biswaroop

I didn't read a long term bet against a face-based device. He focuses on facebook's value as an app-based stimulant, and how VR occupies a completely different (and possibly limited) niche. In fact, Stratechery claims that AR is more promising, and that Apple is well-poised to create a face-based device. His main point is: unless the face-based device joins us in our every day activities like the smartphone, it's not going to be directly useful to facebook.

A cool implementation of a deep ANN using a similar system: https://arxiv.org/pdf/1610.02365.pdf

(the nonlinear units and the backprop are done on a classical computer).

The paper also points out that thermal cross-talk between the thermo-optical phase shifters can limit gate fidelities or conversely the smallest spacing between the waveguides.

I highly recommend Dubliners by Joyce. It's the most beautiful collection of short stories ever written. Ulysses has its roots in Dubliners, and if Ulysses is a daydream, then Finnegans Wake is a nightmare. Leave the nightmares to the hobgoblins.

I'm pretty sure 3M and DuPont sell more than 2 part repair epoxy.

I don't see how systematically accurate image classifiers and facial recognition systems built on deep learning is a 'dead end'. Products are products. If deep learning has led to actual profits in actual companies, it's not a dead end. As to whether this leads to AGI is a completely different question.

Kim by Rudyard Kipling.

Despite Kipling's jingoism in other works, Kim spoke to me in school. It opened me to a world of adventure and fearlessness that I didn't know was possible. I really haven't been the same since I read that book.

But you were talking about the mechanism of consciousness, not verifying whether someone is conscious or not.

It may be impossible to verify if someone is conscious in the same way that I know I am, but we can associate a set of behaviors with an intelligent creature. If we empirically prove that those behaviors emerge from a collection of neural impulses, then it takes a simple assumption (almost a cognitive axiom) that the creature exhibiting those behaviors is conscious, to conclude that emergentism has been empirically verified.

I feel like your position that emergentism is empirically impossible to verify is itself a strong position, and attempts to define a boundary on science when your stated goal is to place doubts on where the boundaries are. Yes it requires a few assumptions, and yes those assumptions are impossible to verify empirically. But then so are the mathematical axioms. We don't go around saying "physics is impossible to empirically verify because mathematical axioms are impossible to prove and are incomplete"

Curious, why is this a bad thing? Also it seems like you can change it to a lower number if you want.

Well it's not completely clear to me that cameras can't look a few cars ahead. If we can, cameras can too. Maybe they're not smart enough yet, but I don't see the fundamental limitation. Apart from that, I agree that cheap + effective LIDAR would help significantly to get autonomous vehicles on the road now.

Seems like it happens whenever a car in front moves out of the way. It goes full stupid-cruise-control mode and just accelerates to the stored speed. Of course there's a lot of noise in terms of things in the stationary reference frame, but still, big red firetrucks don't exactly look like lane markings.

At the very least, as a driver, I'm slightly cautious when I see a car move out of the lane in front. I feel like the solution lies in: (a) separating the signal from the noise in the set of things moving with the ground, and (b) getting the car to 'understand' why other cars change lanes. Sometimes it's so they can avoid obstacles in the road. It almost feels like cars need to have a theory of mind of other cars. I certainly drive with an intuition of driver intentions stuck in my head.

But the Autopilot has cameras, and therefore it has 3+1D data (2D frames, 1D inferred depth from stereo or training, 1D time from frame diffs). Even with a 30fps frame rate at 60mph, two frames from a video feed would see the truck move closer by a meter. The network can be trained to distinguish a truck in the lane from an overpass, so those two frames should be sufficient to signal danger. Instead, the car accelerated over 70 meters before the human, again using just her eyes, applied the brakes.

This is an edge case, and clearly something that requires more training to avoid. The challenge in edge-case learning is to encounter them often enough, or perhaps create them artificially, and then to ensure the learning can be transferred and generalized to all similar situations.

Um, cars kill astronomically more pedestrians than do bikes.

Seconding the recommendations for a cheap trainer + Zwift. Also, if the weather is good, find a local club to ride with outside. Although indoor training is great, nothing releases endorphins like an outdoor ride with some friends.

I think what he/she means is poorly designed visualizations. Just because a plot is grayscale and not interactive doesn't mean it's worse than a cluttered poorly-designed super interactive web widget. It's a poor choice of wording, but I think by "overuse" they might mean "unclear but eye-catching". Besides "overuse" is literally the quantity that is excessive.

This is interesting: the difference in 'screen time' a few decades ago versus today. Is the incredible productization of screen time a recent phenomenon? Could we improve screen time by restricting the access large social media companies have to our brains? Can we convert, or perhaps return, the computer from an interactive television to a digital canvas for hackers?

More on the "scrollytelling" they used here: https://pudding.cool/process/how-to-implement-scrollytelling...

I personally found it a great way to tell the story. It didn't interfere with my reading experience at all.

Profession: Physics graduate student (experimental atomic physics). ~15 people in our group, 5 in our lab.

Workday:

[07:30] Wake up

[08:00] Swim, bike, run or climb for an hour

[09:00] Bike to work, shower, read papers

[10:00] Turn on experiment, align optics

[01:00] Lunch with labmates and advisor

[02:00] Back in the lab, more laser aligning, bug fixing, realizing nothing works.

[07:00] Dinner at student center or Chipotle

[08:00] Data taking starts (sometimes we're lucky and it starts earlier)

[10:00 - 12:00] Data data data. If it's super late (4am), then I'll get up late tomorrow

Meetings can last over 5 hours, especially if it involves writing papers. The schedule usually shifts to later hours as the week progresses. We barely wake up in time before the Friday 11am group seminars. 15 hour days are fairly routine.

Wow I'm a fan of both Satyajit Ray and Herge! It would be amazing to see that sketch. Could you share a picture or scan of it?

>It's not like there is a fixed number of jobs mined from the soil of each country.

This is so true. It applies to any system with competition. People tend to treat things like jobs and startups like natural resources, and tend to assume that competition is necessarily bad. In reality, feedback processes and strong interactions between players turn economies into highly unpredictable emergent phenomena. It's more like a flock of geese than an elementary chemical reaction.

For such large systems, it's hard to tell what the Nash equilibrium is, and whether it matches with the optimum global outcome. Animals (especially humans) have evolved to cooperate well in small groups (e.g. tribes, families). Over larger scales (e.g. countries, between language families), we have a strange balance between distrust, competition, and compassion.

A latex tip: hover over equations in Wikipedia to get the syntax.

This article is about marriage rates, not birth rates.

Also, why are millenials unable to replicate memetically? I don't see any evidence of that. If we're determined to effect change through technology and environmental activism, surely we're capable of spreading these ideas far faster than breeding. Surely working harder to provide a stable human population with access to healthcare, education, and security is a far higher goal than "marry and make babies"

How about "Don't ruin your future"? It doesn't matter what the cause is: "man", "good man", "purple man", "flesh-eating bacteria", etc.

A lot of the tests of single-photon entanglement require coincidence measurements: being able to detect if two photons arrived at the same time. I'm not sure how fast the human brain would be for these kinds of measurements. People usually use FPGA-based amplifiers with fast photodiodes.

Nobody's mentioned Strava yet. (Other sports trackers work as well)

Cycling and running every morning does the trick for me. Starting the day early with the burning desire to complete something in the shortest possible time leaves you applying that mindset to everything else you do. Sadly, this also means I'm noticeably less efficient on days I don't exercise in the morning.

Keeping track of my rides and runs gives me a good perspective on competition. I'm trying to beat my own records, and see how I compare to others. I feel a constant need to improve, and this helps me push myself in a healthy way in other tasks.

I think you're confusing monolingualism with everyone having a lingua franca. Your argument sounds like "We should all just have one common language; therefore, we don't need others."

Bilingualism has existed in countries for centuries without either language dying out. I'm all for everyone speaking English. I want people to know as many languages as possible. What I don't want is everyone speaking only English.

It's not ok to lose some of our knowledge. Uniformity is bland, dull, uninteresting, and sad. There's a reason why we think knowing a lot of languages is cool. There's beauty to be found in other cultures and their languages. I don't want to force everyone to learn languages, but it'd be tragic if we lost French and with it Hugo, Voltaire, Verne, Baudelaire, Dumas, Asterix and the rest. It just isn't the same in English.

Turkey and Indonesia do not use Arabic for communication. There's an Arabic continuum, and Saudi Arabic is rather different from Moroccan Arabic. There are subtle cultural differences preserved in the diversity of language.

England, Kenya, and India all speak English, but English is merely a common language, and a lot of cultural differences come from the linguistic diversity in each of those countries.

Culture is a result of multiple factors, one of which is language. Monolingualism is certainly one large step towards a homogenous culture.

Fascinating. It took me a while to realize that the symbol that looks like 4^E is actually 'ye', or 'the'.

Wow incredible pictures. Also, 4200km is in the exosphere: I think you mean 4200m :)

> all the missed years of stargazing

Thanks! I stargazed nevertheless with a pair of binoculars. It's tough (but sadly possible) to prevent people from doing things they really want.

It's quite ridiculous. In 1997, my parents brought a telescope to India. The Indian customs claimed my parents would "spy on Indian satellites" and never released it. As a young astronomy fan, this made me particularly sad.

In 2013, I went backpacking with a GPS in the himalayas. I was stopped by a park ranger who searched my bag, and claimed that a GPS can be used for 'terrorism'. He took away my ID (but luckily not my GPS), and I could only pick it up when I left the park 3 days later.

Later on the trip, I witnessed incredible amounts of pollution generated by the armed forces in the border regions. Not to mention the consistent human rights violations perpetrated by the Indian Army in Kashmir, all in support of "security".

I don't want to go on a long political rant against the Indian defense system, but I really wish my childhood telescope wasn't stolen.