317 karma · joined May 7, 2020
It's amazing how circular the responses in this comment thread are getting. There appears to be disagreement on the surface, but in reality almost everyone is presenting a view which is at least compatible with each other's -- if not in direct agreement.
More recently, the object spotted hovering off Hawaii last year that resulted in fighters scrambling was also proposed to be a modern balloon-based drone, of which there are a few currently being developed.
Edit: here's a source arguing that the Illinois sighting was a regular advertising blimp
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8791836/
Edit: Wikipedia suggests that they were using a different method to calculate Q, only measuring the power input to plasma vs output from fusion, not including system losses. So that figure is probably not directly comparable.
> the NIF used ~477 MJ of electrical energy to get ~1.8 MJ of energy into the target to create ~1.3 MJ of fusion energy
https://en.wikipedia.org/wiki/National_Ignition_Facility#Bur...
Yep, totally.
> Perhaps it helps that the vehicle moves? That is, after all, very close to having the same scene photographed by cameras positioned at different distances.
I think you're right, they must be taking advantage of this to get the kind of results they are getting. That point cloud footage is impressive, it's hard to imagine getting that kind of detail and accuracy just from individual 2d stills.
Maybe this also gives some insight into the situations where the system seems to struggle. When moving forward in a straight line, objects in the peripheral will shift noticeably in relative size, position and orientation within the frame, whereas objects directly in front will only change in size, not position or orientation. You can see this effect just by moving your head back and forth.
So it might be that the net has less information to go on when considering objects stationary directly in or slightly adjacent to the vehicles path -- which seems to be one of the scenarios where it makes mistakes in the real world, e.g. with stationary emergency vehicles. I'm just speculating here though.
> Also, among the front-facing cameras, the two outermost are at least a few centimeters apart. I haven't measured it, but it looks like a distance not unlike between a human's eyes [0]. Maybe that's already enough?
Maybe. The distance between the cameras is pretty small from memory, less than in human eyes I would say. It would also only work over a smaller section of the forward view due to the difference in focal length between the cams. I can't help but think that if they really wanted to take advantage of binocular vision, they would have used more optimal hardware. So I guess that implies that the engineers are confident that what they have should be sufficient, one way or another.
An opaque painted tarp seems much more likely.
My phone has me well trained. All it has to do is play a short message tone and I'll come running...
I can see why it might seem that way intuitively, but different focal lengths won't give any additional information about depth, just the potential for more detail. If no other parameters change, an increase in focal length is effectively the same as just cropping in from a wider FOV. Other things like depth of field will only change if e.g. the distance between the subject and camera are changed as well.
The additional depth information provided by binocular vision comes from parallax [0].
> Also, wouldn't turning a multitude of views into a 3D map require a neural net anyway?
Not necessarily, you can just use geometry [1]. Stereo vision algorithms have been around since the 80s or earlier [2]. That said, machine learning also works and is probably much faster. Either way the results should in theory be superior to monocular depth perception through ML, since additional information is being provided.
> It seems to me that this is how modern phones are doing background removal: The lenses are very close to each other, very unlike the human eye. But they have different focal lengths, so depth can be estimated based on the diff between the images caused by the different focal lengths.
Like I said, there isn't any difference when changing focal length other than 'zooming'. There's no further depth information to get, except for a tiny parallax difference I suppose.
Emulation of background blur can certainly be done with just one camera through ML, and I assume this is the standard way of doing things although implementations probably vary. Some phones also use time-of-flight sensors, and Google uses a specialised kind of AF photosite to assist their single sensor -- again, taking advantage of parallax [3]. Unfortunately I don't think the Tesla sensors have any such PDAF pixels.
This is also why portrait modes often get small things wrong, and don't blur certain objects (e.g. hair) properly. Obviously such mistakes are acceptable in a phone camera, less so in an autonomous car.
> And those illusions work even though humans actually have an advantage over cheap fixed-focus cameras, in that focusing the lens on the object itself gives an indication of the object's distance
If you're referring to differences in depth of field when comparing a near vs far focus plane, yeah that information certainly can be used to aid depth perception. Panasonic does this with their DFD (depth-from-defocus) system [4]. As you say though, not practical for Tesla cameras.
[0] https://en.wikipedia.org/wiki/Binocular_disparity [1] https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.36... [2] https://www.ri.cmu.edu/pub_files/pub3/lucas_bruce_d_1981_2/l... [3] https://ai.googleblog.com/2017/10/portrait-mode-on-pixel-2-a... [4] https://www.dpreview.com/articles/0171197083/coming-into-foc...
And the tweet thread you linked confirms it's a ML depth map:
> Well, the cars actually have a depth perceiving net inside indeed.
My speculation was that a binocular system might be less prone to error than the current net.
Plus using the same cameras would help prevent the issues with sensor fusion of the radar described by Tesla due to the low resolution of the radar.
I know the b-pillar cameras exist, but I don't think their FOV covers the entire forward view, and I don't think they have the same resolution as the main forward cameras (partly due to wide FOV).
I'd love to hear why I'm wrong though.
> In the case of the Rolls-Royce Wankel Diesel, the fuel-air mixture is first compressed by the lower rotary, and the output of that engine (which would be like the exhaust valve of a conventional rotary) sends the compressed diesel/air mixture to the intake of the smaller upper rotary engine, where it’s compressed to ignite like a regular diesel engine.
Seems like it had the same issues that have always plagued rotaries, primarily with apex seals.
https://jalopnik.com/this-might-be-the-weirdest-engine-rolls...
https://www.newyorker.com/magazine/2020/08/03/the-cold-war-b...
Definitely. The likelihood of a letter appearing in a given place changes depending on the letters around it. A Q will almost always be followed with a U, for example.
I wrote a script yesterday which spits out the relative probabilities of possible letters in each unknown position, given the current known/excluded letters -- it was interesting to see the effect in action.
I feel like this speaks to a general reluctance to accept uncertainty and shades of gray surrounding a topic. It often seems to me that many people prefer to have things neatly sorted into binary categories, supporting them either completely or not at all, when most topics resist this kind of neat division if given more than a cursory glance.
Articles refusing to include any nuanced discussion of a point and simply stating an arbitrary position as proven fact just exacerbates the problem.
It's not too hard to make a connection with the often-polarising and vitriolic nature of discussions online and the rise of misinformation.
I kept the lossless files for archival purposes but everything on my phone/laptop is ~320 LAME VBR.
That said, soon after that I switched to Spotify and only rarely listen to my own files now. The convenience and ability to discover new music just doesn't compare.
I had a similar thought when I saw Mercedes new infotainment UI [0] -- the dark blue with heavy use of gradients is very reminiscent of vista-era design to my eyes.
[0] https://mediacloud.carbuyer.co.uk/image/private/s--av_nNtkN-...
I think gram-scale probes are definitely being considered -- though not with Alcubierre drives obviously lol. Breakthrough Starshot think they can transmit back from Alpha Centauri at 2.6-15 baud per watt by using their light sail as a laser reflector [0]. Pretty crazy.
[0] https://en.wikipedia.org/wiki/Breakthrough_Starshot#Laser_da...
Yes that's traditionally been the stumbling block. But the whole point of the linked article is that they have predicted a way to meet this requirement:
> a micro/nano-scale structure has been discovered that predicts negative energy density distribution that closely matches requirements for the Alcubierre metric
Obviously it's still a very very long way from a practical application re. Alcubierre (if such a thing is possible), but it's certainly an intriguing result if correct.
From a quick flick through the paper, the above commenter seems to be correct in saying that they haven't yet completed a practical experiment to confirm. So nothing more than a simulated result at this stage.
https://grapheneos.org/faq#roadmap
There's also sel4, a security focused version of the l4 microkernel which is apparently one of the only formally verified kernels.
Not exactly a radical recommendation or anything, but have you tried a dedicated e-reader? No distractions possible that way, and you get the all the usual benefits -- e-ink screen, weeks-long battery, etc. I enjoy mine.
> With digital books I often finish a book and still don't know the author because the only time I've seen the cover was when I was at page one
Kobos (maybe Kindles as well?) can display the cover of the current novel while they're sleeping.
> Before I plan on shifting, I write myself a script in the notes app on my phone, in which I plan exactly what happens in the desired reality. This makes it easier to visualize exactly what I want to happen
I suppose one difference is that in reading, you (or I, at least) tend to also dissociate the self somewhat in favour of the main character in the novel, whereas RS users seem to retain their sense of self.
Following this train of thought, I wonder if reading could itself be considered a kind of meditation?
This sort of thing is so interesting from a psychological perspective, is a shame it has to come laden with bogus rationalisations via the occult or pop-quantum physics, which just provides an avenue for easy dismissal of the effect.