Posts by abend0c4 887 publicly visible posts • joined 10 Mar 2023
Ci/CD is hard enough to do when you have a limited number of configurations which you can test comprehensively in a similar timescale to your build process.
It's pretty much impossible when there's an almost infinite number of possible configurations, many of which cannot easily be replicated automatically.
The irony is that few people actually seem to want the new features that this accelerated deployment process is intended to foist upon them.
The thing is, corporations have found increasingly creative ways to hire and fire their meatsacks on demand.
I rather suspect they're going to find it a great deal harder to get out of their contracts with the AI megaliths - or indeed retrieve their data - with similar flexibility.
Still, I look forward to seeing the enthusiastic early adopters suddenly finding themselves to be the party with the least expensive lawyers for a change.
A lot of advertising isn't really aimed at getting you to change the categories of stuff you purchase, it's aimed at getting you to purchase different brands. The international soap powder, food and household conglomerates simply wouldn't exist without advertising - and their increasing globalisation is as much driven by greater marketing efficiency as it is by production efficiency.
But most of what they're pushing is stuff that everyone buys - soap powder, floor cleaner, biscuits, breakfast cereal - so it's not clear that targeting (except by gender, since women typically make purchasing decisions) is terribly useful. However, ever since Lord Leverhulme famously complained "Half the money I spend on advertising is wasted, and the trouble is I don't know which half" the pressure has been on to find the answer. But, from a consumer point of view, it's the wrong question. Without the cost of sales associated with marketing, these products would be significantly cheaper - compare the cost of your supermarket's own brand with the Reckitt-Benckiser-Kraft-Mondelez-Unilver-Procter & Gamble alternative. But that also proves that advertising in some form *does* work - because why else would so many people buy the overpriced alternative? And as long as it works, there'll always be someone claiming to know how to make it more efficient and someone else willing to believe them,
Of course you can - and you'd almost certainly want to in order to distinguish your offering from the alternative.
However, we are potentially faced with the situation in which you're not allowed to make any allusion to the contents of the brackets - in effect not be permitted to mention the project from which yours derives. Though there may be specific cases in which it's obvious, in general that would be a barrier to alternative development paths. Or, indeed, hosting providers if there are, as claimed, outstanding trademark applications for "Hosted WordPress" and "Managed WordPress".
It's entirely possible this is not the intent of the specific legal processes now in train, but it does seem to raise that future possibility.
I'm not a lawyer and I'm sure I don't understand the merits of the respective legal arguments which will ultimately be decided in court.
However, if it were the case that one could (a) trademark the identity of an open source project in such a way that commercial organisations offering hosting solutions would need a licence even to mention the name of the project they hosted and/or (b) there were no requirement to issue such licences on a FRAND basis then I think it's reasonable to suggest we would have to reconsider our notions of "open source".
... that most Netfilx costs arose at the edge - pushing large video files from CDNs to telly boxes - and the bit in the middle processing payments and making it impossible to find anything worth watching would be comparatively much less costly.
But, I suppose, a lot could be analytics. Everyone is so convinced that collecting ever-more data about your customers is the key to economic success that no-one is ever going to question the cost of collecting and processing it, so maybe it's better simply not to know.
power is now the limiting factor
It would be the limiting factor if the predicted demand actually exists and the "productivity gains" justify the real world operating costs.
I still think reality could turn out to be the limiting factor because even in the unlikely event this stuff turns out to be useful to people, no-one seems really to have thought about the prices that will be needed to ensure a return on the almost unbounded investment.
I think that's too narrow a lens.
The problem is not so much that humans also make driving mistakes, but, arguably, that we excuse those mistakes more readily for motoring than for other potential causes of injury or death. It took best part of a century to have reasonable safety standards for vehicles and in many places you're still held to a lesser standard of responsibility if you kill someone with a vehicle through your avoidable conduct than by other means. That historical laxity doesn't mix well with transferring whatever responsibility might remain onto the shoulders of software. You have to look at the social context in which the technology operates.
And there's a similar problem with AI. It's of course true that people have double standards and prejudices and make mistakes. However, ultimately, they can be called upon to account for their actions and explain them. AI may make fewer mistakes and have fewer prejudices (though the evidence for that, so far, is not promising) but it can't display its working. That would be fine if there was an implicit assumption that you would need robust review procedures by qualified people - but if you need those people, the economic case for the technology is suddenly rather weaker so there's an incentive to "believe" the technology is infallible.
I think it's wrong to imagine we can dispassionately reduce this to Benthamite principles: for better or worse we are not utilitarian creatures and we will only willingly adopt technology that can accommodate our natural contradictions. It's not a coincidence that some of the most enthusiastic proponents of both self-driving cars and AI have been accused of having sociopathic tendencies. The only way you can remove the human element from technology is to remove the humans.
... but not quickly enough, it seems.
I'm not sure how technology they can only trust to operate in a "hands-off, eyes-on" environment is actually going to help in "delivering the best driving experiences". It seems to be the worst of both worlds - you need to remain permanently alert as a driver without any of the interactions that would help maintain your alertness.
The original idea was that self-driving cars would reduce accidents and help manage congestion and the technology components were all about ready, but just needed integrating. Then it became obvious that the problem was harder than it looked, so maybe they could concentrate on urban transport in carefully-defined areas: like buses, only occupying vastly more road space. Now it seems like it's never going to work, but they've lost all this money so consumers are going to have to pay for it anyway - and be entirely responsible for its manifold failures until the carmakers have recovered their "investment".
The technology that was promised doesn't work. The technology on offer seems to have very little practical use. Meanwhile, the practical alternatives still seem to be based largely on Victorian engineering. Maybe we can just join up all the individual cars and pull them with an engine.
Given that Microsoft is phasing out third party drivers and that vendors will have to be Mopria-certified to ensure future Windows support, there probably couldn't be a more appropriate moment to reflect on the end-user benefits of swapping a myriad of proprietary drivers maintained by device manufacturers for a single point of failure managed by Microsoft.
I think my favourite is "Maigret" - it perfectly encapsulated the style of the production. And while "Tales of the Unexpected" was rather insistent, it was unmistakable from the first note.
It's interesting how tastes (and technology) have changed. The early versions of the Doctor Who theme are quite spare. Since the reboot, the sound density (and volume) seems to have risen to sense-bludgeoning levels. I blame the youngsters and their popular beat combos.
I find it difficult to keep up these days, but free trade is apparently so last century. However, it's not as if the precedent doesn't already exist in China - where it's quite normal for overseas investors to be required to form joint ventures with local partners: Tesla is one of the few examples that managed to dodge that bullet. And it's a similar picture in many developing countries that want some of that "investment" to stick and not merely move on to the next source of cheap labour when local wages start to rise.
I think there's an argument that the Anglosphere has been rather too lax in its pursuit of foreign investment at any cost with consequences such as Thames Water and the dispiriting sight of Keir Starmer touring the despots of the Middle East with his begging bowl exended.
EDIT: Sorry, Paul Crawford, posts crossed. Basically, what you said...
read the popup dialog
I've lost count of the number of conversations that have taken the form:
- There's an error message
- What does it say?
- Something about a failure
- What are the exact words on the screen?
- Oh...
There's something about computer error messages that seems to induce temporary dyslexia.
Until recently, I'd used Evolution successfully for years on Mint and found it more usable than the alternatives. However, a few months back, most of my mail disappeared - not physically as it's all still on the IMAP server - but nearly all messages and folders disappeared from the UI. If I deleted all the local mail directories and settings and reinstalled, the mail would reappear - for all of half an hour before disappearing again. Couldn't get to the bottom of it. A number of posts online from people reporting similar experiences.
So I'm on Thunderbird now and have configured it more or less as I want it and my experience so far is it seems to be more robust: Evolution would also periodically fail to authenticate and require a background process to be killed to recover. Still haven't found a Linux (or Android) mail client I actually like, though!
The original intention was for the Netscape scripting language to be Scheme (or a derivative). It was Netscape's alliance with Sun, in the face of Microsoft's belated conversion to the Internet, that led to the two collaboratively to attempt to create a competing, unified platform - but the short market window (for Netscape which was approaching its IPO) and the complexities of Java meant that was more of a marketing fiction than a technical reality. There's a quote from Brendan Eich saying I was under marketing orders to make it look like Java... and a later one in which he describes the JavaScript name as a big fat marketing scam. It wasn't simply a retrospective request to use the name: the language syntax was specifically created to support the relationship with Sun.
JavaScript is the result of a collaboration between Sun and Netscape around 30 years ago. I think the original idea was to have Java run in the browser, but various constraints meant they had to settle in the end for something whose syntax looked a bit like Java and which could be passed off as related for marketing purposes. During its short development period it was known as Mocha and then LiveScript before the name JavaScript was chosen. But it wasn't Netscape seeking permission - they were actively supported by Sun at the time.
Clearview doesn't have customers in the EU: and now can't. Bluesky's EU users may not be paying for the service, but there's not much point offering them the service unless someone does. And in the absence of someone outside the EU with deep pockets funding it to make a gesture of political defiance that's most likely to be advertisers within the EU's jurisdiction.
The thing is, it's not "independent" if it's, well, dependent on the cash from a supposed rival. And competition-as-a-service is not competition.
If you have entire industry sectors that depend mostly on stuff that has significant development costs being (apparently) given away, it is inevitably going to lead to the dominance of megacorporations and billionaires - and a complete lack of genuine privacy for the end users whose data is the only currency that changes hands.
The only effective solutions would ultimately lead to Google et al having to charge for their services in cash and I suspect that the general public would, unfortunately, prefer the status quo - and the regulators know it.
Somewhat ironic this article is immediately adjacent to Google blocked 1,000-plus pro-China fake news websites.
Would we require the technological equivalent of a wooden stake and interment at crossroads
Given Google's involvement, I think you can rely on their return to the cemetery in fairly short order.
I don't think there's much harm in someone communicating with facsimiles of the dead if it comforts them, just as long as they accept it's only a facsimile. And, in that regard, I suspect the uncharacteristic tendency of their "loved one" to remind them of the bargains on offer today at various sponsoring merchants will keep them on the right side of reality.
The only time I've seen pictures of him is on a TV report where he was draped in a loose robe, exposing a muscular torso and sucking on a rather phallic cigar. Surely, few young men would begrudge $50 a month for such an unequivocal confirmation of their heterosexuality?
But it must be quite a challenge to spend that amount of money in Romania. Especially if you don't get out much.
Personally, I could probably get away with one of these (if it were available). And if I needed regularly to travel longer distances, it would still probably be cheaper to have one for local journeys and a conventional ICE vehicle too than an electric vehicle that tries to do everything. Of course that won't be universally true. However, we probably need (needed) to think of lower carbon transport options beyond simply swapping out the traction technology of existing vehicle types.
Transcribing Meetings? Who on Earth does this?
People who need a record of what was said and by whom for the avoidance of subsequent doubt or denial. Though in those cases you'd probably need each of the participants to review the transcript and accept its accuracy, so it's not clear AI would save much time in those circumstances.
You will continue to have the choice to use Google Chrome in the same way that other people today have the choice not to: theoretically.
I don't think anyone is suggesting this is an ideal solution or that it will lead to a new dawn of browser development. It's just that your satisfaction with the limited choice available to you is outweighed by the market dominance that accompanies it. Your choice presupposes a cost to other people that they have no current choice but to accept.
Tricky business, free will (and the illusion thereof).
It could utlimately, as you suggest, simply come down to a matter of numbers. If the aggregate data indicates that globally more Jitendras have been hired than Jims then maybe it will self-reinforce.
The intention is always that the hiring process be biased in favour of the putative ideal candidate. The trouble with AI is that all you can really say is that it's biased in favour of someone, but you can't work out why. And no employer is going to conduct the ultimate trial and simply hire a bunch of people at random, so we'll never know whether all this effort really makes much difference.
Considering how much money we now spend in those subscriptions, I hope that will change...
Perhaps the most effective way to change it would not be to hand over significant sums of money in exchange for an ephemeral asset.
However, in reality, most of our purchases of physical media have always in practice been ephemeral too. Most households no longer have the means to play Edison cylinders, 9.5mm film, 78rpm records, laser discs, video cassettes or minidiscs. Even my old paperback books are starting to deteriorate. Even charity shops will turn away a lot of physical media for which there is no significant demand.
Having been an executor on a number of occasions, as I head through old age my priority is going to be to minimise the burden for those who have to clean up after me - not having unnecessary contracts and not burdening them with artefacts they probably don't want and will struggle to dispose of.
Me, too. Slightly earlier using HP Time-Shared BASIC. It used to be possible simply to wander into the Computer Lab at Newcastle University and sit yourself down in front of a KSR33 teletype (or an ASR33 if you wanted to save your program on paper tape) and use the system without any account or authorisation. There was even a built-in language tutorial. As a schoolboy, I wouldn't otherwise even have thought of a career in IT. So I'm grateful to everyone who contributed to that possibility.
Times have changed in so many ways.
rotting the minds of ordinary people
Back in the Victorian era they discovered that increasing literacy did not necessarily make people more highbrow or worthy. It was even suggested by some that literacy had gone too far. The greater the ease of mass communication, the more it seems to amplify human nature for better or worse. The technology may have changed, but human nature hasn't.
I think it's possible that a new complier (or indeed pre-processor, since ending up with C would be a lot less trouble than having to write compiler backends for multiple platforms) could by default flag up pntentially unsafe uses and allow incremental declaration annotations and minor amends to gradually make the warnings go away. Though how far people would actually bother is a different matter.
I'm not entirely sure about a lack of ABI compatibility. I see the reason, but I'm not sure the overhead would be acceptable in kernel code, for example.
I think a good test would be whether you could ultimately write an arbitrary part of the Linux networking stack in the revised language and have it interoperate with the other components. There's a lot going on there in terms of byte-alignment, endian-ness, objects of different lifetimes (eg: sk_buffs may have a different lifetime to the data they point to) and changing ownership (as they move from queue to queue) and if you can crack that, you may well be on to something.
From the perspective of a (mostly) software boy, trying to make the hardware faster might just have been part of the problem.
Unfortunately, most of the complexity is in analysing the data, determining who should have access to what parts of it and then subsequently testing to see the access controls are correct.
Regardless of how the controls are implemented, it seems in these cases that none of the above was actually done. And that, ultimately, is the problem - giving the ability to publish potentially private date to people who haven't got the slightest clue about the possible consequences. If the default configuration was for all the data to be inaccessible to everyone, it wouldn't help - it would rapidly be altered to the simplest setting that appeared to "work", probably giving access to everything to everyone.
In other words, they don't have people on the ground who could collect this information, or the mechanisms to take account of local circumstances, and don't want the expense associated with them.
It's probably as neat an illustration as any that accuracy, transparency or accountability aren't financially compatible with a model based on the algorithmic identification of aggregate characteristics of the source and the recipient.
Google seems only allows you to alter your country of account registration once per year, but you can have multiple accounts in different countries on one phone, which rapidly gets rather convoluted if you move around a bit but sort of works.
Apart from the need to establish a jurisdiction for your agreement - and the convenience of filtering out of your view all of the region-specific apps you're never going to want - the main reason for geographic specificity is enforcing the boundaries of rights-licensing territories. This is something the EU keeps blowing hot and cold about - they can't seem to decide if geoblocking is anti-competitive or vital to protect minority cultures. A more consistent view about where the single market rules apply might be easier to implement and enforce.
while this might enable a new fourth operator to enter the UK market, it is not clear that any such business could compete effectively
If it's basically a market that's closed to new entrants, there's no real incentive for the incumbents to compete: the weakest won't be allowed to fail and reduce competition further and the strongest won't want to grow and risk being broken up because it's gained too much market share.
If it's not going to end up like the water industry, the network infrastructure needs to be run as a single public utility - though its provision and operation could be subcontracted.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
