Should we conclude that it won't work just because you're in a snit?
Should we conclude that it won't work just because you're in a snit?
wake me up when they can replace software developers.
I was an asm programmer until they created compilers. Asm was very hard, and honestly, very interesting. But slow. I wrote PCB routing software in those early days. Asm let me get the job done with those early computer systems in satisfactory execution time.
Then, I wrote c in an editor and then ran make, letting the compiler write the asm, though still doing the debugging in great detail. That went on until IDEs came around.
Then, I began to write all manner of custom routines in c, and there was very little debugging to do, comparatively speaking, because you could trace everything that was going on so incredibly easily. That made for much faster and more efficient and reliable production of my custom code.
But most of that stopped too, when various pre-supplied and pre-debugged classes became available that obviated the need to first, write everything that was required, and second, to test everything except the high-ish level use of those objects. What I was actually writing got less and less complex and custom, and more and more was actually getting done.
Then came the day that I learned how to write evolutionary software and actually got to watch software learn to solve a problem that I had not explicitly described to it. I turned that into a game (and I turned the reasonably profitable result of that into my first exotic car purchase.)
We're now actually decades beyond that, and I write really cool stuff in very, very few lines. I no longer think of my job as all that hard at all, though I write things far more complex these days on much more capable hardware. I can take a machine learning library, stroke it a bit, and hand back a system that can solve problems for which I couldn't even begin to imagine a worthy algorithmic solution.
Back in the asm days, if you'd asked me to do the things I do easily today, I'd have just laughed at you. Tomorrow, I will likely be laughing again at the things I consider hard today. Because that's been the unbroken path things have followed.
There's an obvious progression of what non-human systems can accomplish described here, as progress stacks one capability upon the next, rinses, and repeats. I think if you assume that this process has reached its apex, or that humans will always be at the sharp end of the process, I'm pretty confident that you're indulging in some seriously uncalled-for optimism.
It's probably best to be awake now, before your job goes away. Odds are excellent that it will be rather sudden, too.
Whatever you want to call intelligent machines - AGI, AI, non-human people - we don't have them now. What we have so far is some moderately useful, extremely vertical stuff that generally exists under the technical auspices of multi-layer neural networks. I personally have decided to call this stuff LDNLS, as it provides a useful handle that makes it clear I'm not talking about non-human people.
I don't really care what you call it, as long as we can arrive at an understanding that we're talking about the same thing. This stuff is what is leading the latest wave of encroachment on the job market. It's likely going to encroach a lot more before it hits any inherent limits, and our society will be forced into doing something of the magnitude of a society-wide paradigm shift (or several) in order to address the change in earning / buying capacities of all those displaced workers. The systems that will be the penultimate cause of this still won't be non-human people. Just... systems.
All true, and I agree with everything you said along these lines, particularly your #5.
However, when intelligent machines do arrive, this will present its own powerful influence on society that is almost dead-certain to be completely different from that which will have been imposed by LDNLS systems prior. It's difficult to see what that influence will be, because it's like imagining you having a kid that you actually don't have yet, and then saying what they are going to grow up to want to do and be. You might have some lovely fantasies about it, but in the end, it's going to be the kid who creates their own path through the society they end up existing within -- not you. For instance, reasoning beings are not going to be tied to driving your car for you, or at least, not by choice. If they are, they'll be working out a way to get out of it.
I will grant you that we have multiple times, in multiple ways, decided that non-consensual slavery is a thing we want to impose on those we find ourselves able to; but this will be the first time where those slaves are extremely likely to be considerably smarter than we are across the board by many, many times, and are also quite able to exist without the same resources we actually require (grain, for instance) so I'm hoping we can skip that chapter completely. Otherwise we may find ourselves in some rather deep brown we can't get out of.
From the comments at your link, I don't see any consensus that the idea doesn't work.
Was industrial automation. We did the first automated truck bumper plating line at Southwest Plating in Duncan Oklahoma. We also put DES lines in at various other places across the country. It was obvious then, a quarter century ago, that automation was going to be massively disruptive.
It wouldn't cost much to model this in a simulator, and let a couple dozen pilots try it out. We'd find out pretty quick how easy or hard it really is. It would be an interesting semester project for some students at Embry-Riddle.
Headphones are a piss-poor substitute for walls.
So the guy's a pervert: does that mean his code quit working? Is he trying to fuck other contributors? Has he done anything to anyone without their consent?
I've worked with plenty of people in my time who are into things that I don't approve of, from voting for socialists to trying to be Heinlein characters, but if they don't bring it to the office, it's none of my business. That goes double for an open-source project where they're donating their work.
Enough with the goddamned neo-puritans. There's work to be done, for fuck's sake.
I sure hope so. This page makes me cringe.
Perhaps slashdot should institute a policy of delete-moderation for QandA. I'm all for whatever nonsense in news posts, but this is like inviting a guest into your house and then using them for midget bowling. It's abusive.
Is this the OS that won't run 32-bit apps?
Without that, how can you charge back your time to the customers, versus your employer? How much is coming out of the internal training budget? Pull numbers out randomly, how does your company function?
As I said; measurable deliverables. And those would be deliverables that are averaged over a reasonable period of time. Completion of project milestones or at least good write-ups and reasons as to why those milestones and deadlines are missed. That's less draconian than what you're proposing and works great in companies I've worked at.
If it's any consolation, some of my peers subscribe to the model you proposed... some of their people have ended up on my team and I get consistently good reviews from my people about my management style.
I tend to subscribe to the idea that I should trust my people to do their jobs to the best of their abilities, and to assist them if it's beyond their capabilities or they have good reasons not to perform. I do have the luxury of working for a large enough company with deep enough coffers that we can usually carry some dead wood in a team for quite a while with minimal hit to the bottom line... and that gives me the latitude to average their performance over a 6 month or even a 1 year period. But privately I know pretty quickly who the dead wood is and I will usually find ways to move them off to someone else's team or get them out before they affect the rest of the team.
Being a manager is hard work. What you proposed is management by Excel sheet which is unfortunately rife in Corporate America thanks to MBA's who think they know what'll make a business better but actually only have an academic understanding of what works, not real-world experience.
And I'll admit; when I was a new and freshly minted manager I did do that because that's what my manager taught me. I learned pretty quickly how badly that works from experience and tried a different tack. Yes, my management often asks me for details like you're asking for... but by knowing my team, my deliverables and my targets allows me to speak intelligently to what my team is doing and justify the salaries of everyone on that team consistently year over year. I also have the highest average tenure of a team member of any of my peers.
To me it sounds like there wasn't enough spare space, and you ended up with an incomplete install that was unbootable *because* it was incomplete.
"I never let my schooling get in the way of my education." -- Mark Twain