Major research breakthroughs in the last 10-15 years | Hacker News Reader

Major research breakthroughs in the last 10-15 years

19 pointsaxiom59 comments

When the transistor was first introduced in the late 40s there was quite a bit of hype surrounding it, but I wonder if anyone fully appreciated the significance of the development at the time.

Same goes for the world's fair in 1893, when Tesla's AC generators were first introduced to the wider public. I can't imagine that anyone realized how significant this was. Well, maybe Tesla.

So what are some major research results from the last 10-15 years that you feel will have a large impact?

View on HN

Without PCR (http://en.wikipedia.org/wiki/PCR) almost none of the biotech research of the past 20 years would have been possible. In the past few years PCR has got absurdly fast and convenient. I think RT-PCR is kind of mind-blowing. You can get an off-the-shelf solution that will (in about half an hour) screen your blood for a set of known virii. Or snoop or your food for toxins. Or whatever else you want to do with DNA...

1) Measuring the age of the universe to a few percent ~14 billion years. 2) Determining the dark energy (70%) / dark matter (25%) / regular matter (5%) breakdown of the universe to a few percent. 3) The Sloan Digital Sky Survey, Hubble Space Telescope and Wilkinson Probe

I'm purposely being biased :)

> Determining the dark energy (70%) / dark matter (25%) / regular matter (5%) breakdown of the universe to a few percent.

Translation: We've figured out in even more detail how we have no idea wtf is going on in the universe

Forgot one really important one:

The discovery of planets outside our solar system.

You couldn't be more right.

The Human Genome Project, or perhaps just its completion since it originated outside your window.

Does Internet qualify? :) Probably the greatest thing that has happened in our lifetime.

The internet is definitely a great example. But I guess what I had in mind was more along the lines of things that haven't borne their fruit quite yet. Things that in retrospect people will see as seminal, but that are currently still hidden in research journals.

> things that haven't borne their fruit quite yet.

Well, what about web 3.0?

Machine translation of language. In 10-15 years time, given improvement and dissemination of the technology, we might massively open up the world's communication. This could be significant as we massively increase the marketplace of ideas and communication.

Are you a MT practitioner? Email me (email in profile).

Nope. Purely an interested bystander on this one. While I hate saying leave it to Google they are working quite hard on this one and they have quite a lot of data.

Are you in a startup? I'm working on something in this area... that's why I ask. You were the first one who mentioned this (that I saw) on n.yc.

I'm biased: functional MRI

We're just now scratching the surface as to how humans derive their intellect from 3.5 pounds of squishy grey stuff. Some questions we'll never have licked, but the future looks good.

I totally agree, even more so considering recent advances in real-time fMRI (http://paincenter.stanford.edu/research/rtfmristudy.html),

I also nominate EROS (http://www.answers.com/topic/event-related-optical-signal) as a revolution in the making. I think it will deliver some huge discoveries in the activity of the human cortex.

Real-time fMRI isn't that different - you just analyze the data on-line rather than after the scan. The most promise for me is millimeter and millisecond resolution with new scanning and analytic techniques. And that keeps getting better and better.

As for EROS (from your link): "Its biggest downside is the inability to detect activity more than a few centimeters deep, but it is a new, relatively inexpensive technique that is not invasive to the test subject."

That's a big problem it will need to overcome to compete.

While RT-fMRI is an incremental technological advance (it's just faster data collection with some spiffy signal processing) it hugely broadens the scope of fMRI's usability. The hemodynamic delays of traditional fMRI are atrocious. For the first time fMRI is giving us access to when brain activity happens, which is nearly as importnat as knowing where. If you look at the biofeedback work being done with RT-fMRI you'll see it's extremely exciting.

Also, I don't view the depth restriction of EROS as a very big problem. There are enough mysteries lurking in the top 3 centimeters of your brain to keep academia busy for a hundred years.

I don't know about the BOLD delay being "atrocious". We get good temporal resolution and it's getting better and better (hundreds of milliseconds for some paradigms). I agree though - the biofeedback stuff is very exciting.

As for below the first few centimeters, that's where the control network is in mid-brain structures. That's kind of important, and if EROS can't penetrate deeper and with precision it's not likely to be widely adopted.

I think part of the reason neuro-imaging is so preoccupied with midbrain structures at the moment is because their activity is simpler and easier to correlate to the brain's inputs or outputs. With the huge boost in both spatial and temporal resolution EROS provides we can put more focus where the cognitive action is (the cortex).

0. Sophisticated manipulation of nanotubes.

1. Metamaterials (discovered in the 40s, but only confirmed in the late 90s).

2. Weird properties of DNA. Your DNA, for example, contains encodings of many of the viruses you are immune to; in certain cases they might even "spontaneously generate".

Probabilistic Robotics (think the Darpa Car Races)

This is the one.

No more big rigs, ups drivers, fed ex drivers, taxi drivers, pizza drivers.

The tech seems extensible to store shelf stocking, so a pretty big chunk of all transportation of goods should be automated.

It's easy to envision a retailer that has automated delivery from the factory floor to the store shelf.

This will be huge.

That's one that I definitely had in mind!

Some of the work done in computer vision in particular is pretty incredible. The SIFT result by Lowe for example has just barely started to become applied (take a look at Photosynth by Microsoft - that's entirely made possible through SIFT.) Particle filters in their various forms will become more and more widespread as computation becomes cheaper (the Condensation algorithm is a good example.) All of this is just now starting to add up to allow people to create truly intelligent systems that can interact with the world.

Better wireless data. I know this isn't really new, but I know it's a recent deployment in countries like Bangladesh. It gives you the opportunity to lay a communication foundation throughout an entire third-world country at a significantly reduced cost than laying down cable. I also remember reading that in some countries, copper wire would be dug up by surrounding people to sell. It's nearly free infrastructure.

Stem cell research and cloning. Growing replacement organs and/or cell sets to "repair in place" damaged parts of the human body is going to have the greatest impact in the long haul.

The mechanical heart is another wonder from medical science as well.

In 5-10 years, I'd say personal fabrication.

In 10-15 years, I'd say perhaps Human Genome

I think he meant last 5 - 10 years?

He is talking about the last 5-10 years... Personal fabrication has been available in a limited form for a few years ago, and Human Genome Project started in... 1990 iirc.

Wireless electricity.

That's one Telsa himself was working on 100+ years ago.

Yes, but he didn't get far enough with it to be feasible for everyday life.

How about some gene therapy http://en.wikipedia.org/wiki/Gene_therapy

I think this will be the one, but we've got a couple years, at least.

Almost everything will have a great impact. Alone those small achievements don't have great effect, but together all of this is driving us to singularity (http://en.wikipedia.org/wiki/Technological_singularity). Ray Kurzweil had a talk about this on TED (and his books are interesting too).

synthetic biology.

here's a fascinating TED talk from craig venter. (if you haven't heard of him, he was instrumental in the human genome project.) http://www.ted.com/index.php/talks/view/id/6

the part that was new to me was the idea of synthetic bio solving our energy problems.

Support vector machines and other practical realizations of statistical learning theory research that began in the 1970s.

Why single out support vector machines? Sure, the kernel trick is neat but there's really nothing about them that stands out above other classification methods. It would make more sense just to say "the field of machine learning".

It's a practical (especially for small to medium sized data sets), effective way of doing machine learning that has a highly sound theory behind it and doesn't require arcane tricks that are rarely or never written down, unlike old-style artificial neural networks.

The interest in kernel methods is proof that there's nothing more practical than a good theory. :)

At the same time, I think it's astounding how often k-nearest neighbors (a method considerably simpler than SVMs) is either the best or second-best classifier for a dataset. I'm not trying to pit SVMs vs. neural networks...rather I'm saying most methods of machine learning (neural nets, SVMs, gaussian mixture, kNN, what have you) can be effective and none stand out above the rest.

Woops...forgot the squeeze an "often" before the "either"...

"When the transistor was first introduced in the late 40s there was quite a bit of hype surrounding it, but I wonder if anyone fully appreciated the significance of the development at the time."

Uh. From what I've read, it was on page twenty-something of the NYT, with a blurb about how it might be used in hearing aids.

Why are you so sure the transistor and the AC generator were not expected to change the world? Inventors often envision sweeping changes as a result of their inventions.

I'm surprised no one's mentioned it: Dexter.

http://www.paulgraham.com/anybots.html

1. Internet

2. Human Genome

3. Nano-technology

Nano-technology is one of these buzz words that's been applied to everything from MEMS to car polish creams. It's so amorphous and vague that the only definition I can think of is "anything that's pretty damn small." I'm really not sure if it's a useful term at all anymore, or if it just obscures whatever it's used to describe.

One of my good friends is doing his PhD on MEMS micro-manipulators. But his work falls under non-linear control systems. Can it be described as nano-technology? sure, but what does that tell you? it hardly clarifies what his research is about, it just makes it sound like he's seeking funding from gullible VC's :)

Yeah, nanotechnology was coined to mean molecular manufacturing -- mechanical chemistry. Unfortunately, it's been applied too widely, and so we now have to use these clumsier terms to describe Drexler's idea.

The concept of SENS and the "engineering" approach to ageing. http://www.sens.org

photovoltaics, slow but steady. not much else counts if we don't have the juice to run it. continuing with the energy theme, the fusion reactor is approaching unity with the next iteration predicted to >= unity, if that bad boy can generate more out than in, we'll be in a very different world 50 years from now.

Quantum teleportation.

Genetic algorithms.

Older than 15 years:

http://en.wikipedia.org/wiki/Genetic_algorithm#History

Cloning and aerogel

Semantic web. I'm suprised no one mentioned it.

I think this is more of a planned research breakthrough. I.e. the type that doesn't happen.

I think every research breakthrough is planned. Impact on civilization is what counts. I just have an impression people are not aware of the huge impact Semantic web will have. Why there are no more startups dealing with SW problems?

I think pg's point is deeper than that. Research breakthroughs are made by one or two (usually one) very smart guys. Breakthroughs don't happen because consortiums and commitees declare "let the web be sematic!", which seems to be the case with the sematic web idea.

Christmas Tree Machines. We've seen the beginnings of home fabrication (including fabricating entire homes!) but as this technology progresses, it will impact the entire supply chain and wholesale/retail sector. May be a hundred years, though. But we've seen the beginnings already.

less is more (all-time record)

The internet is just a more sophisticated version of the cave drawing. The Genome is a discovery, not a breakthrough. The last true breakthrough was splitting the atom. Perhaps creating life from Methane gas and electricity would qualify.

Reply on news.ycombinator.com