Today, Negroponte is back to open TED, to reflect on predictions he’s made in the past and to spin some new ones for the future. His point is that when someone tells you that you are “dead wrong,” that you just might be onto something.

Negroponte takes us on a lightning-paced tour of his career, beginning in the 1960s when he worked on computation and creating Habitat-like structures inspired by Moshe Safdie’s Montreal buildings. He says that at the time, he and his colleagues were “considered sissy computer scientists, not the real thing.” This was still the perception in 1976, he says, when he and several others were asked to build a model of Entebbe airport so that the military could practice a rescue simulation. He and his colleagues rigged a camera to a truck to create a detailed, image-based map. It’s a set-up that will remind anyone today of the cars that roam streets, powering Google Maps.

In his very first TED Talk, Negroponte shared his belief that we’d move from the computer mouse to using our fingers to control interfaces. “We picked fingers because everybody thought it was ridiculous,” says Negroponte. “They were low-resolution, there was concern that the hand would occlude what you want to see, and that the finger would get the screen dirty.” Of course, 30 years later, these concerns sound ridiculous. Or at least they do when we take out our smartphones and tablets.

“One of the things about aging is that I can tell you with great confidence that I’ve been to the future. I’ve been there actually many times,” says Negroponte. “How many times in my life have I said, ‘In 10 years this will happen … and then 10 years comes?'”

Negroponte tells the story of a Ph.D student at the MIT Media Lab who, in the early days of GPS, envisioned a system called “Backseat Driver” that would give audio directions to the driver. There were many, many challenges involved in the project but the student was beginning to work some out. However, this student was advised not to patent the technology because the liability surrounding it was just too big. “It shows you how people don’t really look at what’s already happening,” says Negroponte.

As recently as 1995, Negroponte was called crazy for predicting “that we’ll soon buy books and newspapers straight over the internet.” (By Clifford Stoll!) He says, “That gives me great pleasure.”

But there is another thing that can happen with predictions—that people follow a crazy idea, but without quite enough grand vision. Negroponte says that recent reasearch on “the Internet of Things is tragically pathetic.” He says that the idea wasn’t just to put an oven control panel on your phone — but to make the oven intelligent. “You want to put the chicken in the oven and it realizes you’re cooking it for Nicholas and he likes it this way,” Negroponte says.

Negroponte continues by looking at his work since 2000, which shifted to experiments in learning. “Learning is best approximated by computer programming,” he says. “When you write a program—and all programs have bugs—you have to debug it. And you iterate. That iteration is a good approximation of learning.” He brings up 1 Laptop Per Child, the initiative he ran for seven years that distributed 3 million laptops in 40 countries at a cost of $1 billion. The effort led Negroponte to an experiment in Ethiopia, one which resembles TED Prize winner Sugata Mitra’s “hole in the wall.” Like Mitra, Negroponte wondered: “Can learning happen where there are no schools?”

“We dropped off tablets with no instructions and let the children figure it out,” he says. “They were using 50 apps in five days. They were singing the alphabet songs in two weeks. And they’d hacked Android within 6 months.”

Negroponte sees the great challenge of the future to connect the last billion people on the planet to the internet. On a slide, he poses a bold question: “Will internet access be a human right?”

Negroponte’s latest project is using a stationary satellite to connect 100 million people living in rural areas to the internet. The effort will cost $2 billion, he says, which sounds like a lot, but is actually what the United States was spending in Afghanistan in a week.

Negroponte makes a final prediction—and a fun one. “My prediction is that we are going to ingest information—we’re going to swallow a pill and know English and swallow a pill and know Shakespeare,” he says. “It will go through the bloodstream and it will know when it’s in the brain and, in the right places, it deposits the information.”

It sounds shocking now, but will it in 30 years?

In conclusion, Negroponte leaves us with a fascinating way to think about what will come next. “The best vision to use to see the future is peripheral vision,” he says.