Majorana 2 Explained: The road to 1000x improvement — Transcript
Full transcript
- 0:00If you think about the way
- 0:01we process information,
- 0:03it fundamentally hasn't
- 0:04changed since the abacus.
- 0:062,400 years of calculating by
- 0:08moving ones and zeros around.
- 0:10Quantum computing changes that.
- 0:13We've worked
- 0:14on topological quantum
- 0:15computing for 20 years.
- 0:17And we've known that
- 0:19it can work, in theory.
- 0:20There was a period in which
- 0:22we were
- 0:22completely lost.
- 0:23We didn't know which way was up.
- 0:25We did get to a point
- 0:27when things just
- 0:28started to click into place,
- 0:29from knowing what you want to do
- 0:31to actually starting to build it.
- 0:33When what was only theory is now real.
- 0:37We went from Majorana 1 and almost
- 0:39completely reinvented it in Majorana 2.
- 0:42We've seen a thousand x improvement
- 0:43in the stability of our quantum computer.
- 0:53The last year has been quite a wild ride
- 0:56as we've gone from Majorana 1 to Majorana 2.
- 0:59We changed the material stack.
- 1:00We changed the design.
- 1:02We changed
- 1:02the way we were doing the measurements.
- 1:04Somehow, they all came together
- 1:06at the same time and all landed.
- 1:09The superconductor in our device was
- 1:11aluminum previously,
- 1:12and the semiconductor was
- 1:13indium arsenide.
- 1:14We changed the
- 1:15material stack significantly.
- 1:16We replaced aluminum with lead,
- 1:18and we introduced antimony
- 1:20into the semiconductor.
- 1:21That gave us a new material with
- 1:24properties that were
- 1:24even better and led to this
- 1:261000x improvement in that lifetime.
- 1:28And one big thing was the breakthroughs
- 1:30we made on the qubits,
- 1:33the bigger gaps,
- 1:34the longer lifetimes,
- 1:36thanks to lead.
- 1:37When you do an
- 1:37operation every microsecond
- 1:39and your qubit lives
- 1:40for about a millisecond,
- 1:42then it means there's one error every
- 1:45thousand operations.
- 1:47When it is
- 1:48a lifetime of tens of seconds
- 1:50and one operation takes a microsecond,
- 1:52that means you can do tens
- 1:54of millions of operations
- 1:55before an error.
- 1:56And now we're seeing it
- 1:57on 20-second timescale.
- 1:58And sometimes it's
- 1:58like minutes in between flips.
- 2:00And it takes the quantum world and puts
- 2:02it in units of like minutes, you know.
- 2:05It's kind of mind-blowing to
- 2:06have it in like a human scale.
- 2:08The larger that scale is,
- 2:09the more robust your qubit,
- 2:11the better its performance.
- 2:12And that dependence,
- 2:14that functional form, is exponential.
- 2:16It wasn't that we made a thousand x
- 2:18improvement in the underlying properties.
- 2:20It was
- 2:20improvements by a
- 2:21multiple of a few, two to three.
- 2:23But those multiples of on the order of
- 2:24two to three led to a thousand-fold
- 2:27improvement in this
- 2:28measure of performance.
- 2:31We've worked on
- 2:32topological quantum computing
- 2:34for 20 years.
- 2:35But for the last 20 years,
- 2:37that has been theory.
- 2:39What we see now is that
- 2:40lead was a wow moment.
- 2:42Lead was when we really saw
- 2:44what we theorized actually works.
- 2:48The quantum computer is
- 2:49much more than the qubits,
- 2:50just like a classical
- 2:51computer is more than
- 2:53the transistors.
- 2:54We have to control the qubits.
- 2:56We have to be able
- 2:57to measure them.
- 2:58We have to orchestrate them,
- 3:00and we have to be able
- 3:01to run our programs on them.
- 3:03So, we know how to build
- 3:05programs on quantum computers.
- 3:06We know how to put together this sequence
- 3:08of quantum operations,
- 3:10often called gates,
- 3:11to build out the circuits that
- 3:13implement an algorithm or
- 3:14allow us to run an application.
- 3:16We also have some fantastic tools
- 3:19in the form of
- 3:20the Quantum Development Kit
- 3:21and the Quantum Resource Estimator.
- 3:24So, we created the
- 3:25Quantum Resource Estimator
- 3:26to allow people to
- 3:27explore the trade-offs between
- 3:30a particular application
- 3:31or circuit that you need to run,
- 3:32the algorithms that you might
- 3:34use to implement that application,
- 3:36and the architectural considerations
- 3:38about the system it's going to run on.
- 3:41The quantum computer works on the principle
- 3:42of quantum mechanics.
- 3:44Quantum mechanics is the theory that
- 3:46describes how nature works at
- 3:48the smallest scales.
- 3:50And if I want to calculate the properties
- 3:52of a material, of a molecule, if I want
- 3:54to understand the chemical reaction,
- 3:56then I need to understand
- 3:58the behavior of the electrons
- 4:00in this molecule.
- 4:02And doing that exactly on a classical
- 4:03computer is exponentially hard.
- 4:06AI has helped us to increase
- 4:07the scale and speed
- 4:09with which we discover new materials
- 4:11or explore the details
- 4:13of chemical reactions.
- 4:14Our AI models right now,
- 4:16they're based on classical data,
- 4:18but when the system gets too big,
- 4:19they have to start making approximations.
- 4:22I put approximate data in,
- 4:23I get approximate answers out.
- 4:25Quantum changes all that.
- 4:27Quantum allows us to put the exact,
- 4:29ground-truth answer into the AI model
- 4:32and get first-time-right predictions
- 4:34out the other side.
- 4:35But those calculations will still take
- 4:38potentially hours or days.
- 4:40What we just want to do is we want to
- 4:43train an AI model to predict the results
- 4:45of the calculations.
- 4:47Then it will just take seconds and not
- 4:49hours to get the results.
- 4:52What that will enable is that your Copilot
- 4:54can not just write an essay,
- 4:58create an image, create a movie,
- 5:00but it can also then
- 5:02create materials and molecules
- 5:03because it understands quantum physics.
- 5:07That will totally change
- 5:09the way we make things.
- 5:12So, there are questions in chemistry,
- 5:14material science, physics
- 5:16that we just can't answer because
- 5:18we don't have the computing power.
- 5:20No classical computer on the planet
- 5:22can solve it.
- 5:23If we can unlock the answers,
- 5:25we will have a new path to discovering
- 5:27brand new materials.
- 5:28For example,
- 5:29superconductors that operate at
- 5:31room temperature
- 5:32or enable levitating
- 5:34magnetic transportation.
- 5:35Quantum computers are going to help us
- 5:37answer questions we just
- 5:38couldn't answer before.
- 5:40Skepticism is a valid
- 5:41and important part of science.
- 5:42I'm glad people are questioning and
- 5:44looking seriously at the work, right?
- 5:45I don't expect everyone to just, like, take
- 5:47everything at face value.
- 5:48You know, there's many things that are
- 5:50proprietary, that we can't share
- 5:51all the details.
- 5:52There have been people that have
- 5:53read our papers, looked at our stuff.
- 5:55Especially when they do it
- 5:56in good faith, where they're really
- 5:57just trying to get to the truth,
- 5:58that's always a good thing.
- 6:00Everyone who's looked
- 6:01very carefully
- 6:02and seriously at our results
- 6:03has seen that they are not only solid
- 6:06but really actually quite revolutionary.
- 6:08And, you know,
- 6:09if people are still skeptical,
- 6:10all I can say is, you know, there's more
- 6:12where that came from.
- 6:13And
- 6:13maybe we haven't knocked their socks off
- 6:14just yet. Pretty sure the next one will.
- 6:17I think we've just learned
- 6:19a lot along this journey.
- 6:20We were six to 12 months early,
- 6:21I think, in the material change we made.
- 6:24We've incorporated the
- 6:25use of AI-based tools.
- 6:26We've tightened the collaboration across
- 6:28various areas of the team.
- 6:29And very importantly,
- 6:31on the algorithms and application side,
- 6:32we've brought in some of those
- 6:34key applications that
- 6:35have enabled us to
- 6:37bring in those timelines.
- 6:38Right now
- 6:38we're targeting 2029
- 6:40for quantum computers that can
- 6:44solve some problems
- 6:45that haven't been
- 6:46solved before by other means.
- 6:47In the future, we'll build Majorana 3, 4,
- 6:49and 5, and the
- 6:50challenge now is scaling up.
- 6:53Just tremendous to think that we're a few
- 6:55years away from being able to start to
- 6:57use the types of scaled, fault-tolerant
- 7:00quantum computers that will unlock a
- 7:02whole new generation of applications in
- 7:04chemistry, material science, and beyond.
- 7:08I've been a part of this program since it
- 7:10was not even quite a program yet.
- 7:12Yeah, since the very beginning.
- 7:13I think I'm like employee two of
- 7:14the quantum program, basically.
- 7:17I've been blessed to work with some
- 7:18incredibly dedicated and brilliant
- 7:20people, which I find honestly humbling.
- 7:23We have been able to
- 7:25make devices and uncover properties
- 7:27that I wouldn't have
- 7:28dreamed of 20 years ago.
- 7:30And we're doing computation with it now.
- 7:32And that, to me, it's just blowing me away.
- 7:35So, at my core, I'm a technology nerd.
- 7:37I love working with new technologies.
- 7:40It's amazing to be part of an effort
- 7:42that's building a new form
- 7:44of quantum computing.
- 7:46There is
- 7:47nothing more transformative
- 7:48than quantum computing.
- 7:50It got to the point in my career
- 7:52when I said I need a quantum computer to
- 7:54do the research that I want to do.
- 7:57And that's why I joined Microsoft,
- 7:58to actually make it real
- 7:59and accelerate the path
- 8:01to getting it.
- 8:02There's much to discover.
About this transcript
This page contains the full transcript of Majorana 2 Explained: The road to 1000x improvement by Microsoft, generated from the public captions YouTube serves with the video. The transcript has 1,379 words across 275 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.