Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Google thinks the sun is 15.81 light years from earth (google.com)
125 points by ryanatallah on Jan 23, 2018 | hide | past | favorite | 64 comments


Google was also quite confident that the alphabet starts with J and ends with G. https://twitter.com/ridiculous_fish/status/86609697117344972...

Just tried it again, now it both starts and ends with A. 50% improvement ain't bad.


Distance Earth-Sun = 1 AU

1e6 AU = 15.81 light years.

Somebody did not pick up the 'e6' part...


Pretty much. The sidebar references Wikipedia's List of brightest stars [1]. That table has the distance from Earth to the Sun in lightyears. So maybe a parsing problem with 0.000015813. On the January 21st this was changed from 0.000 015 813, so it may have been parsing that wrong.

[1] https://en.wikipedia.org/wiki/List_of_brightest_stars#Main_t...



Huh... with JavaScript off, I get the answer "Distance from Sun: 92.96 million " (presumably miles) on the right hand side. With JavaScript on, I get the wrong answer from the title.

Edit: Actually, it gets the same answer on the right side with JS enabled; didn't notice that earlier. So just the instant answer is missing in the JS-disabled case.

Also, here's a screenshot for posterity (with JS enabled): https://i.imgur.com/8kyI523.png


With JS disabled I get "Distance from Sun: 149.6 millio...", presumably kilometers. Incidentally, thank you Google for filling half of my monitor with whitespace...


Screenshot from Brazil: https://i.imgur.com/ODEtS0Q.png



I tried now on my Google Home, same query, same result.


Google seems to think that Earth is Gliese 412

Gliese 412 is a pair of stars that share a common proper motion through space and are thought to form a binary star system. The pair have an angular separation of 31.4″ at a position angle of 126.1°.[12] They are located 15.8 light years distant from the Sun in the constellation Ursa Major. Both components are relatively dim red dwarf stars.

https://en.wikipedia.org/wiki/Gliese_412


It's still cloudy here in Austria, I'll count the number of suns later when the clouds are gone.


I in the UK and I can't believe people still fall for the "sun" hoax. I don't know anyone who's ever seen it.


Lived in Seattle. We had a nice day of summer once.


Wait, what about the whole sum... ohhhhh, right, clouds. Clouds. Always clouds.

No summer ever.

Clouds always.


Just looked outside in Seattle. I can confirm there is no sun.


No clouds over here in 'Straya. Bit of a scorcher though. Don't let 'em fool you - my mate Bazza reckons he counted four suns once. He loves a good yarn but he's dinky-di and I don't reckon he's taking the piss. I've never had the balls to count the suns me-self - Bazza's built like a brick shit-house but he was right up shit-creek when he done it. Came good after a week or so though. Reckon I'd cark it well before the ambos had time to come get me.


Not so cloudy in Burgenland, I only see one. But I don't have a telescope with a sun filter to scientifically prove it.


I got confused at your mention of "Burgenland" together with your nick, which led me to find out that "Burgenland" and "Burgenlandkreis" (Burgenland county) are two different counties in two different countries:

https://en.wikipedia.org/wiki/Burgenlandkreis https://en.wikipedia.org/wiki/Burgenland


Well that's new to me too. Thanks for the links!


Isn't the earth flat in burgenland?


No, it's an ellipse with infinite radius.


I rather think it's a decimal error. The distance to the sun is 15.813 μly away from earth.

https://www.wikiwand.com/en/List_of_brightest_stars#/Main_ta...


As of this posting, if you put a "what" in front of the query, it returns the correct average distance. I suspected the parser, even if it was learned, is sensitive to the query format.

Despite what the press teams of large companies will tell you, our ability to model language is still in its early infancy.


If they're gonna push their silly instant answers, they'd better make sure the results are correct...


Google Search so aggressively massages my queries that it's become almost unusable.

I recently searched "80's rom-coms" and an instant answer came up on top. It was a list of 90's rom-coms. Similar, yes -- but not at all what I typed, and completely useless to me.


That’s just general search. When looking for code issues and snippets it’s worse

Don’t even get me started with github. Half of humanity’s code seems to be in some bucket that doesn’t understand punctuation when you search


bing seems to get this right in both instant search and otherwise.


I'm slowly switching to Bing--not because Bing is getting better, but because Google is getting worse. Google consistently ignores key terms in my queries, and searches are getting slower and slower. Remember when Google used to have that little timer in the top-right corner of the search results? Not anymore...


I'm glad it's not just me.


If you look at my post history, at some point I used google to determine the average words spoken per day. I initially trusted Google's answer indicating women speak 2 to 3 times as many words per day on average as men (which is not true).

A quick Google search on my phone for "average number of words spoken per day" gives similar results right now. Although the text starts with an ambitious sounding phrase "previous research" it ends in present tense and has 7000 in bold for men and 20,000 in bold for women.



let's not be so quick to be so harsh. Remember the glee WolfAlpha provides us with their interactive visualizations https://www.youtube.com/watch?v=Slop2Vi8cKE


Bing also seems to give something sensible.


WA thinks the distance is 0.9843 AU...


That's actually correct at the time of this post :)

AU is an average distance from Earth to the Sun. Since Earth's orbit is elliptical, it will be closer than 1AU at certain times of year, and further than 1AU at other times in the year.


That's because it is. The Earth's orbit is elliptical, varying between 0.983 and 1.017 AU at its minor and major axes.


That is probably the current distance. 1 AU is the mean distance.


https://en.wikipedia.org/wiki/Astronomical_unit:

”Originally conceived as the average of Earth's aphelion and perihelion”

https://www.britannica.com/science/astronomical-unit:

”a unit of length effectively equal to the average, or mean, distance between Earth and the Sun”

I wondered whether the two are equal. A bit to my surprise, they are (http://www.farmingdale.edu/faculty/sheldon-gordon/RecentArti...)

(If we take into account that earth moves faster near its perihelion, I think that will break down)

And, nitpick, nowadays, the AU apparently is exactly 149,597,870,700 meters.


The distance between the sun and the earth is not constant, whereas the AU is.


Not only is Google "borrowing" content from others to show directly in the search results, more often than not it's completely wrong. Either because they have failed to parse the data correctly, or because it's from some shady web page (happens when googling stuff related to vaccines for instance).


"Earth distance to sun" gives the correct answer: https://www.google.com/search?q=earth+distance+from+sun&oq=e...


Interesting. It says 149.6 million km, which is correct.

[distance earth to sun in km] gives 149.6 trillion km, which is 15.81 light years.


the magic of machine learning, everybody


they are using AI, not machine learning.


I'm not sure I like Google's attempt to solve global warming.


My son likes to ask our Google Home the distance to different planets. The numbers, though generally technically correct (sun has worked in the past), are misleading as some are minimum, some average distance and some maximum.

For example "distance to mars" says 54.6 million km, which is the theoretical minimum. "distance to venus" says 261 million km which is maximum. I believe it was Jupiter that previously gave an average distance but now I'm seeing minimum.


By the "[maximum] distance to different planets" do you really mean literally that, or do you mean the maximum distances to their orbits?


Wouldn’t those be the same? The periods of each planet’s orbit aren’t factors/multiples of one another, so eventually any two will be at opposite points in their orbit.


The maximum distance between two ellipses is not the same as the maximum distance between all pairs of points on them right? Like just imagine measuring the maximum distance from a circular orbit to itself. One measure would give zero (zero separation between the orbits) and the other would give twice the radius (individual points can be 2R away from each other).

If I'm doing this correctly I think the max distance between two points would come out to max_x max_y ||x - y|| whereas the max distance to the orbit would come out to max_x min_y ||x - y||.


There are orbital resonances, but you are right that none involve planets. https://en.wikipedia.org/wiki/Orbital_resonance

The closest is the 2:3 ratio of Pluto:Neptune. (It counts if you consider Pluto a planet. ;)

Otherwise, there's a nice table showing "near-integer-ratio relationships between the orbital frequencies". If I understand it correctly, each of the bigger-than-a-dwarf-planet planets will be at their furthest possible distances from each other in timescales no longer than about 50,000 years.


What's the difference?



A little parsing error here, a little parsing error there...


Someone should win a bug bounty for that one !


But it seems to get it right if you ask the same question in the opposite order https://www.google.com/search?q=Earth+Distance+from+Sun


Have a glimpse at the kind of errors we will be facing with the machine-learning AI in the near future.

How do we debug them and how do we know there is an error to correct in the first place?


Google's systems do not, "think."


Having interacted with a lot of computers over the years (including Google's for the last several of my employment), the machines at Google come as close to thinking as I've ever encountered. Two or three times a day they come up with answers that I cannot begin to comprehend how they arrived at. Sometimes they are brilliant, sometimes they are dumb, and sometimes they are merely mad.

They give every coworker I've ever had a run for their money on catching my fuckups.


Would you really prefer an accurate title like, "Google's machine-learning instant-answer algorithm's output is '15.81 light years from earth' when the input is 'how far away is the sun from the earth'"?


We should send satellites to Alpha Centauri before Google figures this out.


If you google distance between Sun and Earth, Google gets it right.


the AIs are coming to kill us any day now


At this rate, they're more likely to jump off a cliff than kill anyone.


okay, but only for at least as a grave mistake as google's.

wait. when people die for no reason whatsoever, that counts, right?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: