John Markoff
New York Times science writer
Discussion
M
John Markoff
@markoff · reporter
I’m John Markoff and I write for the science section of the New York Times. I have been at the Times since 1988 and covered Silicon Valley from 1992 until 2010. I have written about computer security and more recently focused on artificial intelligence and robotics. I wrote the first article describing the World Wide Web in December 1993. Machines of Loving Grace is a book about the impact of robots and AI.
upvotesharetweet・
Michael Nelson
@mikenelson
Okay, let me ask the other BIG, HAIRY question related to AI. Should robots and AI systems be given enough autonomy so they can decide to kill humans in a battlefield or law enforcement setting? And how will we prevent some country from deciding to deploy such autonomous "killer" systems? We are already halfway there with some of the drone systems.
M
John Markoff
@markoff · reporter
@mikenelson The autonomous weapon question may be the most significant one we face. I think we should do whatever is possible to NOT delegate killing decisions to machines. I think we are already on the verge of a dramatic arms race and that it is likely to accelerate. That said, I believe it is possible to limit the deployment of such systems in the same way we have limited CBW and nuclear weapons.
Michael Nelson
@mikenelson
@markoff @mikenelson I definitely agree with your first sentence. But, CBW bans didn't work real well in Assad's Syria and Saddam's Iraq.
Erik Torenberg
@eriktorenberg · Former Product Hunt
Hey @markoff - Welcome!
What are some of the biggest misconceptions people have around AI?
P.S. loved your a16z podcast episode.
M
John Markoff
@markoff · reporter
@eriktorenberg I also think that we have this amazing capacity to anthropomorphize our interactions with almost everything including machines. It is not clear that we need to design these systems to be human like, but that seems to be where the technology is headed.
Hash_tag_jeff
@jeffumbro · Book Marketing and PR - get in touch
Hi @markoff - can you tell me where your thinking differs from folks like Ray Kurzweil or Jerry Kaplan?
M
John Markoff
@markoff · reporter
let me start at the beginning with my reaction to Jerry Kaplan and Ray Kurzweil. My biggest difference is that the evidence for continuing "acceleration" just isn't there.
M
John Markoff
@markoff · reporter
Moore's Law is supposed to mean that computers get faster faster and cheaper faster. Those aspects of the basic technology of Silicon Valley are slowing today. As a result I don't see any chance of the "Singularity" arriving on time. Or at least when Ray Kurzweil has predicted.
M
John Markoff
@markoff · reporter
@jeffumbro Jerry Kaplan and I differ on how disruptive the current wave of technologies is. He believes that we are headed for a "jobs-apocalypse, while I think that there is little evidence in the developed world that automation will cause huge amounts of structural unemployment.
Dylan La Com
@dylanlacom
Thanks for doing this AMA @markoff!
Can you share any specific AI or deep learning research that you're particularly excited about?
M
John Markoff
@markoff · reporter
There is machine learning/deep learning research everywhere these days! One interesting research effort that I have been watching is the work that Pieter Abeel's group has been doing on dextrous manipulation at UC Berkeley.
M
John Markoff
@markoff · reporter
They have pushed in the direction of robot arms beginning to finally match human abilities to pick and place objects.
Corley
@corleyh · COO @ Product Hunt
Hi @markoff thanks for taking the time to answer our questions. Do you remember the moment you fell in love with science? And which came first, love of science or love of writing? And final question - what is your tip for making complicated concepts easy to understand?
M
John Markoff
@markoff · reporter
@corleyh wow. so for me, it was all about science fiction first, that led to a love of astronomy. I realized in my freshman year of college that my math skills would never be good enough and so I moved on to other interests. But it was astronomy first. I had a 4" reflector telescope that I was in love with when I was a kid.
Michael Nelson
@mikenelson
And do you worry that much of what we read in 5 to 10 years will be written by robots? I've seen some pretty good sports journalism done by algorithms.
M
John Markoff
@markoff · reporter
@mikenelson Actually I don't worry very much about machine journalism. I hated writing earnings reports and I'm glad that machines are doing that. Also I'm skeptical about your point that there is good sports journalism. I think what they've done so far has been to wrap natural language around box scores. When are there are going to be robot interviewers? that's what I want to know! :)
Ben Tossell
@bentossell · Community Lead, Product Hunt
@markoff who are some of the thinkers you most admire in the AI space?
M
John Markoff
@markoff · reporter
@bentossell Well, he is an economist, but I would suggest David Autor at MIT. Also I just read David Mindell's new book on designing robots. (He is an MIT roboticist). A very influential book that I read decades ago was called the Micro Millennium, by Christopher Evans. It is what convinced me that computing technology would change the world. I also respect Janier Lanier a good deal.
Michael Nelson
@mikenelson
You could write a book on encryption (and just recycle all those New York Times articles you wrote on Clipper Chip back in the mid-90s!!)
Michael Nelson
@mikenelson
@mikenelson When it comes to the Obama White House and encryption, it's deja vu all over again, isn't it?? http://thehill.com/blogs/congres...
Alex Carter
@alexcartaz · Operations @ 60dB. Ex-PH Podcasts 😻
Hi @markoff! Many prominent technologists, including Elon Musk, have grave concerns about the dangers of AI in the near future. Are you concerned? What scenario do you think is most likely to negatively impact civilization as a result of advances in AI? Thanks!
M
John Markoff
@markoff · reporter
@alexcartaz I disagree with Elon Musk on the rate of progress, but I believe he has raised important issues about autonomous systems. There are many risks associated with disconnecting machine decision making from human control, short of self aware machine