Sunday, December 04, 2016

Donald Trump Jeopardizes Cyber Privacy And National Security

President-Elect Donald Trump recently released a video in which he promised to work with the Department of Defense and Joint Chiefs of Staff on a “plan to protect Americas’ vital infrastructure from cyber attacks.” This promise reflects Trump’s ignorance of how cyber warfare works — calling in the Marines to secure the nation’s computers is about as effective as exterminating cockroaches with a shotgun.

On the vast, interdependent internet, evolving technologies and best practices must be adopted across the ecosystem for anyone to be secure. An effective cyber defense requires long, hard years of continued investment in research, education, strong encryption, standards, regulations, enforcement, and global cooperation. Unfortunately, Trump’s stated policy goals promise to halt and even reverse the hard-fought progress made in recent years defining and enforcing new cyber standards. The impact on national security will be dire.


Furthermore, Trump’s call to boycott Apple for refusing to break their iPhone encryption and his plan for “closing that Internet up” expose a disregard for cyber privacy and freedom of expression that threatens to undermine our rights and our prosperity.

Stop-and-Frisk in Cyberspace

The US is a cyber superpower, alongside China, England, Israel and Russia. While Edward Snowden’s revelations suggest that the U.S. likely harbors the most potent cyber weapons, the agencies that develop and wield them have a clear mandate to use them only on foreign targets — for example, to retaliate against Russia’s repeated pattern of cyber aggression.

To Trump, however, Vladimir Putin is a friend — the nation’s true enemies lurk within the American homeland: illegal Mexican immigrants, Muslim jihadist refugees, obstructive protesters, and conspiring journalists. Echoing Rudolph Giuliani, Trump has touted stop-and-frisk as a legitimate exercise of “law and order” so we should expect the same in cyberspace, as federal agencies redirect their formidable arsenals away from foreign and toward domestic surveillance. No wonder Peter Thiel supported and now advises Trump — his company Palantir sells the software used by intelligence agencies to monitor large populations; investors plowed another $20 million into the Palantir just last week.
Peter Thiel, co-founder of Palantir

Judicial and legislative oversight bodies normally protect US citizens from mass domestic surveillance. But Trump’s tweets and campaign rally warnings about ISIS have escalated American fear of the terrorist threat to the highest point since 9–11, when Congress passed the Patriot Act. The Republican Congress and Trump-appointed judges may give the President broad leeway.

The Danger of Deregulation

Preventing cyber attacks is impossible without regulation, because cyber neglect is like polluting, drunk driving, or refusing to vaccinate — it endangers not only the reckless, but everyone else as well. The security of every online transaction depends upon the integrity of all the vendors in the ecosystem who handle payments, network traffic, email delivery, cloud servers, and more. Furthermore, any infected computer or device can be used to attack others (as we saw in the October DDoS attack that caused massive internet outages). Without broad regulations and enforcement, internet commerce cannot be secured.

Donald Trump’s campaign speeches and web site have consistently promised to reduce the rules, headcount, and overall spending in the SEC, FTC, CFPB, FCC and IS Oversight Office — the very federal regulatory agencies that have taken the lead in defining and enforcing cyber standards. (His adviser Mark Jamison openly plans to nearly eliminate the FCC.) In addition to the budget savings, Trump sees this as a key element in his plan to promote business and increase jobs. By design, these cuts will relax the rules and enforcement of cyber standards for the public companies, banks, consumer-facing merchants, and network carriers that these agencies regulate. We should expect similar cuts in other regulatory authorities such as the Center for Medicare and Medicaid Services (which enforces HIPAA rules for the healthcare industry) and the Federal Energy Regulatory Commission (which oversees NERC standards for the power grid).

Cyber deregulation will empower American businesses to sell our data to anyone collecting profiles of US citizens. Meanwhile, with a U.S. president who actually invited and benefited from Russia’s intervention in the election, Russian cyber attackers feel they enjoy free rein in American cyberspace. With the rollback of cyber regulations, consumer-facing businesses will slash their own cyber security budgets, leading to weaker systems that further accelerate the growth and severity of information breaches. With our private information exposed, brace for a dramatic rise in identity theft and cyber stalking.

In contrast, the European Union has set the standard for privacy laws that limit how businesses and government agencies can use our information. Once disdained by the business community, these laws now give Europe the competitive advantage. In the wake of Snowden’s revelations, mistrustful Europeans moved their data from US clouds and services to EU alternatives — during Trump’s presidency, Americans will join them. While some Americans look to Switzerland as a safe haven for money, and Canada as a safe haven for our families, many will look to Germany as a safe haven for data.

Cyber 9–11

President Trump’s deregulatory policies will jeopardize not only privacy, but also national security. Our homeland’s greatest vulnerability may well be the cyber threat to our critical infrastructure, potentially disrupting life-support services like power and water. Furthermore, a single breach of a water treatment facility, dam, or nuclear reactor can directly kill millions of people — a cyber 9–11. And yet today most of the nation’s utilities run unpatched software on industrial control systems that remain defenseless, awaiting NERC cyber regulations to kick in next year. A four-year reprieve from these rules by Trump’s administration will expose the U.S. to a massive terrorist attack, and open the door for Russia or other nations to embed cyber bombs in our machinery for future activation. Even if the Defense Department can accurately attribute such attacks, they can only retaliate — they cannot prevent them.

The election of Donald Trump has profound implications for the security of cyberspace. Unless Trump reverses his positions on deregulation, government surveillance, and the Russian threat, his administration will dismantle the safeguards of cyberspace, threatening America’s commercial prosperity, individual privacy, and national security.

Tuesday, September 13, 2016

Investment Recommendation: Claroty Series A

Today, Claroty came out of stealth, announcing a Series A financing led by Bessemer. $32 Million is  is a lot for Series A, but this is an important company for our nation and our planet. To explain why, I thought I'd share this excerpt from our internal investment memo.

EXCERPT from APRIL 2016:

The Need for Industrial Security


The physical infrastructure of modern civilization runs on machinery: traffic lights, railroad switches, nuclear reactors, water treatment, electricity distribution, dams, ship engines, draw bridges, oil rigs, hospitals, gas pipelines, and factories depend upon mechanical elements such as pressure valves, turbines, motors, and pumps. These actuators (like the ones in the original Bessemer steel smelting process) were once manually configured, but today these machines are controlled by software running on directly-attached, single-purpose computers known as Programmable Logic Controllers (PLC). PLCs, in turn, are connected in aggregate to computers running Human Management Interfaces (HMI) through closed, vendor-proprietary Supervisory Control & Data Acquisition (SCADA) protocols like DNP3 and Profibus. Industrial manufacturers provide the machines, the PLCs, and the HMIs, and so Operations Technology (OT) teams typically need to use a mix of controllers and interfaces. This is collectively known as an ICS. 

During the PC revolution, many of these ICS components migrated to cheap, standard PCs, and their SCADA connections migrated to LAN switches and routers that leveraged the connectivity benefits of those PCs’ standard Ethernet ports. The security implications were relatively minor until the Internet came along; but now, if any computer in the building is connected to the Internet, all the machines are potentially exposed. ICS security had once depended upon an air-gap between IT and OT networks, and where absolutely necessary devices like one-way diodes were used to send data out of the OT network to the outside world. However, trends like remote management, cloud, IoT, and the adoption of open standards are eroding the network segmentation and creating new attack vectors.

The threat of ICS attacks is very different from threats plaguing other computer networks. First, there is little valuable data to steal from a PLC (with the theoretical exception of pharmaceuticals), and yet the consequences of an attack are potentially catastrophic; the worst doomsday scenarios of cyber warfare arise from compromised machinery such as gas relays, dams, reactors, and water treatment facilities that can kill millions of people when they malfunction. To get a taste of the kind of damage we’re talking about, watch this video from 2007, where members of the Idaho National Laboratory hacked some of its own machinery.

Second, the fear of unexpected downtime also makes OT teams less willing to experiment with new hardware and software updates. These factors create an environment of older computers running older software that is never patched despite the accumulation of known vulnerabilities.

Finally, OT teams will not run encryption or conventional cybersecurity software on their computers, lest the security processes interfere with the precise and fragile timing of their network; they would rather be infected than incur downtime. And evidence of infections is mounting:

      The Stuxnet worm, allegedly developed jointly by NSA and the Israeli Army’s intelligence arm (Unit 8200), crippled the Iranian nuclear program by destroying their centrifuges;
      Iran crippled the operations of the most valuable company on Earth, Saudi Aramco;
      According to BVP-funded iSIGHT Partners, the Russia-based Sandstone Team developed the Blackworm malware that shut down power for 700K Ukrainians;
      For two years, an Iranian group controlled malware inside a dam in Rye, New York (near BVP’s  Larchmont office).

The malware behind these attacks likely lay dormant for some time, and there is no comprehensive way to know how much more already lurks in critical ICS just waiting to be activated. According to the ICS-CERT, we discover more and more infections every year in US infrastructure.So, at a time when nation-states, terrorists, and criminal organizations are scrambling for an advantage in cyberspace, society’s most critical infrastructure remains exposed and undefended.

Claroty’s Origin

Although our small investment in cyber foundry Team8 is gaining market value, we originally invested for more strategic reasons. Following our roadmap principle of “following the attackers,” we have long known that ICS would develop into a significant target, and hoped Team8 would provide us the best opportunity to invest in this market. They did just this with Claroty (fka Team 82), which is the second spin-out. Claroty is one of two dozen companies addressing cyber attacks on ICS. While Claroty is a newer entrant in this relatively nascent space, we believe deep the experience of its team makes it the likely winner.

Recall that retired Israeli General Nadav Zafrir had founded Team8 to focus the world’s best nation-state cyber warriors on the biggest challenges of cyber security. Zafrir recently commanded Unit 8200, considered Israel’s equivalent to the US National Security Agency (NSA). But unlike the NSA, which employs career-minded employees, Unit 8200 draws and trains the smartest draftees from the Israeli population, who, like everyone else, typically resign their military commission after three years. Naturally, several of them founded cybersecurity companies like Check Point, Palo Alto Networks, and NICE. But now Zafrir, along with the Unit’s former Head of Cyber (Israel Grimberg) and former Chief Technology Officer (Assaf Mischari), recruit and commercially train the top 1% of those graduates, re-purposing them in cybersecurity startups.

A principal skill set attributed to Unit 8200 is blind protocol analysis. If, for example, you wished to hack a Siemens centrifuge, you’d need to deconstruct the packets sent back and forth between the HMI and the PLC, or between the PLC and the actuator. Most protocols were cobbled together decades ago and were rarely well documented, and in some cases the vendors themselves treat them as holy writ. Unit 8200 is reputedly the best in the world at quickly and accurately understanding and parsing them down to the individual bit level. Team8 recruited the best, most experienced ICS thought leaders in Unit 8200, led by their team leader Benny Porat (CS PhD), to staff Claroty.

When Team8 starts a new company, it marries a technical team with an entrepreneurial founder. In the case of Claroty, Team8 recruited Amir Zilberstein, who founded the successful Waterfall Security and Gita Technologies. Waterfall develops ICS security products (unrelated to Claroty’s product); Gita’s technology remains undisclosed. Team8 also recruited Galina Antova, the former head of Siemens’ Industrial Security Services division, to run business development. Antova is a super impressive executive - highly connected, brilliant, and fast-moving. [See Appendix: Due Diligence for summaries of the team reference calls.] Next step is to recruit a CMO – we hope to get Patrick McBride, who was a star at iSight.

Beyond Security 

With meaningful Operations Technology (OT) experience on the team, Claroty is taking a different approach to the market than its competitors who generally come from cybersecurity backgrounds. Rather than lead with the cybersecurity benefits of their product, Claroty has developed an OT visibility platform that first and foremost surfaces operational issues. By deconstructing the proprietary vendor protocols, Claroty has delivered the first heterogeneous HMI with analytics that span an ICS network. Seeing as how most OT teams today care more about downtime than infection, we believe this approach will enjoy a far better reception in the near-term.

Monday, September 12, 2016

It's Time for Robots to Mine the Asteroids

Phil Metzger at University of Florida has just published an important and compelling article titled Space Development and Space Science Together, an Historic Opportunity about the need to develop a Self-sufficient Replicating Space Industry that uses robots to harvest space-based resources . The article is detailed, well-cited and fully attentive to the objections often raised.

Metzger calculates that it would take only a third of Earth's national space program budgets over the coming decades to deploy and complete the industrial infrastructure we need for harvesting resources from space that address major challenges we face in economic development, science, climate change, energy needs and other dwindling mineral resources.

Metzger specifically prescribes an initial focus on mining water for the purpose of fueling steam-based propulsion systems. Robust water deposits on the moon, asteroids, Europa, and elsewhere in the solar system promise bountiful supplies that will propel us to the stars. Another benefit of hydro-propulsion, explained to me this week by Deep Space Industries @GoDeepSpace CEO Dan Faber, is that water would be easy and safe for entrepreneurs integrating propulsion into their satellites today. Metzger has focused his attention and efforts on developing a lunar mine, Faber's company looks to mine water from Near Earth asteroids since their negligible gravity makes it easier to extract the water without escaping lunar gravity. (See DSI design, right.)

Metzger outlines other important projects as well, such as a Space-Based Solar Power system and extraterrestrial compute facilities, sorely needed infrastructure that we simply cannot scale on Earth:
"The primary benefit of space is real estate that biology does not need. Earth is the one special place in the solar system required by life, but machines can function anywhere else."
Why now? Metzger argues that AI has reached the points of maturity and acceleration that we need to pull it off, citing Bill Gates that robotics "is developing in much the same way that the computer business did 20 years ago."

Counter intuitively, the primary obstacles are not technical. Rather it is government inaction, in both funding and regulation. Peter Diamandis' startup Planetary Resources employs JPL veterans who know how to prospect Near earth asteroids today, but PR's mission awaits space-faring nations to legally recognize asteroid mining rights (other than the U.S. which did so last year). And government funding is hard to come by for what the Chairman of the Senate Subcommittee on NASA dismissed as a "nutty fantasy." Metzger lays out strategies for overcoming these obstacles.

Enough said. Click through and read Metzger's important, fascinating paper.

Saturday, October 10, 2015

"Brief Candle in the Dark: My Life in Science" by Richard Dawkins

Oxford Zoology Professor Richard Dawkins is finishing up a whirlwind book tour through the U.S., addressing sold-out venues of free-thinking fans who flock to him as much for his sermons on Reason and Science as they do for a signature on his memoirs.

One of Richard's favorite stops is always Kepler's Bookstore in Menlo Park, where I had the pleasure of interviewing Richard about his memoirs before a crowd that sold out four weeks in advance. Richard graceed his audience by reading several excerpts I selected -- chosen to give a sense for his writing but, like any good trailer, not to reveal crucial plot lines.

So rather than write a review of the book (which the NY Times and Guardian have already done quite well) I'm here to share a little preview of the story, which covers the second half of Richard's illustrious life so far. With this taste of the book, you can relish how Richard crafts every message with subtle detail and humor that, in Silicon Valley parlance, delights the user.

The first excerpt gives a glimpse into life at the hallowed institution of Oxford University, featuring brilliant but eccentric personalities who mix profound wisdom with the backseat bickering of children.  As Richard recounts his unwelcome rotation as Sub-Warden, the setting seems less like Oxford and more like Hogwart's.
Although the Sub-Warden doesn’t have to seat people and their guests (as the presiding fellow does in some other colleges), he is expected to beam the role of genial host at dessert. I did my best, but there was one awkward evening. As I was helping people to find their seats I became aware, from a sort of ominous rumbling, that all was not well. Sir Michael Dummett, immensely distinguished philosopher, Wykeham Professor of Logic in succession to Freddie Ayer, stickler for grammar, conscientious and passionate campaigner against racism, world authority on card games and voting theory, was also famously choleric. When angered he would go even more than usually white, which somehow seemed – though this may be my fevered imagination – to make his eyes glow a menacing red. Pretty terrifying . . . and it was my duty as Sub-Warden to try to sort out whatever this problem was. 
The rumble grew to a roar. ‘I have never been so insulted in my life. You have the most atrocious manners. You obviously must be an Etonian.’ The target of this damning sally was not me, thank goodness, but our quirkily brilliant classical historian, Robin Lane Fox. Robin was hopping with anxiety and bewildered apology: ‘But what have I done? What have I done?’ I didn’t immediately succeed in discovering what the problem was, but in my hostly role I saw to it that the two of them were seated as far from each other as possible. I later learned the full story. It had begun at lunchtime that day. Lunch is an informal, self-service meal and fellows sit where they like, although it is conventional to fill up the tables in order. Robin noticed that a new fellow was hesitantly looking for a place. He courteously motioned her to sit, but unfortunately the chair he indicated was the very chair for which Sir Michael was heading himself. The perceived slight rankled, simmered up through the afternoon and finally boiled over after dinner at dessert. The story had a happier ending, as Robin told me when I asked him recently. A couple of days after that distressing incident, he was approached by Professor Dummett who offered the most gracious apology, saying that there was nobody in the college whom he would less wish to insult than Robin. Thank goodness I was never the target of his ire, although I might have been vulnerable as he was a devout Roman Catholic with the zeal of the convert.
Here is a memory of Richard's biogeographic expedition to Barro Colorado Island in Panama with John Maynard Smith:
This party was also memorable because of the firework display on a huge ship passing through the canal just beyond the trees. Actually falsely memorable, because for years I have been utterly convinced that we saw in not just a new year but a new decade: 1 January 1980. So detailed and full were my recollections of ‘seeing the new decade in’, it took multiple documentary evidence, kindly sent me by Ira Rubinoff, Ragavendra Gadagkar and Nancy Garwood, to finally convince me that what I had thought to be a crystal clear memory was faulty. It was actually 1 January 1981, not 1980. I was quite shaken to discover this, because it made me worry how many other clear memories actually never happened (and the reader of my memoir is, I suppose, duly warned). 
The dreamlike presence of large tankers deep in the jungle was one of the most vivid memories I took away from the place. On several afternoons I had joined the resident scientists swimming off a raft, and it was a surreal experience to see those gigantic vessels gliding calmly, and surprisingly quietly, through the still, clear water, just a few yards away behind the trees. Some of the women scientists liked to sunbathe, and I couldn’t help wondering what the tanker crews thought about the undraped feminine pulchritude diving off the raft deep in the jungle. If those mariners were Greek, did they think Sirens; or if German, Lorelei? Or – peering through the lush tropical vegetation – did they see a vision of Eve’s innocence before the Fall? They had no way of knowing that these tropic nymphs had PhDs in science from some of the top universities in America.
Among my favorite passages is RIchard's recollection of the BBC-televised Christmas Lectures, an annual lecture series for children replete with physical demonstrations. Michael Faraday had launched the London tradition and the honor on has since passed on to the greatest science educators like Carl Sagan, Sir David Attenborough, and Richard.
One agreeable and unanticipated feature of the Christmas Lectures was that the very name was a golden key to unlock goodwill whichever way I turned. ‘You want to borrow an eagle? Well, that’s difficult, I honestly don’t see how we can realistically, I mean do you seriously expect… Oh, you’re giving the Royal institution Christmas lectures? Why didn’t you say so before. Of course. How many eagles do you need?’  
 ‘You want an MRI Scan of your brain? Well, who is your doctor, have you been referred to the MRI department on the National Health Service? Or are you going privately? Do you have health insurance? Have you any idea how expensive MRI scans are, and how long the waiting list?... Oh, you’re doing the Christmas lectures? Well, of course, that’s different. I’m sure I can slip you into a research run, no questions asked. Can you come to the radiography department on Tuesday during the lunch hour?’
 By just dropping the name of the Christmas lectures, I managed to borrow an electron microscope (big, heavy, and transported at the lender’s expense), a complete virtual reality system, an owl, an eagle, a magnified circuit diagram of a computer chip, a baby, and a jactitating Japanese robot capable of climbing walls like a much enlarged, ponderously hissing gecko.  
 I chose, as the overall title of my series of five lectures, Growing Up in the Universe. I meant ‘growing up’ in three senses: first, the evolutionary sense of lice growing up on our planet; second, the historical sense of humanity’s growing out of superstition and towards a naturalistic, scientific apprehension of reality; and third, the growing up of each individual’s understanding, from childhood to adulthood.
It was a tradition to call up volunteers from among the children, which is what Richard proposed to do in preparation for an experiment that involved injecting a human eye.
This being the Christmas Lectures, the next thing to do was to call for a volunteer . . . I produced a huge veterinary hypodermic syringe, fit to sedate a rhinoceros, and asked who would like to take part in the experiment. Normally, the children at the Royal Institution Lectures fall over themselves in their eagerness to assist in demonstrations. Surely nobody would volunteer in this case, and I was about to reassure everyone that it was only a joke when one little girl of seven, probably the youngest in the audience, hesitantly raised her hand. It was my darling daughter Juliet, sitting shyly by her mother. I still choke up a little at the memory of her uncomprehending loyalty and courage in the face of the monstrous hypodermic that I was brandishing.
When Richard does call up a volunteer, it turns out to be a plant -- his friend Douglass Adams. Later in the story he recounts how Douglass introduced Richard to his current wife, Lalla, to whom Richard dedicated the book.
This was 1992, when Douglas Adams reached his fortieth birthday, and his party was memorable for a particular reason. It was there that he introduced me to the actress Lalla Ward, whom he had known from the days when Doctor Who was at its wittiest because he was the script editor and she and Tom Baker gave added value to the wit by their inventively ironic playing of the two leading roles. At the birthday party, Lalla was talking to Stephen Fry when Douglas led me over and introduced us. Douglas and Stephen are both absurdly much taller than Lalla and me, so it was natural that we should find ourselves facing each other under a Gothic arch formed by Douglas and Stephen as they exchanged lofty witticisms high above us. Through the archway I shyly offered to refill Lalla’s glass, and when I returned we rapidly reached agreement that the party was too noisy for conversation. ‘I suppose, by any faint chance, it wouldn’t just possibly be a good idea to go out for a quick meal and – of course – return later?’ We discreetly slipped away and found an Afghan restaurant off the Marylebone Road.  
That Lalla had read The Selfish Gene and watched my Christmas Lectures was gratifying. That she had read The Extended Phenotype (and Darwin) as well was too good to be borne. I subsequently discovered that, in addition to Doctor Who’s companion, she had played a beautiful Ophelia to Derek Jacobi’s Hamlet in the BBC TV production, and was also a talented and versatile artist, published author and book illustrator. As I said, too good to be borne. We didn’t return to the party.  
I mentioned to Lalla that I was about to embark on my American journey, having added to the itinerary a visit to John Brockman. She said she was about to set off for a holiday in Barbados, with a girlfriend from the theatrical world. Impulsively she asked if I would take her to America with me, although it would mean letting down her friend in Barbados. Equally impulsively I agreed.  
Slight embarrassments then opened up. I was due to stay with Dan and Susan Dennett on first arriving in Boston, and later with the Brockmans in Connecticut. In both cases one house guest was expected, not two. How could I broach the subject? Lalla and I fretted that our hosts would ask – it is, after all a perfectly normal question to ask of a couple – ‘How long have you known each other?’ and we would have had to answer, ‘A week.’ As it turned out, they didn’t ask, and it was only years later that Lalla confessed to Dan the truth. ‘Really?’ said Dan, with possibly mock innocence. ‘I thought you’d known each other for years.’
In this excerpt, Richard recounts some of the backlash he faced from his most successful and controversial book God Delusion:
Opposition from religious apologists was predictable, and I’ve already mentioned the flea books. But attacks came, too, from fellow atheists, sometimes in outspokenly belligerent terms. One well-reputed reviewer went so far as to say that The God Delusion made him ashamed to be an atheist. His reason seemed to be that I didn’t take ‘serious’ theologians seriously. I dealt fully with those theological arguments that purport to support the existence of a deity. But I was entirely right not to bother with those that assume the existence of a deity as a starting point and go on from there. 
I have tried but consistently and failed to find anything in theology to be serious about. I certainly take professors of theology seriously when they use their expertise to do things other than theology: jigsaw the fragments of the Dead Sea scrolls, for instance; or minutely compare Hebrew and Greek texts of scriptures, or sleuth out the lost sources of the four gospels and the other gospels that didn’t make it into the canon. That’s all genuine scholarship, fascinating to read and deserving of respect. It’s even true that historians need to study theological logic-chopping in order to understand the disputes and wars that have stained European history, for example the English Civil Wars. But the vacuous deepities (Dan Dennett’s splendid word) of ‘apophatic theology’ (Karen Armstrong’s obscurantist smokescreen), or the expenditure of precious time arguing with other theologians over the precise ‘significance for us today’ of Original Sin, Transubstantiation, the Immaculate Conception, or the ‘mystery’ (sorry, ‘Mystery’) of the Trinity, none of that is scholarship in any respectable sense of the word, and it should have no place in our universities. 
Theological gymnastics over the ‘significance for us today’ of nonsensical ideas from the past like transubstantiation lend themselves to satire – positively beg for it. A gem that I recently met: ‘Of course we don’t literally believe the story of Jonah and the whale. But it is symbolic of Jesus’ death and resurrection . . .’ Suppose science worked like that. Suppose that (to take a most unlikely hypothetical) future scientists were to find that Watson and Crick were completely wrong, and the genetic molecule is not a double helix at all. ‘Ah well, of course nowadays we no longer literally believe in the double helix. But what is the significance of the double helix for us today? The way the two helices twine intimately around one another, though not literally true in the crude, materialistic sense, nevertheless symbolizes mutual love, don’t you feel? The precise, one-to-one pairing of purines with pyrimidine is not literally true, nothing so crude as that, but it stands for . . . When you contemplate the Watson–Crick model, don’t you get an overwhelming feeling – I know I do – . . . etc. etc.
The book features many wonderful cameos in by Richard Leakey, Francis Crick, Chris Hitchens, and so many other great thinkers. For the final excerpt, here are Richard's impressions of Carolyn Porco.
Carolyn came to visit us in Oxford, and has been friends with Lalla and me ever since. She is a planetary scientist, in charge of NASA’s Cassini imaging team – the team that has brought us those stunning pictures sent back from Saturn and its many moons. But she is more than just a good scientist; she is inspired by the poetry of science, especially the romance of the spheres that share our sun. She is the nearest approach I know to a female Carl Sagan, a poet of the planets and singer of the stars. Whether or not the heroine of the book Contact was actually modelled on her, it is a fact that Carl Sagan invited her to be the character consultant on the film version. The scene where Ellie first hears the unmistakable communication from far space still gives me goose bumps when I think of it. The slender, clever young woman, woken up by the mind-shattering signal, bouncing back to base in her open car, exultantly yelling the celestial coordinates into the intercom for her dozing assistants: numbers, numbers, the spine-tingling poetry of those numbers and their arc-second precision. And how poetically right that the hero of the numbers should have been a woman. A role model, just like Carolyn. 
An anecdote displays the poetry of Carolyn, and I related it in the Oxford Playhouse when introducing her Simonyi Lecture. A beloved professor from her days at Caltech was the geologist Eugene Shoemaker, co-discoverer, with his wife and David Levy, of the famous Shoemaker–Levy comet. A pioneer of astrogeology, Shoemaker was part of the Apollo space programme. He was in the running to be the first geologist on the moon, but to his sad regret had to drop out for health reasons, and he turned to training astronauts instead of being one. In 1997 Shoemaker was killed in a car crash in Australia. Carolyn, in her grief, raced into action. She knew that NASA was about to launch an unmanned craft, which was programmed to crash-land on the moon after its mission was accomplished. She managed to persuade the mission manager, as well as the head of the planetary exploration programme at NASA, to add her teacher’s ashes to the spacecraft’s payload. Gene Shoemaker’s ambition to be an astronaut was denied him in life, but his ashes now lie on the moon’s surface where no wind stirs them (it is said that Neil Armstrong’s footprints are almost certainly still intact), and with a photographic engraving bearing these words that Carolyn chose, from Romeo and Juliet: 
      . . . and, when he shall die  
      Take him and cut him out in little stars, 
      And he will make the face of heaven so fine 
      That all the world will be in love with night, 
      And pay no worship to the garish sun. 
I have dined out on that story from time to time, but I usually cannot manage to recite the Shakespeare, and turn to Lalla to rescue me. When she speaks the lines from memory in her beautiful voice, I think I am not the only one around the table to choke up. 

Richard takes a break from book signings to appear on Bill Maher with Neil dGrasse Tyson

Tuesday, June 30, 2015

The Imitation Game (or, Why I Invested in Distil Networks)

In 1950, the journal Mind published Alan Turing’s seminal paper, Computing Machinery and Intelligence, in which he proposed a behavioral definition of artificial intelligence. After all, if a machine can demonstrate intelligence, how can it not be said to possess intelligence? Turing’s test challenged computer scientists to create a thinking machine that, through conversation, could fool a person into believing that it, too, is human; Turing’s challenge continues to drive AI researchers today.

With the proliferation of computers in modern life, the prospect of identifying thoughtful machinery takes on more than just theoretical or philosophical interest. Back in Turing’s day, a thinking machine connected only to a “teleprinter” (as Turing envisioned) would have lived a lonely life, but today there are billions of people online with whom to converse, promising profound implications for society. For example, we increasingly find the machines who answer customer service calls to be more productive and thoughtful than human agents.

Machines who demonstrate intelligence can communicate not only with people, but also with other machines designed to communicate with people – specifically, over 100 million web servers that invite human visitors to browse, learn, chat, transact, and share and with them. If a machine can demonstrate human intelligence in the eyes of a human judge, then no doubt it can win over these other machines on the internet, who are naturally less skilled at spotting other humans.

Or are they? If, say, the human judge in a Turing test can distinguish the smartest machines from humans with 60% accuracy, how well could a machine do at judging them? I call this the Turing Judge Test, a corollary to Turing’s Test that marks a subsequent milestone in the development of AI. If a machine conversing with other parties can outperform the human judges in identifying the machines, that right there’s some mighty good thinking.

With the benefits of shared learning and infinite storage, machines only get smarter over time, and so it seems inevitable that they will eventually pass the Turing Judge Test. On the other hand, as artificial judges get smarter, so do the artificial contestants. Even when machines do pass Turing tests with flying colors, how can they ever out-think other best-in-class machines? Or is there a way of distilling human intelligence into a single line of questioning that distinguishes silicon from gray matter?

Such a distillation would have more than theoretical value – indeed, it’s arguably critical for the safety of any information society. This is not just a theory – machines are already smart enough that they account for most web traffic, successfully posing as human visitors to perpetuate fraud on the government and business web servers they talk to. That’s why many sites use Completely Automated Public Turing tests to tell Computers and Humans Apart (CAPTCHAs).

xkcd
But CAPTCHAs create a nuisance for users and an outright obstacle for some disabled users; even worse, they can now be defeated in various ways – in other words CAPTCHA servers are machines who once passed the Turing Judge Test, but only until the machines they judge got smarter!

As a result, malicious bots wreak havoc on the web, perpetuating data theft, account hijacking, application DDoS attacks, form spam, click fraud, and any other naughty action they can scale up through tireless automation.


And that’s why I just invested in, and joined the board of, Distil Networks. Distil is run by a world class team of machine learning experts whose thinking machines can now distinguish other machines from people with over 99% accuracy. Staples, AOL, Dow Jones, StubHub and many others depend upon Distil’s cloud-based service to immediately eliminate entire classes of attack (and free up all the infrastructure they ran to serve the whims of robotic imposters). The Turing Judge Test has a winner!

At least for now.