AI:
Wilson, Jim and Metz, Cade. “Busting the Myths about A.I. Invading
Our Lives.” The New York Times. 2017 Dec. 13.
I was attracted to this article based on the title - my mind
instantly associates A.I. with robots gradually outsmarting humans and taking
over the planet. Hence, it is refreshing to read a NY Times article assuring
that the dawn of the A.I. anarchy will not be occurring soon. The article is
written as a question and answer with a technology reporter, Cade Metz, who is
located in San Francisco and reflecting on the tech/A.I. that he uses and has
written about. Metz marvels at the advances in the voice assistants in the past
five years, but his ultimate conclusion is that “the improvements have been
enormous. But there is still a very long way to go.” He discusses that the goal
of voice assistants to understand conversational speech, and the current
assistants do a decent job at recognizing speech, but not understanding. With
regard to visual A.I., Metz discusses that computer vision techniques can be
flawed in recognition which is problematic if A.I. is to be used for security
cameras or autonomous cars. As opposed to my initial fears of robot overlords
(which is discussed in the article to a degree), the issue that Metz presses to
be more urgent is the shift in job markets that will result from automation. We
saw an example of this in class with the masonry-laying robot. However, what
scares me most about A.I. ebbing into the workforce is the prospect of A.I.
overtaking design positions. Structural engineers rely more and more heavily on
structural analysis software, and although it is years away from fruition, the
thought of software autonomously producing a complete set of design documents
seems plausible to me and is very unnerving. I do not believe the construction
industry will reach a point where A.I. is trusted enough to stamp and send out
a set of drawings, but I do think A.I. will progress to the point of completing
and detailing an entire design autonomously.
Computer:
Wod, Marcus. “The Plan to Build a Skyscraper that Doesn’t
Cast a Shadow.” Wired.com.n 2015 Mar. 13.
Without saying, computers have become a stable in nearly all
applications, especially in the field of engineering. Under the computer tag,
an article was listed that discusses the design of a building that will reflect
light in such a way to minimize/eliminate its shadow. What I found most
impressive in the article is that the architects simply entered their desired
restraints into a software, such as building area and maximized light to the
ground. In turn, the software iteratetd through all the possible outcomes and
spat out the viable shapes that satisfy the restraints. Design at the press of
a button! I think this example shows the immense power that computers hold in
crafting solutions to particular problems, especially with buildings.
Software:
Tanz, Jason. “Andy Rubin Unleashed Android on the World. Now
Watch Him Do the Same with AI.” Wired.com. 2016 Feb. 09.
Andy Rubin is an incredible developer who created Android and
headed Google’s mobile internet efforts, until he left Google to start his own
company. Tanz’s article largely recounts Rubin’s constant innovation and
unwavering determination to bring his visions of the future into the realities
of today. One of the more eye-catching aspects of the article to me was the
included timeline that cites Rubin to be pioneering preliminary versions of
social networking (1981), mobile phones (1992), self-driving cars (2004) all
before the research and development of these technologies were mainstream. As
the article calls him, he is an impatient futurist. Pairing with the discussion
of AI above, Rubin foresees A.I. being the focus of the next technological era.
However, Rubin is aiming to build his neural network database differently than
the other tech players. His vision is to send a fleet of robots (sensing
machines) out in the world to feed his A.I. database firsthand with immersive
data. To do this, Rubin’s company, named Playground Global, invites developers
to use Playground’s advanced engineering sensors and hardware to fulfill their
ideas. With this setup, Rubin can fuel the creativity of other developers while
sending his sensors out into the world to build his database of data. I am
anxious to see the A.I. that is developed from Rubin’s platforms; his thoughts
and goals sound good in theory, but I have yet to be convinced that they will
work in reality. Regardless, it was quite interesting to read about a developer
with such an excitement and vision for where tech can be in the future.
Future:
Metz, Cade. “Building A.I. That Can Build A.I.” The
New York Times. 2017 Nov. 05
In the same A.I. spirit as the other articles, Metz writes about a
prospective tool that is in development for the future of A.I. The article
states that only 10,000 people in the world have the expertise required
to develop A.I. - which is leaving the large tech companies to scramble for
qualified developers. Hence, AutoML is a pursuit of some tech giants such as
Google. ML is explained in the article to stand for machine learning, which
means that AutoML would describe machines that are capable of coding A.I. This
does not currently exist. The most prevalent concern discussed with A.I. was
the automation of jobs, which would cause a shift the workforce. If AutoML
becomes advanced enough, coders would be coding themselves out of a job. I
appreciate that Google’s solution to the lack of qualified computer experts is:
“we can code computer experts!” - however, coding computers with artificial
intelligence to generate more computers with A.I. will propagate into a fleet
of self-improving A.I. I find this to cross the line of what should be done with
A.I.
Comments:
To Andrew Maita:
Most of the articles I read for this round of posting pertained to
A.I., one of which focused on building a sensory database for robots to
encounter and understand the world around them. I don’t know too much about the
3D printers used for buildings, but I believe they are typically stationed in
one location and set up to print around themselves. I could easily see 3D
printers advancing into a mobile system with sensors that allow the system to
move around the footprint of the building and print as needed. I also think it
will only be a matter of time until we can 3D print all of the materials needed
for a complete building - I would not be surprised if eventually we can 3D
print the rebar as the machine builds up the concrete.
To Zac Arnold:
Having programs with the ability to identify code violations is
super useful in my eyes. I always envision designing my own house one day, but
know that I would mess up something with the architectural codes and
restraints. To the same effect, I agree with you that avoiding something that
is obviously catastrophic in a design seems like an intuitive “decision-making”
ability for the software.
As you say in your sociology section, some people welcome the
automation and others have reservations. I typically have reservations. I think
automation can easily be taken too far. One of the articles about A.I. that I
read discussed how Google is aiming to develop A.I. that can program more A.I.
I think this level of automation, where coders are coding themselves out of a
job, takes automation too far; yet, it is the same effect as automating cars on
the street and putting taxi drivers out of a job.
To Jordan Shuster
I agree with you that robotics and automation has the potential
displace large quantities of people from their current jobs. As I have posted
before, A.I. coders are trying to code computers that are capable of developing
more A.I. software (AutoML). In the same effect as eliminating jobs from the
unskilled and lower-class workers, programmers are trying to program themselves
out of a job. From an objective standpoint, a well-coded software and machine
would be the more efficient and economical option for many jobs. If we keep
leaning into reliance on machines, I do not know what will happen to the
workers who are displaced from their job. Who knows... as A.I. for business
applications becomes more prevalent (as you discuss with your Google Cloud ML),
perhaps everybody will have the business chops at their disposal to thrive as
successful entrepreneurs.
AI:
ReplyDeleteI share to an extent a fear of AI not taking up our jobs. I understand that for now, majority of robots that are created or even in the testing stages are only created to do work that humans do not want (for various reasons) but what defines what humans want and not want to do. Right now its manual labor but soon it might actually take on the design aspects of our trade. Again this reiterates our class discussion of how AI has yet to achieve the ability to create or decided on its own. So yes I agree that AI is getting scarily advanced (Based on the Black Mirror episodes I have watched) but from the perspective of a structural designer, I can say at least for the next decade, we should be fine.
I have to agree the title “Busting the Myths about A.I. Invading Our Lives.” is attracting, because the idea of having robot replacing us is scary and as you mentioned reading such an article is refreshing, because it leaves me wondering what would happen if automation would take over within the next few years when there is already a high percentage of people who didn’t have their education go beyond high school and their skills are dependent of the jobs they want to replace first such as call centers, retail stocking, etc.
ReplyDelete