MIT Future Compute 2019
I'm gonna tell you a story,
I'm gonna tell you about my town
I'm gonna tell you a big fat story, baby
Aw, it's all about my town
I have always been fascinated by MIT and the values it represents. This dates back to my computer science undergraduate years where I was often searching the web for the best materials on various subjects. Inevitably, a lot of times I ended up with lecture notes from one of MIT’s professors. Just for context, MIT is now 5th place in the ranking of universities for the number of Nobel laureates. The top spot belongs to Harvard University, MIT’s neighbor in Cambridge.
Therefore, I did not think twice when I was asked by Honza, our CTO, if I wanted to attend Future Compute 2019, a conference on the future of computing, held by MIT Technology Review on its own campus.
The States did not really want to impress us upon our arrival. Newark is a medieval, overcrowded airport and even places to sit were scant. Hungry after a long flight, we hijacked a table at one of the mid-aisle restaurants. Here the record-low unemployment manifested itself in an overdressed man whose only job was to circle regularly through the restaurant and push the guests to either order food on the self-service tablets or get the hell out. We adapted pretty well. Once we found a place to sit, we intently scanned for food that would take a long time to cook and a long time to eat. Here I must admit Honza was the clear winner. Whereas my spaghetti with meatballs landed in front of me disturbingly fast, Honza waited for his 2-acre pizza for a good 10 minutes longer.
As the time of our departure was nearing, Honza went to check the information panel to make sure the boarding gate had not changed. Just to be sure. And for sure, our flight was canceled. The long queue at the United Airlines kiosk made it clear there was no point trying to get to some other flight. Quickly googling train connections to Boston, we found out the next one leaves in 30 minutes, plenty of time to get an Uber to Newark Penn Station.
“Two tickets for the train to Boston, please.”
“Ok, the next one then.”
“What is the first available train then?”
“Let me see...tomorrow at 2:30 p.m.”
That would mean missing the entire first day of the conference, something we were not willing to accept. As our last hope, Honza checks Uber. $520 is a lot, but we had no choice. The driver arrives in 5 minutes, so after just briefly visiting the toilets at the train station (I cannot imagine having to spend any more than a brief moment there), we set out to look for our car in the jammed four-lane traffic. There he goes. Actually, it was a she, so I was pretty sure she would kick us out after she finds out the destination.
“Are you really going to take us to Boston?”
“Yeah. It’s the farthest I ever drove anyone, but why not.”
Our hero is a young teacher in private kindergarten. She lives in Queens and her mom is also Ubering.
Of course, our flight was not canceled just for the fun of it. There is a snowstorm raging in Massachusetts and already midway through Connecticut the road conditions on the highway are so bad we are sliding left-to-right in the feet of beaten snow. At least the truck guys enjoy it, racing alongside us as fast as we are slow.
After a five-hour drive with just one brief stop at a place I swear I saw in Fargo, we arrived at 1 A.M. at the hotel, right at the waterfront. Few hours of sleep ahead of us are really good, considering.
Conference, day 1
MIT is located in Cambridge, along the Charles River (I checked, nothing to do with our king). Going to the Media Lab from our hotel in a taxi, we were surprised by how few people were walking the streets in the height of the Monday-morning commute.
The first day of the conference covered a wide range of topics, from the relevance of Moore’s Law to new materials for computer architectures and neuromorphic computing.
Below are some of the facts which stood out for me
“Every technology follows the law of diminishing returns curve. So when somebody says Moore’s law is dead, they say that the current set of innovations are somewhere on that curve. What we lack is the confidence that we have a cascade of these.”
MIT, in collaboration with Massachusetts General Hospital, is using carbon nanotube systems capable of processing terabytes of data per second with millions of gas sensors directly on the chip to detect infections from patients’ breath.
Graphcore’s Colossus is a microarchitecture specifically built for massively parallel AI applications, consisting of 23.6 billion transistors. I wonder if Entity Framework would run at least decently on that.
For me, by far the most interesting talk was that of Thomas Reardon, CEO of CTRL-labs. His company has developed a wristband that is capable of reading the neural signals your brain sends your hands and using those signals to control devices. Like a Jedi.
We had dinner at Parla, one of many italian joints on Hanover Street. It is a very small place of an obvious fame and peculiarly behaving service. Similar to our sports bars, only without a Tipsport sign above the entrance.
Conference, day 2
The second day of the conference was dedicated to all things quantum. We learned that there are many different types of quantum computers, but the ones already available are mostly using superconducting architecture.
Microsoft is developing topological qubits to overcome the fragile nature of qubits and provides developers with a Quantum Development Kit, a set of tools around the Q# language which you can use today to write quantum algorithms and soon hopefully run them in Azure.
Alan Baratz from D-Wave talked about the company’s upcoming 5000-qubit quantum computer, to be the most powerful one available. D-Wave chose a different technology than Google or Microsoft for its computers and announced the first commercially available quantum computer back in 2011. Not for common folks though, you had to pay $10,000,000 and sacrifice a bedroom to house it.
D-Wave also provides a cloud environment to run quantum algorithms on its computers and you better make them fast as the hourly rate is $2,000.
The rest of the talks felt mostly like sales pitches.
From the conference, we went directly to Logan airport where we managed to get a good dinner before the long flight to Zurich.
After some 80 hours, we landed in Prague and I immediately got off to write this blog because that’s how Jan would want it.