Posts by rgjnk 192 publicly visible posts • joined 16 Jan 2024
Wasn't/isn't there a related problem in the US with Doordash(?), pre delivery tipping, and the effect this had on the (non)service people were getting?
All these businesses with the same core model have a similar level of ethical/quality management, attract a similar level of workforce, and all descend into a similar level of mess in the end.
So was this a proposal to turn Local/HCI into something closer to Hub or just another round of confusing partial product that was sort-of Azure but not really enough to be actually useful?
I've look repeatedly at spending on their on-prem/hosted solutions (in fact have spent) and it's a real mess, especially when you get into costs and lifespan with their ever shifting ideas of what's available & stuff that dies of neglect if not outright killed.
The only constant is their desire to drag you into the morass of Azure-proper.
I'd steer well clear if it wasn't for a couple of things that force it to be used; it's not that their stuff is technically bad, more that (like a couple of other vendors) they're deliberately hostile to the customer & mess you constantly around for their gain, instead of it being a nice simple exchange of product for cash.
There's only really one person with the specific motivation, the access and the technical ability/ineptitude to have done this simple fiddling with this half-assed outcome.
A person also well known for regularly plunging off in the deep end at 2 or 3am on weird rants & other stuff after snacking on a bowl of Special K.
I mean it could have been someone else but everything points in one direction.
So is the story at the end that their services aren't going to drive them financially, they've already milked the VCs and their plan/desperate hope is that someone will lend them money secured against a rapidly depreciating stack of GPUs?
It's a bold strategy. Let's see if it pays off for 'em.
BTW nothing wrong at all with the AMD gear, but their problem in this field as with others using GPU is that Nvidia is where the ecosystem support is. Often even if you'd love to move you just can't. The debate for buyers & investors is how long that lockin is sustained.
Let me guess, that means Musk's giant firework Starship?
SLS is a million miles from perfect but scrapping it now it's apparently ready for proper man rated missions in favour of something that definitely isn't doesn't seem that great a tradeoff.
Even worse if it subsequently means the active possibility of further moon missions gets deferred to push a Martian pipedream.
I hope their breakdown of what went where was a little bit more nuanced than just the nominal nationality of the company.
Plenty of US/European company contracts are carried out by divisions based entirely in the UK, and a lot of work by the UK based companies will have included divisions in the US or elsewhere.
The surge in 'French' contracts for example (if it's the ones I'm thinking of) is all/almost all UK based. It's basically the UK domestic industry except for the nameplate.
It's a complicated environment and the domicile of the parent company is only part of it and doesn't necessarily give a fully accurate impression of sovereignty.
I'm sure they've had a lot of 'expert' input from the tech bros about all the wonderful easy improvements that could be done to this legacy gear.
The sad reality is that ATC systems are exotic stuff on a par with avionics (if anything much worse due to uptime and other factors), and the engineering required is not trivial. And there are a lot of things you don't use and don't do because your performance and reliability requirements are on a level the average techy type probably can't even start to comprehend. Something like VOIP for example can be a non-starter.
Doubtless the plan is to try to go quick, cheap COTS on the basis of 'how hard can it be?' and it'll go badly wrong and those responsible won't even begin to understand why.
And it's not like you can afford to blow it all up a few times and claim it was all useful tests, or spend years with it maybe working under endless beta; if this stuff is in service it has to work, first time and every time.
You just know that some of the chancers/self proclaimed geniuses will have stuck their noses into this despite having zero clue and this not being a realm where their bluffs can work.
Browsers have come and gone over the years, so the value of Chrome is not the most solid thing. It has obvious value *to Google* as part of their portfolio but does it have that value by itself?
Is it the engine? 'Cos that isn't the same thing; there are *lots* of browsers based on the same core, so it's hardly a special sauce.
Or is it the popularity? 'Cos how much of that is down to the product, and how much is just down to it *currently* being the default bundled browser on so many devices? Any reason for that to persist post-sale?
Or is it just the data? What is that actually worth when detached from the rest of the Google empire? What can a third party do to monetise just that bit, vs what all the other browser vendors can do too?
It might be the big dominant beast right now but once split off it could get overtaken by something else in exactly the same way predecessors did over the years. Assuming anyone still thinks supplying a browser is a worthwhile end in itself.
So maybe $50 billion is a bit much to ask for just another free browser in a crowded market not exactly flush with massive returns.
Not wishing to be funny but some of us have dev rigs that are bigger than the whole infrastructure of 37signals, and looking after those is a part time job for one person alongside the other rigs plus other tasks.
It's not that complex or expensive a job if the kit is any good.
Funnily enough my latest personal dev rig is a new on-prem cloud because running tests in someone else's raises so many issues including controlling the ongoing costs.
Anyone with basic reading and maths skills could work out that cloud isn't exactly a cheap option, especially for anything persistent; if you need something permanently dont rent it! There are plenty of cases where cloud does makes sense but some of those also apply to just running your own on-prem cloud.
Running your own kit also usually means you know exactly what the cost will be ahead of time and can have tight control, something a cloud subscription doesn't exactly guarantee. Very easy to get burned especially if someone gets careless with a deployment.
37Signals are just lucky they had what sounds like a relatively tiny setup to move and that it was set up so they could port without reinventing anything; very easy to become tied to a specific platform if you start using all those lovely features.
Not sure the blinkered focus on 'memory safety' is worth the rest of the compromises of this current obsession with Rust.
Your inability to design, implement and test adequately to produce robust code to eliminate a subset of non-inherent issues is not a solid justification, any more than your inability to ride a bike would be reason to force training wheels on everyone.
And I especially hate it when people try to mandate their current favorite solution as the One True Answer rather than properly defining the problem and leaving it open how to reduce it.
It's not like Rust is exactly perfectly robust (or even viable) in all scenarios either.
If you dig through the kernel there's support for all sorts of weird & wonderful stuff, and while it might keep those with a tidiness fetish happy to scrap 'obsolete' stuff the benefit is normally minimal - there's rarely any ongoing maintenance effort involved on those chunks of code and as you already have *many* branches of CPU support in the kernel there usually isn't any wider simplification as you still have to cope with all the other processors & their needs & features.
Plus as has been mentioned embedded 486 isn't completely dead yet.
So is this genuinely useful or just a cleanup for the sake of it? Using lines deleted as the metric suggests the latter.
So is your solution to restrict adult access to read social media? Or to make sure only the 'right sort' of people can post or is suitably pre moderated?
Any 'solution' is likely to end up worse than whatever you think the problem is.
Also slightly ironic that this was itself posted on social media, doubtless with full expectation that any study would find all the author's posts perfectly harmless.
What isn't explicitly mentioned is that in implementing these rules to help the poor vulnerable kiddies, you just happen to require an ID check for *all* users.
Funny how often 'think of the children' turns into a general purpose way of imposing onerous registration rules on the adults. Absolutely no scope for that to be abused!
If getting that data wasn't an underlying aim the proposed rules would be set up differently.
"greater economies of scale, more efficient manufacturing, and improved quality and scale" is something you also get with all that 48V stuff that comes from telecoms and automotive already.
Going for 400V/800VDC as a next step just means you can continue to benefit from some of that scale.
Upside of going 400VDC & up is you don't need quite such ridiculously oversized conductors & connectors, though they'll still be substantial at the power levels involved & might even need plenty of cooling too. Big downside is how solidly hazardous a DC supply at that voltage is in multiple ways & will require some serious safety design & changes to working practices.
The other comical bit is going to be cooling the equipment at that power density, it's not exactly trivial. There's one or two fun examples from old compact supercomputer projects (taking an existing design and making it deskside sized for deployment into flying/floating use), and I've done testing myself with kit that needed ridiculously sized coolant feeds to extract the energy; amazingly easy to heat a full flow full bore watermain if you dump enough power into it. The cooling system will probably end up as a serious hazard in itself.
Maybe the better option is to accept density limits and not chase into the realm of silly engineering requirements and safety hazards?
Amazing how some people are so eager to have an official version of the truth and for all things contrary to that to be stamped on.
I guess they haven't learned the many lessons of the past where the official truth turned out to be somewhat less factual than the 'misinformation', or how a new management might decide the new 'truth' is something some people might find unpleasant.
The cheerleaders are also frequently oblivious to how these wonderful powers could be easily turned against them. Again missing the repeated lessons in how these things are inevitably abused.
You can always find an unpleasant event to use as an emotional justification for anything, but the resulting law is rarely good.
Start billing your customers the extra, from the sound of it *every* supplier of these products will be hit so all of you will be raising your prices together and no one specifically loses out.
And it might remind everyone who actually pays tariffs ie. it's not a way of bring new taxes in from outside (which at least one important person doesnt understand). Only thing it generates sort term is price inflation.
As a final thought these must be truly tiny businesses if they're talking about $1000 or $3000 bills. Not being harsh but that's hobby scale and while morally fighting might be a good idea in reality the lawyers are going to cost you a disproportionate amount of money you won't see back. Keep your head down, pause your operations, let the big players pay to fix it as they'll be hurting exactly the same.
Funny how this & a couple of other versions of this story are missing the critical detail of who actually developed & built this thing. It's not secret and is mentioned on the fuller stories.
The MOD might have requested it & paid for it and had one of their people driving it for this test but despite the spin MOD isn't responsible for the work that actually created it.
The people who put the hours into designing and making it are getting sold short to make MOD look more competent.
I thought the full NLRB story was the more interesting one, even if the reports studiously avoided looking too hard at the very obvious reasons His Muskness had for sneaking all that data out the back door.
The majority of these activities look less about the 'official' purpose of DOGE and more about how Musk has been let loose to serve himself.
Other places trialled this & found it increased stress levels as people felt under more pressure to achieve deadlines in the shorter week - you might be doing the same hours overall but a working day is a working day and sometimes the extra day/delay is what saves you. So it's not necessarily great, I guess people see the result they want to.
I've been doing flexible hours and a 4 & a bit day week for years; the Friday hours are only vestigial but it's amazing how often that nominal extra 'day' comes in useful. It's not like it interferes with the day off as it's only a bit of a morning.
Funny thing is that for all the benefits it was supposed to bring, I suspect we only shifted to that pattern because it matched the actual levels of Friday working especially post-lunch.
I guess a 4 day working week appeals to management when you still want people to be working rigid hours 8-6, and it seems tempting to employees for the long weekend, but I honestly prefer the flexibility to have late starts or early finishes when I want and spread the effort out a bit.
I think you've misunderstood why the massive bureaucracy with health provision as a sideline might have cocked this up - it's not down to not doing it themselves.
Plus one of the glorious failures of the 'national' health service is actually being lots of smaller local services that each do a lot of their own thing instead of one single unified organisation with a single common solution. So 'in house' would involve lots of independent mini projects instead of (hopefully!) a smaller number of suppliers who should have done the implementation more than once.
Sounds like they've done what they always do when it comes to change; just go through the motions.
Slow inept minimal implementation at massive cost - almost vindictive compliance of mostly ticking the EPR box while in reality achieving little of the possible gains and not really changing anything. You can bet any delivery metrics have been twisted like a corkscrew to be 'met' on paper while utterly missing what they were supposed to represent.
Fair bet the biggest achieved outcomes were in handing out cash to selected suppliers and employing a whole load more staff to manage the thing.
And doubtless to be followed by a demand for a lot more funding as the whole reason for any failure will be pushed as being a result of inadequate budgets...
Americans built a system of checks & balances that turned out to be easily subverted, and mostly relied on trust that a head of state with a lot of the powers of an old style absolute monarch wouldn't do bad things.
Turns out if you build a naive system with insufficient controls that sooner or later people will take advantage of it.
And given what we've seen as themes in American literature & cinema over many many years not only did they already understand exactly what could happen to their Government, but they also understand it has scope to get much much worse.
Let's not pretend that a big chunk of Ukraine's 'technology industry' isn't still of the sort that you want to steer well clear of. The minor inconvenience of a war inside the border hasn't done much to inhibit that side of the industry any more than it has their equivalents a bit further east.
Also be interesting to hear how and where the legal tech businesses are hosting their operations these days given how easy it is to relocate these sorts of businesses. Guess some just moved further to the west away from most of the action, but you'd imagine the ones wanting to avoid staff getting abruptly conscripted will have workers relocated a bit further away. If it's all remote outsourced services they're selling it's an easy thing to sort.
They fuck over everyone.
Especially anyone who needs to buy their exorbitantly priced top end hardware and then gets another dry poke when they have to add the software licenses to actually use it.
And it's not like the consumer end is getting cheaper or more flexible with time...
Feels like it's a bit of a magic money tree for someone, a bit like how they already managed to spend a major 9 figure sum on other projects that didnt get delivered.
The money seems disproportionate to the outcome even given the sort of specialist facility we're talking about.
Then again certain nuclear regulators in the UK do have a habit of taking what is considered current best practice, slapping a hefty coating of gold plate on top, then adding platinum over that then studding it with diamonds. Then people wonder why the resulting projects end up massively late & overbudget compared to other people doing the same work elsewhere.
A hole that you can trigger by doing nothing obviously bad, and that punches straight out of the sandbox sounds less like a bug and more like a carefully designed 'feature' quietly inserted by a TLA. The combination of an innocuous trigger & a significant effect are pretty classic signs.
Even if the company and it's freaky founders/management didn't scare you off, why would you pay any attention to them about anything?
They're still a fairly small business with middling products and the contracts they won are relatively tiny stuff. The only truly big thing about them is their notoriety and their massively(!!) overinflated stock price.
If any of these wonderful new ideas really did take off then more likely than not it'd be a much bigger & more ruthless fish that'll actually take them forward.
Considering there were >250K reservations (supposedly) and plenty of unsold stock those numbers aren't too impressive. Though does again prove that the supply of tasteless idiots with too much money isn't running short.
Maybe it sold better than the electric F150 or the Hummer EV but in the grand scheme of US truck sales all the electric ones are totally insignificant in a market that numbers in millions of units per year.
Obviously this is the current hot thing, but Java is a great example of the costs and consequences.
Anyone who has had the fun of playing with the JVM settings, working with the garbage collection, the sheer memory footprint of an application and the fun & games that can be involved with 'temporary' memory consumption during stuff like copying will appreciate that the 'safety' comes with a big set of consequences.
C/C++ might let you get yourself into a mess if you're careless but at least being closer to the metal you have full visibility of what's being done and have some options for dealing with it if you need to.
As with all things, memory safety is useful but let's please not lose sight of what's involved in making it happen and that the results are hardly perfect, just different.
"The Congressman also cited concerns that Musk may be using government data to train his own AI model, Grok, although this seems to us speculative with no evidence to back it up."
It might seem speculative but given it's Musk, and given his long history of merrily ignoring all sorts of rules if there's a buck in it, is there any reason for you to *not* think he'd do it?
To me, I look if it's possible it could have happened, and if he'd see a benefit from doing it, and on that basis I'd say it's more likely that not that all sorts of data *has* been pushed into Grok.
Honestly - why wouldn't he do it?
And here was me thinking that a cloud platform and a container one weren't quite the same thing.
Admittedly OpenStack has tried to glue container orchestration alongside the rest of its many bits but in my head it's still an utterly different thing from K8S.
Maybe they'll become a bit closer integrated now they're under the same roof but it's never been an either/or just like it isn't on any of the hyperscalers who offer both a K8S and a VM solution.
Sooner or later everyone will look back at most of this junk and realise that the idea of 'AI supremacy' is about as useful as 'blockchain supremacy' would have been back when that was the hot thing not so long ago.
Some of this stuff is good and useful but as with the many blockchain ideas people came up with, a lot of it - even if perfectly delivered - is at best an inferior way to achieve something that is already possible by simpler means. Many of the AI proposals aren't even close to reaching that low bar.
The wild hype cycle continues but it still doesn't make the crude hype any truer.
(I may be slightly burned out from endlessly dismantling various vague AI->?->Profit! proposals I get that can be usefully & cheaply achieved with other methods that might even still involve some older simple AI instead of an LLM and/or masses of GPUs)
They seem to have landed well enough to survive so all the difficult stuff was OK, the issue seems to be more from creating a design that just isn't stable enough.
Ego can convince an engineer that their design is fine, or proximity to the boss can make them proclaim it as such, but they've done something different to usual practice that looks like it has inadequate supports and/or could topple easily and funnily enough it has now done that repeatedly.
For all the effort they maybe should have started with something slightly more conservative/known to work and only then iterated to something more innovative?
"As we have said many times before, we have never built a backdoor or master key to any of our products or services and we never will,"
Given their response was to just unlock the front door and leave it open, that's both true and yet ignores the result they created.
They could have just ignore the order and dared the government to enforce it, which would have been a much more effective block.
As opposed to the current capitulation which lets them publicly play at resisting while cheaply overdelivering the result the order ultimately wanted.
-40C isn't too bad for the low end and probably has enough tolerance to go sufficiently beyond, but +85C isn't particularly high in reality. In some parts of the world an enclosure can see that just from ambient plus a bit of solar heating.
Much better than plain commercial grade but a way to go for a lot of use potential embedded use cases.
It seems most of those chiming in from all sides think it's something simple to manage and any old rubbish will do for replacing it and the existing situation re. obsolescence and suppliers is either down to incompetence or some conspiracy.
The reality is that these systems have reliability and performance requirements similar to avionics (in many ways beyond); they must do a certain job with incredibly high assurance, determinism and continous uptime.
Engineering these things is an exotic task with only a small group of people able to do it because it's *hard*. Most people aren't even vaguely aware of the sort of stuff required let alone able to do it, a bit like most are utterly oblivious to what's involved in the compute inside avionics or other specialist fields compared to their usual toys.
Seeing people stumble in to mess with it without understanding the how and why of everything from the UI to the processors to the networks isn't a huge surprise but it's truly dangerous.
The failure of officialdom to plan for easily anticipated obsolescence and update cycles isn't a consequence of the original system design and it's definitely no reason to have 'experts' roll up, declare it easy and cock it all up especially if it's part of a grift to sell their own consumer grade crap.
The reason it caused continuity issues is doubtless the same reason people were using them at all - they were a provider of cloud services suitable for certain official levels of security and that isn't exactly a big pool of players as it's a lot of effort.
If you needed 'cloud' that couldn't go on the usual public hosts and you weren't (or couldn't) running something suitable privately then they were where you went, especially if interacting with something already on there.
It's been a while but I seem to remember that they were involved in stuff like secure Azure hosting; if you wanted Azure at a high enough level of security it was actually them running a private instance.
I used them too for a couple of things that fell between what suited the normal public clouds and what we would allow on our various private ones.
Ultimately they filled a useful niche but didn't have the scale, and even if they had managed to last longer they'd still have got squeezed out by the hyperscale players on one side and in-house stuff on the other.
When they're replaced with plastic isn't the normal solution to run the new stuff using the old as a route, either as a straight push through inside the old or using a tool to fracture the old apart & pull the new through the space?
At least that's how I've seen any replacements being done in recent years. Quite a neat process how they manage to incrementally install the new pipe & connections without having to shut it all off.
That's a lot of 'disused' infrastructure that's abandoned in place (as in the metal is still down there) yet still fully occupied by its replacement. If they blindly assume upgraded == old disused available then the estimates will be wildly out.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
