Posts by that one in the corner 5065 publicly visible posts • joined 9 Nov 2021
> I can think of a few things to call Elon, but imbecile is not one of them.
> Seek professional help.
Like Joe, I too used to suffer from "Musk is an imbecile" syndrome.
I did as you advised and consulted a professional[1] and am happy to report that I shall no longer call Musk an "imbecile".
Instead, I shall use one or more of: ninnyhammer, clodpole, cuddy, loggerhead, gander or mome.
[1] lexicographer
> I would have told HR to tell him that they are going to let him go so he can find a position that is a better fit.
*Really* hoping that you are just using that wording because you know how that is how HR would end up phrasing it!
I find that kind of mealy mouthed wording absolutely infuriating and totally spineless.
"Let him go": especially in the US with the "at will" concept, there is no "letting" - saying "bugger off" is frankly politer than hearing that phrase.
"So he can find a better fit": if he felt there was somewhere he could better work, he could find it without being chucked out the door and having to look whilst bring unemployed (which is really going to help his negotiations, isn't it). Just say "we're fed up with you, please go away", at least that has a ring of honesty about it! Saying "please go away within the next month or we'll chuck you out the door" would be best.
By the descriptions given by Musk fans that Frohnhoefer's tweets were "lashing out with emotion", rude and didn't provide any factual answers. "Bad mouthing his boss" - another one of those UK/US things, as I missed where Frohnhoefer used the word "prat" (or equivalent).
Are we getting individualised Twitter results that means those people are reading some other tweets attributed to Frohnhoefer?
Fans also saying that he should have just replied with "Yes, Sir, 100% agree with you, Sir?". So the "disruptive engineers" that Elon hires should be disruptive lickspittles!
Aaargh, no, no, this has got me clicking on the "Show this thread" links now. Can't look away. Must look away. It's too late for me, run, save yourselves.
Aside from tap-to-pay, which by now should be an entirely off-the-shelf module, what is the need for wireless?
There is a honking great cable going from the car to the charge station - surely there is room in that for a little bit (more) data? A nice standardised panel on the dashboard and/or drop you phone into a cradle (if the link from your phone to your car is insecure then the problem is nothing to do with charging!)
Ditto the charger station, especially those that are grouped together (FWIW can't think of anywhere round here that has less than 5 or 6 charge bays) - they already have wires going to them, so wire them to a switch and use whatever is the most secure 'Net connection available, using all the standard encrypted links. Even if that is via radio - 4G - all that a nearby SDR ought to be able to do is disrupt the connection and drop the charger back to local control, not snarf credentials.
Musk has explicitly referred to the US First Amendment when talking about "free speech", so let's keep the discussion within his terms - just makes thing easier. So:
The First Amendment provides that Congress make no law respecting an establishment of religion or prohibiting its free exercise. It protects freedom of speech, the press, assembly, and the right to petition the Government for a redress of grievances.
https://www.whitehouse.gov/about-the-white-house/our-government/the-constitution/
There you go - Right to Free Speech is purely about whether the US government is allowed to stop you.
Committed 17000 commits last week[1], top engineer, keep him.
Only committed twice last month[2], fire her butt.
[1] search and replace to update the date on all the (C) notices in the source
[2] after 3 weeks poring over logs and wireshark dumps, found the chain of events that created a circular reference that prevented memory bring released and caused a key service to restart and reload its data every 3 hours.
> the motor industry has done pretty well at solving the problem of efficient manufacturing.
Lest we forget, Musk also "knows more about manufacturing than any other person on the planet" [1]
[1] sorry, haven't got a URL for that clip, but go watch Thunderf00t's latest YT video...
The specific architecture is the one called RPC - with the behaviour that has been described in many comments here, but put as simply as possible: your code invokes a procedure/function wibble() and it works (or returns a failure code). Your code has no idea if the function is satisfied locally or by invoking network traffic to a remote server: your code simply waits for the function to return. All data conversions required are hidden by RPC library.
This is very different in terms of how your code would use a different architecture, such as triggering an explicit request to a server - and expecting back a response that need not be handled by code anywhere near the code that caused the request to be sent.
I've not spotted anyone claiming that only one specific implementation should be considered *the* means of satisfying that architecture - can you say which one(s) you're thinking of?
And the RPC library is free to decide whether you even need to leave the boundaries of your current thread, process, machine or network; and whether any arguments and return values need to be footled with to be compatible with the function provider.
Your code just makes a simple function call, you compile it into your own library and the support decides the efficient way to satisfy it. Hopefully.
> unless there is one who publicly executes employees.
It wasn't an execution, he was just scheduled to polish the Falcon fairing and happened to be given an unfortunate timeslot. The straps? Oh, all our window cleaners use them, H&S you know. On backwards? Oh dear, such a klutz.
Now, whilst we talk about your code, could you just grab that chamois...
> the first chip NOT based on von Neuman architecture.
Except for anything based on Harvard Architecture, like, say, the AVR family [1] - Arduino ring a bell?
How about all your current x86-64 CPUs, with their Data Flow Architecture that allows out-order-execution to take place?
[1] yes, I know, it is a modified Harvard, but can you name a more well-known example?
> I've never met anyone that has attributed a dinner they cooked to the chef that wrote the recipe book either
Round these parts it is absolutely the norm to tell anyone who asks "Oh, got that from Delia" (or Fanny or Beeton, as appropriate).
Helps with Christmas pressies as well - if they liked the results (and they cook as well, of course) then they'll probably make good use of their own copy of the book. Unless it was the Corner Family Cake recipe - that we keep to ourselves.
Ditto with knitting & crochet patterns, music, circuit diagrams, LEGO models...
Guess it is all just down to the environment you grew up in.
It would be non-trivial to store all those labels on each node as its weighting was changed during the training. And then to sensibly track through what nodes were involved, and how much, in generating this bit of the output, gathering the labels and finally generate some useful results from the whole lot (hey, we can train an ML system to do that).
The real question is, how much extra computation & storage would that require and is it worth it to MS to pay that cost? Given that they could fling those machine resources at building up the main Copilot model.
Or they could take the money needed to pay for that extra processing for X months and buy their way to some legal ruling that says Copilot is totally fine to do what it is doing. And for months X+1, X+2 etc the legal team who managed it get 1/k of the saved costs in their paypacket (k being whatever they can screw out of MS - but those will be expert screwers, so may even end up with 1 > k > 0).
I spy, with my little eye, someone who has absolutely no idea what the concepts of "copyright" and "licensing" mean!
Firstly, except for a few outlier cases (unless you want to bring them up), *all* of the source code you are ever going to interact with is copyrighted.
Secondly, *all* licensing *requires* that the material be copyrighted, otherwise there is nothing that *can* be licensed.
Thirdly, (ignoring courtesy, academic requirements and plain common sense) it is the licence provisions that say under what conditions the code can be copied and whether attribution is required.
Unless they state otherwise (because they can put into their books code from other people, so long as they obey the relevant licence - and that code can include code from yet other people, so long as ...) then all the code in the book is (c) O'Reilly. Your use of that code is controlled by their licence, which itself is included inside the book (and is also included, verbatim, in the online provided-for-your-convenience copy which I linked to and which Richard 12 was kind enough to copy from the book as confirmation).
Finally, at the risk of breaking your brain, even if O'Reilly decided to publish a book that contained nothing more than 200 suitable source code examples - and clearly stated which bits were copyrighted by whom - with nothing more than a contents page and a blurb on the back saying "this is a collection of other people's copyrighted source code, used under the terms of the the MIT licence, arranged so that you can learn by working your way through from start to finish" they would STILL clearly mark that book as "Copyright O'Reilly and Associates". And be totally correct in doing so.
That is true.
Which raises the question whether "enabled" times out and resets, putting any unresponded-to flaws public? As how many people (promptly) formally abandon a project rather than wandering away and forgetting about it...
Maybe I'm overthinking this.
From the article, it sounds like the default for this private reporting is "off" until each repos owner enables it. Which seems a bit odd. In fact, disabling it at all seems odd - unless it has some clear disadvantage (which would be?) why would anyone want to disable it?
The only thing that comes to mind is that this might be a route to spam a project with bogus reports in a way that isn't visible to the rest of the world (compared to creating 10,000 bogus bug reports in the public tracker). But you'd expect at least a warning in the announcement if there was a flaw that big in the system...
Can anyone with more experience than me (not hard) of being a Github project owner shed some light on this?
Leaving aside the US Army not even knowing what is in their own app (!), all the others "knew* they were "surveiling" their users (aka spying on them, data rape etc) but that was ok because they are all Good Guys.
But as soon as there is even a hint that the Russians could be getting the same data, throw their hands up in feigned shock "who could be so dastardly?"
So, are they going to give up spying on their own users or just find a good old US replacement and carry on, business as usual?
Though one bit I liked: members of the NRA don't need to feel paranoid any more, now they know who was watching them.
Also, it took me ages to learn how to say "Novosibirsk" with confidence, glad to have an excuse to trot it out again.
To be fair, these stories are up against some *really* tough opposition at the moment (we see you lurking there, Elon).
Any story sent in at the moment is going to feel like small change; fingers crossed that El Reg knows that and is holding back some good ones to use after we've regained our sense of normality. Please let be the case. Pretty please?
So many PC cases have ludicrous numbers of colourful lights in them that last time I rebuilt my home PC I couldn't even source a motherboard without weird lighting effects (thankfully disabled in BIOS).
Maybe we can convince the youngsters that, if even whitebeards have them, LEDs are totally passe (e acute that is meant to be) that they have to move onto The Next Big Thing - which is, of course, the USB-C powered Jacob's Ladder. That should be good for a few hijinks.
So how long does this ML take to train on a new user's gait? Do you spend the first 100 hours wondering if, this time, your left foot goes forwards whilst the right one puts on full braking or, like yesterday, it will be a luxurious smooth ride - backwards?
Remember, the major problem with seven league boots was the groin strain as one foot suddenly shoots out 38 kilometres past the other.
In response to a question about this in October, the Teesside principal investigator for the Novavax and Moderna COVID-19 vaccine trials said that the latest indications are that immunity from Covid variants is short lived, to the extent that they no longer expect to see "herd immunity" effects.
This is the same whether immunity comes from vaccination or acquired from infection.
Oh, and the presence of testable antibodies does *NOT* mean you are "immune" and certainly not "naturally immune", whatever that is supposed to mean.
Barrier? Will there be red flashing lights and an excited voice shouting:
"1.9 exa-flops; 1.95 exa-flops; she won't make it, we've got to shut it down; 1.99 exa-flops".
Camera flare streaks across the room; an exhaust fan shrieks, winding up past top-C, the sudden gust blowing off the young Assistant Programmer's glasses and hair scrunchy: "My God, Albert, you are beautiful".
"1.999 exa-flops." A loud cracking noise, like a whip wielded by the gods themselves; a visible shockwave passes through the room, rattling the processor cages.
"TWO exa-flops! 2.1! 2.2!" The noise recedes, the room stops shaking, the emergency lighting switches back to white. "Holding steady at 2.2 exa-flops, Sir".
Close up on the main console, as a shaking hand approaches a large button then, decisively, punches it: "Press release away and running. We've done, by Heaven, we've done it!".
> then lift the pen and wander off to another corner apparently unconnected
I spent many hours trying to stop the plotter doing that, starting with code that worked fine drawing on the screen but was sub-optimal when the same commands were sent to the plotter: did you know that changing the pen colour on the screen is faster than on the plotter? Bloke who first wrote this code didn't! [1]
Great fun was had with scaling for various paper sizes, providing trim marks with overlaps so that you could sticky tape multiple plots together without leaving gaps. These were Gantt charts and the like for some rather large projects. In order to try and avoid customers managing to hit untested edges cases, the program was written to cope with pages from approximately A-3 to A12. The latter tested by sending the HP-GL code into a little formatter that printed lots of them on A4 via the laserwriter, then cutting out the postage stamps and sticking them all together. Then handing the results over to our Project Manager as a copy of our own latest progress (he took it well).
All done pre-Windows in TurboPascal, of course. None of your nancy-boy loadable printer drivers here, t'were all done with overlays.
[1] yes, that was also me.
Racism?
I just re-read the article and didn't spot a single reference to any characteristic of any individual at all. Did you perhaps misread "Qatar snoops" as "Qatari snoops"?
No, wait, spotted it: you think that "regime" is an ethnicity and all the members of that group are being labelled as authoritarian!
