Sundar Pichai is hot to invest in AI, his own field, which does not make money. Maps is the real killer app for phones. Maps is the reason Google has an Android business, and Pichai doesn't know it. If Pichai messes up the maps business, that will finally be the end of him. He has already messed up the reputation of search, with china-censored-search, and advertisements on the search home page for google cloud and android (ads on the search home page were things that Larry & Sergey would NEVER allow !!)

I worked on 3 email systems - one of the earliest UNIX systems, delivermail, XNS Email, and X.400 as a protocol inter-operability designer. Our Xerox Clearinghouse system got into distributed deadlock because it tried to overload the email storage buffers, producing a distributed deadlock (A is full of mail and email for B. B is full of mail and has mail for A. Even if they connect they cannot transfer mail because there is no space on either side. They have to wait for the mail to time out & get deleted as "undeliverable" and then they make progress for a few minutes before senders at both ends inject more emails to bring it back into distributed deadlock.) So there is a big problem if sender email transport rates approach the slowest bottleneck link in an email network. At Xerox in 1984 we were working with much smaller systems, 9600 baud transatlantic and 56kbaud norcal-socal links, 1.5MB of email buffer storage, and The Clearinghouse name servers were just sending tiny updates but they used an O(N^2) algorithm (N name servers trying to send to N other name servers, rather than using an O(n) spanning tree) and hung themselves.

Another issue is reliability. The TCP checksums are only 32-bit one's compliment sums. This works okay with super-reliable wired networks with a few tens of megabytes and maybe a few hundred, occasionally, and I don't have time to do the math but as you scale into hundreds of megabytes and gigabytes, the checksums will fail. The checksums will pass corrupted email. You will get into the classic "End to End Argument in System Design" which means only the endpoints know how strong to make the checksum. You'll need a more powerful end-to-end checksum to achieve reliable delivery. If you try to strengthen SMTP email then 90% of mail users will pay an unacceptable price. Don't laugh; checksum failure has caused OSPF routing outages before, breaking internet routing.

I'm not surprised there are some negative comments about this author I see a fundamental misunderstanding / mistake in the opening paragraph of their master's thesis : https://github.com/Gankra/thes...

"As a result, our core infrastructure is constantly broken and exploited. Worse, C and C++ are holding back systems from being as efficient as they could be. These languages were fundamentally designed for single-threaded programming and reasoning, having only had concurrency grafted onto their semantics in the last 5 years. "

This is not true, the idea in C was that no system-level primitives (i/o OR concurrency) should appear in the language. I was writing multi-threaded C in 1983. It was an idea born from observing all the mistakes in language design for the previous 15 years 1960-1975. It's the raison d'etre for stdio.h, and many other design decisions in the language. If you criticize the design motivations of a language, you'd better understand the decisions, first.

A USB-C Monitor can charge your MacBook (hopefully) so the 2 ports can do the work of 3 ports (HDMI, Power, USB) while docked.

They could have at least provided a Lanyard Loop for the 12 dongles you'll need to carry to make your MacBook Pro usable ...

The book is just a collection of pages held together with giant binder clips because Apple has Revolutionized the Design of a Book.

Like all Apple product features, the second edition of the book will be deleted, get one while you can!

This is the end. Their hubris knows no limit, so it is time for them to take a huge fall. Let's not buy any more iPhone 7's, that will teach them!

My favorites (I own 2 or more copies of each) : Chromebook with Celeron 2955U and m2 msata 2242 drive : Acer c720 4GB ; Toshiba Chromebook 2 2015 model. Both are also available in i3 models but I have never been able to tell the difference between the i3 and Pentium-class Celeron models (2955U and 3215U). Chromebooks have a very high standard for sound and horizontal viewing angles and so they are GREAT streaming devices - much better than most sub-$400 windows devices! These models can also do some lightweight 3D games (league of legends at 25-40 fps with 10-12 shaders or 20-23 shaders for the i3 models.)

All the extra portability is canceled out by the 12 dongles you have flopping around whenever you pick up the laptop to carry it to a meeting. END STORY.

Look carefully and you will see tremendous UI improvements; They screen is WAY better, brightest in its class; The sound is WAY better, best of any laptop. The disk is WAY faster, really astonishingly fast; The touchbar is WAY better only because it includes TouchID which means instant logins (no more fumbling with CAPS LOCK key!) The keyboard absolutely sucks, but it sucks no worse than any other stupid-thin keyboard on a device in this class. So overall, this machine is an amazing improvement, completely ruined by a brain-dead mania for USB 3.1c ports ...

Next year they will replace the top row (0-9) with another touchbar and so forth until 2020 when Tim Cook will reveal that the MacBook Pro is just a tablet!

They needed to include ONE regular USB 3.0 port. There are a zillion devices that fit *inside* a USB port, such as a YubiKey, and nobody wants to carry those $50 security items around on a dongle so that they have to "dongle in" whenever they login. Apple blew it bigtime on the MacBook Pro redesign. They had it made in the shade until Tim Cook got this idea : https://www.youtube.com/watch?...

It's pretty clear that fan-films are okay as long as they are crap. Once you raise enough $$$ to produce something good CBS/Paramount will rob you ...

For so many years in a row, Intel has been making faster and faster processors. This year, for a change, they have decided to focus on making only slower processors, and the Broadwell series is the result! This year they are slowing down the CPUs, next year they will slow down the system bus ...