There's a number of things that are alarming me about this, but I'll focus on one thing: the privacy implications. Not just for the user, but for those who will appear in the pictures that Google will be - for sure - ingesting for the purpose of crowdsourcing street view imagery. Suddenly we have this platform that will be capturing millions of faces and potentially know where people have been even if they don't use the app.
Not to mention that this type of app is treating people as incapable of following directions and it opens up the possibility of new avenues of advertising. Google already keeps an archive of your whereabouts and can even tell when you get in and out of a car. Just imagine wielding your phone around and getting "special offers" for that store you go by but never walk into.
> There's a number of things that are alarming me about this, but I'll focus on one thing: the privacy implications. Not just for the user, but for those who will appear in the pictures that Google will be - for sure - ingesting for the purpose of crowdsourcing street view imagery. Suddenly we have this platform that will be capturing millions of faces and potentially know where people have been even if they don't use the app.
Umm, the camera stream of AR isn't transmitted from your phone. The 3D objects are rendered on top of detected features on device.
> How do you know? And even if it doesn't for starters, there's no telling it'll stay that way.
Same way I know there's no pagan god hiding in my phone whispering my secrets to evil aliens: by actually working with the technology, observing the behaviour of my device and reading the privacy policies and activity log published by Google and not finding any proof.
Of course you can argue that Google might lie for whatever reason... but then you probably shouldn't be using Google Maps at all. Or any product built by a corporation.
Also by watching the upload bandwidth. If it streams the live video feed off the device at anywhere near being useful for facial recognition it would be very noticeable.
It doesn't have to be the entire video feed. it could be snapshots from a specific direction. If millions use it, there's a good chance even a handful of snapshots from each user at a location that thousands of people pass by could get it mapped better than how they currently do.
Sundar repeatedly stated in this Google I/O 2019 presentation how they were condensing and bringing AI and ML techniques onboard to phones to provide better responsiveness and lower the security issues of taking private data into their cloud.
A very good result from all the bad press they've gotten lately. If Google can improve its stance on privacy then I think it solves it's biggest problem.
1)ingesting for the purpose of crowdsourcing street view imagery
If they are doing that(Which I strongly suggest they are not, well at least not yet), then having multiple views allows google to remove things that move (people/cars)
2) Depending on how they do this, if its using "visual landmarks" then people are a massive pain. They move, and cause all sorts of problems when doing positioning. So they'll need to be removed for the system to work properly.
3) The actual scary part is not the faces, persay. Its the fact that any picture where google has streetview is now able to be precisely located, regardless of who, when or how it was taken.
AI scares me. Even though I've used neutral networks and even more complex models/networks in projects and have a fair understanding of how they work, they still scare me a lot. Even though we are far away (supposedly) from creating a sentient AI, I can't help but think that AI will turn into something like "I am Mother" rather than C-3PO.
And yes, I understand that AI and machine learning can help us in ways that will advance our abilities in medicine, science, etc but we're human. if something can be used for evil, we should be sure that that will be attempted. We, as a species, have an annoying need of being the first and having control of everything. Only that AI may, in the future, be uncontrollable.
As far as AI are concerned, there is no intelligence in them - at all.
As far as understanding what intelligence is - we have a long way to go on that as well.
What these system can do, is process a large quantity of data and then match against a specific set of criteria. The problem is not the computer system but the actions taken by real flesh and blood people who do not question what these systems put out.
Yeah,and that's the point I was trying to make about right now. But at some point we'll find ourselves in the position of having created general intelligence. Be it smarter than us or not. In the case of this model, it "could accurately simulate how the universe would look if certain parameters were tweaked [...] even though the model had never received any training data where those parameters varied."
This means that it somehow was able to figure something specific out based on data that had nothing to do with it. Whether by accident or because there's a correlation that we haven't yet figured out, this model was able to understand it. And that's amazing and scary at the same time.
All the robots were the same AI and it effectively killed off humanity because it thought that we were not good enough. When a machine starts killing us off because of dubious and intrinsically evil and selfish reasons, then I can't think it's benevolent.
I'm supposed no one has mentioned the - sort of - deterioration in our relationships. back in the day it used to be a lot easier for two people up meet, fall in love, get married and have kids. I'm not saying that there weren't any forced marriages, but that's not my point. In the age of tinder we are exposed to more and more people and possibilities. So basically it seems like the average Joe and Betty keep holding off for the perfect mate, when chances are they already rejected them. I'm just observing that the typical meets are decreasing rapidly and could be to, at least partially, blame.
More to the point, there is a curious social stigma around having children before you enter your 30s that has developed.
The idea that you should "play the field" is part of it, but I'm not sure that tells the whole story. There seems to have been a concerted effort to reduce pregnancies among younger women. In fact, programs like "16 and Pregnant" and "Teen Mom", which follow young mothers in their teens and 20s, were, according to their creator, meant as a "cautionary tale" to reduce pregnancy rates. In modern culture, "rural hick" even conjures up images of a young mother. Not the image most want to portray, which is a powerful social tool.
Equally curious, once you enter your 30s all of a sudden the social norms flip to "why haven't you had children yet?", "the clock is ticking", etc. However, once you are in your 30s, there are some rather hard limits to how many children you can practically have.
As others have said, having children very early is unsafe, and it's a demanding job that requires emotional maturity to do well. Having children before you have economic independence is also a very bad idea.
But your 20s are also critical from a career point of view. Women still have to fight a lot of prejudice from employers, which translates into trying to show commitment to career by delaying children.
The UK now has a very soft "two child policy": you won't normally get paid child benefit for children beyond two. The US has a ludicrous healthcare system where simply giving birth can cost tens of thousands of dollars.
Also, they don't call it labour for nothing: it's a physically demanding, uncomfortable process.
I think its less stigma but economic realities that are driving this. People have too much debt and not enough income and are too early in their careers in their 20s to comfortably take time off to have a kid or pay for childcare.
For what? Less than half of Americans, aged 25-34, have a post secondary education. Less than half of Americans in the same age range own a home. The same age group are much more likely to reject car ownership than generations past. Outside of those purchases, it is not common for young people to take on debt.
You no doubt describe a minority, but the majority aren't putting any effort into picking up the slack. Birth rates are falling in all walks of life.
> not enough income
Or too much income? There is a very strong correlation between being poor and having more children.
At the extreme, the women of Niger have over 7 children each, on average. Even the least paid people in America are living like kings in comparison to the people of Niger.
But even within America, the reason the "rural hick" is oft associated with young mothers is because rural areas tend to be poorer and poorer people are more likely to have more children.
> are too early in their careers in their 20s to comfortably take time off to have a kid
While the average reader on HN is certainly career driven, most people are not. "You need to focus on your career" is a line that has been used to deter 20-somethings from having children. I'm not sure that is the same as them actually focusing on their career. Most people simply find what work is available to them, trudge through the day, and then go home as soon as possible to not have to think about work again until tomorrow.
> For what? Less than half of Americans, aged 25-34, have a post secondary education. Less than half of Americans in the same age range own a home. The same age group are much more likely to reject car ownership than generations past. Outside of those purchases, it is not common for young people to take on debt.
But the job prospects and odds of financial security for people without a college degree have cratered, and so they're still unable to support a family. Pick your poison: either stable income but a shitload of debt, or no debt but no income, and neither is a good place to start having children. College in America now is damned if you do, damned if you don't.
> But the job prospects and odds of financial security for people without a college degree have cratered
Are you sure you are not confusing that with the fact that post secondary attainment is rising, which means that those who are unable to attain a post secondary education, and find work, due to challenges in their life (disability, for example) take a larger share of the lower education segment?
Let me put it another way, as that may be confusing. Take two people. Let's say one is a highly intelligent, hard working, person who had to drop out of high school to care for his ailing parent. The other a drug addict who dropped out of high school because the drugs started to take over his life.
The highly intelligent person has a job. The drug addict does not. His addiction has left him unable to keep employment. Given this scenario, 50% of those who are high school dropouts are unemployed.
Okay, now let's say the first person's parent got better and he was able to return to high school and graduate. He is no longer in the high school dropout category, but now the high school completion category.
That means that 100% of those who are high school dropouts are now unemployed. If you weren't paying attention, you might think that the job market for those who are high school dropouts cratered, but in reality nothing really changed except a rise in high school attainment.
> either stable income but a shitload of debt, or no debt but no income, and neither is a good place to start having children.
There is never a good place to have children. But the data clearly shows that the richer you are, the less likely you are to have children. How are you resolving that discrepancy?
> But the data clearly shows that the richer you are, the less likely you are to have children. How are you resolving that discrepancy?
Without taking a stand on the greater discussion here, I'd just like to add: birth control is not all that cheap, and poorer people are much less likely to be as educated on all the birth control options available to them anyway. I would also reason that the feeling of having limited life prospects would make one more likely to have kids as it becomes a larger milestone at that point.
> Less than half of Americans, aged 25-34, have a post secondary education
Are you sure? Unless you mean post-secondary education means graduating college as opposed to attending college without graduating, the information I could find doesn't support this (with the caveat I didn't find info specific to the 25-34 demographic)
There's certainly lots of people who attend college (with loans) and dropout. They presumably may have the worse of both scenarios, college debt and no degree to show for it.
Pregnancies at 16 have very real health risks and 16 years old are not ready to be parents yet, not ready to be independent either. There is nothing curious or odd about the end.
Have you seen any data to support this? I'm not convinced it's true, based on my own experience. It seems to me like what's changed is the range of possibilities for women, who can now more easily pursue educational and professional ambitions rather than settling down right away. There has also been a striking decrease in the social stigma of being a single "older" woman, which used to be seen as reflecting a woman's desirability.
In my subjective opinion, things have gotten better in this sense, rather than worse. But it's kind of pointless to say so without any kind of study, because my own experience is limited and distorted by my unique perspective.
I don't have tangible information, but what I'm referring to is what I know from experience in my social circle. Maybe, as you say, my perspective is distorted in some way. However, 15 years ago if you wanted to meet someone you had to go out and talk to people, which is hard and you're more open to hearing someone out. Today you open an app and swipe into oblivion and no one seems good enough. I've heard many people say that; both men and women. I've had people show me perfectly nice looking people they met through online dating and they either swiped right or rejected them after one date because they weren't good enough.
And again, it's important to marry someone you want to marry and not marry just because you're getting older or for the sake of it. Those marriages, more often than not, end in divorce. But I see a lot of people either using Tinder as a way to get as many one night stands as possible, or swiping away looking for someone perfect who they'll never find. I also know people who developed severe depression because they were on the receiving end of perpetuated ghosting.
I am, by no means, saying that online dating hasn't worked for people, but it just seems like it's done more harm than it's done good.
There are a lot of interesting things here, although some (like that first graph) are rather subtle. In particular, I'd point everyone to the graph over time of where straight Americans met their partners. To my eye, it has three distinctive elements.
First, finding romance at a traditional "third place" has been almost completely destroyed. Meeting a partner through family, neighbors, church, or college shows an undulation in the sixties, followed by a sharp downturn in the last decade.
Second, meeting partners through work or friends has ceased to be a replacement. The decline in e.g. church meetings was thoroughly offset by professional and social meetings until the mid-nineties, but those meeting sources have been declining since.
Third, the spike in online meetings, and the s-curve in bar/club meetings. Crucially, this happens after meetings through friends and coworkers flatline; they aren't just being displaced by online dating, they stopped prior to that explosion.
Alongside all of those patterns, we can add the rising age of first marriage, the lowering frequency of sex (overall and within relationships), the rising gap in sexual intent versus outcome in both men and women (i.e. how much less sex people are having than they want), and the gap between intended and actual family size. A lot of the most obvious relationship shifts over the last ~70 years seem clearly good, and seem to benefit the people I know. But the data overwhelmingly suggests that people are forming relationships later and with more difficulty, with outcomes further from what they intend.
Isn't the opposite true? That in the age of Tinder, meeting people is easier than ever - and so it should be more likely for someone to also fall in love with someone. I don't see how the limited dating pools of the past provide any advantage on this front. It doesn't make sense to me that people today would "hold out for the perfect mate" when exposed to a larger dating pool - my thinking would be that when people have met someone they have a good relationship with they will settle down. I also don't buy the idea that someone has already "rejected someone when meeting them for the first time, if they weren't interested they wouldn't have gone on the date.
Tinder only works for a subset of the population due to human behavior. Women largely select for traits that are a proxy for alpha status. The same applies to all online dating. There is a paradox of choice going on where there is so much available that everyone ratchets up their minimum requirements.
The bottomline is that I've seen it happen in real life, to people in my social circle and family.
I wrote a reply to another comment that addressed something similar. The best way I can create a correlation in the real world is this way: If you go to a restaurant with 20 choices it takes people much longer to order something as opposed to a restaurant that can fit their menu on one page.
It's some psychological paradox, or research, that I can't remember. Someone actually replied with a TED talk that touches on this matter.
Your Tinder comment made me think of Barry Schwartz and his "paradox of choice" where too much choice ultimately leads to more unhappiness because it sets the expectation that, will all these options, we should expect to find 'perfect'
Greek has funny animal noises. Dog: γάβ (yav - but pronounce the y as yearn). Chicken: πάκ πάκ πάκ (pak pak pak). Bird in general: τσίου (tsiou - pronounced kind of like chew).
Yes, once you start the search a directory corresponding to the date of your run will be created inside the saves directory. Inside this directory all your models and progress will be saved automatically (you can later resume the search by changing save_folder value inside configuration file). Once the search is done you will find best_topology file which will containt the topology with its weights. Furthermore, inside the deepswarm.log file you will see all the previously evaluated models with their loss/accuracy values.
This looks pretty cool, although it's missing a little more explanation on how to actually use it. I think the API section doesn't have enough information.
First, thanks for creating and sharing this. Second, I'd like to offer some thoughts related to what you said above ("some suggestions in detail are welcome"):
You describe it as:
1) An application-layer proxy
2) Which can be controlled by an API
The README and the Swagger API docs do a reasonable job of explaining #2. But it doesn't include much about #1.
First-time visitors might find it helpful if you started off by explaining the basics:
- What are some situation in which you might want to use this?
- What was your motivation for creating this? (What were you using before, and what was missing that made you want something different?)
- An example of how to set it up for a simple use case (not just how to install pichi, but how to point the other software to it (e.g. what protocol do clients use to talk to Pichi?)
For example, let's say I install Pichi, and then use the API to point it to several SS servers. What do I do next? How do connect my web browser to Pichi so that the traffic goes via those servers?
Not similar to haproxy, which works at the transport layer, while pichi works at application layer. I've not really understood what 'redirect' means. Yes if you meant acting as an HTTP proxy.
This. In all fairness an AI could make policy, but probably not good policy. People think anything relating to AI is cool, but forget the scary side of it, like with Gpt-2.
Not to mention that this type of app is treating people as incapable of following directions and it opens up the possibility of new avenues of advertising. Google already keeps an archive of your whereabouts and can even tell when you get in and out of a car. Just imagine wielding your phone around and getting "special offers" for that store you go by but never walk into.