Bits of Books - Books by Title


The Driver In The Driverless Car

How Our Technology Choices Will Create The Future



Vivek Wadhwa and Alex Salkever



More books on Inventions

Politicians fail us bc they can't turn the clock back to better times (where we were actually, in real terms, poorer, less safe, and shorter lived)

There is widespread dissatisfaction, either latent or overt, with the staus quo. We might love individual doctors or teachers, but few of us think the medical or educational systems are doing a great job. We like our cars, but hate having to drive on crowded roads.

Past tech upheavals - gunpowder, printing press, steel, cars, manned flight - all changed the way we live and use resources. But this time, Moore's Law is leading exponential change. Once something can be expressed as data, it becomes subject to the law of accelerating returns. Once the human genome was sequenced it became an information technology, and accelating returns applies.

Siri is an example of narrow AI - it does a much better job than a human can on a very specific range of tasks. They are much better at accessing data stored in complex databases, but can't (yet) combine or interpret the data creatively.

IBM's Watson suggesting diagnoses to doctors, but also building databases from millions of patients and the medications they are taking, to determine what truly does have a positive effect, and which ones led to negative outcomes.

'Flipped model' of education: teachers no longer providing content - students get that then=mselves, at their own pace, from online video lectures, explanations and illustrations. The teachers become coaches and guides. Traditionally, teachers came up with their own lesson plans. But someone else has already prepared that same lesson, and done a better job, and others ahve modified and added to that. The very best is available, so why not use it? Today, the online content is not great - YouTube videos of lectures etc. But outside-the-box teachers are using games and simulations to get students to discover things rather than being told.

Until now, the fight against obesity has involved two factors - diet and exercise. But, for almost everyone, they aren't effective. We simply don't yet understand how our body works. The same thing applies to wider health issues. Some people become diabetic if eat just moderate amount of sugar. Others eat more yet don't.

Our biological responses vary during the day, at different times of the month, who our friends are, where we work and live, and whether or not we are exercising regularly. All this as well as what we eat and drink, and by how our gut bacteria varies. This is where the hope of personalised medical treatment lies - starting with a personal genome sequence, and then using AI to evaluate all the possible diets and pills that could potentially improve our health.

For a woman, pregnancy can be the most dangerous time of your life. And this is true not just in places like India or Vietnam, but also much of rural America. Pregnancy related deaths in the US have more than doubled since 1987.

Hard to teach robots to deal with tasks that don't have xplicit rules. Folding laundry is impossible (today) - a 10 yo child can deal with a balled up Tshirt, but a robot can't. The act of folding comprises multiple steps, many of which are hard to describe to a machine. But, the Jetson's Rosie is coming, for obvious reasons. The cost of vital components is dropping dramatically - the single-axis controller, a core component of a mobile robot - has fallen from $1000 to $10. And the AI software is advancing at a similar exponential rate.

Voice recognition in a similar place. Google and Apple bots do a good enough job of translating speech to text, or from one language to another. They struggle with accents, mis-pronounced words and colloquialisms, and can't yet cope with multiple people talking at once in a noisy room. But these obstacles will melt away as computational power increases (an iPhone will have same power as human brain by 2023) and AI gets more sophisticated.

As always, the first 1% takes as much time as the last 99%.

Google's DeepMindsystem beat the world's best Go player in 2016 by watching humans play the game. But not just brute force; it innovated. Came up withnovel moves which seemd to make no sense at first, but which ultimately gave an advantage.

Intell argument - should robots be allowed to decide for themselves whether to kill people. Author has a cynical view - people don't care about this abstract idea. No public concern about drones being used to attack opponents in Afghanistan etc, even when innocents are collateral damage. In Dallas, police used a bomb on a robot to kill Micah Brown, who had allegedly killed 7 police at a protest rally. Again, public apathy.

Vanity trumps privacy: people are willing to give away incredibly revealing details about themselves in exchange for social validation. Even as growing awareness of identity theft dangers.

We are entering the Drone Age. For the first time in history you can buy toys which are better than what the military is currently using.

Clear benefits for society. If a drone can deliver a pizza it means fewer cars on the road (less accidents) and less pollution. Drones can do dangerous jobs like power line inspections, roof inspections, fire-spotting and security. And they improve access - rural areas can get same-day delivery of spare parts or needed medicine. And precision agriculture, optimising use of fertilizers and weedkiller.

And the danger. Although countries are working on defence against drone attacks, nobody can yet defend against a swarm of drones carrying bombs.

Then the ethical Qs. Should drones be allowed to fly over private property, filming as they go? Or schools? Should people be allowed to shoot down drones if they overfly their property? And if so, does the Second Amendment empower drones to shoot back if attacked?

In past we've classified cancers depending on where they occur - breast cancer or colon cancer. But now recognize that cancer is essentially a genetic diseas - need to select treatments based on the signature of the cancer's mutations.

Your microbiome - bacterial rainforest in your gut. Many children born with a genetic predisposition to type-1 dianbetes, but don't ever contract it. Researchers found that children who became diabetic had a lot less variety in their gut bacteria, and over-abundance of types known to promote inflammation. Correlation doesn't necessarily imply causation, but big advantage is that fixing the problem is easy (a fecal transplant), whereas almost all other interventions are expensive and difficult. And may not need a fecal transplant - the bacteria in cheeses, yoghurt and meats can survive digestion and recolonise depleted gut populations.

AVs will spell the end for most private ownership of cars - the cost of sharing an AV, using it only when you need it, will be dramatically lower than owning your own car. Obvious social benefits - disabled, elderly or children get personal mobility, safe to take a cab late at night, and an end to 'driving while black' syndrome.

And will produce unexpected outcomes. Why not buy a big motorhome instead of a cramped apartment? Travel to a beach or snowfield while you sleep and wake up to a landscape instead of a city.

And will change travel considerations - for many shorter trips it will make more sense to use an AV than to fly, with all its hassles.

2013 32,000 people in US died in car accidents. 92% of them due to human error. In the developing world the casualty rate is twice as high bc difficulty of getting to hospital in the crucial first hour after an accident.








Books by Title

Books by Author

Books by Topic

Bits of Books To Impress