Neutrinos, Book Club, Air Conditioning. July 13, 2018, Part 1
Science Friday
Science Friday and WNYC Studios
4.4 • 6.3K Ratings
🗓️ 13 July 2018
⏱️ 47 minutes
🔗️ Recording | iTunes | RSS
🧾️ Download transcript
Summary
Transcript
Click on a timestamp to play from that location
| 0:00.0 | This is Science Friday. I'm Ira Flato. A little bit later, we'll be having our book club, where we'll be talking about Stephen Hawking's a brief history of time. But first, people like to say that our brain is like a computer, right? It's a usual metaphor. But in reality, no computer comes close to matching the power and complexity of the human brain. |
| 0:21.8 | But that hasn't stopped technologists from trying to build an even better computer that acts even, well, closer to the human brain. |
| 0:29.2 | And one company is closer now to that goal. |
| 0:32.5 | Here to tell us more about this milestone in brain-inspired computing is Amy Nordrum, news editor at IE CREPRIE Spectrum. Well, there's all other selected short subject in the size. Hi, Amy. Hi, Ira. So what's the breakthrough reach this week? Yeah, well, this field of neuromorphic computing, as it's called, has a bunch of big research projects going on right now, and one of the biggest is called Spinnaker out of Europe. |
| 0:55.1 | And this week, a group related to that project announced that they had done the largest |
| 0:58.8 | simulation on their computer to date. So they did a simulation of a neural circuit that contained |
| 1:04.1 | 80,000 neurons and 300 million synapses. So they were able to run this simulation and then |
| 1:09.4 | compare it to traditional software running on |
| 1:11.6 | supercomputers today that's typically used to do these kinds of neural simulations. |
| 1:15.6 | And they found that the results of their new hardware, their new system, were as accurate as this top-of-the-line state-of-the-art neural simulation software that most scientists are using on supercomputers today. |
| 1:29.3 | So this is a good step for the project. |
| 1:31.2 | It means that they're on the right track. |
| 1:35.3 | And then as they continue to scale up their neuromorphic computer called Spinnaker, |
| 1:39.1 | the project will hopefully be able to produce results that are even more accurate and do it with less power and faster than the supercomputers that are used today for these kinds of things. |
| 1:48.0 | And why do they think they need to create such a computer like this? They just wanted to get, to recreate the human brain? Is that the idea? |
| 1:52.0 | Yeah, the thinking is that since our traditional computers really aren't designed, |
| 1:56.0 | similar to the brain, if you really want to study the brain in depth, if you want to run a lot of simulations on it, |
| 2:00.0 | learn more about how the brain learns and adapts and the plasticity elements of it, you really should design a computer that works a lot like the brain. And so in theory, if they continue to scale this project up, they'd be able to do a lot more accurate simulations of our own brains. What if you could get close to that? Think of the ethics and ethics considerations here. Absolutely. I talked with the researcher involved with this project. |
| 2:21.4 | Lost. If you could get close to that. Think of the ethics and ethics considerations here? |
| 2:18.0 | Absolutely. |
| 2:18.6 | I talked with the researcher involved with this project. |
| 2:21.3 | Last night, Sasha Von Albaata and was asking her about this possibility. |
... |
Please login to see the full transcript.
Disclaimer: The podcast and artwork embedded on this page are from Science Friday and WNYC Studios, and are the property of its owner and not affiliated with or endorsed by Tapesearch.
Generated transcripts are the property of Science Friday and WNYC Studios and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.
Copyright © Tapesearch 2026.

