An Architect Whose Wild Structures Imitate Cells

Sabin’s work draws from fields like biology, materials science, and mechanical engineering. MICHAEL GEORGE

Sabin’s work draws from fields like biology, materials science, and mechanical engineering.

MICHAEL GEORGE

ColorFolds responds to its environment, folding and unfolding when sensors detect changes to its own structure or people walking below. MICHAEL GEORGE

ColorFolds responds to its environment, folding and unfolding when sensors detect changes to its own structure or people walking below.

MICHAEL GEORGE

The installation’s color shifts thanks to layered polyester film. Nickel-­titanium wires contract like muscle fibers in response to heat or electricity. MICHAEL GEORGE

The installation’s color shifts thanks to layered polyester film. Nickel-­titanium wires contract like muscle fibers in response to heat or electricity.

MICHAEL GEORGE

Team members (clockwise from left include Martin Miller, Nick Cassab, Giffen Ott (on ladder), Andrew Moorman, Jenny E. Sabin, Andrew Lucia, David Rosenwasser, Jessica Jiang. Not shown: Max Vanatta MICHAEL GEORGE

Team members (clockwise from left include Martin Miller, Nick Cassab, Giffen Ott (on ladder), Andrew Moorman, Jenny E. Sabin, Andrew Lucia, David Rosenwasser, Jessica Jiang. Not shown: Max Vanatta

MICHAEL GEORGE

Architects like to poach design elements from the life sciences—a DNA helix here, a spherical nucleus there. But Cornell architecture professor Jenny Sabin goes even more interdisciplinary. “I'm not interested in just making buildings from beautiful forms in biology,” she says. Her experimental installations draw from just about every department on campus.

One of Sabin's latest projects is a set of hanging panels called ColorFolds that echoes the extracellular matrix, the zone between cells filled with proteins and other molecules critical to cell function. The structure responds to its environment like cells do, folding and unfolding when sensors detect people walking below. As it moves, the colors change, thanks to a polyester film Sabin tracked down after talks with a materials scientist. And for the movement, two mechanical engineers made springs from nickel-titanium wires that shorten in response to heat and electricity, contracting like muscle fibers. Looks like the engineers are borrowing from biology too.

An Architect Whose Wild Structures Imitate Cells

Sabin’s work draws from fields like biology, materials science, and mechanical engineering. MICHAEL GEORGE

Sabin’s work draws from fields like biology, materials science, and mechanical engineering.

MICHAEL GEORGE

ColorFolds responds to its environment, folding and unfolding when sensors detect changes to its own structure or people walking below. MICHAEL GEORGE

ColorFolds responds to its environment, folding and unfolding when sensors detect changes to its own structure or people walking below.

MICHAEL GEORGE

The installation’s color shifts thanks to layered polyester film. Nickel-­titanium wires contract like muscle fibers in response to heat or electricity. MICHAEL GEORGE

The installation’s color shifts thanks to layered polyester film. Nickel-­titanium wires contract like muscle fibers in response to heat or electricity.

MICHAEL GEORGE

Team members (clockwise from left include Martin Miller, Nick Cassab, Giffen Ott (on ladder), Andrew Moorman, Jenny E. Sabin, Andrew Lucia, David Rosenwasser, Jessica Jiang. Not shown: Max Vanatta MICHAEL GEORGE

Team members (clockwise from left include Martin Miller, Nick Cassab, Giffen Ott (on ladder), Andrew Moorman, Jenny E. Sabin, Andrew Lucia, David Rosenwasser, Jessica Jiang. Not shown: Max Vanatta

MICHAEL GEORGE

Architects like to poach design elements from the life sciences—a DNA helix here, a spherical nucleus there. But Cornell architecture professor Jenny Sabin goes even more interdisciplinary. “I'm not interested in just making buildings from beautiful forms in biology,” she says. Her experimental installations draw from just about every department on campus.

One of Sabin's latest projects is a set of hanging panels called ColorFolds that echoes the extracellular matrix, the zone between cells filled with proteins and other molecules critical to cell function. The structure responds to its environment like cells do, folding and unfolding when sensors detect people walking below. As it moves, the colors change, thanks to a polyester film Sabin tracked down after talks with a materials scientist. And for the movement, two mechanical engineers made springs from nickel-titanium wires that shorten in response to heat and electricity, contracting like muscle fibers. Looks like the engineers are borrowing from biology too.

How Graph Databases Can Reinvent Recruiting

lumaxart/Flickr

They say it’s not what you know but who you know. However, the missing implication of that is the importance of who they know. The ways that we are all connected are broadly interesting in a “six degrees of Kevin Bacon” way. However, for recruitment sites trying to find and access important, talented people, it’s much more valuable than that. Endless internet searches and manually trawling through CVs is the common process for many recruiters today — but this is time consuming and quite restrictive in what information it can deliver. Also, once the data is gathered it can be difficult to sit back and actually make sense of it. So, how can the recruitment industry move on from this?

Traditionally, the ways that such data is stored has not been easy to access or analyse. Relational databases, often formatted in tables, could provide plenty of information about individual records in a basic siloed form, but simply weren’t designed to show relationships and patterns between those records. You could unearth some of these connections at a very high level, but the results were extremely slow and lacked real definition. It’s like the difference between understanding two people that live in one house and knowing if they are married, siblings, flatmates or tenants.

Today, far too much time and effort is consumed in the slow process of working with the old style of databases or manual methods like trawling the internet. This works, of course — but it gives you little or no advantage over your competition. And the market for great talent is nothing if not competitive.

For example, tools like LinkedIn are built on a different kind of database in the background that fundamentally tracks relationships. This is the same as Facebook, Twitter — any social network, which revolves around how people know each other. The reasons these companies run on this kind of technology are clear. For a start, it’s much faster and more efficient. You can achieve the same kinds of effects in these ‘graph’ databases with 10 to 100 times less code. But perhaps more importantly, you can also perform different kinds of tasks that just aren’t possible on older tech.

If recruitment sites collect existing data sets from the likes of LinkedIn, other job sites and personal records/CVs to create their own, more comprehensive graph database, they could start to immediately understand and classify candidates based on their relationships with others, their background, specific skills and even their hobbies and interests to see if they have cultural fit. This would not be identified by reading it on profiles, but by asking the right questions with specific search criteria and key words. For example, ‘has Joan Phillips ever worked for Sony or known anyone that has?’ These systems can pick out the important connections for us in a matter of seconds.

When it comes to looking at interests and hobbies, recruitment sites can use this information to proactively approach individuals for roles that may be of interest to them. Graph databases can help source individuals who aren’t necessarily looking to move but have a clear link of interests with another dimension of the role or the company itself. For example, it might be clear from the interest data that the candidate is a football fan — the perfect match for a role at the FA. These sorts of advantages are, at best, laborious and, at worst, impossible to generate manually.

The important detail here is that it makes us more effective at identifying people who genuinely should be the best match for a position. This industry is much maligned for contacting people out of the blue for roles they aren’t interested in — but that’s because the best examples of execution are so effective and such a great match that they remain undercover. Graph databases help create more of these instances.

This doesn’t mean there’s no human component. It just means you’re able to use the power of machines to quickly and at scale, crunch the numbers, answer your queries and give you superhuman ability. The machine does what it’s good at and the recruiter is able to concentrate on what they are good at.

Equally, these tools should increase the opportunity for candidates to understand the possibilities that lie before them. If you can describe to them the caliber of employees who have worked at your company (or your client’s) — even the kinds of things those people have gone on to achieve after leaving, you can add to your arguments with clear, persuasive data. Again, this comes directly from better tracking of relationships in the data.

For a long time, this industry has been built around the power of informal, real world networks and word of mouth recommendations. There’s no longer any reason that the tools we use to pursue those same goals are designed around the same philosophy. Armed with graph databases, recruitment sites will be able to quickly identify the best candidates for their roles and in turn, candidates are much more likely to be happy with their new jobs that have been matched so well.

Emil Eifrem is founder of the Neo4j open source graph database project.

How Graph Databases Can Reinvent Recruiting

lumaxart/Flickr

They say it’s not what you know but who you know. However, the missing implication of that is the importance of who they know. The ways that we are all connected are broadly interesting in a “six degrees of Kevin Bacon” way. However, for recruitment sites trying to find and access important, talented people, it’s much more valuable than that. Endless internet searches and manually trawling through CVs is the common process for many recruiters today — but this is time consuming and quite restrictive in what information it can deliver. Also, once the data is gathered it can be difficult to sit back and actually make sense of it. So, how can the recruitment industry move on from this?

Traditionally, the ways that such data is stored has not been easy to access or analyse. Relational databases, often formatted in tables, could provide plenty of information about individual records in a basic siloed form, but simply weren’t designed to show relationships and patterns between those records. You could unearth some of these connections at a very high level, but the results were extremely slow and lacked real definition. It’s like the difference between understanding two people that live in one house and knowing if they are married, siblings, flatmates or tenants.

Today, far too much time and effort is consumed in the slow process of working with the old style of databases or manual methods like trawling the internet. This works, of course — but it gives you little or no advantage over your competition. And the market for great talent is nothing if not competitive.

For example, tools like LinkedIn are built on a different kind of database in the background that fundamentally tracks relationships. This is the same as Facebook, Twitter — any social network, which revolves around how people know each other. The reasons these companies run on this kind of technology are clear. For a start, it’s much faster and more efficient. You can achieve the same kinds of effects in these ‘graph’ databases with 10 to 100 times less code. But perhaps more importantly, you can also perform different kinds of tasks that just aren’t possible on older tech.

If recruitment sites collect existing data sets from the likes of LinkedIn, other job sites and personal records/CVs to create their own, more comprehensive graph database, they could start to immediately understand and classify candidates based on their relationships with others, their background, specific skills and even their hobbies and interests to see if they have cultural fit. This would not be identified by reading it on profiles, but by asking the right questions with specific search criteria and key words. For example, ‘has Joan Phillips ever worked for Sony or known anyone that has?’ These systems can pick out the important connections for us in a matter of seconds.

When it comes to looking at interests and hobbies, recruitment sites can use this information to proactively approach individuals for roles that may be of interest to them. Graph databases can help source individuals who aren’t necessarily looking to move but have a clear link of interests with another dimension of the role or the company itself. For example, it might be clear from the interest data that the candidate is a football fan — the perfect match for a role at the FA. These sorts of advantages are, at best, laborious and, at worst, impossible to generate manually.

The important detail here is that it makes us more effective at identifying people who genuinely should be the best match for a position. This industry is much maligned for contacting people out of the blue for roles they aren’t interested in — but that’s because the best examples of execution are so effective and such a great match that they remain undercover. Graph databases help create more of these instances.

This doesn’t mean there’s no human component. It just means you’re able to use the power of machines to quickly and at scale, crunch the numbers, answer your queries and give you superhuman ability. The machine does what it’s good at and the recruiter is able to concentrate on what they are good at.

Equally, these tools should increase the opportunity for candidates to understand the possibilities that lie before them. If you can describe to them the caliber of employees who have worked at your company (or your client’s) — even the kinds of things those people have gone on to achieve after leaving, you can add to your arguments with clear, persuasive data. Again, this comes directly from better tracking of relationships in the data.

For a long time, this industry has been built around the power of informal, real world networks and word of mouth recommendations. There’s no longer any reason that the tools we use to pursue those same goals are designed around the same philosophy. Armed with graph databases, recruitment sites will be able to quickly identify the best candidates for their roles and in turn, candidates are much more likely to be happy with their new jobs that have been matched so well.

Emil Eifrem is founder of the Neo4j open source graph database project.

Why Google’s Cancer-Detecting Pill Is More Than Just Hype

Google is hoping to build nanoparticles that can detect cancers inside your body—and notify a wearable computer on your wrist. Google

Before Google started work on a pill that aims to detect cancers and other diseases by sending magnetic nanoparticles into your bloodstream, it talked to Sam Gambhir.

Gambhir is a professor of radiology, bioengineering, and materials science at Stanford University and the director of the university’s Canary Center for Cancer Early Detection—a researcher at the forefront of a movement that seeks to identify cancers far sooner than we do today. Googlers Andrew Conrad and Vik Bajaj approached him about a year and a half ago, not long after the company hired Conrad to oversee a new health sciences effort inside Google X, its “moonshot” research lab.

Basically, Gambhir says, they wanted recommendations on what moonshots they should try for. And as Google built up its health operations, he continued to consult with the company and the health sciences lab it now runs in Silicon Valley.

One of the projects Google eventually settled on was what Conrad calls the “Nanoparticle Platform,” an effort to build a cancer-detecting pill, publicly revealed last week. The idea is that this pill will contain magnetic nanoparticles that can latch onto certain cancer-related molecules in the bloodstream—and that a wearable device could then use magnetic properties to recognize when this happens. As Gambhir points out, this is just one of many efforts to detect cancer in vivo—i.e. within the body, without drawing blood. But he’ll also tell you that Google brings something new to such a project.

Part of it, he explains, is that Google has built an unusually talented team that spans multiple disciplines, including physics, chemistry, and biology. “They have brought on a lot of very smart people that are thinking about these problems in very unique ways, ” Gambhir says, pointing out the company has hired some researchers from his lab. Certainly, some academic centers have built their own multi-discipline teams, but on top of this, Google provides a new kind of corporate leverage. It aims to push this sort of thing into the market at speed.

“Academic institutions aren’t as good at making an actual product. Research has to leave the academic world and move into the industrial world, and most industrial world applications are focused on therapeutics rather than diagnostics—and certainly not diagnostics based on wearable sensors,” Gambhir says.

Conrad and his team have taken much the same approach in building a contact lens that can detect blood sugar levels through the tears in your eyes. The lens would let diabetics track their blood sugar without ever having to draw their own blood, and through various third parties, Google is now working to turn the thing into a product. The lens was originally developed at Microsoft, but it’s Google that’s trying to commercialize the thing.

All that said, it will likely be years before a Google cancer-detecting pill reaches the market—if it reaches the market at all. Google has yet to test its nanoparticles on humans—at this point, it’s sending prototypes into artificial human limbs—and according to Muneesh Tewari, who heads a research lab working on early detection of cancer at the University of Michigan, reaching the market will require not only additional research but some rather significant regulatory wrangling. “The concept is very exciting and has merit—to do more proactive and continuous monitoring in the blood,” he says. “The question is how feasible is this and over what time frame. It’s still quite early days.”

First, he says, Google must demonstrate that its particles can indeed attach to markers in the bloodstream that are highly specific for cancer—and that a wearable device can read information from these particles well enough to make an accurate diagnosis. And then it must win approval from the Food and Drug Administration. Google has said that its particles are similar to those used with MRIs and other clinical procedures, but according to Tewari, the regulatory safety bar may be higher because its pill is intended to be used by healthy individuals

Colin Connolly—a senior scientist with Quantum Diamond Technologies, a company that specializes in biological sensors built with magnetic nanoparticles—says much the same. “It remains to be validated that this would be a safe way to go. With anything that’s inside a person—as opposed to a blood draw—the challenge is greater.”

Indeed, it is. But that’s largely the point. The Google X lab was built for challenges like this. Says MIT’s Robert Langer, another academic who has consulted with the company on its nanoparticles project: “A lot of companies are doing one step in a project like this. They’re doing four steps.”