Hacker News: steve_musk

New comment by steve_musk in "Quantum scientists deliver world’s first integrated circuit at the atomic scale"

steve_musk — Sun, 10 Jul 2022 15:21:26 +0000

Not sure it is really related to the original post but I think this is what they are talking about:

https://www.technologyreview.com/2020/01/04/131018/why-the-q...

Simulating matter on the quantum scale with AI

steve_musk — Thu, 09 Dec 2021 23:58:39 +0000

Article URL: https://deepmind.com/blog/article/Simulating-matter-on-the-quantum-scale-with-AI

Comments URL: https://news.ycombinator.com/item?id=29504821

Points: 7

# Comments: 2

New comment by steve_musk in "Scientists looking for aliens investigate radio beam from nearby star"

steve_musk — Sun, 20 Dec 2020 19:40:03 +0000

Maybe the media headlines are sensational (as they always are), and it’s unfortunate that it leaked before the papers were officially released.

But Jupiter emitting radio waves is really not relevant here. The signal they saw is a very arrow frequency, which pretty strongly indicates a technological origin. See [1] for a better overview of the situation.

[1] https://sites.psu.edu/astrowright/2020/12/20/blc1-a-candidat...

New comment by steve_musk in "Apple Silicon M1 Disruption"

steve_musk — Mon, 07 Dec 2020 01:16:32 +0000

Here they are[0]. The 4 performance cores + 4 efficiency cores of the M1 hold up pretty well against the 8 core AMD chips.

https://www.anandtech.com/show/16252/mac-mini-apple-m1-teste...

New comment by steve_musk in "macOS Big Sur"

steve_musk — Thu, 12 Nov 2020 19:06:48 +0000

With Big Sur safari now supports 4K YouTube videos

New comment by steve_musk in "How close are computers to automating mathematical reasoning?"

steve_musk — Sat, 29 Aug 2020 16:31:12 +0000

A human is no more capable of solving the halting problem than a computer.

And we aren’t doing some kind of magical logic when we humans create proofs. Sure naively general automated proof generation could be exponential in nature, but that hasn’t stopped us from mathematics research. There’s no reason it should stop computers either.

New comment by steve_musk in "TSMC and Graphcore Prepare for AI Acceleration on 3nm"

steve_musk — Thu, 27 Aug 2020 22:21:59 +0000

Intel 7nm doesn’t exist, and TSMC 5nm chips have been in production for months now.

New comment by steve_musk in "Belarus has shut down the internet amid a controversial election"

steve_musk — Fri, 14 Aug 2020 04:56:14 +0000

Or some Russian hardware mysteriously is “stolen” and ends up in Belarus with some little green men to operate it.

Would they need to shoot down the entire constellation? I could imagine them threatening to shoot it down (and maybe a few demonstrations) would be enough to convince SpaceX to stop sending signals into Belarus.

New comment by steve_musk in "Belarus has shut down the internet amid a controversial election"

steve_musk — Tue, 11 Aug 2020 21:56:21 +0000

Shoot down the satellites?

New comment by steve_musk in "Boredom is but a window to a sunny day beyond the gloom"

steve_musk — Sun, 16 Feb 2020 17:13:38 +0000

I assume the “Possible End of Progress” section is what the OP is referring to.

New comment by steve_musk in "WHO declares coronavirus outbreak a public health emergency"

steve_musk — Fri, 31 Jan 2020 01:03:38 +0000

No, I think you are misunderstanding the thread or projecting something onto the statistics.

New comment by steve_musk in "WHO declares coronavirus outbreak a public health emergency"

steve_musk — Thu, 30 Jan 2020 20:52:41 +0000

Just to put it in better perspective:

Coronavirus:

-5,974 confirmed cases

-170 deaths

-143 recoveries

-5,661 still sick

SARS:

-8,098 cases

-774 deaths

-7,324 recoveries

New comment by steve_musk in "My 2020 Hackintosh Hardware Spec"

steve_musk — Tue, 28 Jan 2020 22:08:43 +0000

In a way this isn’t a “similarly spaced Mac” though - it’s using Xeon CPUs, ECC RAM, server motherboard with 6 memory channels, professional GPU, etc.

The problem is if you want a Mac tower desktop they don’t offer anything with consumer components.

DeepMind developed model for mammography screening

steve_musk — Wed, 01 Jan 2020 22:27:07 +0000

Article URL: https://deepmind.com/research/publications/International-evaluation-of-an-artificial-intelligence-system-to-identify-breast-cancer-in-screening-mammography

Comments URL: https://news.ycombinator.com/item?id=21931176

Points: 4

# Comments: 0

New comment by steve_musk in "OpenAI Releases Largest GPT-2 Text Generation Model"

steve_musk — Tue, 05 Nov 2019 21:05:10 +0000

It’s very standard ML abbreviations.

New comment by steve_musk in "Microsoft Wins Pentagon’s $10B Contract"

steve_musk — Sat, 26 Oct 2019 12:58:24 +0000

What point are you trying to make? The employees make up the company.

New comment by steve_musk in "Tesla’s Neural Processor In The FSD Chip"

steve_musk — Thu, 26 Sep 2019 21:36:36 +0000

Sure if you’re driving only on the highway. What about driving around the city at 5-20 mph? Tesla’s end goal is to have autopilot do everything, so if they are serious about it then it makes sense to consider the impact at non-highway speeds.

Let’s say you are driving at an average speed of 20 mph - then by your estimate you’re using ~4666 watts. So if your chip is 500W it’s a significant portion of the power budget.

New comment by steve_musk in "iPhone 11 Pro and iPhone 11 Pro Max"

steve_musk — Wed, 11 Sep 2019 01:36:09 +0000

Look at the history of CPU performance over the past decade. If we were to get a 20% IPC increase in desktop processors it would be a big deal.

New comment by steve_musk in "Supercomputers will start building a 3D map of the world"

steve_musk — Tue, 13 Aug 2019 21:35:05 +0000

There is a difference between parallelizable and embarrassingly parallelizable. The former means that you can get better performance by dividing the work among different processors, while the latter usually implies that the work can be dividing into independent units of work that don’t need to communicate with each other.

A supercomputer typically means that those thousands of cores are connected with fast and expensive interconnects so that the cores can communicate with low latency. A large portion of the budget is usually spent on this interconnect. If you have an embarrassingly parallel problem and you run it on a supercomputer then that expensive interconnect is sitting idle - you would get the same performance on AWS or a more standard compute cluster.

New comment by steve_musk in "Raspberry Pi 4 Review"

steve_musk — Tue, 16 Jul 2019 16:01:22 +0000

>Like your PC and smartphone

This isn’t necessarily true for your smartphone, because some ARM CPUs don’t do the speculative execution that Spectre/Meltdown exploit. For example, CPU in the raspberry pi 3 didn’t have these vulnerabilities which I assume is why the parent was asking this question.