To be fair, Ajit Pai was Asshat Number One at the FCC before Trump, too. It's just that he's in charge now.
Not that Wheeler was all that wonderful either; but Pai seems to regard consumers as an unpleasant side effect of business. He's the sort of Randian who longs for the days of the robber barons.
Hell, the guy used to be Associate GC at Verizon. That's not a job they hand out to someone inclined to protect consumer interests.
It seems obvious that the journalist has copy-pasted the sentence and then proceeded to remove references to Uzbekistan, Estonia, and Belarus. This edit strikes me as odd.
I suppose it would be too much to ask that you spend a minute doing a little research to find out?
Google Perspective - which the developers say, over and over, is an experimental tool which should not be used in production because it's far too inaccurate[1] - evaluates inputs using a logistic-regression model that's built using supervised learning, trained on a handful of corpora taken from major sites (Wikipedia is one) that were labelled by a large panel of human judges.
In other words, they don't have a definition for "toxic". They let the human judges decide for themselves, and each judge's input influences the model somewhat, but the overall model reduces the impact of outliers (by using logistic regression, obviously).
Yes, the Jigsaw team is taking a particular rhetorical stance by using the term "toxic". That has little or no bearing on the success or failure of their approach. Currently I'd rate that approach moderately interesting, evolutionary,[2] and not ready for production.
What's actually more interesting is how they've packaged it. There's a github repository (with associated research materials and whatnot), an easy-to-use web service, and an experimental "evaluate as I write" web application. Various researchers have been playing with that last sort of thing for over a decade (Google itself had the great auto-plagiarizing tool Google Writer), but we haven't seen it in the mainstream yet. That sort of computer-assisted writing is coming - it's relatively easy to implement, it's very easy to use, it fits with our existing infrastructure, it offers quick rewards - so it's good to be aware of it in these early days.
[1] In terms of both precision (too many false positives) and recall (too many false negatives). They note specifically that it doesn't recognize many patterns identified by human judges, and that it incorrectly matches many inputs against patterns it does know.
[2] I know academic researchers have been discussing applying these sorts of approaches to evaluating comments for various metrics such as "helpfulness" at least since 2008, because I was one of them then. So this is not an unexpected development.
If I was in charge, there would be a whole lot of people being marched out the door
Yes. Including a number of people in HR, based on Fowler's account - and she says she has evidence (email logs and the like). Not that I'm inclined to doubt her anyway; her story is completely in keeping with previous Uber scandals.
The responses she reports getting from HR are not just unethical; several of them are outright illegal. This is systematic lawbreaking, not simply bad corporate culture. I wouldn't be surprised if some enterprising legal firm isn't looking up former female employees and trying to put together a class-action suit.
there certainly are revolutions in scientific advancement, typically called "paradigm shifts"
Kuhn's version of the history of science is popular, but it's far from universally accepted. Contrast, say, Paul Thagard's, or Paul Feyerabend's, or David Stove's.
While scientific thought, in its modern Western conception,[1] has certainly gone through any number of revisions, some of them dramatic,[2] the paradigm-shift model is not "certainly" an adequate or accurate description of those changes. Maybe it's a pretty good one; some people argue it's the best we have. But it's not a settled matter.
[1] So qualified because talking about the history of "scientific thought" in general is hopelessly vague. It is, of course, possible to discuss scientific methodology as a putatively culturally-neutral realm, but in discussions of how a particular body of thought has developed, it's unproductive to pretend that's not historically and culturally specific.
[2] There are any number of popular examples: Darwin-Wallace evolution, special and general relativity, the Copenhagen interpretation, dismissal of luminiferous ether, and so on.
Surely, without Thales' work, Aristotle blah blah & cetera.
There's plenty of room in the past.
(And we can regress down other lines too. We wouldn't have miniaturization without the various techniques for precise machining, or without quality steel. We wouldn't have semiconductors, integrated circuits, or high-capacity batteries without the work of all those chemists who isolated elements and explored their properties. We wouldn't have cheap commodity computers without video games.)
You have to hand it to the PHP crew - they actually get shit done, no matter what.
Indeed. Shit is precisely what they get done. Indeed, despite PHP's lack of originality in most areas, it has been quite inventive in the production of shit.
The next version will support "====" for things are really, *really* equal.
We won't be safe until PHP adds the ===== operator, which is always false.
PHP also could really use the stricter ====== operator, which terminates your program and wipes the disk regardless of its operands.
Of course I jest. PHP will never get better, we will never be safe, and the former is only one of many causes of the latter. But given the PHP designers' inclination to add everything plus a kitchen sink with swapped hot and cold valves and a clogged trap to the language, it's surprising that they haven't added operators for doxastic logic:
x ?= y is true if the program believes x is equivalent y.
x ?== y is true if the program believes x and y are the same entity.
x ??= y is true if the program believes that it believes that x is equivalent to y.
Aside from some contemporary proprietary dialects, COBOL doesn't use = for assignment either; you use the MOVE or SET verb, or an assignment clause such as GIVING with verbs that have such clauses.
LISP and descendants have SET, SETQ, and SETF for assignment per se, and possibly others (e.g. Scheme's set!), plus the binding forms LET and LETREC (and possibly others).
APL uses left-arrow for assignment, obviously.
Forth uses !.
Don't they teach programming language history these days?
When I look at my smartphone I see the fucking Eye of Sauron.
Scary little fucking things.
Easily fixed. Just use an Android phone and wait until it gets into the inevitable state of running the battery down in a couple of hours, regardless of what's on it and what's enabled. Deniable privacy!
I taught myself some APL[1] using a standard keyboard and an APL implementation that ran on OS/2, if memory serves. Wasn't bad. Yes, you had to compose all the APL characters using Alt-key sequences[2], but APL is so terse that you didn't have to do it very often.
J uses ASCII rather than APL's grab-bag of symbols. I'm not sure I'd call it more readable, though.
It would be interesting to see a less-opaque Iverson-Backus language, with the matrix operators and tacit-programming features.
[1] Because I believe it's useful to try as many programming-language paradigms as possible, and because I'd read Iverson's and Backus' Turing Award addresses.
[2] Can't remember whether Alt-Right gave you the GoTo operator or the NarcissisticPersonalityDisorder operator.
when you pay a high price for something, you tell your self [sic] it's worth it, otherwise you have to admit to yourself your [sic] an idiot!
Care to cite a methodologically-sound study that shows a high correlation between price and consumer satisfaction for automobiles?
Surveys suggest there isn't one. Last year, the highest-rated brand was luxury Lincoln, but number two was Honda, and Toyota outranked its luxury marque Lexus. For that matter, Honda's luxury line Acura was the lowest-scoring brand in the survey.
If people did feel compelled to defend their choice to pay more for a luxury brand, then they should rate Acuras (largely just rebadged Hondas) higher than the cheaper equivalent Honda models. They don't - quite the reverse.
The profit margins on Tesla autos indicate that they're not Veblen goods, so people aren't buying them simply for conspicuous consumption. Tesla unit profit margins are pretty good - around 19% - but they benefit from selling directly rather than through dealers (who take a cut, and who discount cars from their MSRP to get sales). Of course selling directly isn't an unalloyed benefit (which is why other manufacturers aren't doing it); it means lower volume, so less economy of scale and less market penetration.
Indeed it appears that, as with many commodity goods, consumers generally treat car prices as signaling quality, and are therefore more likely to be more critical of expensive models.
I'm no fan of the Tesla vehicles - they're completely unsuited to my needs, and I'm not impressed by gadgets and technological gee-whiz features that offer little benefit. But independent reviews suggest that at least some Tesla models are indeed quite good, for the use cases they satisfy.
And the marquee element. Bob Whipple has a nice little essay on those two in From A to <A>.
Animated GIFs are definitely up there too.
But to be fair, when the web began, there was no audio, video, blink, or marquee. I'm trying to remember what we complained about the most back in the day. The fact that it was slower than gopher and largely duplicated WAIS, maybe. The lack of persistent connections and consequent always-in-slow-start behavior. Starting Lynx and discovering your $TERM is wrong, or the system you're using doesn't have the right termcap or terminfo. Eternal September.
Sounds like what we need is a Greasemonkey script that finds audio and video elements, removes them from the DOM, and replaces them with simple visual elements with onclick listeners that restore the original audio or video element. Probably it should remove the autoplay attribute in the process, and force the control attribute on. There may be such a script already at userscripts.org. (I'd write one myself if I didn't have plenty of real work stacked up. Though if I get sufficiently annoyed at some A/V-using site I might do it anyway.)
Then all you need to do is use Firefox, install Greasemonkey, and install that user script.
Of course, some sites will inject audio/video out of band using scripting, iframes, and other trickery. It'll be an arms race, as it always is. But even a relatively simple implementation like the one described above would cut a great deal of this crap. (Probably not in the specific case of Facebook, but that's the price you pay for using Facebook. I know, I know, for many people it's an excellent way to keep in touch with friends and family. I'm thinking the only real solution is a custom proxy that reads m.facebook.com and strips out as much crap as possible before passing it on.)
Circa 1989 I was working at IBM, on a graphics project. We were using IBM PC RTs, the first IBM UNIX desktop line, before the RS/6000 (which was about to start shipping at that point; we had a number of pre-release ones around the office, with the Texan badges).
Since we were working on graphics, we'd given all the machines the names of artists for their hostnames. Mine was matisse, so I had a poster of Matisse's Blue Nude hanging over my desk. And one day, to avoid doing real work, I changed my/etc/motd to an ASCII recreation of the Blue Nude I painstakingly crafted by hand, as one did Back in the Day. I took the liberty of inscribing "IBM" across her navel, and captioned it "Big Blue Nude".
Well. Some weeks later one of the IBM High and Mighty came to the office, and I was demonstrating some feature we'd implemented, and was told that nudity was Not Appropriate in an IBM office.
After the representative of the aristocracy departed, I stuck Post-It notes over the lascivious portions of my Matisse poster, and edited motd to put a diagonal "CENSORED" over the BBN's nipples. My manager decreed that sufficient and the BBN motd remained until I left the company a couple of years later (and perhaps longer).
I think I still have a copy somewhere. Of course it is preserved in Usenet of blessed memory, and anyone half-competent at searching should be able to find it.
I liked the Telebit Trailblazers myself. Robust, fast, could spoof UUCP and X-Modem and various other protocols... I used to use a pair to work remotely between Ohio and Massachusetts, sometimes with a regular getty shell, sometimes UUCP file transfers, sometimes SLIP and full IP. Even at V.42bis speeds it was quite usable.
Good to know the Internet's Cult of Tesla is alive and well. Y'know, Matthew Inman drawing a funny picture doesn't automatically make something true.
Sure, when I read a biography of Tesla, as a lad decades ago (and, yes, before Tesla became cool again), I thought he was pretty nifty. I'd've been a little leery even then about "genius" and "brilliant engineer", though. Tesla had some excellent ideas, and many unworkable ones. He wasn't able to bring many of his more famous ideas to fruition - generally because his process wasn't sufficiently rigorous and his attention wandered, not due to the machinations of the mustache-twirling villain Edison.
Was Edison a bastard who played dirty tricks on his competitors? Sure. Did he take credit for the work of his employees and others? Absolutely. But he was also a knowledgeable and capable engineer who worked very hard himself.
And the simplistic evaluation of the relative merits of their ideas is sophomoric. Take electricity transmission. At the time, the Tesla/Westinghouse A/C system had undeniable advantages over Edison's plan for low-voltage DC with a generating plant on every city block. These days, we wish we weren't stepping point-of-use voltages down quite so far - the US would be better off with ~220V, like Europe. But at the time that might have strained insulator technology. And these days, for long-distance transmission, we'd really rather be using HVDC - an approach that neither Tesla nor Edison mooted, as far as I know. Meanwhile we have Tesla's various stabs at wireless transmission using resonant coupling, which is only now becoming practical and then only for short-range charging of batteries, because of inefficiencies and range limitations. Turns out you really want a waveguide for power transmission.
They were both trying shit out. Some of it was workable; much of it wasn't.
It is not money that gives Elon credibility, but his list of accomplishments. He has a track record of turning ideas into reality, so people pay attention.
Texas Sharpshooter Fallacy.
Musk has had success (according to certain metrics) with some ideas. He has a lot of ideas. Attributing success to some of them after the fact says nothing about the potential of others.
And then there's the question of originality. How this latest babble is newsworthy is beyond me - it's been an SF cliche for decades. The earliest that I can think of off the top of my head is Bester's The Computer Connection, which is 40-odd years old, but the conceit has roots at least as far back as Smith's "Scanners Live in Vain" (1950).
I know of only two people using PGP for personal email purposes
There's little point in doing so, because few recipients have PGP. The presence of PGP signatures just confuses or worries many recipients, and of course if they don't have a PGP-compliant MUA (or they do but don't have a key pair, or you don't have their public key) you can't send them encrypted email.
I've had a PGP key for over a decade. In fact I have both personal and work keys. I think the only time I use them for email is for communicating with external security researchers or with my fellow security-team members.
But I agree that PGP (that is, implementations of OpenPGP, such as PGP and gpg) has usability issues, and most of the various wrappers around PGP implementations that I've seen - including Enigmail - aren't that much better.
S/MIME has some usability advantages over PGP implementations, but many MUAs don't support it, and it has all the PKI issues of the X.509 certificate hierarchy (which are legion). The OpenPGP PKI is a mess, of course, with its chaos of web-of-trust, keyservers, arbitrarily publishing keys on web pages and the like, sending keys by email, sending keys by mental telepathy, etc; but that chaos lowers the barrier to entry. These days getting a personal email certificate signed by a widely-recognized public CA isn't too bad (though baffling for non-experts, and generally done with poor key hygiene), but for most of its 15-year history getting personal certificates was a mysterious and non-free process for ordinary users.
And then there's PEM. Oh, PEM, we hardly knew ye. (Except for Base64, of course.)
Slashdot Mirror
They should try one of those metal detectors that you see advertised in comic books. I understand those things always find valuable metal objects.
To be fair, Ajit Pai was Asshat Number One at the FCC before Trump, too. It's just that he's in charge now.
Not that Wheeler was all that wonderful either; but Pai seems to regard consumers as an unpleasant side effect of business. He's the sort of Randian who longs for the days of the robber barons.
Hell, the guy used to be Associate GC at Verizon. That's not a job they hand out to someone inclined to protect consumer interests.
I'm dubious. They say the Niels Bohr page is constantly in flux from bot edits, but every time I look at it, it says one thing or another.
It seems obvious that the journalist has copy-pasted the sentence and then proceeded to remove references to Uzbekistan, Estonia, and Belarus. This edit strikes me as odd.
A bot did it.
I suppose it would be too much to ask that you spend a minute doing a little research to find out?
Google Perspective - which the developers say, over and over, is an experimental tool which should not be used in production because it's far too inaccurate[1] - evaluates inputs using a logistic-regression model that's built using supervised learning, trained on a handful of corpora taken from major sites (Wikipedia is one) that were labelled by a large panel of human judges.
In other words, they don't have a definition for "toxic". They let the human judges decide for themselves, and each judge's input influences the model somewhat, but the overall model reduces the impact of outliers (by using logistic regression, obviously).
Yes, the Jigsaw team is taking a particular rhetorical stance by using the term "toxic". That has little or no bearing on the success or failure of their approach. Currently I'd rate that approach moderately interesting, evolutionary,[2] and not ready for production.
What's actually more interesting is how they've packaged it. There's a github repository (with associated research materials and whatnot), an easy-to-use web service, and an experimental "evaluate as I write" web application. Various researchers have been playing with that last sort of thing for over a decade (Google itself had the great auto-plagiarizing tool Google Writer), but we haven't seen it in the mainstream yet. That sort of computer-assisted writing is coming - it's relatively easy to implement, it's very easy to use, it fits with our existing infrastructure, it offers quick rewards - so it's good to be aware of it in these early days.
[1] In terms of both precision (too many false positives) and recall (too many false negatives). They note specifically that it doesn't recognize many patterns identified by human judges, and that it incorrectly matches many inputs against patterns it does know.
[2] I know academic researchers have been discussing applying these sorts of approaches to evaluating comments for various metrics such as "helpfulness" at least since 2008, because I was one of them then. So this is not an unexpected development.
Yes, your anecdote definitely proves the general case.
I know, I know. Critical thinking is hard.
If I was in charge, there would be a whole lot of people being marched out the door
Yes. Including a number of people in HR, based on Fowler's account - and she says she has evidence (email logs and the like). Not that I'm inclined to doubt her anyway; her story is completely in keeping with previous Uber scandals.
The responses she reports getting from HR are not just unethical; several of them are outright illegal. This is systematic lawbreaking, not simply bad corporate culture. I wouldn't be surprised if some enterprising legal firm isn't looking up former female employees and trying to put together a class-action suit.
there certainly are revolutions in scientific advancement, typically called "paradigm shifts"
Kuhn's version of the history of science is popular, but it's far from universally accepted. Contrast, say, Paul Thagard's, or Paul Feyerabend's, or David Stove's.
While scientific thought, in its modern Western conception,[1] has certainly gone through any number of revisions, some of them dramatic,[2] the paradigm-shift model is not "certainly" an adequate or accurate description of those changes. Maybe it's a pretty good one; some people argue it's the best we have. But it's not a settled matter.
[1] So qualified because talking about the history of "scientific thought" in general is hopelessly vague. It is, of course, possible to discuss scientific methodology as a putatively culturally-neutral realm, but in discussions of how a particular body of thought has developed, it's unproductive to pretend that's not historically and culturally specific.
[2] There are any number of popular examples: Darwin-Wallace evolution, special and general relativity, the Copenhagen interpretation, dismissal of luminiferous ether, and so on.
Surely, without Thales' work, Aristotle blah blah & cetera.
There's plenty of room in the past.
(And we can regress down other lines too. We wouldn't have miniaturization without the various techniques for precise machining, or without quality steel. We wouldn't have semiconductors, integrated circuits, or high-capacity batteries without the work of all those chemists who isolated elements and explored their properties. We wouldn't have cheap commodity computers without video games.)
And watch this get modded "Troll" since blah blah blah I'm so fucking daring.
Sigh. The "call me a troll" prolepsis was a tired, trite cliche on Usenet in 1990.
Eternal September remains eternal.
You have to hand it to the PHP crew - they actually get shit done, no matter what.
Indeed. Shit is precisely what they get done. Indeed, despite PHP's lack of originality in most areas, it has been quite inventive in the production of shit.
The next version will support "====" for things are really, *really* equal.
We won't be safe until PHP adds the ===== operator, which is always false.
PHP also could really use the stricter ====== operator, which terminates your program and wipes the disk regardless of its operands.
Of course I jest. PHP will never get better, we will never be safe, and the former is only one of many causes of the latter. But given the PHP designers' inclination to add everything plus a kitchen sink with swapped hot and cold valves and a clogged trap to the language, it's surprising that they haven't added operators for doxastic logic:
x ?= y is true if the program believes x is equivalent y.
x ?== y is true if the program believes x and y are the same entity.
x ??= y is true if the program believes that it believes that x is equivalent to y.
A few more rules, and PHP becomes Godel-complete!
Pascal got the := operator from ALGOL.
Aside from some contemporary proprietary dialects, COBOL doesn't use = for assignment either; you use the MOVE or SET verb, or an assignment clause such as GIVING with verbs that have such clauses.
LISP and descendants have SET, SETQ, and SETF for assignment per se, and possibly others (e.g. Scheme's set!), plus the binding forms LET and LETREC (and possibly others).
APL uses left-arrow for assignment, obviously.
Forth uses !.
Don't they teach programming language history these days?
When I look at my smartphone I see the fucking Eye of Sauron.
Scary little fucking things.
Easily fixed. Just use an Android phone and wait until it gets into the inevitable state of running the battery down in a couple of hours, regardless of what's on it and what's enabled. Deniable privacy!
Damn but I miss Symbian S60.
APL was effectively superseded by the J language.
I taught myself some APL[1] using a standard keyboard and an APL implementation that ran on OS/2, if memory serves. Wasn't bad. Yes, you had to compose all the APL characters using Alt-key sequences[2], but APL is so terse that you didn't have to do it very often.
J uses ASCII rather than APL's grab-bag of symbols. I'm not sure I'd call it more readable, though.
It would be interesting to see a less-opaque Iverson-Backus language, with the matrix operators and tacit-programming features.
[1] Because I believe it's useful to try as many programming-language paradigms as possible, and because I'd read Iverson's and Backus' Turing Award addresses.
[2] Can't remember whether Alt-Right gave you the GoTo operator or the NarcissisticPersonalityDisorder operator.
Plato is a fictional shadow-puppet made up by that guy standing at the mouth of the cave.
when you pay a high price for something, you tell your self [sic] it's worth it, otherwise you have to admit to yourself your [sic] an idiot!
Care to cite a methodologically-sound study that shows a high correlation between price and consumer satisfaction for automobiles?
Surveys suggest there isn't one. Last year, the highest-rated brand was luxury Lincoln, but number two was Honda, and Toyota outranked its luxury marque Lexus. For that matter, Honda's luxury line Acura was the lowest-scoring brand in the survey.
If people did feel compelled to defend their choice to pay more for a luxury brand, then they should rate Acuras (largely just rebadged Hondas) higher than the cheaper equivalent Honda models. They don't - quite the reverse.
The profit margins on Tesla autos indicate that they're not Veblen goods, so people aren't buying them simply for conspicuous consumption. Tesla unit profit margins are pretty good - around 19% - but they benefit from selling directly rather than through dealers (who take a cut, and who discount cars from their MSRP to get sales). Of course selling directly isn't an unalloyed benefit (which is why other manufacturers aren't doing it); it means lower volume, so less economy of scale and less market penetration.
Indeed it appears that, as with many commodity goods, consumers generally treat car prices as signaling quality, and are therefore more likely to be more critical of expensive models.
I'm no fan of the Tesla vehicles - they're completely unsuited to my needs, and I'm not impressed by gadgets and technological gee-whiz features that offer little benefit. But independent reviews suggest that at least some Tesla models are indeed quite good, for the use cases they satisfy.
And the marquee element. Bob Whipple has a nice little essay on those two in From A to <A>.
Animated GIFs are definitely up there too.
But to be fair, when the web began, there was no audio, video, blink, or marquee. I'm trying to remember what we complained about the most back in the day. The fact that it was slower than gopher and largely duplicated WAIS, maybe. The lack of persistent connections and consequent always-in-slow-start behavior. Starting Lynx and discovering your $TERM is wrong, or the system you're using doesn't have the right termcap or terminfo. Eternal September.
Sounds like what we need is a Greasemonkey script that finds audio and video elements, removes them from the DOM, and replaces them with simple visual elements with onclick listeners that restore the original audio or video element. Probably it should remove the autoplay attribute in the process, and force the control attribute on. There may be such a script already at userscripts.org. (I'd write one myself if I didn't have plenty of real work stacked up. Though if I get sufficiently annoyed at some A/V-using site I might do it anyway.)
Then all you need to do is use Firefox, install Greasemonkey, and install that user script.
Of course, some sites will inject audio/video out of band using scripting, iframes, and other trickery. It'll be an arms race, as it always is. But even a relatively simple implementation like the one described above would cut a great deal of this crap. (Probably not in the specific case of Facebook, but that's the price you pay for using Facebook. I know, I know, for many people it's an excellent way to keep in touch with friends and family. I'm thinking the only real solution is a custom proxy that reads m.facebook.com and strips out as much crap as possible before passing it on.)
ASCII pr0n
Ah, yes. That reminds me...
Circa 1989 I was working at IBM, on a graphics project. We were using IBM PC RTs, the first IBM UNIX desktop line, before the RS/6000 (which was about to start shipping at that point; we had a number of pre-release ones around the office, with the Texan badges).
Since we were working on graphics, we'd given all the machines the names of artists for their hostnames. Mine was matisse, so I had a poster of Matisse's Blue Nude hanging over my desk. And one day, to avoid doing real work, I changed my /etc/motd to an ASCII recreation of the Blue Nude I painstakingly crafted by hand, as one did Back in the Day. I took the liberty of inscribing "IBM" across her navel, and captioned it "Big Blue Nude".
Well. Some weeks later one of the IBM High and Mighty came to the office, and I was demonstrating some feature we'd implemented, and was told that nudity was Not Appropriate in an IBM office.
After the representative of the aristocracy departed, I stuck Post-It notes over the lascivious portions of my Matisse poster, and edited motd to put a diagonal "CENSORED" over the BBN's nipples. My manager decreed that sufficient and the BBN motd remained until I left the company a couple of years later (and perhaps longer).
I think I still have a copy somewhere. Of course it is preserved in Usenet of blessed memory, and anyone half-competent at searching should be able to find it.
I liked the Telebit Trailblazers myself. Robust, fast, could spoof UUCP and X-Modem and various other protocols... I used to use a pair to work remotely between Ohio and Massachusetts, sometimes with a regular getty shell, sometimes UUCP file transfers, sometimes SLIP and full IP. Even at V.42bis speeds it was quite usable.
Most students, however, are just plain disastrous
Citation needed.
You can qualify your bogus generalization as much as you like. It's still a bogus generalization.
Good to know the Internet's Cult of Tesla is alive and well. Y'know, Matthew Inman drawing a funny picture doesn't automatically make something true.
Sure, when I read a biography of Tesla, as a lad decades ago (and, yes, before Tesla became cool again), I thought he was pretty nifty. I'd've been a little leery even then about "genius" and "brilliant engineer", though. Tesla had some excellent ideas, and many unworkable ones. He wasn't able to bring many of his more famous ideas to fruition - generally because his process wasn't sufficiently rigorous and his attention wandered, not due to the machinations of the mustache-twirling villain Edison.
Was Edison a bastard who played dirty tricks on his competitors? Sure. Did he take credit for the work of his employees and others? Absolutely. But he was also a knowledgeable and capable engineer who worked very hard himself.
And the simplistic evaluation of the relative merits of their ideas is sophomoric. Take electricity transmission. At the time, the Tesla/Westinghouse A/C system had undeniable advantages over Edison's plan for low-voltage DC with a generating plant on every city block. These days, we wish we weren't stepping point-of-use voltages down quite so far - the US would be better off with ~220V, like Europe. But at the time that might have strained insulator technology. And these days, for long-distance transmission, we'd really rather be using HVDC - an approach that neither Tesla nor Edison mooted, as far as I know. Meanwhile we have Tesla's various stabs at wireless transmission using resonant coupling, which is only now becoming practical and then only for short-range charging of batteries, because of inefficiencies and range limitations. Turns out you really want a waveguide for power transmission.
They were both trying shit out. Some of it was workable; much of it wasn't.
It is not money that gives Elon credibility, but his list of accomplishments.
He has a track record of turning ideas into reality, so people pay attention.
Texas Sharpshooter Fallacy.
Musk has had success (according to certain metrics) with some ideas. He has a lot of ideas. Attributing success to some of them after the fact says nothing about the potential of others.
And then there's the question of originality. How this latest babble is newsworthy is beyond me - it's been an SF cliche for decades. The earliest that I can think of off the top of my head is Bester's The Computer Connection, which is 40-odd years old, but the conceit has roots at least as far back as Smith's "Scanners Live in Vain" (1950).
I know of only two people using PGP for personal email purposes
There's little point in doing so, because few recipients have PGP. The presence of PGP signatures just confuses or worries many recipients, and of course if they don't have a PGP-compliant MUA (or they do but don't have a key pair, or you don't have their public key) you can't send them encrypted email.
I've had a PGP key for over a decade. In fact I have both personal and work keys. I think the only time I use them for email is for communicating with external security researchers or with my fellow security-team members.
But I agree that PGP (that is, implementations of OpenPGP, such as PGP and gpg) has usability issues, and most of the various wrappers around PGP implementations that I've seen - including Enigmail - aren't that much better.
S/MIME has some usability advantages over PGP implementations, but many MUAs don't support it, and it has all the PKI issues of the X.509 certificate hierarchy (which are legion). The OpenPGP PKI is a mess, of course, with its chaos of web-of-trust, keyservers, arbitrarily publishing keys on web pages and the like, sending keys by email, sending keys by mental telepathy, etc; but that chaos lowers the barrier to entry. These days getting a personal email certificate signed by a widely-recognized public CA isn't too bad (though baffling for non-experts, and generally done with poor key hygiene), but for most of its 15-year history getting personal certificates was a mysterious and non-free process for ordinary users.
And then there's PEM. Oh, PEM, we hardly knew ye. (Except for Base64, of course.)