations in the stray light — cars moving, trees swaying — are even a bigger problem.

[Twig]

I'm referring to algorithmically filtering noise from signal, given the data received by the camera. A bit more than a frequency filter. Ideally, you'd also have more than one camera/beam, operating at different frequencies, allowing for more information generated and received.

Why can't computers perform bandpass filtering (as I understand it from a quick internetting)? Assuming the camera can supply the necessary information, I don't see any reason why a computer couldn't just ignore irrelevant data. And, even if the camera can't supply that information, there definitely exist cameras (or is it just the lens) that only receive certain frequencies of light - ala infrared cameras. I admit I have very little knowledge of optics. U:

Anyway!

I'm not saying it'd be easy, or even possible (certainly not with our current level of technology). But you're straight up denying the possibility? Technology advances in ways we never see coming, sometimes. Whether it's someone developing a trillion-frames-per-second camera or discovering a new way to simulate an already heavily-researched lighting technique in graphics or finding the newest, known, farthest-away galaxy or using a warped piece of glass to SEE that galaxy or what-ev-er... How long ago was it that no one even thought high-speed photography would be possible? And now we have this? I mean, robotics - a field in which this problem probably could easily fit - is hard. It's just incredibly hard. But if people sat around talking about how impossible shit was all the time, we wouldn't have running robots or robots that can navigate obstacles or flying robots or whatever. These things are kind of impressive, but also kind of super simple. These are things we, as human beings, take for granted - 'cept maybe flyin' - because our brains are magical. But it still somehow manages to be impressive! It's only going to get better as time goes on. How, given everything we've done with science thus far, how many supposed limits we've surpassed, how can you be so pessimistic? At the very least, it never hurts to try. Failure breeds just as much knowledge and experience as success (if not more!) needed to actually make it WORK the next time around.

BURRRRR

I hope this doesn't sound too hostile! It might. I'm tired, and I'm also playing a bit of a devil's advocate (although, strangely, on the optimistic side, rather than the pessimistic). U:

[Nappi]

Ehhh... what?

I have said it multiple times now that I think the technology is super impressive and that there are certainly applications where that thing can be of real use. I'm also very much aware of how technology advances. If the fact that I can't for the life of me imagine that type of system ever being incorporated in a car (where it solves a very minor issue; one that can be easily, and cheaply, overcome by traffic lights, mirrors, cameras, tagging etc. even now), makes me a technology pessimist in your eyes, so be it. Electron microscopes have been around since the 1930s but they have still not invaded our homes, even though they would be super interesting toys to play with. The fact that technology is constantly advancing, does not mean that certain things will ever get small, cheap, practical or even mature.

I also don't understand where you got the idea that I think they shouldn't even try. Of course they should try. They should develop the technology as much as they can, and they will. I only believe that they should focus on applications that are worth the effort, i.e. ones where there is a real problem to be solved. And they will.

EDIT: Pure computational bandpass filtering of the pixel data (time-strength) at the optical frequencies would require a sampling rate to be in the petahertz range. This cannot be achieved. That was just a side note, as I wasn't sure what you were suggesting earlier. It's not an issue, as it is much easier to use filters or other techniques to lower the sensitivity of the detector outside the wavelength range of interest.

Anyway, nevermind.

[Twig]

I also don't understand where you got the idea that I think they shouldn't even try.

Well, you called the car example "bullshit", which sort of implies it would be a complete waste of time. If that wasn't your intention, then I retract that notion!

This cannot be achieved.

You sure about that? If we can do 1,000,000,000,000, why couldn't we add three more zeroes to the end, some day? Is there some physical limiting factor of which I'm unaware?

Anyway, sorry again if that post was overly hostile. I kind of get a little overexcited when thinking about future technology.

[Nappi]

Well, you called the car example "bullshit", which sort of implies it would be a complete waste of time. If that wasn't your intention, then I retract that notion!

Ahh.. ok. My bad. I think they should definitely develop this technology, namely ultrafast imaging and "seeing around corners" stuff, further. The car thingy, which I still think is bullshit, would only be a byproduct of this research (except it won't, in my opinion ). A lot of work is required before this technology can be used in any commercial application, and I don't think that they should, or that they will, focus on making that crazy car thing a reality.

You sure about that? If we can do 1,000,000,000,000, why couldn't we add three more zeroes to the end, some day? Is there some physical limiting factor of which I'm unaware?

The title of the clip is somewhat misleading. The camera isn't actually tracking a single photon package as it travels through the bottle. The video description actually states "For that, they built a camera and software that can visualize pictures as if they are recorded at 1 trillion frames per second." He mentions in the clip that the photon packets are launched and recorded millions of times, and the video is generated from this vast amount of raw data. This can, crudely, be compared to taking a photo of a bicycle tire after every 361 degrees: it may look like you have taken 360 photos during one full rotation, while there has actually been many full rotations.

I don't know what sort of (continuous) detection rates can be achieved with modern methods but they are nowhere near those required for optical filtering. One would have to be able to record actual waveform of light, whose frequency is in the hundreds of terahertz range, and not just the information it carries, e.g. ones and zeros in the optical fiber. Solid state detectors (which are based electron-hole pair generation caused by photon absorption) are fundamentally not up to the task. For what it's worth, I'm not aware of any potential detection scheme that may achieve this in the future either, but obviously I cannot be absolutely certain that there never will be one. I'm almost willing to bet that there won't be, though.

Anyway, I don't think that computational optical filtering is something that scientist are dreaming of, nevermind striving for. There are many ways to perform wavelength resolved measurements and various physical filtering schemes that can be utilized already.

[Twig]

The title of the clip is somewhat misleading. The camera isn't actually tracking a single photon package as it travels through the bottle. The video description actually states "For that, they built a camera and software that can visualize pictures as if they are recorded at 1 trillion frames per second." He mentions in the clip that the photon packets are launched and recorded millions of times, and the video is generated from this vast amount of raw data. This can, crudely, be compared to taking a photo of a bicycle tire after every 361 degrees: it may look like you have taken 360 photos during one full rotation, while there has actually been many full rotations.

Ahhhh, well, I completely misunderstood how it worked. Thanks for explaining it. U:

Actually, thinking on it with this new clarification, I now better understand the reverse ripples. I should've realized it, then, as I am familiar with the concept, but I just didn't bother thinking about it at the time. I'm dumb! That's what I get for watching the video while distracted by other things.

[Nappi]

Yup. He did not emphasize that point enough, in my opinion. I probably wouldn't have paid attention to it myself, if 1 Tframes/s hadn't struck me as such an outrageous claim.

It's still incredible tech, though.

[Twig]

Yeah it is!

[Merus]

I like how they've basically built a survey bot like you see in science-fiction.

I don't know if everyone else has noticed yet but we live in the future. Flying cars and jetpacks turned out to be bad ideas, but still: we live in the future.

[ysbreker]

To get annoyingly back on topic: The conference I helped organize just put the video's of the talks on line. If you happen to work with HTML/CSS/JS or are interested in that kind of stuff go check em out. You might learn a thing or two Highlights for me this year:

- Alex Russell, One of the developers of Chrome at Google talks about what is keeping the web from doing awesome stuff with the current technology.

- Phil Hawksworth a designer at a big London design company gives a very entertaining talk about how shitty modern day CMS's really are.

[ysbreker]

[part 2 because the forums are being mean and not let me post 4 links ]

- Marcin Wichary works at Google making those awesome Google logo Doodles. He gives a talk about how his team tackles the challenges they come across when making these doodles.

Bonus talk:

- This is an interview with three people who have various disabilities talking about how they experience the web. And how we as developers can improve the experience with relatively simple tricks.

[ThunderPeel2001]

This is right up my street!

I have to say that I think Phil Hawksworth's talk was extended to about ten times the length it needed to be, but I agree that existing CMS's are of the devil.

One thing I don't understand is how rare a simple and lean fully-bespoke CMS is. It's not that hard to construct something that only does what you need it to from scratch. And it takes no time at all, plus it doesn't lock the client in (any developer worth their salt can come in and make changes).

That way your markup and design are fully protected, and site remains fast. In a perfect world, that would be the solution (except when it turns out a blog would do everything they require).

Trying to convince a client of the evils of existing CMSs is an uphill battle though. I recently lost out on a big job because my rival pitcher convinced the client that a big complex CMS was exactly what they needed.

*sigh*

[toblix]

Marcin Wichary's talk was just... wow.

[MrHoatzin]

This is a long one but a fun one. I've been listening to podcasty stuff as I draw. David Graeber starts with a ridiculous premise of Where are all the awesome things that 1960s science fiction promised we'd have by now? …and manages to be show that it is not at all an unreasonable question for us to ask. Along the way he drive-bys some fun interpretations of pop culture he touches upon.

Anyway, you should all also read Debt: The First 5000 Years. Good book.

[osmosisch]

I'm sure this is old news to a lot of people but I found this discussion of how fructose is metabolised both enlightening and a bit scary:

There's a lot of hoo-hah and kerfuffle surrounding it and the speaker gets a bit obnoxious with his 'mkay, but the juicy biochemical part in the middle seems pretty unambiguous.

[Roderick]

Very interesting, I watched it in total. I actually find this a comforting message, because it confirms some things (sugar is bad, multiple diets work but not for reasons advertised), denounces others (the oft heard notion that carbs are the root of evil).

[osmosisch]

I found this to be an excellent accompanying article.

e: to the sugar video I posted on the previous page.

[tegan]

This is a long one but a fun one. I've been listening to podcasty stuff as I draw. David Graeber starts with a ridiculous premise of Where are all the awesome things that 1960s science fiction promised we'd have by now? …and manages to be show that it is not at all an unreasonable question for us to ask. Along the way he drive-bys some fun interpretations of pop culture he touches upon.

I listened to most of this last night, and now that I've had some time to digest it... while his resulting argument about how we've bureaucratization has slowed technological advancement to a crawl is interesting, his initial premise of "kids in 1900 got everything from science fiction by 1950, but kids in 1960 didn't get anything from science fiction" is full of crap.

For one thing, they didn't. Although it's quaint to think that the only things they envisioned were submarines and television, they wanted all of the same insane shit that we wanted in the 1960s, just with a certain Victorian sensibility to them. Second: his examples of saying that we got submarines but not flying cars is very odd. We got both. It's true that flying cars aren't available for the public sector, but they certainly exist. But apart from sheer quantities, that situation is absolutely no different from submarines. In fact, most of the things he touches on are either existing technologies or in development. NASA began development of a faster-than-light warp drive last year. A Japanese robotics firm started mass-producing

, as well. Although I'm not aware of a robot that does laundry, we most definitely have robots that do chores. I'm not familiar with the tech behind Siri, but by all accounts it's a pretty damned advanced AI for something intended for the average consumer, and I'd say it (only just) fits the description of an AI that you can talk with.

Yeah, it's disappointing that I'll probably never get my Wipeout car or Iron Man suit, but the idea that we don't get any of the things that science fiction has promised us is bullshit. Saying that the internet is simultaneously the most important invention of the last few decades and also unimpressive as shit is true, but no less so than television or the automobile. The important, world-changing inventions have always been defined by being utilitarian and accessible to the masses. They will always be, by necessity, boring.

(I'll cut him some slack, though, since that talk seems to be from around 2011 and would have just barely predated most of those counterexamples I just posted)

[MrHoatzin]

Well sure, but it was really just a flamboyant trick question that allowed him to talk about something else. It is not even a premise as such. He does make a solid case for why poetic technologies are no longer pursued with as much fervor as they have been for the entirety of modernity.

[ThunderPeel2001]

For those who don't want to watch the whole 90 mins, here's the four steps those turn obesity around in kids (and presumably adults):

1. No sugary drinks -- Only milk and water. Juice is no good for you.

2. Always have fibre with your carbohydrates -- this helps your body counter any bad effects from eating carbs.

3. Wait 20 minutes for second portions.

4. Equal every minute at the screen, doing exercise (pretty damned tough to maintain, he admits).

Fructose is the same as ethanol, but without the buzz. All the same side-effects and damage to your body.

[brkl]

Seriously, people in the 60s didn't even have the imagination to dream up the cool stuff we have now. The dream of flying cars: no traffic jams. That's a damn boring idea compared to the awesome technologies we use to help us masturbate.

60s dream technologies are based on the premise that we would still be doing the same shit we were doing fifty years ago, only better. Instead, we are doing amazing things nobody could have anticipated then.