Skip to navigationSkip to content
Maya Ish Shalom

Good evening.

Team Trump in trouble

Trump's tax returns will become public if the Supreme Court follows precedent. The high court must decide whether to grant the president's new petition for review. If it does, that won't necessarily bode well for Trump.

If the Supreme Court follows precedent, Trump’s tax returns will become public

It’s a weird one... on the one hand, the long history of state vs federal reach and powers is a struggle that shaped the constitution by the framers. And on the other hand US presidents have also traditionally voluntarily disclosed their tax returns...

Sitting president aside, this is one to watch

It’s a weird one... on the one hand, the long history of state vs federal reach and powers is a struggle that shaped the constitution by the framers. And on the other hand US presidents have also traditionally voluntarily disclosed their tax returns...

Sitting president aside, this is one to watch: it’s far reaching in the implications towards the ever expanding executive powers...

Yesterday Trump filed a petition to the US Supreme Court seeking to block a subpoena that will force his accountants to turn over his financials. The prior cases on related matters didn't turn out well for presidents Richard Nixon or Bill Clinton. But Trump is making much of the distinctions here and

Yesterday Trump filed a petition to the US Supreme Court seeking to block a subpoena that will force his accountants to turn over his financials. The prior cases on related matters didn't turn out well for presidents Richard Nixon or Bill Clinton. But Trump is making much of the distinctions here and hoping the justices will agree that this case, arising from a state grand jury investigation, is unconstitutional.

The mind of the millennial

Psychologically speaking

What the health?

Inside Apple

Judicial affairs

Building a sustainable future

Flooding is difficult to predict and prepare for. A hydrologist in the UK wants the government to be more aware of the risks of building homes and businesses in floodplains—or at least improve design standards. There have been devastating floods in the north of England this week.

Why flooding is so difficult to predict and prepare for

Flooding is a tragic reality for many parts of the UK today – and flooding will also become more common and more extreme with climate change. The Met Office believes that intense rainfall associated with flash flooding could become almost 5 x more frequent by the end of the century.

In the UK, greater

Flooding is a tragic reality for many parts of the UK today – and flooding will also become more common and more extreme with climate change. The Met Office believes that intense rainfall associated with flash flooding could become almost 5 x more frequent by the end of the century.

In the UK, greater investment needs to be made in developing more innovative water management systems and new, more radical, forms of flood defence. There are some fascinating infrastructure innovations in this area, particularly from Asia, which is home to some of the world’s wettest countries.

In Tokyo, a massive “underground cathedral” is part of the Metropolitan Area Outer Underground Discharge Channel – a system of dams, levees and tunnels defending Tokyo. In extreme flood cases, the system takes in water from the 5 rivers crossing Tokyo, holds it temporarily, and then discharges it into the largest river when it’s safe to do so. The pumps in this system can push 200 tons of water per second (which equates to emptying a 25m swimming pool in 2-3 seconds)

Another example is the work led by Professor Kongjian Yu, developing the concept of “sponge cities” – a city designed (or retrofitted) to passively absorb, clean and use flood water. There are now hundreds of sites around the globe that use sponge city concepts, including permeable pavements, wetlands, and rain gardens all with the aim of absorbing excessive rainfall through soil infiltration and/or retaining it in underground tunnels and storage tanks until flooding recedes or it can be safely channelled.

It will be interesting to see whether the UK looks to greenlighting more ambitious approaches to flood management over time. Projects of this scale will of course come with a significant price tag.

No business like showbusiness

Food under threat

That's it, time for bed.

Close
If you have dark skin, a self-driving car may be more likely to hit you

If you have dark skin, a self-driving car may be more likely to hit you

Read more on Vox

From Our Members

  • This is such an important point, and I'm glad journalists are covering it. Algorithms are going to manage so much of our lives in the future, from simple traffic lights to operating room procedures. We need to make sure that these algorithms are not unconsciously biased against women and minorities now

    This is such an important point, and I'm glad journalists are covering it. Algorithms are going to manage so much of our lives in the future, from simple traffic lights to operating room procedures. We need to make sure that these algorithms are not unconsciously biased against women and minorities now when the standards are being created. It's harder to fix problems when they've been codified into a system.

  • Another of the incredibly crucial reasons why are working Innovation and new technologies absolutely must intentionally include people from as many different types of heritage, experience, color, gender expression, etc as at all possible. I think it's safe to guess that if AI design team included not

    Another of the incredibly crucial reasons why are working Innovation and new technologies absolutely must intentionally include people from as many different types of heritage, experience, color, gender expression, etc as at all possible. I think it's safe to guess that if AI design team included not only more people of color, but more people with many kinds of experiences, those implicit assumptions would probably be much more likely to have anticipated that failure. No matter how smart we are, we don't know what we don't snow and we need others who can see what we cannot.

  • So they’re basically white cops.

  • Or, as the article eventually gets to, maybe not.

    The algorithm the researchers used was trained on a data set of primarily white people.

    They controlled for some factors, but not a lot of others.

    The manufacturers don’t share their data sets with researchers for good reason. They’re afraid of biased

    Or, as the article eventually gets to, maybe not.

    The algorithm the researchers used was trained on a data set of primarily white people.

    They controlled for some factors, but not a lot of others.

    The manufacturers don’t share their data sets with researchers for good reason. They’re afraid of biased articles accusing them of things which don’t exist.

    The manufacturers are aware of the importance of self-driving cars recognizing all people. And also of recognizing everything else that it would be a good idea to not hit with a vehicle. They can’t afford bias against a group of people. They also can’t afford bias against any groups of inanimate objects. They understand that self-driving cars can’t roam around smacking into things. And any smacking into will be subject to the most intense of scrutiny.

    Racism is a problem. But it doesn’t help anyone or anything to create racism where it doesn’t exist.

  • « The study’s insights add to a growing body of evidence about how human bias seeps into our automated decision-making systems. It’s called algorithmic bias. »

  • Another example of the life and death consequences by not funding diverse founders and teams.

  • Racist machines. The future of the past is now.

  • The developers of autonomous vehicles are in too big a rush. We know from fifty years experience with commercial aircraft and ships that AI and machine vision are not enough. Self-driving vehicles should have specific lanes, away from other vehicles. There should be beacons on everything that doesn’t

    The developers of autonomous vehicles are in too big a rush. We know from fifty years experience with commercial aircraft and ships that AI and machine vision are not enough. Self-driving vehicles should have specific lanes, away from other vehicles. There should be beacons on everything that doesn’t move.

    The AI for self-driving vehicles is not good enough. The machine vision is not nearly good enough. The infrastructure is not ready. We can fix all of these things, but that would require a few hard choices and a lot of investment. Who is in?

  • shocking. and, of course.

  • To the many people commenting that self driving cars being more likely to hit dark skinned people is a consequence of poor diversity in tech industry, here is a bit of information: self driving vehicles use image recognition algorithms. They way these work is they analyzing the pixel pattern on every

    To the many people commenting that self driving cars being more likely to hit dark skinned people is a consequence of poor diversity in tech industry, here is a bit of information: self driving vehicles use image recognition algorithms. They way these work is they analyzing the pixel pattern on every frame captures and compare these patterns to a data bank the algorithm used to learn what a person is for example. Dark skin makes you harder to see, for you and for a camera thus more likely to go undetected by algorithms. Plain and simple. Solution: use infrared.

  • You may want to think twice about getting a summer tan.

  • This is less a function of algorithm bias an has more to do with spotting dark-skinned people in low light situations. Let’s not make a mountain out of a molehill.

  • Wow. This is pretty scary. Yet another reminder that there are SO many things we need to carefully consider before new tech like this is unleashed!

  • There may come a time when cars become self driving but the decision to service and how much to pay will still be a personal choice. How many commercial aircraft have crashed due to poor maintenance and cost cutting? There are reasons why jetliners have 2 pilots so why at this stage trust a driverless car.

  • Wouldn’t Infrared cameras feeding the algorithms erase AI bias in this case?

  • Be careful the wrong people don't get a hold of this article and make a huge racial deal about this. There are some very good points, opinions and thought in this article, however, the one thing that comes to mind.....

    How does the system differentiate inanimate objects or even other vehicles, or animals

    Be careful the wrong people don't get a hold of this article and make a huge racial deal about this. There are some very good points, opinions and thought in this article, however, the one thing that comes to mind.....

    How does the system differentiate inanimate objects or even other vehicles, or animals of different colour shades?

    This tells me that AI still has a long way to go with learning, algorithms, and success, which is acceptable purely by default that it is artificial. Is it really because the intelligence inventing and teaching, is unintentionally racially biased, or even unable to thing of the colour and race of objects??? I don't think so.

  • I feel this is a very uneducated excuse for an argument to start and if you didnt make your technology well enough to detect all humans its insufficient and most of those vehicles detect objects not colors and has no feelings so cant possibly be racist