hongzi

Link to the artwork: https://i.imgur.com/p2lCwef.gifv

It happens so often that I stumble upon some breathtaking photos on Instagram, immediately have an urge to add to my bucket list, but struggle so hard to find where the photo is taken. I have also seen so many people asking for precise location whenever I post some good photos. To help this need, me and my friends build Miru. We want it to be a platform for people to discover, share or just directly ask where a photos is taken.

On Miru, the current main functionalities are

- Discover, choose and zoom in/out any location, search this area, and scroll through the photos and see where they are taken.

- Ask & find, see what other users are asking for. Each circle is a rough location of the photo. Help answering with precise location if you are a local expert. Successfully found photos will join the discover page.

- Collections, everyone can build a bucket list of places to go.

- Upload, share or ask for the precise location of photos from a local file or from an Instagram link.

- (Weather, a related project that we built, predict/notify epic sunset/sunrise)

We are currently building a reward mechanism to recognize people who share or answer the location of a photo. For example, it can come in a form of bounty. Multiple users are interested in the same photo, jointly pay some points to who gives a correct answer. Using the points, one may unlock photo locations that only few knows.

Let us know if you have any thoughts and feedback :) We can be also reached at contact@miru.app

Thanks a lot for the suggestion!

I'm a hobbyist landscape photographer. There have been too many times I missed the chance of going out for stunning sunset/sunrise. An accurate notification of epic sky would have helped me plan ahead much better.

So during the pandemic I wrote a python program. The pipeline is to (1) download daily cloud formation data from NOAA (https://www.nco.ncep.noaa.gov/pmb/products/nam/); (2) get a minute level extrapolation (NAM only has hourly prediction); (3) use a physical model to get the sun location and calculate the cloud color based on Rayleigh scattering (https://en.wikipedia.org/wiki/Rayleigh_scattering); (4) visualize the sunset/sunrise color on matplotlib Basemap; and (5) send out an email notification when the cloud is colorful enough for long enough.

Me and friends tested it for the summer and it was pretty accurate (good true positive). Some photos we got are in https://crispysky.com/#demo

It turns out the entire calculation for the US continent was pretty fast (each tile on the map can run the computation in parallel). After downloading the data from NOAA, serving the notification for me and my friends or serving thousands of people should almost cost the same. So I wrote this simple website to share the service :) The viewing experience is optimized for desktop users for now.

It's a small web service for fun but I would love to learn what you think we can further improve, for this website and for data-driven photography in general. Thanks!

I think the most interesting visualization is the length distortion around here https://youtu.be/udqihUBGuZ8?t=521

More history: https://en.wikipedia.org/wiki/The_Hardest_Logic_Puzzle_Ever, https://www.technologyreview.com/2012/06/13/185681/the-harde...

Very clean mechanical OR structure

A nice technical inside look of the product: https://ai.googleblog.com/2019/06/an-inside-look-at-google-e...

Project website: http://nrg.cs.ucl.ac.uk/mjh/starlink/

A nice business insider article: https://www.businessinsider.com/anyscale-berkeley-databricks...

I really like the central theme of this essay. Can't agree more -- problems are more important than ideas. This also applies to doing good research in academia. Research shouldn't be about impress other people but should be about solving important (agreed by others too) problems.

It's actually quite non-trivial given that Tesla's autopilot is purely vision based

It looks like the spec would be similar to high-end PC?

I'm really impressed by the "surgery" operations to keep reusing models as opposed to toss old models and retrain when some small part of the game changes. Appendix B has some pretty good dive-in.

Counter argument from Bill Gates: https://www.treehugger.com/corporate-responsibility/solar-en...

Video: https://www.youtube.com/watch?v=ZRCdORJiUgU

IMO though, Google's 2011 year review still remains the best Zeitgeist video of all time: https://www.youtube.com/watch?v=SAIEamakLoY

I guess someone somewhere maintains a database of all leaked username passwords and this feature compares a credential with this database? Is that how this works?

Now that 2020 is around the corner, I'm wondering the latest technical development.

The reorganizational plan: https://www.eecs.mit.edu/news-events/announcements/eecs-reor...

Nature Astronomy paper: https://www.nature.com/articles/s41550-019-0930-9

Live stream: https://www.youtube.com/watch?v=ZfPoUROLw3M

Thanks for sharing! Julie is amazing!

Creativity, Inc. [1] It's not a textbook but a really fun read for the inside stories that pushed Pixar to its success. Maybe it's just me personally but I found it hard to follow some syllabus to learn management --- instead, I enjoy being inspired by other great managers through their concrete stories.

[1] https://www.amazon.com/dp/B00FUZQYBO/

Science report: https://science.sciencemag.org/content/366/6469/1121

Very heated reddit discussion https://www.reddit.com/r/theydidthemath/comments/e2ii8m/requ...

This related article about Moon's habitation condition is also pretty fun to read: https://www.nytimes.com/2019/07/08/science/apollo-moon-colon...

If the gaming experience is similar to what the trailer shows, this would be the closest to what a VR killer app should be

It would be super cool if the cover of the back truck is a solar panel: https://www.tesla.com/xNVh4yUEc3B9/06_Desktop.jpg

Interesting, the no rear-view mirror design is already approved by regulations and being deployed. Roadster shouldn't be far now +.+

This lab is doing a whole series of amazing stuff---

500ps hand gesture recognition system: http://www.k2.t.u-tokyo.ac.jp/perception/zSpace/index-e.html

Robust tracking for moving objects: https://www.youtube.com/watch?v=p7IL0Gvux7U