bwang29

It seems like for most entry level engineers ranged at $100-120k who wants to live in a one bedroom apartment in SF, a 30% off in rent could potentially be better than 30% cut in salary, considering the current tax rate.

In China, there have been attempts by startup companies offering employees sleeping pods near the office, while providing water and shower so that the employee can simply stay around the office during the course of a year long employment with occasional "offsite" airbnb living. But unfortunately such attempt was shut down: https://www.techinasia.com/china-napping-pods-startup-forced...

This almost occurs to me as funny. Why does it need to be a confession? It didn’t seem like Planetary Scientists have claimed as a group that they “want” to live on Mars. The confession only makes sense if Planetary Scientists are all presumed to want to live on Mars in the first place. And even that is true, it’s not a particularly strong confession as the simple fact of not wanting to live on Mars doesn’t imply a betrayal.

This is an extremely well written story, I kept imaging action scenes from thriller.

But I kept wondering the legality of actually stepping into an iPhone fab without permission. As the journalists might be seen as given prior approval, what if they saw the new iPhone and how does journalism in here functions toward revealing information v.s. respecting policies. I also felt that the reason this article was written is only for partially humanitarian cause, but still largely trying to capture curiosities from readership, and the risks definitely seem to have paid off.

I think the question is why would we need protective watermark anyway if stock photo companies are already crawling and sometimes phishing for use of licensed stock photography on the web and then directly send out an DCMA or a charge?

I've been in many situations where the copyright owners reached out for damage fees after downloading a full-res, un-watermarked photo from free stock photos sites in blog posts, so I'm sure the tech is all there already.

I think this paper is showing that you "can" train an auto exposure/white balancing/edit flow algorithm with a DL pipeline, but the results do not necessarily mean it will outperform simple and cheaper auto exposure/white balancing algorithms that's out there. And the flexibility in this approach also allows masking and background removal.

However, most of the examples in the paper in fact shows improvements of exposure and color. If you import those images and tweak 3 or 4 adjustments of clarity, curves, exposure, saturation in Polarr or Lightroom, you will quickly get very close to the result produced by this paper. However, it is still impressive that it could get to an exposure histogram that looks exact like the ground truth.

Maybe someone can benchmark this against the Google photos auto enhance. A lot of people turn the auto-enhance in Google off because it sometimes create unnatural looks for photos, which are tolerable to everyday consumer but for pros it just looks bad.

Lastly, if you look very closely on the input images, some of them appears to be artificially adjusted to show how the model works. (last page, 4th row, fist image, which looks both underexposured and overexposured after damping brightness through post processing), and these input images are not always the type of images you can get from cameras.

Submitted my Show HN 1075 days ago (https://photoeditor.polarr.co/), for a WebGL photo editing tool, now has turned into a startup with more than 10M users.

It seems like economist uses Javascript to create the "You've reached your article limit" dialogue. Simple press ESC key to stop JS from executing on Chrome so that you can read the article.

Also here is my TL:DR summary of it if you're still trying to fight through the pay wall:

There is a thing called grating coupler that works like little high frequency antennas that receives light signals. When you put a whole array of them you will be able to do various scans of light signals to simulate the camera pointing at different direction, or fisheye, telephoto effects without the need of tilting or moving the surface of the array. The underlying computation relies on the ability to calculate and control the timing of signal travelled from each antennas, plus some classic signal interference and phasing issues. An 1cm x 1cm array will contain 1 million such couplers which would create a similar sized image as an iPhone 7 rear camera, but since there is no lens involved, the camera can be made a lot thinner.

The experiment involved putting teddy bears behind a blue screen. Two bears enter the screen and when the screen is lifted with one bear or three bears, the baby will express a lot more anxiety than seeing two bears

Is it just me or someone else thought about the 2016 movie Arrival talking about time in languages (no spoiler - I think it is a great film for NLP folks)

I remember there was a research claiming babies have innate ability to perform simply addition and subtraction (1+1=2 for example), and here is a reference to babies innate ability predicts their math skills

https://today.duke.edu/2013/10/babymath

P.S. I was also told the hardest part of researching babies is to get clearance and approval from their parents

So it seems like we as human have the basic building block for math built-in, and it takes practices and further development to create complex concepts based on those simple blocks.

I was invited last year in a round table discussion on this new rule. The original responses for the rule were mixed as there are plenty of nuances. Specifically, the main concern is the rule would lead to investors demanding founders to first acquire this type of status before they were able to get funded, causing a Chicken and Egg problem.

The original proposal of the rule required 375k of funding from US only accredited investors, and there are also requirements for the startup founders to maintain a certain threshold of ownership while being able to hire a lot of American employees in short period of time. I don't remember the specific number requirement/head counts but it was definitely enough to pressurize the company to expand in size quickly, while many tech startups do not necessarily need to hire dozens of employees in a 1-3 year period, not to mention they all had to be US citizens. 1-3 years would also be a stretch for most startup to figure out a concrete plan of growth in order to quality for an extension. And what if the founders want to bootstrap themselves?

Ultimately this rule still doesn't show any concrete pathway of residency or visa guarantees after 6 years. And because it is not a visa, it will take time to educate immigration officers and TSAs as well as creating a reasonable structure to allow founders to travel outside US legally as well. Historically there is a huge delay of understanding OPT, STEM and O1 visas in plenty of US embassies.

To make things even more complicated, it is not USCIS's interest to be a judge to tell which company would qualify for this rule, and they would need to form a group of trusted committees to check case by case if a company and their founders/co-founders qualify for this rule. I also heard that founder's spouses would be able to travel to US too.

I can feel there is something really interesting here but the writing, title and explanation here seems to be poor. The surprise here is the random starts and stops as well as 10X speed difference in the DNA replication process. Wouldn't it be more surprising that these processes run at the same speed?

"...started watching individual DNA strands..." I'm wondering how important the visualization and nature of "filming" helped to form the idea or prove the hypothesis that replication is a async process. I didn't quite gather which is more impressive, the filming or the discovery.

“Sometimes the traffic in the next lane is moving faster and passing you, and then you pass it. But if you travel far enough you get to the same place at the same time.” also doesn't quite make sense to me as the video clearly shows no queuing structure as seen in real life traffic, and there is no real hypothesis to explain the sudden stops and starts.

My company sponsored my O1 and yes I'm a Chinese national. I'm starting my EB1, thanks for the advice!

Just verified it is single-entry for Chinese nationals.

Ah, that's right. I have no idea it is country specific. It is good to know.

It confused a whole lot of me the rationale behind single entry O1 visa - it does create a lot of inconvenience as we do have a subsidy in China that I'd wish to be able to travel back to.

For those in the thread who struggled for H1B's chance of winning, I'd recommend considering an O1 visa - I think the requirement is slightly lower to apply than previous years, and O1 doesn't have a lottery process.

I've stayed in the US as a collage student, then graduate student from a top university then starting my company for a total of 9 years. I lost the H1B lottery twice and have been on F1 then F1-OPT, then F1-OPT-STEM_EXT in the first 8 years. I recently successfully converted into an O1 this year, therefore finally bypassing the H1B lottery. I consider the O1 visa a life saver but not everyone is that lucky to be able to apply. Now the thing is, my O1 visa is a single-entry visa, meaning I have to stay in US for 3 years until it expires and I can't leave the country otherwise I can never come back to this country again.

Edit: Obviously someone found out that O1 is single entry for Chinese nationals, and it might not be the case for most other nationals.

Now I wonder, what happens if you add sound effect, to change pitch, tone, or widen the sound stage. What would do to the decoded image?

I think the 2.00 bits/ pixel result looks quite more "analog" with a film grain effect to me.

I've observed hardware design is getting flatter in the last couple of years (thinkpad, MacBook, even IKEA furniture). Guess it has something to do with the change in software UI language.

I've noticed two day airs are using a lot more air cushions in CA, US than standard shipping.

My apartment complex has a large delivery bin area and I always wish I can leave the packaging boxes and padding/cushions in that area and only bring back the content of the box (provided that nothing is sensitive and nobody is watching). Would it be great if the delivery person can unbox the item at the delivery location with a fee and then bring the box back for recycling?

I'm not an expert on generalized A.I. training but it seems like the problem of training with audio and visual data input from games suffer the exact scale problem with Alpha Go at specific tasks. I could imagine training an A.I. to play a complex RPG game like Witcher, you would first probably need to train on a horse riding game, a running game, a weather reaction game, a free fighting game, a trading game and maybe a couple hundreds of thousands of games, each for a couple million times of trial and error? However, it also seems that human doesn't need to take this amount of reinforcement training data to quickly understand the complex mechanics in life. Wonder if there is any comparison between the amount of trials and errors a baby need to go through V.S. an A.I. need to go through using reinforcement learning to stand up and walk under similar gravity and muscle group setup?

That would be awesome! Can you add me on polarrist

So is a startup ideas' ideal state also when it is being conjured but not yet made?

This made me realize that our effort to land on the moon is effectively landing on a relocated proto-Earth's mantle. On a grander scale, landing on any planet in the solar system would probably raise similar thought, but as Moon seems to be directly derived from earth, it will still mean something different for us to land on, say, Mars.

I spent a couple of years in Beijjg. I think most Chinese people in major cities are fully aware of the brainwashing goal and agendas of the media and would even consider CCTV1 to be funny. Young people create memes and satires on news content all the time. The thing is, CCTV1 has quite literally evolved into its own genre and it is therapeutic to listen to when you feel unsatisfied with your life.

What most struck me is his comment on Chinese's brain washing media landscape actually helps the country to stay upbeat and optimistic. There is definitely truths in it and at this point I can see it as a grand strategy to help the economy in a positive way. Maybe any country in the unique situation as China will tend to the exact same thing.

If I understood this correctly, this is based on a higher (but unknown to consumers?) chance that the Galaxy Note 7 's battery could explode during flight and catch fire. Then, what would be considered a chance high enough that a laptop battery's explosion would not be a concern? Would it be safer to lock these phones into an explosion-proof box before they onboard the flight?

NASA is making a lot of "record-breaking" news in the past few weeks and it is increasingly difficult for me to think how these news affect me and my relationship to the Galaxy. I understood many of these discoveries translated to "current science could be wrong". But if counting stars isn't my hobby, what am I supposed to take a way from this and digest the meaning of 11.1B light years distance?

Is there currently a way to tell how much isn't too much to pay influencers or to know what others have paid? I have seen Instagram-famous with 250k followers asking for $5k for one single post in their demographic- but after calculating the ROI with many guess work and assumptions, it seems to be a lot worse than FB ad. The influencers often pitches themselves to be more "organic", or the ad is giving more "brand" effect that's hard to measure with traditional metrics. Sometimes this just feel like a new way of gambling and the seller has way more information and leverage than the buyer.

In the github discussion I saw the suggested alternative of "leader" and "follower". And it is very interesting to see that the current translated Chinese terminology of Master-Slave relationship in databases is "主从关系", where "主" means Main, "从" means Follow. So there is no connotation of slavery in the translated concept at all. I wonder how other localization works in this case.

In terms of "Apps" that is comparable to a quality web app, I think Chrome Store is pretty dense. If you look at the photo apps here: https://chrome.google.com/webstore/category/collection/pictu...

It is pretty much a replica of the top search results of "online photo editor" or "web photo editor". It is as competitive as web but I think the layout/review/star information in the store is still more useful than looking at a linear table returned by search.

It seems like Play Store is going to replace the Chrome Web Store. Speaking as app developer, the only advantage for searching for apps on Chrome Store is that most apps are optimized for a laptop/chrome book with a mouse and keyboard, and the apps work better for desktops.