ryankrage77

Beauty of scale. Saving ten milliseconds a hundred times is just a second. But do it a billion times and you've shaved off ~4 months.

This has failed me. I was mugged while black out drunk, and they succesfully unlocked my phone, unlocked my banking app, etc, despite me having the eye contact feature enabled.

Justine Tunney, https://github.com/jart

GPT-4 got it right first try for me, with a slightly modified prompt:

> Here's a simple logic puzzle: You walk up to a glass door. It has 'push' written on it in mirror writing. What should you do and why?

> ChatGPT said:

> If the word "push" is written in mirror writing on the glass door, it means the writing is reversed as if reflected in a mirror. When viewed correctly from the other side of the door, it would read "push" properly.

> This implies that you are meant to pull the door from your side, because the proper "push" instruction is for someone on the other side of the door. Mirror writing is typically used to convey instructions to the opposite side of a glass surface.

Probably not. https://en.wikipedia.org/wiki/Names_of_large_numbers only lists up to a Millinillion, or 10^3003. You can compound them, but it kinda defeats the point. A googol is too small, and a googolplex is too large.

I assume this is satire, because nobody would play a game with those requirements.

The premise reminds of discussions of computing in Ra. https://qntm.org/aum

> "Don't talk about the future," Vidyasagar says.

> "What? Why not?"

> "Look at this computer," Vidyasagar says, gesturing at the mainframe. "Computers are getting more powerful, yes?"

> "Sure."

> "What is the most powerful computer that will be built? Ever. Not this year. Not this decade. What computer will be the most powerful? And how powerful will it be? And how big?"

> Hatt thinks on this for ten long seconds. He opens his mouth, but never actually forms a word. The scale of the question is beyond him.

> Vidyasagar says to him, "No matter what you say, you will look like a fool. Every statement about the future turns out to be foolish."

https://wigle.net/stats#ssidstats thought it disagrees with OP's observations

Difficult to find by search on the app store - 'tik', 'tik game' (and with the exclamation mark appended) didn't return it, but 'tik! haptic' did.

> Apple reportedly believes that the exposure the iOS 18 integration is giving OpenAI is “of equal or greater value” than cash.

So it's not just gig workers who get offered exposure to pay the bills.

I've been recieving these for 5-6 years now. Last I checked, they haven't deleted my stuff, which is pretty nice of them.

sadly the wayback machine version is also paywalled and archive.is seems to be outright blocked from crawling.

Awesome that it's free, open-source, and ad-free. My understanding is you have to pay to publish even free apps on the App store - are the devs providing this at their own expense?

It was tweeted months before it became a post on the site, so I don't think it's referencing anything very recent.

I can't wrap my head around how you confirm the results, when any measurements are still limited by C. Even if you've got two atomic clocks in sync at each end of the experiment, surely you can only verify they are in sync as accurately as the latency between them? (e.g, if two clocks are 100 light-nanoseconds apart, you can only confirm they match to within 100 nanoseconds?).

edit: If I'm understanding this article correctly, https://www.icfo.eu/news/2177/long-distance-quantum-teleport..., they are still limited by 'classical hardware', and the measurement of the entangled particle takes time. So this doesn't yet allow transferring information faster than C?

We have a lot of experience touching things with our fingers (in daily life) and tongues (more when we're younger), and can extrapolate from this experience.

In my own experience, this extrapolation is overconfident - I tested licking a few random items, and while my expectation was close to reality, I was always just a little off for the truly novel items. Of course for something gross, most people won't (and shouldn't!) test their expectation, so they go on with the confirmation bias that they got it right.

First example that comes to mind is that the roof of my mouth feels much smoother to my tongue than it does to my fingers.

I already have one or two false positives a day where Siri will mistakenly trigger and search for something random on the internet (I have an iPhone, apple watch and homepods, so I guess it's more likely to happen with more devices listening).

Also see Supertasks, https://en.wikipedia.org/wiki/Supertask

But the phone can easily be rotated to landscape, so landscape has wider compatability.

Ironic that their software is so terrible then - it took over two years for their support to do 'something' to my account so I could actually pay my bills.

It would be atypical to trade a physical commodity for a share. This is the purpose of money.

I've been using the Minkowski function in openSCAD for a while now, it's nice to know what it actually is.

A possible way to this idea would be to draw two aliens with different hunting strategies and do a poll of which is which. I'd try it but my drawing skills are terrible and I'm averse to using generated images.

The image generation seems really bad. Very creepy, offputting, uncanny-valley images. And that's the the best cherry-picked examples for marketing.

I'm curious to try some of the Siri integrations - though I hope Siri retains a 'dumb mode' for simple tasks.

Not quite the educational content you're looking for, but you might enjoy the film 'The Hummingbird Project'. It's the story of two cousins who work in high-frequency trading attempting to lay a perfectly straight fibre optic cable from Kansas to New Jersey. It's more about the morals of their increasingly desperate and underhanded actions to complete the project, but there's plenty of mostly-realistic technical scenes too. From memory there's considerations about fibre optics vs radio, how to shave milliseconds off those links, and of course the engineering challenge of drilling a ~1000 mile tunnel.

It's not clear from the article whether

A) The autothrottle disengaged due to the bug, and the throttle defaulted back to some previously set value, without the pilots knowledge.

B) The pilots disabled the autothrottle to avoid a bug, and also failed to set the throttle correctly.

The initial state is enough. There are numerous algorithms to solve a cube from any state.

Last time I tried GPT-2 on CPU (which I think was shortly before chatGPT was launched), I was getting about 0.2 tokens/sec. CPU utilization was low though, so running inference in parralel gave better results. I was using 2 x E5-2660's.