Yes, I'm aware of that. But that post doesn't contain the word 'voice' even once :-)
23 posts • joined 1 Dec 2008
Re: Pretty sure
Yup, and IMHO it doesn't really make sense for Google to keep the Assistant confined to the Pixel phones. Presumably increased adoption of the Assistant would be a good thing for them.
I can see them thinking that the camera optimizations (image processing algorithms, etc.) are a USP of the Pixels though. It's arguable whether they could easily be ported over to phones by other manufacturers, though. Sounds like the hardware-software integration runs pretty deep.
There is an analysis error in the article
Let's compare and contrast two paragraphs from early on in the article:
"Despite having been at the receiving end of years of criticism for its overwhelmingly white male workforce, the social media giant's latest figures show that only 4 per cent of its workers are Hispanic and just 2 per cent African-American. The percentage of women at the company crept up a single percentage point to 33 per cent."
"But fear not! Change is coming, and Facebook noted that of the new senior leadership hires in the past 12 months, 9 per cent are black, five per cent are Hispanic, and 29 per cent are women. Which apart from the small increase in black hires – from 2 to 9 per cent – is a rounding error improvement in Hispanic and actually worse figures for women."
We can see the author is comparing the two sets of numbers, e.g. "small increase in black hires", "a rounding error improvement in Hispanic", "worse figures for women".
But whereas the first paragraph described the percentages of the entire workforce, the latter describes "senior leadership hires".
While I'm always happy to see exposure to lightweight solutions for Linux GUIs, I found this article rather confusing as it mixes and matches window managers (OpenBox) with distros (Crunchbang++, SparkyLinux), strategies (DIY Debian/Arch), and a bonafide desktop environment (LXDE). I'm not sure if the decision to present these 5 different beasts as a list is a helpful one, especially to those who are unfamiliar with the difference between a window manager, desktop environment, and distro.
Other than that, agreed with most of the points!
Re: The importance of complexity
I would say that the vast majority of programmers never encounter NP-hard problems. In terms of algorithmic problems, most probably sorting and searching problems form the bread and butter cases.
Most CS graduates in my neck of the woods end up doing enterprise information systems, which is where all the jobs are, and I would argue that knowledge of NP-hard problems is almost irrelevant (with the possible exception of those working in operations research).
Don't get me wrong, I'm firmly on the side of advocating the teaching of CS as a scientific field, but the truth is that once they enter the real world, most of that stuff quickly becomes a distant memory.
However, perhaps all this talk of the next wave of Big Data(TM) and Data Science(TM) and Predictive Analytics(TM) and whatnot will change the landscape somewhat in the near future...
"After all, legitimate users of Megaupload probably thought it was OK until the day the Feds closed it down."
Anyone who saw the garish design, the banner ads, the constant upsell to the premium account, etc. and *still* thought Megaupload was 'probably OK' for cloud storage deserved to lose their data IMHO. Call it some sort of 'digital Darwinism' at work.
Can I have John G.'s and Jonathan Hogg's babies?
If the proposed initiative will be more like those horrible "learn to code in a day" tutorials, then yes, I agree it could be very harmful. But surely the solution should be to campaign for compulsory programming that focuses on the aforementioned problem solving? This is what LOGO was designed for all those years ago. Is there something more up-to-date and friendly for this day and age?
Also, hats off to Andrew for quoting a response that included the line "I'm incredibly disappointed in your article" -- I was, too.
Standards compliance is the key
I'm with commenters Adam Azarchs and skelband and others on this... I don't think there's anything wrong with "native HTML5 support". <b>As long as Microsoft faithfully comply to the HTML5 standard</b>, they're free to remove whatever layers they want between HTML code and x86 instructions...
Just as there are many implementations of Java Virtual Machines -- as long as they comply faithfully to the Java language spec, more power to them, and more power to us!
Of course, we all know the real dangers: differing implementations/interpretations of the HTML standards by different browser makers. But this is true whether they're "native" or not.
*NOT* Eliza redux
"shouldn't take a great deal of processing to map the lexical items and grammatical structures from the given info in order to arrive at a possible answer or at least a range of good guesses."
This is exactly the type of comment Adam Azarchs above was predicting.
Yes, yes, lexical items and grammatical structures are all well and good when dealing with lab-controlled sentences. Try dealing with the sort of ill-formed linguistic mess us humans produce everyday...
"and another article whose conclusion was determined long before copy was written."
Umm, I don't know about you, but I prefer to read articles by people who actually have a clear conclusion and message they'd like to convey before they start writing the article!!
Paris, because she knows a thing or two about conclusions. Or is that happy endings?