pinkyand

Maybe the networks now are quite weak.

But in a future where a large percent of rides are shared by 3 passengers on average, or maybe even more via SUV(for cost savings of course), being the dominant provider is tied directly achieving optimal routes and average share ratios.

In such a future , the network effects over riders is much stronger , and it's probably a "winner takes most" market.

Excatly. Nothing. The computer should be able to figure out what to do on it's own.

My guess is that one thing that future software will have is automatic generation of user interfaces.

I think "naked objects" or "apache isis" are the best examples we've got for that for now , but they are rooted in somewhat complex java code, instead of being rooted in an easy to use tool, such that lets the business analyst who had some course , sit with the client and fully define a running system , step by step. In some cases the auto-generated UI would be good as it is , and the system will be used as-is.

In other cases , we might need easy tools to customize the UI.

> There's still nobody around who can do a product release like Apple.

But they might just reusing old fame. another one or two non-revolutionary product releases and , and their launches might become like everyone else's.

Maybe for those people, someone should design a plan in which you cash for signing to a low cost plan for a higher price?

They are mainly interested in the phone as sales devices. To sell stuff you need color.

You didn't read the article and it shows.

They compared treatment-as-usual to treatment-as-usual and the online intervention. An my guess is that treatment-as-usual does include meds.

I think they've improved in this regard: they've added an independent microcontroller inside the core(which will probably run the arduino layer).This will solve compatibility and real time issues.

This is in general targeted to a bunch of chips that integrate a microprocessor and a microcontroller and are usually fabricated in 65nm-45nm. With it's process advantage(22nm finfet) , intel can offer orders of magnitude lower sleep currents, some nice decrease on power(depending on x86 vs arm details), and more power in general. All interesting to embedded guys.

But they still have to create a large library of peripherals and support it(which they might solve by creating a few speedy cores and let the crowd code and share/support peripherals), add analog blocks, and gain the trust of the embedded community as a reliable long term supplier(which might be the one thing intel couldn't solve - because of it's past as an unreliable embedded supplier).

The second option is that this is mainly to motivate other embedded chip companies guys to use it's fabs.

Anyway ,i'm grabbing the popcorn.

One use case in production i might see for this:

1. Using callbacks as commonly done in javascript, might offer a very easy to use way to write low power code(because the chip can automatically sleep between callbacks). This is nice versus today's high complexity low power design.

2. The author used regular mcu for this board. But low-power mcu needs orders of magnitude less power. And of course low-power pcb design can save even more.

Combining the 2 together, maybe we can see a good enough option for building 1st version, low-power products(say kickstarter guys) where they can spend a bit more on chips and there's less competition on specs - but things like low cost of development, time to market, and high domain knowledge with less technical knowledge are important qualities.

That sounds great. Did you teach them mostly what's on better explained?

Also I wonder how would math studies look if we teach people the intuition, a bit of the formalism and lot of how to use numeric and symbolic tools - instead endless rote learning.

With that kind of thinking you might like the site/book at betterexplained.com.

Interesting comment. And i might have quoted wrong, it might speak of "engagement".

I wonder: In the article they mention that the reasons for such effectiveness are "rapid personalized feedback" and "entertaining content keeping kids focused".

Those reasons could apply to many educational fields. So i wonder: are there any other apps/sites showing very strong gains versus standard teaching?

The noise cancellation in these is so much better than competitors, according to reviews.

So i wonder whether machine learning can learn such noise cancellation algorithms and let others compete with bose ? Because lower cost "noise bubble" seems like something that many will buy.

Assembly is bad for auditability, which is important in crypto to prevent subtle errors.

Great analysis with lots of details.I wonder thought: what percentage of u.s. eCommerce orders are done through amazon's warehouses? and what percentage is using their store(even though it's 3rd party) ?

> outside of the US Amazon has just barely started to expand beyond books.

Not sure it's correct. For example amazon has a 2-day service across Europe from any amazon warehouse in the continent. But maybe what they do is not enough for Europe?

UBER is far from a commodity.

The ride sharing business(especially if we talk about 2 or more people sharing a ride with the same driver) has strong network effects - the more passengers you have ,you can offer cheaper prices and share rides optimally.

In business that enjoys a network effects we usually see one winner controlling the market, and i suspect that's what we'll see here, and that's why UBER got such a huge valuation.

There's a leak[1] talking about windows 9 possible being partly free, but some features are subscription based. That would give them the incentive to solve the security issues.

[1]http://arstechnica.com/information-technology/2014/04/the-in...

He could have posted it through the mobile browser.

There was a company(maybe a VC) that used big data to predict startup sucsess.

IBM used watson with some emr company to predict heart disease in thousands of patients.

> Who said decisions should be easy? I sweat over each and every one.

True medicine is a really hard job.

So maybe the way decide in medicine should be changed ? maybe they be done remotely but through a recommendation system so that each doctor would get plenty of nearly similar cases - so he could practice and hone it's intuition ?

Or since the complexity is really too great (and as a patient hearing that "this is art not science" isn't that encouraging , even if true) maybe the overarching goal of medicine should be to remove the art, and we finally have the tools to do it(watson, machine learning ,remote diagnosis , etc...) ?

And if we agree that that's a good goal, maybe the current way we try to do it (evidence based medicine) , is a bit clumsy due to both the science and the resistance by doctors , but at least it's going in the right directorion.

Yes i agree, sometimes "production" goes too far. In some of those complex case it's even wiser to allocate much more time in advance from doctors and nurses, to prevent future complications - which are much more costly.

But i wonder:

On what sources do you rely when it comes to managing complex cases ? is it only known hard coded medical data in books and databases ? or also some intuition ?

And if it's intuition , how can you form intuition for such complex cases where really each case is pretty unique , as you said ?

> The problem is that the industrialization of medicine makes it increasingly difficult to be helpful.

Atul gawande, which is a doctor and a healthcare journalist ,and knows the medical system from inside, thinks that industrialized medicine is exactly what's needed:

http://www.newyorker.com/magazine/2012/08/13/big-med

And in general the positive results from decision support systems seem to agree with him.

He quotes 2 scientific research papers as claiming ‘spending large amounts of time online for social purposes may increase social distress and have negative impact on self-esteem.’ or similar stuff.

But look at the details, the first article is about correlation, not causation. http://europepmc.org/abstract/med/21685655

Stopped reading after that.

Sorry for the late response - i'm across the Atlantic.

I wonder, with the API , would it be possible for a startup to build such "second opinion service" , with reasonable costs ? will you block it?

That's definetly progress, but still , the point made by spitfire is true: The real challenge is adoption, not tech.

We even see it with the low acceptance of Watson in medicine and IBM's move into africa.

I wonder thought: why hasn't IBM offered a "second opinion" service directly to consumers ? It could might have sped adoption.

Thanks you!

It sound very impressive and i'm very curious.

A few questions:

1. It feels a lot like literature-based-discovery. Is it ? if so , how is watson better than current literature based discovery methods that did create some scientific hypothesis AFAIK?

If not please share more details(if you can of course) about how it reaches hypothesis).

2. Does it work well in non biological literature ? Are there any examples ?

True, that's the tech behind it.

Is there any reason most programmers should know any of this ?

The code should look more like this(and should use a standard library):

import Password_protector

P = Password_protector(date_of_first_use) #library upgrades algorithms every once in a while, only a single algorithm available per date. If there's a need for automatically upgraded hashing library should handle covertly.

protected_password = P.protect_password(password)

if P.compare_password(password, protected_password) == True ...

Or something similar. That at least should cover the basics - safely.