Q&A with Vint Cerf, father of the internet
On September 13, the Baker Institute hosted Dr. Vint Cerf, vice president and chief internet evangelist at Google, for its annual Civic Scientist Lecture Series. The lectures, an initiative sponsored by the Baker Institute Science and Technology Policy Program, host leading scientists from around the country who have impacted public policy. The goal of the series is to expose scientists and future scientists to the notion that they can have an impact outside the laboratory.
Dr. Cerf’s lecture, titled “The Future of the Internet,” discussed the risks and opportunities of the digital age and the acceleration of technological innovation by placing the present in a historical context. Speaking with firsthand experience in building the internet from its inception, Dr. Cerf outlined continuing challenges of keeping it secure and equitable for users across the globe.
The lecture included twenty minutes for questions and answers. However, that proved to be far too short a time for the curiosity of the audience. Dr. Cerf generously agreed to answer additional written questions submitted from audience members.
1. How does the role of big data and analytics of that data play a role for the future as data creation and collection accelerates?
Plainly the quantity of information we collectively generate is increasing thanks to online tools for its manual and automatic generation and collection. Analytics help us catalog, curate, assess and understand data and information in general. Douglas Engelbart and J.C.R. Licklider both predicted that computers would eventually become partners in the processing of information and that they would facilitate cooperative and collaborative group knowledge work. They will be tools we use to handle increasingly large quantities of information. Google’s search engine is just one example as are the collaborative Google “docs” that allow multiple parties to interact through them concurrently.
2. What one book would you recommend?
John L. Hennessy: Leading Matters, 2018.
3. Is Blockchain the next Internet?
NO. Ok, I will elaborate. It’s a distributed ledger whose entries are intended to be immutable. It is not necessarily a fast process; the participating computers have to select one among them to sign the next block. The rate at which transactions can be entered into the chain may thus be limited. There are ways to cope with this using parallelism, but it isn’t clear that the blockchain is necessarily superior to more conventional, replicated, digitally-signed databases. The anonymous blockchain version makes be uncomfortable. I like to know who I am depending on, so I prefer the “permissioned blockchain” version in which the participants are known, if I have to use blockchain at all.
4. What’s the prospect for Quantum communications Internet?
The transmission of single, entangled photons that preserve their quantum state is increasingly feasible. So far this method has only been used to transmit keying material in such a way that eavesdropping is detectable. The transmission of entangled photons with which to continue a quantum computation has not been achieved as far as I am aware. It is not immediately clear how this would be a benefit — for example, one could perform a quantum computation, obtain a binary result, send that result by conventional means, then set up a next quantum computation using these results.
5. What about the future of Net Neutrality and the throttling of the Internet as a function of $$$? Thanks.
Under the present US FCC rules, there is no basis for FCC intervention in the event of anti-competitive behavior by Internet Access Providers. Internet is treated as an unregulated information service. Actions that erode user choices appear not to be prosecutable, at least not by the FCC. Ideally a new title for the Telecommunications Act should be enacted that re-introduces an enforceable set of rules the limit anti-competitive behaviors. Whether this Congress or a later one is prepared to do that is unclear.
6. What are your ideas on how to incentivize behavioral changes to ensure that people are making safe choices when it comes to technology, including the Internet?
Tough problem. Early exposure to the weaknesses of high technology, the risks and hazards (such as phishing, malware, spoofing) seems essential. We teach kids to look both ways before crossing the street. We make them take driver training classes before they can apply for a license. We need to teach critical thinking to deal with “fake news” and misinformation. We need to provide people with tools to protect privacy, to support strong personal authentication (to avoid account hijacking), two factor authentication devices, etc.
7. What advice would you give students interested in pursuing Internet-related research or careers?
Internet is all about software, computing and communications. It is about information sharing, multi-party business models, easy access to communication services. There is a tidal wave of new applications coming and an Internet of Things with devices in the billions. Studying software reliability, safety, security, privacy and ethics are all topics worthy of attention. Better software development ecosystems to limit mistakes would be very welcomed.
8. Is enough being done to stop Russian’s interference?
No. We need a much more active and collaborative effort across domestic and international law enforcement, intelligence gathering and sharing, improved defensive technology, better surveillance of our networked systems, increased attention to infrastructure defenses… the list is long.
9. Why is India larger than the USA in IT?
The simplest answer is that they graduate 4X as many engineers as we do because their population is 4X that of the US. The Indian Government has made a point of encouraging engineering studies and has actively promoted outsourcing which has been embraced by American business as a way to save costs. However, India is not as heavily invested in Internet and online operations relative to the US. The main competitor there is China.
10. Google has been in the news lately post the 2016 election with its leaders seemingly leaning LEFT. Thought?
As a matter of policy, Google has adopted a neutral stance with regard to politics. As individuals, Google employees are free to follow their political preferences but it is forbidden to introduce these into Google policy and products. Like other platform providers, Google is under significant pressure to filter content — consider the European Right to be Forgotten requiring web index filtering as well as the General Data Protection Regulation (GDPR). As is evident, extreme content is being removed from social networks and the US Congress continues to press companies to act as censors.
11. The Internet to the average person is a lot more than the transport of bits/packets. What are your thoughts about the repeal of Net Neutrality, GDPR and the most recent EU decision on copyright?
I have referred already to NN and GDPR. Regarding copyright, it seems to me that those pressing for longer terms for copyright protection, criminalization of analysis of cryptographic systems for copyright protection, and increased erosion of the ability to reference copyright material will visit long term damage to the legitimate discovery and sharing of information and, in particular, the power of public domain accessibility after expiration of copyrights.
12. Clarify/quantify Net Neutrality and AI.
I have already responded to NN and the loss of ability to enforce its protections. Regarding AI, especially machine learning, this is a powerful but narrow technology. It can achieve amazing things in narrow fields such as Go and Chess. It has proven extremely useful in machine translation, in optimized operation of our data center cooling systems, image recognition and many other applications. I reject the extreme alarms about robotics and AI although I agree that we should be wary of relying on AI methods without human intervention. I see AI and ML as tools to be used to empower human ability to cope with information overload.
13. Address Bit Rot.
This is a big concern of mine. Digital information is recorded on media that may not last very long or for which readers may eventually no longer be available (think 5 ¼” floppy disks, 3 ½” floppy disks, DVDs, CDROMs, 9-track tapes…). Moreover, digital information often requires software to be correctly interpreted. Think about video games and spreadsheets. If the software doesn’t work, the preserved data may be uninterpretable. We may need to preserve hardware descriptions of computers, copies of old operating systems and copies of old application software to correctly interpret the digital information of the past in the future.
14. Tim Berners-Lee recently addressing the re-participation of equal Internet. What are his chances?
This is Tim’s effort to re-distribute the World Wide Web. He thinks that companies operating large scale data centers such as Amazon, Microsoft and Google have centralized the Web. He has developed a suite of software he calls SOLID. The reason that we have large companies is that there is an economy of scale in the operation of data centers. Depending on individuals to hold other people’s data reliably and in perpetuity may not be prove to be workable. I think it remains to be seen.
15. Do you have any theories of ideas for long-term data storage and libraries?
I think we will need to copy data from older media to newer ones and we will need to preserve software for future use with older content. I believe libraries and archives have a clear role to play but they must be supported so business models are needed to make it possible for them to serve in the future as they have in the past with earlier media. There are some very speculative ideas about using DNA to encode binary information: it’s slow to encode and searching might be difficult. I think we will continue to develop new media and methods for coding and indexing. AI and Machine Learning may have a role to play there.
16. What’s your attitude towards the Named-Data Networks? Will NDN replace the current network architecture?
I am still skeptical that NDN will prove to be a scalable, general purpose mechanism but I do think there are some applications for which the idea works. Data stored in the network (e.g. map information) could be shared on demand as needed to multiple parties needed access to it (subscribers in effect).
17. Internet is becoming more and more about a lot of personal data collected over time and accessible very easily in a consolidated fashion (Google Search). Who should own that data?
I think we should distinguish searching of the WWW (Google Search) from the “owner” of the data. If people put information up on the Internet/Web in publicly accessible form (think Facebook, Twitter, Instagram…) it isn’t clear how control should be applied. Digital information is easy to reproduce so it isn’t consumable in the usual sense of the term. Privacy is hard to protect in this online age. Think of all the photos that might contain your image that you didn’t put up on the net, someone else did (and then tagged it). We share a lot of data with others because they need it to provide services (delivery agencies need our addresses and phone numbers to do their job for instance). I think your question is very hard to answer and probably falls into the “it’s complicated” category.
18. Opinions on Google Project Zero?
This is a project to identify “zero day” vulnerabilities. See https://googleprojectzero.blogspot.com/. I am extremely pleased to see this level of candor from Google trying to improve the quality of software and to alert the community to potential hazards.
