Saturday, August 31, 2019
The Mysterious Vaping Illness That’s ‘Becoming an Epidemic’
August 31, 2019 at 07:33PM
A surge of severe lung ailments has baffled doctors and public health experts.
Friday, August 30, 2019
YouTube Said to Be Fined Up to $200 Million for Children’s Privacy Violations
August 31, 2019 at 02:38AM
The fine, from the Federal Trade Commission, could have significant repercussions for other social media platforms that are popular with children.
Twitter C.E.O. Jack Dorsey’s Account Hacked
August 31, 2019 at 12:54AM
Mr. Dorsey’s account began posting racial epithets, profanities and bomb threats on Friday afternoon. A company spokesman confirmed it was hacked.
It’s Not Easter, but There Might Be a Surprise Hidden on This Article
August 30, 2019 at 10:27PM
The New York Times is no stranger to the trend of hiding code in plain sight.
How Tainted Gold May Have Ended Up in Your Phone
August 30, 2019 at 06:55PM
A New TV Show from The New York Times on FX and Hulu
Don’t Use Bootleg or Street Vaping Products, C.D.C. Warns
August 30, 2019 at 06:34PM
With more than 200 cases of respiratory illnesses possibly related to vaping, and no single culprit, public health experts are advising young people not to buy cannabis and other e-cigarette products off the street.
The Week in Tech: Are You Ready for Facebook’s Future?
August 30, 2019 at 04:00PM
We got a glimpse of what the social network’s privacy-first focus might look like in practice. You may not love it.
What Does VSCO Think About the ‘VSCO Girls’?
August 30, 2019 at 12:00PM
We spoke to the brands behind Gen Z’s highly aestheticized (and widely parodied) personal brand.
Thursday, August 29, 2019
Hospitals Should Replace Infection-Prone Scopes With Safer Models, F.D.A. Says
August 30, 2019 at 01:27AM
Reusable duodenoscopes, used for internal examinations of patients, have been linked to a series of hospital outbreaks.
Uber and Lyft to Spend $60 Million to Maintain California Driver Status
August 30, 2019 at 01:03AM
A state legislative proposal called Assembly Bill 5 could could give drivers more employee rights. The companies said they are willing to start a ballot initiative to fight it.
Organoids Are Not Brains. How Are They Making Brain Waves?
August 29, 2019 at 09:46PM
Clusters of living brain cells are teaching scientists about diseases like autism. With a new finding, some experts wonder if these organoids may become too much like the real thing.
India’s Restaurants Rebel Against Food Delivery Apps
August 29, 2019 at 09:21PM
Frustrated by the steep commissions and discounts that apps like Zomato and Uber Eats push on restaurants, thousands have started a movement to #Logout from them.
Those People Starting Successful Tech Companies? Most Are Middle-Aged
August 29, 2019 at 07:18PM
Some of tech’s biggest names had founders in their teens. But new research shows that for fast-growing start-ups in the U.S., the founders’ average age was 45.
Wednesday, August 28, 2019
Voices in AI – Episode 94: A Conversation with Amy Webb
August 22, 2019 at 03:00PM
[voices_in_ai_byline]
About this Episode
Episode 94 of Voices in AI features Byron speaking with fellow futurist and author Amy Webb on the nature of artificial intelligence and the morality and ethics tied to its study.
Listen to this episode or read the full transcript at www.VoicesinAI.com
Transcript Excerpt
Byron Reese: This is Voices in AI brought to you by Gigaom, and I’m Byron Reese. Today, I’m so excited. My guest is Amy Webb. She is a quantitative futurist. She is the founder and CEO of the Future Today Institute. She’s the professor of strategic foresight at NYU’s Stern School of Business.
She’s a co-founder of Spark Camp. She holds a BS in game theory and economics, and an MS in journalism from Columbia, and she’s a Nieman Fellow at Harvard University. And she- as if all of that weren’t enough, she’s the author of a new book called The Big Nine: How the Tech Titans and Their Thinking Machines Could Warp Humanity. Welcome to the show, Amy.
Amy Webb: Hello. Thank you for having me.
So, I always like to start off with definitions. Maybe that can be tedious, but start me off with how do you define intelligence?
Well, I think intelligence in different context means different things. As it relates to artificial intelligence, I think, generally, what we’re trying to describe is the ability for a system to make decisions and choices that would mirror or come close to the way that we make decisions and choices. So, I think, when the term Artificial Intelligence was originally coined at Duke University in the 1950s by Marvin Minsky and John McCarthy, back then the thinking was to try and build a machine that came close to human cognition and whose intelligence could be measured against our own. Five decades later, I think it’s clear that, while we’re able to build incredibly powerful machines, there’s a lot that we don’t understand about how the–a lot we don’t understand about our own intelligence and our own cognition, and I think, Minsky and McCarthy, if they were around today still working very intently, that they would say that intelligence, AI was probably the wrong term to use.
Actually, McCarthy did say that pretty quickly after he coined the term. He said he thought it set the bar too high. But I guess the question I really want to ask is, in what sense is it artificial? Do you think AI is actually intelligent, or do you think it’s something that mimics intelligence the way that, say, artificial turf mimics grass?
Well it’s certainly been billed to mimic what we think we know about human intelligence. And that extends from hardware architecture. The current deep neural net system is a layered approach to machine learning that is built to physically mimic what we think the sort of synapses in our own brains–how they’re transferring and parsing and understanding information. That being said, I think that–I also think that intelligence is probably the wrong term because it’s fairly loaded, and when we talk about intelligence, we tend to–I think we talk about intelligence because we’re trying to quantify in some way our own cognitive power. I don’t know if that’s the right term for AI because I think AI inherently is different. Just because we’ve built it to mimic what we think we do doesn’t mean that from here out it continues to behave in that manner.
Yeah, it’s funny. I’ve always thought all of the analogies between how AIs function, how the brain functions are actually just kind of techno marketing. The brain is still largely a black box, right? We don’t know how even a thought is encoded or anything like that, and so, I’ve often suspected that’s just–that machine intelligence really doesn’t have anything in common with human intelligence. Would you agree, disagree, defer?
Yeah, I mean, I think that some–we’re talking in broad strokes and generalities. So, I think that there are certainly some areas of the AI ecosystem that do a fairly good job of mimicking what we do, so reinforcement learning and hierarchical reinforcement learning is present in dogs and children and adults. So basically, you’ve got a, say, toddler, and you’re trying to teach the toddler correct and incorrect, right? So the correct- I don’t know. The correct word is mommy for your mom. The correct term for the color that is the color of the sky is blue, stuff like that. And through praise or correction, the child over time learns. So we are reinforcing the correct answer, and we are, hopefully, in a very gentle loving way punishing the–guiding the child away from the wrong answer. So, the same is true in how some AI systems learn. And in fact, for anybody who’s been following AlphaGo and all of the work that the DeepMind team has been doing- a lot of what they’re doing is self-improving. It was reinforcement learning and then self-improvement, so.
Listen to this episode or read the full transcript at www.VoicesinAI.com
[voices_in_ai_link_back]
Byron explores issues around artificial intelligence and conscious computers in his new book The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity.
6 Google Tricks That Will Turn You Into an Internet Detective
August 27, 2019 at 09:42AM
Even if you’re already a Google pro, these tricks will get you to your desired results even faster.
Peloton Is a Phenomenon. Can It Last?
August 28, 2019 at 07:53PM
Exercise manias like the Thighmaster and Tae Bo have all come and gone. Peloton is trying to defy that trend.
When Children Use Technology, Let Common Sense Prevail
August 28, 2019 at 04:00PM
All things in moderation is a good way to think about it, says Jessica Grose, the editor of NYT Parenting.
The Joys of Being a Late Tech Adopter
August 28, 2019 at 12:43PM
New-gadget season is almost here, but our columnist has concluded, after testing hundreds of gizmos over the years, that it pays to wait.
Facebook Tightens Rules on Verifying Political Advertisers
August 28, 2019 at 01:00PM
Bracing for the 2020 U.S. presidential election, the company adds more steps for buyers of political ads. Disinformation experts aren’t sure it is enough.
6 Google Tricks That Will Turn You Into an Internet Detective
August 27, 2019 at 09:42AM
Even if you’re already a Google pro, these tricks will get you to your desired results even faster.
Tuesday, August 27, 2019
Philip Morris and Altria Are in Talks to Merge
August 28, 2019 at 04:00AM
With Altria’s investment in Juul, a combination of the tobacco giants would dominate the international market for e-cigarettes.
Telegram Pushes Ahead With Plans for ‘Gram’ Cryptocurrency
August 28, 2019 at 01:50AM
When Facebook announced plans for a similar effort this year, the regulatory pushback was fierce. But Telegram appears undeterred.
Former Star Google and Uber Engineer Charged With Theft of Trade Secrets
August 27, 2019 at 09:07PM
Anthony Levandowski, one of Silicon Valley’s foremost technologists on autonomous cars, was charged with theft and attempted theft of trade secrets from Google.
The Baroness Fighting to Protect Children Online
August 27, 2019 at 07:25PM
Baroness Kidron says tech giants exert too much leverage over young users and has pushed laws to change that. “It’s little Timmy in his bedroom versus Mark Zuckerberg in his Valley,” she said.
Netflix Falls Short in Big Screen Debut of ‘The Irishman’
August 27, 2019 at 06:27PM
Martin Scorsese’s ‘The Irishman’ will debut in select theaters three weeks ahead of its Netflix release, a shorter run than Netflix had hoped.
Is It Time to Upend the Periodic Table?
August 27, 2019 at 12:00PM
The iconic chart of elements has served chemistry well for 150 years. But it’s not the only option out there, and scientists are pushing its limits.
How to Turn an iPhone Into a Work-Only Tool
August 22, 2019 at 05:59PM
To prevent distractions, Conor Dougherty, an economics writer, dumped social media and anything fun — even his browser — from his smartphone.
Cadillac’s Last Stand? Storied Brand Aims (Again) for Revival
August 27, 2019 at 10:00AM
Once the epitome of luxury, the high-end General Motors line is attempting another makeover to catch up with rivals that have overtaken it.
Hi, Alexa. How Do I Stop You From Listening In On Me?
August 22, 2019 at 06:12AM
Humans help train artificial intelligence, so someone may be listening to what you tell Siri, Google Assistant or Alexa. Here’s what to do if you have a problem with that.
Monday, August 26, 2019
Lawsuit Over Computer Chips Invokes Trade War With China
August 27, 2019 at 12:52AM
Globalfoundries has accused T.S.M.C., a competitor in Taiwan, of infringing on 16 patents in a suit that could also affect companies like Apple and Google.
Automating HIPAA Compliant Hosting By Using DevOps
August 26, 2019 at 07:11PM
DevOps is no longer just an industry buzzword. It has become the standard development practice irrespective of the size of the business and across a range of industry verticals. The healthcare industry which historically has been a laggard when it comes to adopting new technologies is quickly catching pace and has started to see the advantages of hosting healthcare data in the cloud using DevOps.
Managing compliance in healthcare software
There has been a huge surge in the amount of healthcare data being generated in recent times. Apart from the EMRs and EHRs, there has been a tremendous increase in the amount of healthcare-related data produced by the health monitoring devices and other health applications as well.
Compliance with the regulations is a must when storing health data. The sensitive patient data is subject to several laws and regulations like HIPAA, HITECH, GDPR and so on. With new and stringent compliance norms being rolled out and increasingly aware consumers who are alert about their privacy and data safety, adherence to the data compliance norms decidedly takes the main stage when dealing with healthcare software development.
What is HIPAA?
HIPAA or the Health Insurance Portability and Accountability Act is a legislation by the government of the United States that provides provisions for data privacy and security of sensitive patient data and safeguarding medical information.
The act aims at modernizing the flow of healthcare information and protecting any personally identifiable information maintained by healthcare providers, insurers or any associated organization.
Why take the DevOps approach for ensuring HIPAA compliance?
The benefits of DevOps for organizations are multifold. By correctly applying the DevOps principles, you ensure that your organization starts off on the right foot for ensuring compliance. Instead of worrying about compliance after the application development phase is completed, following the DevOps practices necessitates the incorporation of compliance into the pipeline right from the beginning of the development process.
Since DevOps breaks down the silos that exist between the development and operations teams, it brings everyone together to think of compliance to norms. Unlike traditional processes where the only people worrying about HIPAA compliance are the members of the security team and possibly the lawyers, DevOps offers the opportunity to every team member from developers to testers to the IT ops guys, to work collaboratively and become stakeholders for ensuring compliance to the HIPAA norms.
Automating the process to ensure HIPAA compliance has the potential to improve the speed, security and reliability to the norms stated. Moving data and workflows to the cloud gives you accessibility to data and promotes interoperability but the security of cloud-hosted data remains a challenge. Ensuring that the data stored on cloud stays HIPAA compliant is a challenge that DevOps can make manageable.
How to automate HIPAA compliant hosting using DevOps
Automating a secure cloud-based healthcare organization has much more to it than simply choosing a HIPAA compliant cloud service provider. In addition to covering the prerequisites for HIPAA compliant hosting, you also need to ensure that the security of the sensitive patient data is given due importance. Taking the DevOps approach in software development can facilitate compliance by design.
Design the process for incorporating security
In the traditional development practices, the security team’s role came into the picture at the end of the development cycle which lasted for months or even years. The healthcare applications that work on the DevOps principles have a much higher development velocity than the traditional development cycles. The security review at the end of the development cycle no longer suffices. It has to be baked into the process itself.
Infusing an aspect of security into the DevOps process, popularly termed as DevSecOps, requires a shift in the mindset when developing healthcare applications. Security needs to become a shared responsibility with an end to end integration approach.
Security needs to be considered right from the beginning when designing the application infrastructure. Automation of the security gateways is another important component for ensuring that the DevOps workflows do not slow down.
Secure coding
Developers have access to tools such as open-source code vulnerability scanners and static code analyzers which are capable of testing the code for security lapses as soon as the changes are manifested. With secure coding in place, the errors are caught earlier and rectified sooner resulting in greater compliance and a more secure database.
Version control
Application development with version control requires the developers to periodically check in the changes they made to a common repository. This ensures that every member of the team has access to the latest version of the application. It also helps with automated deployments and DevOps results in a more secure development process. This also brings transparency into the process by ensuring that the changes made to the application are clearly reflected in the system.
Continuous Integration
Once version control is established, continuous integration results in an automated build as the code changes get checked in. Unit tests and early feedback ensures that the code and its related resources are integrated regularly and the teams can detect problems, if any, at an earlier stage. In case any particular build fails, the teams can fix the errors and re-test it allowing a stable build to be always available.
Release management
The deployment and release of the artifacts need to be done against a staging database. A database architect needs to review it to determine its production readiness. The actual deployment may be manual, automated using a release management tool or managed using a staging environment depending on the level of control required in the project.
Ensure security of cloud-hosted data
There are two aspects of security that need to be looked out for- security of the data and security of the CI/CD pipeline. The development and operations environment need to be scrutinized individually.
Cloud-native technologies like microservices and containers are a major part of the DevOps initiatives and these need to be covered under security to ensure that the cloud-hosted data remains compliant to HIPAA. Containers allow for a dynamic infrastructure at a larger scale. Static security policies aren’t sufficient to ensure the security of sensitive data. The security needs to be continuous and ingrained into every stage of the healthcare app life cycle.
Standardization of the environment
Standardization and automation of the environment is essential to minimize the chances of unauthorized access. In the case of microservices, tight access control and centralized authentication mechanisms are required for ensuring security.
Data encryption and secure API gateways
A container orchestration platform with the required security features like encryption of data between apps and services reduces the chances of damage as a result of any unauthorized access. Secure API gateways increase routing visibility and promote authorized access. By limiting the number of exposed APIs, the security of the data increases.
Isolate the containers
Containers are a target for valuable data, both in transit and at rest. These need to be isolated from one another as well as from the network to address the security concerns. Additionally, integrating security scanners for the containers is a measure in the right direction to amp up the security.
Automate the testing
Security analysis tools should be incorporated as a part of the build. The input validation tests, authentication and authorization need to be automated. Automation of security updates via a DevOps pipeline results in formation of a well-documented change log that is beneficial during the audit process as well.
Monitor for compliance
Monitoring of the performance and availability of the application is crucial in DevOps. This not only ensures that the features are released to the users at a faster pace, but it also makes sure that the quality and security standards are upheld and in compliance with the regulatory guidelines.
Any unusual blockage or deadlock into the process has to be resolved in real-time. The effect the deployments have on the performance also has to be communicated during the ongoing production cycle. Healthcare organizations are required to monitor and manage access to data.
The sensitive data stored within the database has to be available and identifiable and breaches if any have to be reported promptly. You also need to have access to the records of data management and server use so that any inherent performance issues can be quickly resolved.
Closing words
HIPAA compliance is a mandatory requirement if your organization stores database, infrastructure or workflows on the cloud. Irrespective of whether you are building a web application or mobile app, DevOps has the potential to help you with seamless application delivery.
The application developed need to give full visibility for audits and maintaining regulatory adherence to the norms. Integration of security in DevOps requires an upheaval in the mindsets, processes and tools being used for it to be effective in the cloud environments.
About The Author
Vinati Kamani writes about emerging technology and their applications across industries for Arkenea. She is an avid reader and self-proclaimed bibliophile. When not at her desk penning down articles or reading up on the recent trends, she can be found traveling to remote places and soaking up different cultural experiences.
Friday, August 23, 2019
Voices in AI – Episode 94: A Conversation with Amy Webb
August 22, 2019 at 03:00PM
[voices_in_ai_byline]
About this Episode
Episode 94 of Voices in AI features Byron speaking with fellow futurist and author Amy Webb on the nature of artificial intelligence and the morality and ethics tied to its study.
Listen to this episode or read the full transcript at www.VoicesinAI.com
Transcript Excerpt
Byron Reese: This is Voices in AI brought to you by Gigaom, and I’m Byron Reese. Today, I’m so excited. My guest is Amy Webb. She is a quantitative futurist. She is the founder and CEO of the Future Today Institute. She’s the professor of strategic foresight at NYU’s Stern School of Business.
She’s a co-founder of Spark Camp. She holds a BS in game theory and economics, and an MS in journalism from Columbia, and she’s a Nieman Fellow at Harvard University. And she- as if all of that weren’t enough, she’s the author of a new book called The Big Nine: How the Tech Titans and Their Thinking Machines Could Warp Humanity. Welcome to the show, Amy.
Amy Webb: Hello. Thank you for having me.
So, I always like to start off with definitions. Maybe that can be tedious, but start me off with how do you define intelligence?
Well, I think intelligence in different context means different things. As it relates to artificial intelligence, I think, generally, what we’re trying to describe is the ability for a system to make decisions and choices that would mirror or come close to the way that we make decisions and choices. So, I think, when the term Artificial Intelligence was originally coined at Duke University in the 1950s by Marvin Minsky and John McCarthy, back then the thinking was to try and build a machine that came close to human cognition and whose intelligence could be measured against our own. Five decades later, I think it’s clear that, while we’re able to build incredibly powerful machines, there’s a lot that we don’t understand about how the–a lot we don’t understand about our own intelligence and our own cognition, and I think, Minsky and McCarthy, if they were around today still working very intently, that they would say that intelligence, AI was probably the wrong term to use.
Actually, McCarthy did say that pretty quickly after he coined the term. He said he thought it set the bar too high. But I guess the question I really want to ask is, in what sense is it artificial? Do you think AI is actually intelligent, or do you think it’s something that mimics intelligence the way that, say, artificial turf mimics grass?
Well it’s certainly been billed to mimic what we think we know about human intelligence. And that extends from hardware architecture. The current deep neural net system is a layered approach to machine learning that is built to physically mimic what we think the sort of synapses in our own brains–how they’re transferring and parsing and understanding information. That being said, I think that–I also think that intelligence is probably the wrong term because it’s fairly loaded, and when we talk about intelligence, we tend to–I think we talk about intelligence because we’re trying to quantify in some way our own cognitive power. I don’t know if that’s the right term for AI because I think AI inherently is different. Just because we’ve built it to mimic what we think we do doesn’t mean that from here out it continues to behave in that manner.
Yeah, it’s funny. I’ve always thought all of the analogies between how AIs function, how the brain functions are actually just kind of techno marketing. The brain is still largely a black box, right? We don’t know how even a thought is encoded or anything like that, and so, I’ve often suspected that’s just–that machine intelligence really doesn’t have anything in common with human intelligence. Would you agree, disagree, defer?
Yeah, I mean, I think that some–we’re talking in broad strokes and generalities. So, I think that there are certainly some areas of the AI ecosystem that do a fairly good job of mimicking what we do, so reinforcement learning and hierarchical reinforcement learning is present in dogs and children and adults. So basically, you’ve got a, say, toddler, and you’re trying to teach the toddler correct and incorrect, right? So the correct- I don’t know. The correct word is mommy for your mom. The correct term for the color that is the color of the sky is blue, stuff like that. And through praise or correction, the child over time learns. So we are reinforcing the correct answer, and we are, hopefully, in a very gentle loving way punishing the–guiding the child away from the wrong answer. So, the same is true in how some AI systems learn. And in fact, for anybody who’s been following AlphaGo and all of the work that the DeepMind team has been doing- a lot of what they’re doing is self-improving. It was reinforcement learning and then self-improvement, so.
Listen to this episode or read the full transcript at www.VoicesinAI.com
[voices_in_ai_link_back]
Byron explores issues around artificial intelligence and conscious computers in his new book The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity.
Google Tries to Corral Its Staff After Ugly Internal Debates
August 23, 2019 at 11:21PM
The tech company, telling its employees to “stop disrupting the workday,” issued new workplace rules designed to rein in its freewheeling culture.
Qualcomm Wins Reprieve in F.T.C. Antitrust Case With Appeals Court Ruling
August 23, 2019 at 11:09PM
A federal appeals court ruled that the chip maker did not have to modify key business practices while the court reviews a lower-court ruling against the company.
As Amazon Fires Spread, So Do the Misleading Photos
August 23, 2019 at 11:20PM
The fires are real. But the photographs that celebrities and politicians are sharing? Well, it depends.
Facebook Bans Ads From The Epoch Times
August 23, 2019 at 09:39PM
The social network has struggled to implement a consistent political advertising policy as groups appear able to get around its transparency rules.
How Uber Got Lost
August 23, 2019 at 05:11PM
The once-swaggering company is losing more money and growing more slowly than ever. What happened?
The Week in Tech: Facebook’s First Step Toward Treating Our Data Better
August 23, 2019 at 04:00PM
The social network’s latest privacy feature at least lets us imagine how we could better control our data in the future.
Snapchat’s Disappearing Act Leaves Venice Beach Searching for Its Future
August 23, 2019 at 12:00PM
Snap changed social media — and Venice, Calif. — for good. What’s next?
H.D.R. Is Coming to a TV Near You. Here’s Why You Should Care.
August 22, 2019 at 04:04PM
If you’re shopping for a new TV, a new acronym should be on your radar: H.D.R., or High Dynamic Range. Soon it’ll be everywhere, but here’s why it’s worth a look now.
Thursday, August 22, 2019
Newsroom: Netflix Losing US Share as Rivals Gain
August 21, 2019 at 07:01AM
Hulu will surpass 40% of paid OTT subscription users   August 21, 2019 – This year, 182.5 million people in the US will view content via subscription over-the-top (OTT) services, […]
YouTube Disables 210 Channels That Spread Disinformation About Hong Kong Protests
August 23, 2019 at 12:58AM
Days after Facebook and Twitter cracked down on China-backed accounts that were spreading such disinformation, YouTube did the same.
Voices in AI – Episode 94: A Conversation with Amy Webb
August 22, 2019 at 03:00PM
[voices_in_ai_byline]
About this Episode
Episode 94 of Voices in AI features Byron speaking with fellow futurist and author Amy Webb on the nature of artificial intelligence and the morality and ethics tied to its study.
Listen to this episode or read the full transcript at www.VoicesinAI.com
Transcript Excerpt
Byron Reese: This is Voices in AI brought to you by Gigaom, and I’m Byron Reese. Today, I’m so excited. My guest is Amy Webb. She is a quantitative futurist. She is the founder and CEO of the Future Today Institute. She’s the professor of strategic foresight at NYU’s Stern School of Business.
She’s a co-founder of Spark Camp. She holds a BS in game theory and economics, and an MS in journalism from Columbia, and she’s a Nieman Fellow at Harvard University. And she- as if all of that weren’t enough, she’s the author of a new book called The Big Nine: How the Tech Titans and Their Thinking Machines Could Warp Humanity. Welcome to the show, Amy.
Amy Webb: Hello. Thank you for having me.
So, I always like to start off with definitions. Maybe that can be tedious, but start me off with how do you define intelligence?
Well, I think intelligence in different context means different things. As it relates to artificial intelligence, I think, generally, what we’re trying to describe is the ability for a system to make decisions and choices that would mirror or come close to the way that we make decisions and choices. So, I think, when the term Artificial Intelligence was originally coined at Duke University in the 1950s by Marvin Minsky and John McCarthy, back then the thinking was to try and build a machine that came close to human cognition and whose intelligence could be measured against our own. Five decades later, I think it’s clear that, while we’re able to build incredibly powerful machines, there’s a lot that we don’t understand about how the–a lot we don’t understand about our own intelligence and our own cognition, and I think, Minsky and McCarthy, if they were around today still working very intently, that they would say that intelligence, AI was probably the wrong term to use.
Actually, McCarthy did say that pretty quickly after he coined the term. He said he thought it set the bar too high. But I guess the question I really want to ask is, in what sense is it artificial? Do you think AI is actually intelligent, or do you think it’s something that mimics intelligence the way that, say, artificial turf mimics grass?
Well it’s certainly been billed to mimic what we think we know about human intelligence. And that extends from hardware architecture. The current deep neural net system is a layered approach to machine learning that is built to physically mimic what we think the sort of synapses in our own brains–how they’re transferring and parsing and understanding information. That being said, I think that–I also think that intelligence is probably the wrong term because it’s fairly loaded, and when we talk about intelligence, we tend to–I think we talk about intelligence because we’re trying to quantify in some way our own cognitive power. I don’t know if that’s the right term for AI because I think AI inherently is different. Just because we’ve built it to mimic what we think we do doesn’t mean that from here out it continues to behave in that manner.
Yeah, it’s funny. I’ve always thought all of the analogies between how AIs function, how the brain functions are actually just kind of techno marketing. The brain is still largely a black box, right? We don’t know how even a thought is encoded or anything like that, and so, I’ve often suspected that’s just–that machine intelligence really doesn’t have anything in common with human intelligence. Would you agree, disagree, defer?
Yeah, I mean, I think that some–we’re talking in broad strokes and generalities. So, I think that there are certainly some areas of the AI ecosystem that do a fairly good job of mimicking what we do, so reinforcement learning and hierarchical reinforcement learning is present in dogs and children and adults. So basically, you’ve got a, say, toddler, and you’re trying to teach the toddler correct and incorrect, right? So the correct- I don’t know. The correct word is mommy for your mom. The correct term for the color that is the color of the sky is blue, stuff like that. And through praise or correction, the child over time learns. So we are reinforcing the correct answer, and we are, hopefully, in a very gentle loving way punishing the–guiding the child away from the wrong answer. So, the same is true in how some AI systems learn. And in fact, for anybody who’s been following AlphaGo and all of the work that the DeepMind team has been doing- a lot of what they’re doing is self-improving. It was reinforcement learning and then self-improvement, so.
Listen to this episode or read the full transcript at www.VoicesinAI.com
[voices_in_ai_link_back]
Byron explores issues around artificial intelligence and conscious computers in his new book The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity.
Soap, Detergent and Even Laxatives Could Turbocharge a Battery Alternative
August 22, 2019 at 06:36PM
Researchers are trying to develop options to lithium-ion and other batteries in a quest for quick bursts of power and extended energy storage.
Facebook’s New Tool Lets You See Which Apps and Websites Tracked You
August 21, 2019 at 05:22AM
The tool is a response to criticism that Facebook has faced over how it safeguards its users’ privacy.
How Should Big Tech Be Reined In? Here Are 4 Prominent Ideas
August 21, 2019 at 05:20AM
Some critics of the industry want to see the businesses broken up. Others would like to strengthen regulation of Silicon Valley. And there are shades of gray.
Alibaba Postpones Hong Kong Listing as Protests Roil Markets
August 22, 2019 at 03:07PM
The Chinese e-commerce giant delayed plans to sell potentially billions of dollars worth of shares in the market, which has fallen as tensions have worsened.
Voices in AI – Episode 94: A Conversation with Amy Webb
August 22, 2019 at 03:00PM
[voices_in_ai_byline]
About this Episode
Episode 94 of Voices in AI features Byron speaking with fellow futurist and author Amy Webb on the nature of artificial intelligence and the morality and ethics tied to its study.
Listen to this episode or read the full transcript at www.VoicesinAI.com
Transcript Excerpt
Byron Reese: This is Voices in AI brought to you by Gigaom, and I’m Byron Reese. Today, I’m so excited. My guest is Amy Webb. She is a quantitative futurist. She is the founder and CEO of the Future Today Institute. She’s the professor of strategic foresight at NYU’s Stern School of Business.
She’s a co-founder of Spark Camp. She holds a BS in game theory and economics, and an MS in journalism from Columbia, and she’s a Nieman Fellow at Harvard University. And she- as if all of that weren’t enough, she’s the author of a new book called The Big Nine: How the Tech Titans and Their Thinking Machines Could Warp Humanity. Welcome to the show, Amy.
Amy Webb: Hello. Thank you for having me.
So, I always like to start off with definitions. Maybe that can be tedious, but start me off with how do you define intelligence?
Well, I think intelligence in different context means different things. As it relates to artificial intelligence, I think, generally, what we’re trying to describe is the ability for a system to make decisions and choices that would mirror or come close to the way that we make decisions and choices. So, I think, when the term Artificial Intelligence was originally coined at Duke University in the 1950s by Marvin Minsky and John McCarthy, back then the thinking was to try and build a machine that came close to human cognition and whose intelligence could be measured against our own. Five decades later, I think it’s clear that, while we’re able to build incredibly powerful machines, there’s a lot that we don’t understand about how the–a lot we don’t understand about our own intelligence and our own cognition, and I think, Minsky and McCarthy, if they were around today still working very intently, that they would say that intelligence, AI was probably the wrong term to use.
Actually, McCarthy did say that pretty quickly after he coined the term. He said he thought it set the bar too high. But I guess the question I really want to ask is, in what sense is it artificial? Do you think AI is actually intelligent, or do you think it’s something that mimics intelligence the way that, say, artificial turf mimics grass?
Well it’s certainly been billed to mimic what we think we know about human intelligence. And that extends from hardware architecture. The current deep neural net system is a layered approach to machine learning that is built to physically mimic what we think the sort of synapses in our own brains–how they’re transferring and parsing and understanding information. That being said, I think that–I also think that intelligence is probably the wrong term because it’s fairly loaded, and when we talk about intelligence, we tend to–I think we talk about intelligence because we’re trying to quantify in some way our own cognitive power. I don’t know if that’s the right term for AI because I think AI inherently is different. Just because we’ve built it to mimic what we think we do doesn’t mean that from here out it continues to behave in that manner.
Yeah, it’s funny. I’ve always thought all of the analogies between how AIs function, how the brain functions are actually just kind of techno marketing. The brain is still largely a black box, right? We don’t know how even a thought is encoded or anything like that, and so, I’ve often suspected that’s just–that machine intelligence really doesn’t have anything in common with human intelligence. Would you agree, disagree, defer?
Yeah, I mean, I think that some–we’re talking in broad strokes and generalities. So, I think that there are certainly some areas of the AI ecosystem that do a fairly good job of mimicking what we do, so reinforcement learning and hierarchical reinforcement learning is present in dogs and children and adults. So basically, you’ve got a, say, toddler, and you’re trying to teach the toddler correct and incorrect, right? So the correct- I don’t know. The correct word is mommy for your mom. The correct term for the color that is the color of the sky is blue, stuff like that. And through praise or correction, the child over time learns. So we are reinforcing the correct answer, and we are, hopefully, in a very gentle loving way punishing the–guiding the child away from the wrong answer. So, the same is true in how some AI systems learn. And in fact, for anybody who’s been following AlphaGo and all of the work that the DeepMind team has been doing- a lot of what they’re doing is self-improving. It was reinforcement learning and then self-improvement, so.
Listen to this episode or read the full transcript at www.VoicesinAI.com
[voices_in_ai_link_back]
Byron explores issues around artificial intelligence and conscious computers in his new book The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity.
Newsroom: Netflix Losing US Share as Rivals Gain
August 21, 2019 at 07:01AM
Hulu will surpass 40% of paid OTT subscription users   August 21, 2019 – This year, 182.5 million people in the US will view content via subscription over-the-top (OTT) services, […]
Inside India’s Messy Electric Vehicle Revolution
August 22, 2019 at 01:13PM
A million electric rickshaws sprang up out of nowhere and are now being used by 60 million people a day. The government and vehicle makers are struggling to catch up.
A Business Built on Disinformation: Highlights From Our Investigation
August 22, 2019 at 10:08AM
Driven by Facebook, The Western Journal has had one of the biggest audiences in the news media, but little is known about the publication. Here’s what we found out.
Wednesday, August 21, 2019
Instagram’s Chain-Letter Uprising
August 22, 2019 at 12:28AM
Hordes of confused Instagrammers are trying to declare sovereignty from Instagram. Do they accidentally have a point?
Newsroom: Netflix Losing US Share as Rivals Gain
August 21, 2019 at 07:01AM
Hulu will surpass 40% of paid OTT subscription users   August 21, 2019 – This year, 182.5 million people in the US will view content via subscription over-the-top (OTT) services, […]
YouTuber Finds Lost GoPro Video of Man Who Drowned and Returns It to His Family
August 21, 2019 at 09:46PM
Rich Abernathy was filming a dive for YouTube when he found an old GoPro partly buried in the muck. The camera had captured the last moments of Richard Lee Ragland III’s life.
How to Turn an iPhone Into a Work-Only Tool
August 21, 2019 at 09:28PM
To prevent distractions, Conor Dougherty, an economics writer, dumped social media and anything fun — even his browser — from his smartphone.
2 Prominent Academics to Cut Ties to M.I.T. Media Lab Over Epstein Link
August 21, 2019 at 06:50PM
The academics, Ethan Zuckerman and J. Nathan Matias, said they were protesting the institute’s connection to the financier, who was facing sex-crime charges when he killed himself this month.
Hi, Alexa. How Do I Stop You From Listening In On Me?
August 21, 2019 at 01:08PM
Humans help train artificial intelligence, so someone may be listening to what you tell Siri, Google Assistant or Alexa. Here’s what to do if you have a problem with that.
BI Consolidation Roundup: Who is best positioned to win after the recent M&A activity?
August 17, 2019 at 01:54AM
Read Andrew Brust’s Full Report
Report Summary: Business Intelligence Market Consolidation
What took place from April through June, 2019 was the most significant BI market consolidation in more than a decade. Sure, market events like the emergence of Tableau and self-service BI in general were significant on their own. But these were nonetheless gradual and discrete developments, part of an evolving ecosystem that welcomes new vendors, and new offerings from existing vendors, over periods of years. The recent consolidation contrasts markedly with such typical calm and orderly progression. With a slew of deals in a period of just 10 weeks, the market calculus has been transformed. To understand this better, we need to examine the deals.
- The race is on, in the white-hot areas of cloud, analytics, and AI. The recent BI market consolidation demonstrates this and it’s the reason even more mergers are likely to come.
- The bigger mergers beg questions around the parties’ compatibility, in terms of corporate culture, messaging, cloud orientation, and, of course, technology. Existing customers beware.
- The BI consolidation of 2007-2008 sucked oxygen out of the then-acquired platforms and created opportunities for new ones to emerge. That M&A round provides lessons for vendors and cautions for customers this time around
- The fierce loyalty of Tableau customers, based in no small part on the company’s independence and personality, is at risk, as those customers watch their hero company’s vaunted brand subjugated to a new parent.’s
- Microsoft stands to gain from the churn and uncertainty, as the company’s stability offers sanctuary from the market upheaval.
- Disruption-averse customers may wish to re-platform, migrating to vendors whose stability is more assured.
- AI is the next frontier, both for innovation and M&A activity. And BI platforms are in position to bring AI to the mainstream.
How Many Triangles Are There? Here’s How to Solve the Puzzle
August 21, 2019 at 12:00PM
Counting will get you nowhere. Try a little combinatorics instead.
Netflix and Big Theater Chains Haggle Over Release of Scorsese’s ‘Irishman’
August 21, 2019 at 12:00PM
Will “The Irishman,” a mob-world epic starring Robert De Niro, come to a theater near you? After months of talks, there is still no resolution.
Where Did North West Learn How to Sew? At Camp, of Course.
August 21, 2019 at 12:00PM
In Hollywood, there’s a glamorous extracurricular program for fashion-savvy kids as young as 6.
Inside America’s Dysfunctional Trillion-Dollar Fighter-Jet Program
August 21, 2019 at 12:00PM
The F-35 was once the Pentagon’s high-profile problem child. Has it finally moved past its reputation of being an overhyped and underperforming warplane?
The Collective Memory of American Shoppers
August 21, 2019 at 12:00PM
On Facebook and Reddit, retail-focused groups remember the heyday of malls and share stories of shuttered department stores.
Tuesday, August 20, 2019
FogCam Is Signing Off in San Francisco
August 20, 2019 at 11:04PM
Long before streaming video, it captured images of campus life every 20 seconds. The quirky project, believed to be the longest-running public webcam, will shut down at the end of August.
‘I’m Sick of Seeing My Face,’ Says the Internet’s Kombucha Connoisseur
August 20, 2019 at 08:37PM
A woman’s taste test captured the fancy of the internet. Her rapid evolution of expressions may be why.
Welcome to McDonald’s. Would You Like a Podcast With Those Fries?
August 20, 2019 at 08:05PM
With podcasts rising in popularity, it’s no surprise that companies are producing their own. What’s more surprising is that people are actually listening to them.
How Should Big Tech Be Reined In? Here Are 4 Prominent Ideas
August 20, 2019 at 07:23PM
Some critics of the industry want to see the businesses broken up. Others would like to strengthen regulation of Silicon Valley. And there are shades of gray.
Facebook’s New Tool Lets You See Which Apps and Websites Tracked You
August 20, 2019 at 07:08PM
The tool is a response to criticism that Facebook has faced over how it safeguards its users’ privacy.
China’s Soft-Power Fail: Condemning Hong Kong’s Protests
August 20, 2019 at 04:55PM
Beijing wants greater sway over global public opinion. Instead, its propaganda outlets make Chinese leaders look like bullies.
In New Facebook Effort, Humans Will Help Curate Your News Stories
August 20, 2019 at 01:00PM
A news section inside Facebook’s mobile app will be run by journalists and not just algorithms, the company said.
Ransomware Attack Hits 23 Texas Towns, Authorities Say
August 20, 2019 at 12:00PM
The state declined to say which towns were affected by the coordinated cyberattack. But one expert said it could signal more such attacks in the future.
Monday, August 19, 2019
State Attorneys General Said to Be Near Formal Investigation of Tech Companies
August 20, 2019 at 03:50AM
Their bipartisan effort comes after a smaller group of the state representatives met last month with antitrust officials at the Justice Department.
Art Disappears in Private Hands. Can Social Media Resurface It?
August 16, 2019 at 12:24PM
Private art collections are notoriously secretive. A collective website aims to make them viewable by all.
Amazon Uses a Twitter Army of Employees to Fight Criticism of Warehouses
August 16, 2019 at 06:31AM
They love working there. It’s great. They love it. It’s great. They love it. It’s great. They love it. It’s great. They love it. It’s great. They love it.
Alibaba’s Strong Results Suggest Chinese Consumers Are Still Spending
August 15, 2019 at 11:22PM
Though growth has slowed, the e-commerce giant’s results signal that the country’s consumers still feel comfortable enough to open their wallets.
How Flat Earthers Nearly Derailed a Space Photo Book
August 15, 2019 at 06:58PM
What a photographer’s struggle to raise money for his book of images tells us about Facebook and conspiracy theorists.
Hong Kong’s Economy Shudders After Protests Plunge Airport Into Chaos
August 15, 2019 at 07:46AM
Clashes at one of the world’s busiest transportation hubs could cast further doubt on the city’s future as a business capital on China’s doorstep.
Capital One Hacking Suspect Had Data From Other Targets, Officials Say
August 15, 2019 at 03:07AM
Prosecutors said they found the stolen material from 30 organizations on servers in the bedroom of the accused hacker.
WeWork Takes Key Step Toward I.P.O., Citing Heady Growth and Huge Losses
August 15, 2019 at 12:15AM
A public offering by the company, which leases shared office space, will be a major test of investors’ appetite for fast-growing but unprofitable start-ups.
Newsroom: Facebook Losing Younger Users in Germany and France
August 14, 2019 at 07:01AM
August 14, 2019 – eMarketer has downgraded Facebook user growth in both France and Germany, due to declines among younger age groups as they shift their usage to other platforms. […]
Facebook and Twitter Say China Is Spreading Disinformation in Hong Kong
August 19, 2019 at 11:58PM
The social media companies removed accounts and said they were sowing divisive messages about the Hong Kong protests.
Ride Sharing Adds to the Crush of Traffic at Airports
August 19, 2019 at 11:52PM
The record number of travelers and the popularity of Uber and Lyft have left airports scrambling to ease congestion.
Alibaba’s Strong Results Suggest Chinese Consumers Are Still Spending
August 15, 2019 at 11:22PM
Though growth has slowed, the e-commerce giant’s results signal that the country’s consumers still feel comfortable enough to open their wallets.
U.S. Companies Working With Huawei Get More Time to Stop
August 19, 2019 at 05:05PM
A 90-day extension is intended to give rural telecommunications time to wean themselves off the Chinese company, which supplies parts and equipment.
To Power A.I., Start-Up Creates a Giant Computer Chip
August 19, 2019 at 03:30PM
The chip could improve how quickly artificial intelligence systems can learn tasks. But its complexity and size (as big as a plate) could be a challenge.
Paging Big Brother: In Amazon’s Bookstore, Orwell Gets a Rewrite
August 19, 2019 at 12:00PM
As fake and illegitimate texts proliferate online, books are becoming a form of misinformation. The author of “1984” would not be surprised.
Sunday, August 18, 2019
Terrorists Turn to Bitcoin for Funding, and They’re Learning Fast
August 18, 2019 at 09:12PM
The authorities have begun to raise alarms about a steady uptick in the number of militant groups using the hard-to-trace digital currency.
G.E. Is Accused of Major Fraud, and Stocks Take a Beating
August 18, 2019 at 02:00PM
Plus, WeWork takes steps to go public, and investigations into the Jeffrey Epstein case continue.
Newsroom: Facebook Losing Younger Users in Germany and France
August 14, 2019 at 07:01AM
August 14, 2019 – eMarketer has downgraded Facebook user growth in both France and Germany, due to declines among younger age groups as they shift their usage to other platforms. […]
Saturday, August 17, 2019
Closing of Pacific Standard and Topic Shows Perils of Depending on a Rich Patron
August 17, 2019 at 09:10PM
With the loss of the two award-winning publications, digital media got a little less brainy in the summer of ’19.
Friday, August 16, 2019
BI Consolidation Roundup: Who is best positioned to win after the recent M&A activity?
August 17, 2019 at 01:54AM
Read Andrew Brust’s Full Report
Report Summary: Business Intelligence Market Consolidation
What took place from April through June, 2019 was the most significant BI market consolidation in more than a decade. Sure, market events like the emergence of Tableau and self-service BI in general were significant on their own. But these were nonetheless gradual and discrete developments, part of an evolving ecosystem that welcomes new vendors, and new offerings from existing vendors, over periods of years. The recent consolidation contrasts markedly with such typical calm and orderly progression. With a slew of deals in a period of just 10 weeks, the market calculus has been transformed. To understand this better, we need to examine the deals.
- The race is on, in the white-hot areas of cloud, analytics, and AI. The recent BI market consolidation demonstrates this and it’s the reason even more mergers are likely to come.
- The bigger mergers beg questions around the parties’ compatibility, in terms of corporate culture, messaging, cloud orientation, and, of course, technology. Existing customers beware.
- The BI consolidation of 2007-2008 sucked oxygen out of the then-acquired platforms and created opportunities for new ones to emerge. That M&A round provides lessons for vendors and cautions for customers this time around
- The fierce loyalty of Tableau customers, based in no small part on the company’s independence and personality, is at risk, as those customers watch their hero company’s vaunted brand subjugated to a new parent.’s
- Microsoft stands to gain from the churn and uncertainty, as the company’s stability offers sanctuary from the market upheaval.
- Disruption-averse customers may wish to re-platform, migrating to vendors whose stability is more assured.
- AI is the next frontier, both for innovation and M&A activity. And BI platforms are in position to bring AI to the mainstream.
Older People Need Rides. Why Aren’t They Using Uber and Lyft?
August 16, 2019 at 11:36PM
Seniors need transportation alternatives more than ever, but many are intimidated by ride-hailing apps.
A.I. Is Learning From Humans. Many Humans.
August 16, 2019 at 09:11PM
Artificial intelligence is being taught by thousands of office workers around the world. It is not exactly futuristic work.
The Man Who Wrote the Guidebook for Civilian Bomb-Squad Technicians
August 16, 2019 at 05:36PM
At War talked to Tom Gersbeck, a retired Marine Corps bomb-disposal officer, who recently published the second edition of his book, “Practical Military Ordnance Identification.”
The Week in Tech: Are Lawmakers Too Eager to Weaken Big Tech’s Legal Shield?
August 16, 2019 at 04:00PM
Revising Section 230 could change the internet as we know it. How about policy decisions based on sound evidence?
Newsroom: Facebook Losing Younger Users in Germany and France
August 14, 2019 at 07:01AM
August 14, 2019 – eMarketer has downgraded Facebook user growth in both France and Germany, due to declines among younger age groups as they shift their usage to other platforms. […]
Newsroom: Facebook Losing Younger Users in Germany and France
August 14, 2019 at 07:01AM
August 14, 2019 – eMarketer has downgraded Facebook user growth in both France and Germany, due to declines among younger age groups as they shift their usage to other platforms. […]
Thursday, August 15, 2019
Amazon Uses a Twitter Army of Employees to Fight Criticism of Warehouses
August 15, 2019 at 10:48PM
They love working there. It’s great. They love it. It’s great. They love it. It’s great. They love it. It’s great. They love it. It’s great. They love it.
Alibaba’s Strong Results Suggest Chinese Consumers Are Still Spending
August 15, 2019 at 05:50PM
Though growth has slowed, the e-commerce giant’s results signal that the country’s consumers still feel comfortable enough to open their wallets.
I Shared My Phone Number. I Learned I Shouldn’t Have.
August 15, 2019 at 01:00PM
Our personal tech columnist asked security researchers what they could find out about him from just his cellphone number. Quite a lot, it turns out.
Newsroom: Facebook Losing Younger Users in Germany and France
August 14, 2019 at 07:01AM
August 14, 2019 – eMarketer has downgraded Facebook user growth in both France and Germany, due to declines among younger age groups as they shift their usage to other platforms. […]
Wednesday, August 14, 2019
Capital One Hacking Suspect Had Data From Other Targets, Officials Say
August 15, 2019 at 03:07AM
Prosecutors said they found the stolen material from 30 organizations on servers in the bedroom of the accused hacker.
India Shut Down Kashmir’s Internet Access. Now, ‘We Cannot Do Anything.’
August 15, 2019 at 12:25AM
Pharmacists can’t restock medicines; workers aren’t being paid. But the government still loves to block the internet for “peace and tranquillity.”
Art Disappears in Private Hands. Can Social Media Resurface It?
August 14, 2019 at 07:52PM
Private art collections are notoriously secretive. A collective website aims to make them viewable by all.
How Flat Earthers Nearly Derailed a Space Photo Book
August 14, 2019 at 06:58PM
What a photographer’s struggle to raise money for his book of images tells us about Facebook and conspiracy theorists.
Burner Apps and No Selfies When Reviewing Restaurants
August 14, 2019 at 04:00PM
It’s harder to remain incognito in the digital age, says Tejal Rao, our first California restaurant critic. She prefers dining tech that’s understated, too.
WeWork Takes Key Step Toward I.P.O., Citing Heady Growth and Huge Losses
August 14, 2019 at 05:17PM
A public offering by the firm, which leases shared office space, will be a major test of investors’ appetite for fast-growing but unprofitable start-ups.
Hong Kong’s Economy Shudders as Its Airport Descends into Chaos
August 14, 2019 at 03:22PM
Clashes at one of the world’s busiest transportation hubs could cast further doubt on the city’s future as a business capital on China’s doorstep.
The Rise of the Virtual Restaurant
August 14, 2019 at 01:20PM
Food delivery apps are reshaping the restaurant industry — and how we eat — by inspiring digital-only establishments that don’t need a dining room or waiters.
Voices in AI – Episode 93: A Conversation with Pieter Abbeel
August 08, 2019 at 03:00PM
[voices_in_ai_byline]
About this Episode
Episode 93 of Voices in AI features Byron speaking with Berkeley Robotic Learning Lab Director Pieter Abbeel about the nature of AI, the problems with creating intelligence and the forward trajectory of AI research.
Listen to this episode or read the full transcript at www.VoicesinAI.com
Transcript Excerpt
Byron Reese: This is voices in AI brought to you by GigaOm, I’m Byron Reese. Today I’m super excited we have Pieter Abbeel. He is a professor at UC Berkeley. He’s the president, founder and chief scientist the Covariant.ai. He is the founder of Gradescope. He holds an undergraduate degree in electrical engineering and an MS degree from Stanford, a PhD from Stanford in computer science and probably a whole lot more. This is gonna be an exciting half hour. Welcome to the show, Pieter.
Pieter Abbeel: Thanks for having me Byron.
There are all these concepts like life and death and intelligence that we don’t really have consensus definitions for. Why can’t we come up with a definition on what intelligence is?
Yeah it’s a good question. I feel like traditionally we think of intelligence as the things that computers can’t do yet. And then all of a sudden when we manage to do it, we understand how works and we don’t think of it as that intelligent anymore, right? It used to be okay, if we can make a computer play checkers that would make it intelligent, and then later we’re like, ‘Wait, that’s not enough to be intelligent, and we keep moving the bar,’ which is good to challenge ourselves, but yeah it’s hard to put something very precise on it.
Maybe the way I tend to think of it is that there are a few properties you really want to be true for something to be intelligent, and maybe the main one is the ability to adapt to new environments and achieve something meaningful in new environments that the system has never been in.
So I’m still really interested in this question of why we can’t define it. Maybe… you don’t have any thoughts on it, but my first reaction would be: if there’s a term you can’t define, maybe whatever it is doesn’t actually exist. It doesn’t exist; there’s no such thing, and that’s why you can’t define it. Is it possible that there’s no such thing as intelligence? Is it a useful concept in any way?
So I definitely think it’s a useful concept. I mean we definitely have certain metrics related to it that matter. I mean, if we think about it as like absolute, is it intelligent or not? Then it’s very hard. But I think we do have an understanding of what makes something more intelligent versus less intelligent. Even though we might not call this is an intelligence because it can play checkers, it’s still more intelligent when it’s able to play checkers than when it’s not. It’s still more intelligent, if let’s say, it can navigate an unknown building and find something in that building, than when it cannot. It’s more intelligent if it can acquire maybe the skill to play a new game it’s never seen before—you just present it with the rules and then figures out on its own how to play well. Which is essentially done by AlphaGo Zero, right? It was given the rules of the game but then just played itself to figure out how to play it maximally well. And so I think all of those things can definitely be seen as more intelligent if you can do them, than if you cannot do them.
So we have of course, narrow intelligence to use this construct which is an AI that we train to do one thing, and right now a technique we’re using that we’re having some success in is machine learning, a method which philosophically says “Let’s take data about the past and project it into the future.”
And then there’s this idea of a general intelligence which is somebody as versatile as you and [me], it’s what we see in the movies. Is it possible those two technologies have nothing in common at all? They share no code whatsoever, because there’s a vague sense that we get better and better at narrow and then it gets a little less narrow then you know it’s AlphaGo, then it’s AlphaGo Zero, then it’s AlphaGo Zero Plus and eventually it’s how? But is it possible they aren’t even related at all?
That’s a good question. I think, the thing about more specialized systems whether it’s in let’s say learning to play games or a robot learning to manipulate objects, which we do a lot of at Berkeley. I think often what we can get to succeed today tends to be somewhat narrow. And if a neural net was trained to play Go, that’s what it does; if it was trained to stack Lego blocks, that’s what it does. But I think at the same time, the techniques we tend to work on and by we I mean not just me and my students, but the entire community, we try to work on techniques where we have a sense that it would be more generally applicable than the domain we’re currently being able to achieve success in.
So for example we look at reinforcement learning and the underlying principles. We could look at individual successes which is where a neural net was trained through reinforcement learning for a very specific success, and of course neural nets are very specific to those domains, whereas games or robotics or another domain and within those domains [are] very specific and like the game of Go or Lego blocks stacking or peg insertion and so forth.
But I think the beauty still is that these ideas are quite general in that the same algorithm can then be run again, and the same algorithm can be run again to have a robot learn to maybe clean up a table, and so I think there is a level of generality ‘under the hood’ that’s doing the training of these neural nets even if the resulting neural net often ends up being a little specialized.
However you know I just heard an interview you gave where you were talking about the case that if you gave a narrow AI a bunch of data about planetary motion, it could predict the next eclipse and the next one and the next million. But if all of a sudden a new moon appeared around Jupiter and you said “What’s that going to do to planetary motion?” it wouldn’t know because all it can do is take data about the past, make predictions about the future, and it isn’t that simple idea: take data about the past, make projections about the future not really the essence of what intelligence really is about?
Yeah. So what you’re getting at here is, to be fair, it’s not something that humans figured out very easily either. I mean it’s only when Newton came about [that] we started as humanity to understand that there is this thing called gravity and it has laws and it governs how planets and stars and so forth move around in space. And so it’s one of those things where, definitely right now I suspect if we just gave a massive neural network (without putting any prior information in there about what we already learned about how the world works), a bunch of data about planetary motion, it’s not very likely it would discover that.
I think it’s not unreasonable that that’s hard to do because I mean humans didn’t discover it ‘til very late either in terms of time of our civilization and it took a very kind of exceptional person at that time to figure it out. But I do think that those are the kind of things that are good motivators for the work we do because since sometimes what it points out to is, something that’s called Occam’s Razor, which says that the simplest explanation of the data is often the one that will generalize the best. Of course ‘define simple’ is not easy to do, but there is a general notion that the [fewer] equations you might need, the [fewer] variables might be involved, the simpler the explanation and so the more likely it would generalize to new situations.
And so I think laws of physics are kind of extreme. A really nice example of coming up with very very simple, low dimensional description of a very large range of phenomenons. Then yes. I don’t think neural nets have done that yet. I mean of course there’s work going in that direction, but often people will build in the assumptions and say “Oh it does better when it has the assumptions built in.” That’s not a bad thing to solve one problem but it’s not necessarily the way you have intelligence emerge in the sense that we might want it to emerge.
Listen to this episode or read the full transcript at www.VoicesinAI.com
[voices_in_ai_link_back]
Byron explores issues around artificial intelligence and conscious computers in his new book The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity.
Tuesday, August 13, 2019
Voices in AI – Episode 93: A Conversation with Pieter Abbeel
August 08, 2019 at 03:00PM
[voices_in_ai_byline]
About this Episode
Episode 93 of Voices in AI features Byron speaking with Berkeley Robotic Learning Lab Director Pieter Abbeel about the nature of AI, the problems with creating intelligence and the forward trajectory of AI research.
Listen to this episode or read the full transcript at www.VoicesinAI.com
Transcript Excerpt
Byron Reese: This is voices in AI brought to you by GigaOm, I’m Byron Reese. Today I’m super excited we have Pieter Abbeel. He is a professor at UC Berkeley. He’s the president, founder and chief scientist the Covariant.ai. He is the founder of Gradescope. He holds an undergraduate degree in electrical engineering and an MS degree from Stanford, a PhD from Stanford in computer science and probably a whole lot more. This is gonna be an exciting half hour. Welcome to the show, Pieter.
Pieter Abbeel: Thanks for having me Byron.
There are all these concepts like life and death and intelligence that we don’t really have consensus definitions for. Why can’t we come up with a definition on what intelligence is?
Yeah it’s a good question. I feel like traditionally we think of intelligence as the things that computers can’t do yet. And then all of a sudden when we manage to do it, we understand how works and we don’t think of it as that intelligent anymore, right? It used to be okay, if we can make a computer play checkers that would make it intelligent, and then later we’re like, ‘Wait, that’s not enough to be intelligent, and we keep moving the bar,’ which is good to challenge ourselves, but yeah it’s hard to put something very precise on it.
Maybe the way I tend to think of it is that there are a few properties you really want to be true for something to be intelligent, and maybe the main one is the ability to adapt to new environments and achieve something meaningful in new environments that the system has never been in.
So I’m still really interested in this question of why we can’t define it. Maybe… you don’t have any thoughts on it, but my first reaction would be: if there’s a term you can’t define, maybe whatever it is doesn’t actually exist. It doesn’t exist; there’s no such thing, and that’s why you can’t define it. Is it possible that there’s no such thing as intelligence? Is it a useful concept in any way?
So I definitely think it’s a useful concept. I mean we definitely have certain metrics related to it that matter. I mean, if we think about it as like absolute, is it intelligent or not? Then it’s very hard. But I think we do have an understanding of what makes something more intelligent versus less intelligent. Even though we might not call this is an intelligence because it can play checkers, it’s still more intelligent when it’s able to play checkers than when it’s not. It’s still more intelligent, if let’s say, it can navigate an unknown building and find something in that building, than when it cannot. It’s more intelligent if it can acquire maybe the skill to play a new game it’s never seen before—you just present it with the rules and then figures out on its own how to play well. Which is essentially done by AlphaGo Zero, right? It was given the rules of the game but then just played itself to figure out how to play it maximally well. And so I think all of those things can definitely be seen as more intelligent if you can do them, than if you cannot do them.
So we have of course, narrow intelligence to use this construct which is an AI that we train to do one thing, and right now a technique we’re using that we’re having some success in is machine learning, a method which philosophically says “Let’s take data about the past and project it into the future.”
And then there’s this idea of a general intelligence which is somebody as versatile as you and [me], it’s what we see in the movies. Is it possible those two technologies have nothing in common at all? They share no code whatsoever, because there’s a vague sense that we get better and better at narrow and then it gets a little less narrow then you know it’s AlphaGo, then it’s AlphaGo Zero, then it’s AlphaGo Zero Plus and eventually it’s how? But is it possible they aren’t even related at all?
That’s a good question. I think, the thing about more specialized systems whether it’s in let’s say learning to play games or a robot learning to manipulate objects, which we do a lot of at Berkeley. I think often what we can get to succeed today tends to be somewhat narrow. And if a neural net was trained to play Go, that’s what it does; if it was trained to stack Lego blocks, that’s what it does. But I think at the same time, the techniques we tend to work on and by we I mean not just me and my students, but the entire community, we try to work on techniques where we have a sense that it would be more generally applicable than the domain we’re currently being able to achieve success in.
So for example we look at reinforcement learning and the underlying principles. We could look at individual successes which is where a neural net was trained through reinforcement learning for a very specific success, and of course neural nets are very specific to those domains, whereas games or robotics or another domain and within those domains [are] very specific and like the game of Go or Lego blocks stacking or peg insertion and so forth.
But I think the beauty still is that these ideas are quite general in that the same algorithm can then be run again, and the same algorithm can be run again to have a robot learn to maybe clean up a table, and so I think there is a level of generality ‘under the hood’ that’s doing the training of these neural nets even if the resulting neural net often ends up being a little specialized.
However you know I just heard an interview you gave where you were talking about the case that if you gave a narrow AI a bunch of data about planetary motion, it could predict the next eclipse and the next one and the next million. But if all of a sudden a new moon appeared around Jupiter and you said “What’s that going to do to planetary motion?” it wouldn’t know because all it can do is take data about the past, make predictions about the future, and it isn’t that simple idea: take data about the past, make projections about the future not really the essence of what intelligence really is about?
Yeah. So what you’re getting at here is, to be fair, it’s not something that humans figured out very easily either. I mean it’s only when Newton came about [that] we started as humanity to understand that there is this thing called gravity and it has laws and it governs how planets and stars and so forth move around in space. And so it’s one of those things where, definitely right now I suspect if we just gave a massive neural network (without putting any prior information in there about what we already learned about how the world works), a bunch of data about planetary motion, it’s not very likely it would discover that.
I think it’s not unreasonable that that’s hard to do because I mean humans didn’t discover it ‘til very late either in terms of time of our civilization and it took a very kind of exceptional person at that time to figure it out. But I do think that those are the kind of things that are good motivators for the work we do because since sometimes what it points out to is, something that’s called Occam’s Razor, which says that the simplest explanation of the data is often the one that will generalize the best. Of course ‘define simple’ is not easy to do, but there is a general notion that the [fewer] equations you might need, the [fewer] variables might be involved, the simpler the explanation and so the more likely it would generalize to new situations.
And so I think laws of physics are kind of extreme. A really nice example of coming up with very very simple, low dimensional description of a very large range of phenomenons. Then yes. I don’t think neural nets have done that yet. I mean of course there’s work going in that direction, but often people will build in the assumptions and say “Oh it does better when it has the assumptions built in.” That’s not a bad thing to solve one problem but it’s not necessarily the way you have intelligence emerge in the sense that we might want it to emerge.
Listen to this episode or read the full transcript at www.VoicesinAI.com
[voices_in_ai_link_back]
Byron explores issues around artificial intelligence and conscious computers in his new book The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity.
Blog Archive
-
▼
2019
(4141)
-
▼
August
(178)
- The Mysterious Vaping Illness That’s ‘Becoming an ...
- YouTube Said to Be Fined Up to $200 Million for Ch...
- Twitter C.E.O. Jack Dorsey’s Account Hacked
- It’s Not Easter, but There Might Be a Surprise Hid...
- How Tainted Gold May Have Ended Up in Your Phone
- Don’t Use Bootleg or Street Vaping Products, C.D.C...
- The Week in Tech: Are You Ready for Facebook’s Fut...
- What Does VSCO Think About the ‘VSCO Girls’?
- Hospitals Should Replace Infection-Prone Scopes Wi...
- Uber and Lyft to Spend $60 Million to Maintain Cal...
- Organoids Are Not Brains. How Are They Making Brai...
- India’s Restaurants Rebel Against Food Delivery Apps
- Those People Starting Successful Tech Companies? M...
- Voices in AI – Episode 94: A Conversation with Amy...
- 6 Google Tricks That Will Turn You Into an Interne...
- Peloton Is a Phenomenon. Can It Last?
- When Children Use Technology, Let Common Sense Pre...
- The Joys of Being a Late Tech Adopter
- Facebook Tightens Rules on Verifying Political Adv...
- 6 Google Tricks That Will Turn You Into an Interne...
- Philip Morris and Altria Are in Talks to Merge
- Telegram Pushes Ahead With Plans for ‘Gram’ Crypto...
- Former Star Google and Uber Engineer Charged With ...
- The Baroness Fighting to Protect Children Online
- Netflix Falls Short in Big Screen Debut of ‘The Ir...
- Is It Time to Upend the Periodic Table?
- How to Turn an iPhone Into a Work-Only Tool
- Cadillac’s Last Stand? Storied Brand Aims (Again) ...
- Hi, Alexa. How Do I Stop You From Listening In On Me?
- Lawsuit Over Computer Chips Invokes Trade War With...
- Automating HIPAA Compliant Hosting By Using DevOps
- Voices in AI – Episode 94: A Conversation with Amy...
- Google Tries to Corral Its Staff After Ugly Intern...
- Qualcomm Wins Reprieve in F.T.C. Antitrust Case Wi...
- As Amazon Fires Spread, So Do the Misleading Photos
- Facebook Bans Ads From The Epoch Times
- How Uber Got Lost
- The Week in Tech: Facebook’s First Step Toward Tre...
- Snapchat’s Disappearing Act Leaves Venice Beach Se...
- H.D.R. Is Coming to a TV Near You. Here’s Why You ...
- Newsroom: Netflix Losing US Share as Rivals Gain
- YouTube Disables 210 Channels That Spread Disinfor...
- Voices in AI – Episode 94: A Conversation with Amy...
- Soap, Detergent and Even Laxatives Could Turbochar...
- Facebook’s New Tool Lets You See Which Apps and We...
- How Should Big Tech Be Reined In? Here Are 4 Promi...
- Alibaba Postpones Hong Kong Listing as Protests Ro...
- Voices in AI – Episode 94: A Conversation with Amy...
- Newsroom: Netflix Losing US Share as Rivals Gain
- Inside India’s Messy Electric Vehicle Revolution
- A Business Built on Disinformation: Highlights Fro...
- Instagram’s Chain-Letter Uprising
- Newsroom: Netflix Losing US Share as Rivals Gain
- YouTuber Finds Lost GoPro Video of Man Who Drowned...
- How to Turn an iPhone Into a Work-Only Tool
- 2 Prominent Academics to Cut Ties to M.I.T. Media ...
- Hi, Alexa. How Do I Stop You From Listening In On Me?
- BI Consolidation Roundup: Who is best positioned t...
- How Many Triangles Are There? Here’s How to Solve ...
- Netflix and Big Theater Chains Haggle Over Release...
- Where Did North West Learn How to Sew? At Camp, of...
- Inside America’s Dysfunctional Trillion-Dollar Fig...
- The Collective Memory of American Shoppers
- FogCam Is Signing Off in San Francisco
- ‘I’m Sick of Seeing My Face,’ Says the Internet’s ...
- Welcome to McDonald’s. Would You Like a Podcast Wi...
- How Should Big Tech Be Reined In? Here Are 4 Promi...
- Facebook’s New Tool Lets You See Which Apps and We...
- China’s Soft-Power Fail: Condemning Hong Kong’s Pr...
- In New Facebook Effort, Humans Will Help Curate Yo...
- Ransomware Attack Hits 23 Texas Towns, Authorities...
- State Attorneys General Said to Be Near Formal Inv...
- Art Disappears in Private Hands. Can Social Media ...
- Amazon Uses a Twitter Army of Employees to Fight C...
- Alibaba’s Strong Results Suggest Chinese Consumers...
- How Flat Earthers Nearly Derailed a Space Photo Book
- Hong Kong’s Economy Shudders After Protests Plunge...
- Capital One Hacking Suspect Had Data From Other Ta...
- WeWork Takes Key Step Toward I.P.O., Citing Heady ...
- Newsroom: Facebook Losing Younger Users in Germany...
- Facebook and Twitter Say China Is Spreading Disinf...
- Ride Sharing Adds to the Crush of Traffic at Airports
- Alibaba’s Strong Results Suggest Chinese Consumers...
- U.S. Companies Working With Huawei Get More Time t...
- To Power A.I., Start-Up Creates a Giant Computer Chip
- Paging Big Brother: In Amazon’s Bookstore, Orwell ...
- Terrorists Turn to Bitcoin for Funding, and They’r...
- G.E. Is Accused of Major Fraud, and Stocks Take a ...
- Newsroom: Facebook Losing Younger Users in Germany...
- Closing of Pacific Standard and Topic Shows Perils...
- BI Consolidation Roundup: Who is best positioned t...
- Older People Need Rides. Why Aren’t They Using Ube...
- A.I. Is Learning From Humans. Many Humans.
- The Man Who Wrote the Guidebook for Civilian Bomb-...
- The Week in Tech: Are Lawmakers Too Eager to Weake...
- Newsroom: Facebook Losing Younger Users in Germany...
- Newsroom: Facebook Losing Younger Users in Germany...
- Amazon Uses a Twitter Army of Employees to Fight C...
- Alibaba’s Strong Results Suggest Chinese Consumers...
- I Shared My Phone Number. I Learned I Shouldn’t Have.
-
▼
August
(178)