Talk at the Xconomy and the World Frontiers Forum’s Net@50 event started with a simple phrase about taco toppings and ended with discussion of a brain-computer interface that could perhaps enable a collective intelligence.
Net@50 last week explored the internet’s past, present, and future with thoughts from a rare group of engineers who imagined and built the internet’s precursor; executives grappling with the challenges and opportunities baked into today’s internet; and thinkers predicting what the next half century of connected computers could bring.
The star-studded lineup of speakers included internet pioneers Robert Kahn, Radia Perlman, Vint Cerf, Elizabeth “Jake” Feinler, Bob Metcalfe, and Leonard Kleinrock; executives from Akamai Technologies (NASDAQ: AKAM) and Dell EMC; researchers and thought leaders from MIT, Cornell Tech, and the University of Washington; and musician and internet evangelist Peter Gabriel.
Without further ado, here’s a rundown of the discussions. Video of the talks is available here.
—Juan Enriquez, managing director of Excel Venture Management (pictured above), laid the groundwork for the day with a presentation arguing the assembled group of internet pioneers have been short-changed by society and not celebrated enough for the tectonic impact of their technological advances.
The reason to call together the internet engineers of the past goes back to a Mexican phrase, Enriquez said: “There’s a saying in Mexico: You’re not putting enough cream on your tacos. And that means, you’re being too modest. This event is about celebrating some of the most important people in the world, people who made our world. We in Boston, and we in the US, and we in the world haven’t recognized what some of the people in this room have done.”
“You get what you celebrate,” Enriquez said, highlighting society’s focus on superficial celebrity rather than the researchers who drove the internet’s transformation of the modern world and eventual reorganization of the economy from “atoms to bits.”
“What the folks in this room did, and the ones before and the ones after them, they changed the language. They took every ABC and put it into ones and zeros,” he added. “They created a single language.”
—Robert Kahn, who worked at Bolt, Beranek & Newman on the ARPANET contract and later invented TCP/IP with Vint Cerf, said everyone should keep an eye on the future as they gather to commemorate the past.
Kahn, who now heads the Corporation for National Research Initiatives, pushed to center the internet around information, rather than locations where information is stored, by something he developed decades ago called digital object architecture. The shift would help users share and retrieve information stored for long periods of time with less worry that the location of the information had shifted over the years, he said.
“This is a good way and a convenient way to build the internet up in a collaborative level,” Kahn said. “We need to change our perspective and take the long view. How are we going to manage information over the long haul?”
—Leonard Kleinrock, whose UCLA lab oversaw sending the ARPANET’s first message in October 1969, retold the story of that historic day, noting he and his students hadn’t given the content of the message much thought ahead of time.
“We had no idea we should make a good message,” Kleinrock said.
It all worked out, though, when the intended message “Login” failed halfway through transmission to a computer at Stanford and the computer scientists were left with the simple, prophetic, “Lo.” Kleinrock told Xconomy more about the episode here.
Kleinrock recalled the story on stage with Cerf, Radia Perlman, Elizabeth “Jake” Feinler, Dave Clark, and Kahn.
—Akamai CEO Tom Leighton spoke about the increasing importance of the internet’s edge, which includes devices like phones and computers as well as the “last mile” of connectivity. The stakes for security on the edge have never been higher, he said.
“It’s a very serious challenge today,” he said, adding that Akamai’s technology attempts to “absorb attacks at the edge before they get to the core of the internet.”
Leighton—in a conversation with ClearSky Data CEO Ellen Rubin—said Akamai’s identity technology, which tries to eliminate people hacking into accounts with stolen credentials, has been largely successful by focusing on the exact ways people log in with phones or how they type passwords on computers. Akamai boosted its identity management capabilities this year with its acquisition of Janrain.
“The way we do it is [looking at] how you hold your device, how you log in, how you tap it, how you move your mouse,” Leighton said. “We are recording this whenever you log into any major bank or commerce site, and we look to see which ones are human or which ones are bots.”
—MIT Media Lab founder Nicholas Negroponte contrasted the prevailing culture of collaboration of the 1960s and 1970s that underpinned the development of the early internet with the current tech culture where the US is attempting to wall itself off from Chinese tech giants Huawei and ZTE due to national security concerns.
“We can’t cut ourselves off, especially since as a nation we have pushed telecom very low on the spectrum,” he said. “In terms of sharing basic knowledge, we cannot cut off a company like Huawei. … They are doing it better than we are.”
Negroponte also argued for the nationalization of basic internet access, saying competitive market forces haven’t done enough.
“The cost of doing this is trivial,” he said. “It is really not a big number—in the $5 to $10 billion range. … You really could connect everybody, and it doesn’t have to be broadband.
—Neil Gershenfeld, director of MIT’s Center for Bits and Atoms, led a wide-ranging discussion about pushing the boundaries of today’s internet. His talk was about the shrinking distinctions between the digital and physical worlds. “This leads to, there is no edge,” Gershenfeld said. “Bits become atoms and atoms become bits, at a fundamental level.”
He described how researchers are taking lessons from one to improve the other. For example, “we’re learning all of life is made from 20 parts—amino acids,” Gershenfeld said. “We wrote up an interesting proposal to the Department of Defense to reduce their entire supply chain to 20 parts. We’re learning how to build all technology to a small set of building blocks, analogous to the role of the personal computer.”
—As computing has advanced, data processing has moved closer to the edge, nearer to the source of the data, said Deborah Estrin, professor of computer science and associate dean at Cornell Tech. That’s partly to save energy and decrease latency.
“Most recently, I think the driver for doing that local processing is the ‘P’ word, which is privacy,” Estrin said. One of the keys going forward will be for technology developers and users to figure out how to balance “the opportunities of personalization and use of having instruments that know so much about me, and that privacy and need to push against the tides of surveillance.”
—As an early-stage startup investor with Founder Collective, it’s Eric Paley‘s job to try to predict which inventions might achieve commercial success on a large scale. Often, things that are “quite trivial,” or at least appear that way, reach that commercial success. (He mentioned Twitter as an example.) Meanwhile, seemingly innovative and amazing creations often don’t make it—at least not right away. Sometimes it takes decades before “the technology from the edge is ready,” Paley said.
—Nadya Peek, an assistant professor in the University of Washington’s College of Engineering, noted that the internet has changed a lot in the past 50 years, but manufacturing equipment hasn’t evolved much. She’s working to change that by making re-configurable machines, modular controls, and other advances in digital fabrication tools.
“Machines are still made application-specific—you still only can do one kind of thing with each machine,” she said. “Maybe for the diversity of production to match the diversity of application, what we need is a completely different kind of infrastructure to produce parts and products.”
—For several years, Gershenfeld, Cerf, Hunter College psychology professor Diana Reiss, and musician Peter Gabriel have been working toward creating an “interspecies internet.” Gabriel said one of the highlights of his massively successful career was playing with a bonobo named Panbanisha, part of an effort to communicate with apes through music.
“So, we’re improvising together here,” Gabriel said as a video of the cross-species jam session played on the screen behind him. “Musicians know when they connect.”
The experience “just blew me away and started the conversations that led us into thinking about non-human intelligence coming online,” Gabriel said.
Reiss, a cognitive psychologist who studies dolphin cognition and communication, said her goal is to use the internet and other technologies to give animals a voice. “We now know that other animals think,” she said. “How can we connect minds?”
The pursuit brings interesting biological challenges. Dolphins, for example, don’t have hands; they do have large brains like humans, but very different brains, Reiss said. And then there’s the octopus, with a brain in each of its tentacles. “It’s so challenging, and it’s exciting,” she said.
Ultimately, she hopes the effort will “lead to improved conservation efforts by knowing more about these other minds we share this planet with,” Reiss said. “Hopefully, we’ll get more excited and protect [animals] more.”
Cerf said he’s intrigued by the idea that we could perhaps use machine learning to translate languages back and forth between species. Is it possible “we could translate what a dolphin has to say to what a chimpanzee might want to hear? What possible context could they share?” Cerf wondered. “It reminds people that we share a common planet. We better do something to hang onto it because otherwise our species may disappear.”
Gershenfeld said that Cerf has suggested that “if we can learn how to talk to [other] cognitive species on this planet, we would be better equipped to talk to cognitive species on other planets.”
“Maybe that’s the furthest point beyond the edge and a good point to end the session,” Gershenfeld concluded.
—In a panel about fixing social media, security expert and Grip Mobility Technology CEO Juliette Kayyem said her largest concern right now with social media is radicalization. Social media gives people the ability to find community or a network to radicalize, and it also amplifies a toxic public discourse that, among other things, does not condemn racism and “in fact promotes it,” she said. “I would like it to be shamed. I want these people to go back inside.”
“Autoplay on YouTube might be one of the most dangerous things on the planet,” said Raffi Krikorian, managing director of Emerson Collective and former vice president of engineering for Twitter. Autoplay will “send you down a rabbit hole of things that are driven by your particular engagement. … You’re no longer forced to see moderate oppositional opinion and have discourse on it because that’s not what these platforms are set up to do.”
Asked by moderator Wade Roush, an Xconomy contributor and Soonish podcast host, to offer possible solutions to social media’s problems, Kayyem and Krikorian said it will take a mix of approaches. Legislation could help by requiring more transparency around advertising purchases, Kayyem argued. Social media companies also could do more; Roush pointed out that some have suggested YouTube turn off the autoplay feature. Krikorian also suggested that social media platforms should try to slow things down and “dampen virality.”
And users could more proactively seek out diverse voices and news sources and get better at spotting misinformation.
“We shouldn’t blame the victim,” Krikorian said. “A lot of this should fall on platform providers” and governments. But, he added, “I want to believe people are educatable.”
—During the closing panel discussion, internet pioneer and University of Texas innovation professor Bob Metcalfe proposed what he calls “composable branded filters” for social media users—“a way for each of us to construct our own filters.” (It’s a suggestion he discussed in more detail with Roush in an earlier interview, which you can read here.)
On the panel, which was focused on predicting the next 50 years of the internet, Metcalfe said an easy prognostication is that many people will be “perpetually monitored.”
“This is the internet of things, only we’re the things,” Metcalfe said. “I view this as good news.”
“If we were all constantly monitored,” he continued, “like I think we will be, that’s a huge medical trial. [We’d track] all these phenotypes out there. We could do scientific research on information being collected from a population of 7 billion.”
John Roese, global chief technology officer for Dell EMC, thinks that technology will get better at not only enabling communication between humans, but enhancing it. For example, a combination of sensors and augmented reality devices could notify someone during a conversation that the other person is feeling stressed, Roese said. And perhaps advances in virtual reality will one day be able to deliver “full, deep human interaction” in a digital environment, he said.
“We’re nowhere near that today, but we’re on track to get there,” Roese said. “I think we just have to build a better user experience for the internet.”
Danny Hillis, the parallel computing pioneer and co-founder of Applied Invention, took things a step further and argued that the next significant phase of the internet will be hooking up our brains to it directly via implantable devices. He expects the development and implementation of such brain-computer interfaces will play out over 50 years. The idea is people will become extensions of the network, which could open up the possibilities of humans thinking together in new ways, thinking together with computers, and thinking together with other species.
“It’s a new kind of intelligence that spreads itself across the network,” Hillis said. “Unless we think we’re the end of evolution, there is going to be another level. I think it’s an inevitable way that intelligence evolves and the universe evolves.”
This new mode of intelligence could help solve massive global challenges such as climate change, Hillis argued. “I don’t think we’re any place close to the inherent limits of how smart you can be, and how smart you can be as a society,” he said.
Nevertheless, Hillis realizes being directly wired into the network is “not an appealing idea to everyone.” “I’m kind of squeamish about it myself,” he said. “But I think it’s pretty much inevitable that once the power to do this [is available], people will want that power.”
Moderator Kara Miller, host of WGBH’s “Innovation Hub,” wondered if that huge increase in connectedness and data sharing might also result in a much higher level of vulnerability for people. Hillis responded that improvements in technology and security could help with that, though he admitted tech improvements might not fully solve the “social problems.” In the evolved internet he envisions, “things don’t talk to each other by default, they talk to each other because they agree, by policy.” “Then I think we can build a better level of trust on the internet,” he added.
Metcalfe said his worry is the opposite: If people struggle to handle all this connectivity, it’s possible Luddism will increase, and there could be efforts to suppress technological advances on the internet “out of fear.” “We should keep solving problems, but not give up—not pass laws that make the internet illegal,” he said.
Xconomy deputy editor of tech Jeff Bauter Engel contributed to this report.