Posts by UK DM 40 publicly visible posts • joined 11 Dec 2020
Yes this is true.
But the 'look up' comment was a reference to, why are you still buying his products. When you can have trained your AI to do everything his business did and outcompete him.
Anyway the shareholders will fire him, maybe you assume he is the shareholder in your comment, I did not, as the labels themselves clearly indicate the different roles and different powers of influence the role has.
Yes it turns out that while AI is 60% as good as a junior developer.
If you feed the AI all the information that a/all-the C-level execs have to hand it turns out it makes better decisions. So it could be said to be 125% as good. There is no bias, politics, corruption, willful ignorance and a multitude of human negative moral capital in the process. Won't shareholder's be pleased to hear this?
Stop worrying about it taking over the lower level jobs when now a low level thinking person can make entrepreneur level billion dollar decisions, look up the glass ceiling has gone.
I think you missed my point.
I really did mean sit in a room.
The assumption is everything (all content) has already been ingested by the AI. So what is important to ingest today is what was created yesterday because that is the only new content that exists.
I addressed the how do you do that, by putting humans in a room to create content and that is the only content you let the AI ingest today.
Hey the sit in a room (and dance for the AI) helps pay their salaries. That should be seen as a benefit, how many aspiring musicians would love the idea to make hourly money doing something they like in an adhoc way, to get paid to practice.
The idea you need official stamp of authenticity is an idea but really the outcome is an AI that works instead of an AI that socks. So the project will pay for itself in higher qualities.
I mean look at hugging face verses the various AI big players.
Radiation particles in steel are difficult to remove. Due to our lack of mastery of the universe at the quantum level.
Content created by humans is easy to keep creating, just ask humans to sit in a room and create content.
Yes I understand the concern with AI feedback loop of ingesting their own output with hallucinations included. But getting a source of data entirely created by humans remains as easy as it always was. Taping the creativity of humans is as easy as it always was.
Even if some creative industries loose their financial value because AI can generate a good enough equivalent at a cheaper cost (or rather this is really a collaboration of a human and machine to produce something at lower cost).
Even if those creative industries are worth less humans will still be involved in those creative arts. For example the number of guitar players that exist in the world few of them are famous names, many can enjoy their creative spirit around a camp fire just like humans always have for 1000s of years.
What they mean is wayland is better for modern hardware for the use cases the maintainers dream about achieving.
These use cases appear niche gimmicks in retrospect and not what users actually want. So maybe the entire vision needs to be reworked into a new project that allows for a separate focus across new boundaries 2D and 3D API.
Don't they call it premature optimization to design the architecture so limiting for the needs of today; because in their dreams they want their creation to be capable of a future feature that does not exist or is not used enough for users to care to want it in the future (because you didn't resolve the basic requirements)
AI has yet to prove it can generate secure source code of a large project from text generation. The CEO is in charge of a company that needs to do security better.
The proverbial monkey typing on a keyboard can bang out Mosart does seem to have been accomplished. Only because it ingested Mosart and then imitated what it saw. Music has many pleasant to the human ear possible outcomes that can all be considered an acceptable solution.
Doing specific and hard engineering (such as security) has not been proven to be possible yet, where many answers can appear correct but few of them are correct. Where many facets of requirements exist, many unspoken and need to be solved simultaneously.
As others in the field of computer programming have pointed out by now, code generation was not the problem that needed solving. Banging out more lines of code faster is not a solution to any destination but having more legacy code that no one understands and no one needs to be resposible for (someone is of course, better make damn sure it is the CEO)
Vibe coding means you are never allowed to look at the code, the AI must always been in charge of modifying it in the future and you are never allowed to look at the code, otherwise it is not vibe coding. This future does not exist yet but where is the oversight of the AI injecting a foreign actor's interests into the changes it makes.
AI is the new way for the CEO to make short term gains, take their bonus today, pat themselves on the back and leave the business encumbered into the future.
Given the Copyright law and provenance of works has been broken the world needs to call for all IP law to be aboloshed. No longer does the state protect any claim to any idea or works. The only people these laws now protect are the interests of large corps such at the one this CEO belongs to. Why can't I just 3D print a device, make PCB and copy firmware? One that has been security audit by my people. They don't have monopoly anymore.
At some point they will be left holding expensive IP on chips that the world no longer needs. As a suitably useful alternative will be available that doesn't require royalty payments to be made and the complete supply chain will exist entirely outside of American control.
Circle (the minor dialect change to the C++ LLVM compiler) by Sean Baxter's excellent research claims to have proven all the best bits of the borrow checker are possible with C++ today without substantially changing the whole language.
You can mark the security critical parts to under go this stricter checking with keyword that are in effect the opposite of rusts unsafe. This would obviously allow a suitably configured compiler to only allow 'safer' code or otherwise error.
So a path does exist that seems compatible with what exists, but the existing standards committee structure and viewpoint still has their head buried in the same. So maybe this recent announcement by Bjorne is to force them to chose a credible path. At the moment is is unclear what real world problem the profiles idea will be capable of solving or if it is enough to cover everyones needs. Is it still a talking point vapourware or is it implemented in at least one of the bug three compiler suites. So we can take it for a spin?
Its just a shame that Sean's work appears to be been self funded and has failed to see a commercial backer or see it being made available as open source available to all under existing compiler licenses.
Agree just use IntelliJ. I had have been a 10y suffering Eclipse user before making the switch. Suffers from too much design by committe whereas IntelliJ feels is designed by developers for developers.
As a polygot software developer, I consider myself as primary a Java developer, despite spending over 300% more time developing in Scala and Kotlin over the past few years, all non-Andriod projects.
I don't really see much of a distinction or separation now as it all just works and interops.
Maybe they should add the Kotlin into the Java ranking numbers instead of treating it as a separately ranked figure, those headline number don't tell the whole truth.
I welcome the diversity of effort from the VS camp to improve the experience there.
So far this ISA extension hell you speak of has not occured. This situation already exists for intel/AMD over MMX though to AVX. Solutions exist in that space already to manage software. I see little difference in the specification and bickering with choping and changing of available extensions between intel/AMD than the vision of a potential future with RiscV you are so fearful of.
I can see RiscV on budget phones working well, budget because they under perform on user expectations, not budget just because they are RiscV.
What really is not clear, is if RiscV as an architecture can get anywhere near ARM on the metrics what count in ARMs highend market segment. It may become apparent within the next 5 years that the high performance general computing goal isn't within architectural reach.
However it looks to eat the ARM low end segment for breakfast as things stand, so maybe performance mobile computing and performance cloud computing is relatively safe. Maybe ARM in cloud will face competition, since while ARM provides cost efficiency, there is plenty of cloud use that could use even better cost efficiency RiscV can bring (again this comes from other factors like performance per watt, datacentre density, not because it is free and open-source).
Being open-source will only financialy benefit those that can extract value from that aspect of the proposition and is not expected to significantly improve end user cost of ownership. The first movers are pocketing the gains from their risk. (Yes that was a pun)
"Chinese kitchenware is OK though. As long as it doesn't have WiFi."
Agreed, but I thought Huawei's main product line was 5G antennas, which is about as close to kitchenware you can get in the 5G stack.
Yet we must rip out and replace this metal.
At least some other counties policies have been worded with more understanding of the risks, such antenna end ok , core/backhaul/backend/control not ok.
Maybe you can confirm if microwave ovens would be allowed (in the kitchen) ? Asking for a friend
But thats is also what humans do. Someone shows them a thought in person or written. They then plagiarise that thought later in another context.
Just because the AI is more able to point towards its sources people want to label it as something bad. Humans are all standing on the shoulders of giants and clinging onto the notation they are something special.
AI is proof of the proverbial monkey banging on a typewriter able to churn out Shakespeare.
Maybe the real issue with outsourcing now is the UK customer knows AI exists and doesn't want their trade secrets in the form of instructions and prompts leaked into an AI model for rest of the world to know.
As all the outsourcers are surely trying and doing this even if you write a contract prohibitting it. There comes a point where, you've got to consider, but I can just do that myself.
Leaking referrs to routes being advertised when they should not be. This is a bad thing. This causes traffic to take an incorrect route towards destination (shifting congestion points) or a wrong turn and never getting to its destination (blackhole).
In this case that statement is confirming that the cause of the disruption is not bad BGP configuration that has been accepted by BGP neighbours concerning leaked prefixes.
Which leaves the cause squarely with the management team at VM for simply dropping off the Internet via such things as equipment failure or incorrect configuration of BGP with regards to their own AS and network.
Huh, crosswind and shear are easy things to manage. Detectable, weather quality is reported, corrected with flight control input reactions, all things already going on with such a system.
The article talks of the system managing the flight controls so the pilot can manage the critical incident with the aircraft.
But surely that incident has a high chance of causing the system to disconnect itself because it is not operating within the allowed envelope that has a better than 1 in 1 billion flying hours of failure, which I understand to be a benchmark all good aircraft manufacturers demonstrate they achieve for certification.
Onshoring the means of production is better protected.
While the fab itself might not be able to get anywhere close to the volume for all USA internal demand, non-industrial, non-millitary etc.
The talk of it being for higher end, maybe or maybe not bleeding edge time will tell makes
There needs to be 2 sides in court to challenge rightful ownership, if one side does not show up or is not known, and the side that does show up (the prosecution) they provide enough evidence (that is true or fictional including witnesses confirming timelines) who is going to win legal custody?
Your strawman Moby Dock claim would be a scenario where prior art exists in a timeline that predates copyright law in that jurisdiction or exceeds copyright law protection timelines which maybe kind of based around author life expectancy.
Who looking to defend a works that may over 25y, 40y, 50y or 70y old (different jurisdictions have different timelines) ? When the value of the works today is probably not worth the legal costs to defend it ? Let alone the extraction of montery value once possession is taken.
What is more important is works have value to someone today and protection from plagiarism from claimants with spurious and fictional claims against a new publication that appeared in recent times but have no authorship name attached but are marked as public domain.
Just like this comment, don't believe everything you read on the Internet (get qualified legal council), just because a works is marked public domain does not mean it is.
A chain of custody with legal accountability is still needed in most jurisdictions, to grant a works public domain status.
Public domain can not really exist, except in terms of facts about the universe.
Some entity created the works, some entity has to claim the legal title for the works, to then from that legal position of authority, grant other entities rights to the works, that might be described as putting the works in the public domain. This process is due to the automatic legal status of copyright.
Otherwise another entity could claim they are the legal copyright holder of the works and they did not consent to another party publishing th works.
You mean, I own this *instance* of an NFT.
This NFT may or may not have a registry able to verify an attachment to your claim, that can prove provenance (of the NFT itself) which might be conceptually rated to an artifact that has more value (like digital artwork)
So geez, you have the paid significant money for the provenance of being the first sucker^Wbuyer to pay for the thing.
The rest of us can just right-click Save As ... your copy.
Enjoy :)
The future if NVMe to GPU data transfer looks rosy, be it for computation or gamers.
https://devblogs.microsoft.com/directx/directstorage-api-available-on-pc/
This promises to gamers a future for loading textures and other graphics artifacts from NVMe to GPU with lower CPU overhead.
But still needs GPU vendors to provide support at their end both in revised hardware and drivers to achieve the objectives.
So this ElReg article appears to mimic for Linux what Microsoft have had in beta over past 18 months.
Here is to an open source and spec future. Hip hip..
While I support the principal of data ownership the GDPR is trying to legally address.
In this case it is not clear how standard use of Google analytics puts the defendant in obvious breach of GDPR.
This is because technically the EU data subject (the website user) provided all the personally identifiable data directly to Google themselves. In effect Google is the data controller for the data the data subject handed over to Google.
This is because although the defendant's website may have acted at an 'introducer' it did not act as an obvious middleman relaying the data. The website user interacted directly with a Google website when it was 'invited' to do so by the website configuration. I say invited because there is numerous methods the website user could have taken to have prevent that from happening if they were so concerned.
The data subject is the party that gave Google the data.
Understanding the actual method and act of conveyance of the data is an important part of how the GDPR law is constructed.
I am not trying to troll with this comment, but it is important the legal system understands and clarifies such important technical facts when it is being attacked by the GDPR law.
Now if the defending company had given Google a copy of the personally identifying data themselves, that would in clear breech territory.
This is equivalent to have the 3 parties in the same meeting room, having the customer recite their personally identifying data out loud. Then claiming there is a GDPR breech by the defending company, because Google wrote the information down.
Bring on the turnover tax.
This is basically a tax 'where the customers are', but unlike VAT would not be passed on and is paid by every middleman. Now providing you keep it simple (few excemptions), it becomes easy to collect and easy to audit. Middlemen are fewer in the globalised world and the rate expressed as a percentage would be under 1% maybe under 0.001% during initial implementation.
You can make it more attractive for companies to pay corporation tax instead of turnover tax by offsetting the these taxes against each other such that it is not at 1:1 equivalence. Because as things stand (current corp tax law) forcing a company to pay a useful amount of corporation tax has other significant gains to the way their money is structured. This also means to some degree companies already paying corp tax may find they have no turnover tax to pay at the end of the year.
I think we can all agree, to the OPs rhetorical point, continually offshoring profits away from the customers nation state environment. An environment that facilities the customer and supplier to transact in relative peace and security has a cost to maintain.
Maybe Intel will be along demonstrating there is no supply problem.
When they flood the market with low and mid range cards before the end of Q2, as they kick off their multi year GPU product line up refresh.
Then attempt to push on the door of team red and team green, by eating into the bottom end of their high spec gaming lineup they launched within the last 18months.
8 cents?
Great, I wish to pay for 250 licenses please.
Gimme a selection of all CPUs you have please. Say 4 of each type.
Where do I pay ?
Where do I download all the IP in verilog/RTL format so I can work on my designs and organise a fab. Yes that is possible :)
I already have access to your ARM DesignStart(tm) but you only provide 2 very low end processors, with problemaic limitations. I assume because I paid I don't have these limitations now.
Also the recommended Kali ARM compiler is around 3k GBP just for toolchain. GCC and LLVM seems better. The website looks like it is from the 90s.
Thanks very much.
Where do I sent that 20 quid now?
You post demonstrates a significant level of ignorance about the state of previous ISAs and about how and why RISC-V architecture is looking towards todays requirements and those anticipated in the future.
Your comment on tool chains is nonsense, the same tool chains and the their optimizations are already being utilised by the new ISA. However there will still be variations, like with Intel/AMD have concerning ahead of time optimizations like instruction scheduling because Intel Netbust works differently to Skylake. This area of optimization is set to explode because there will be a zillion different implementations in silicon of RISC-V.
Luckily they will all still run the same code ok just some specific silicon may do something in one clock cycle and another take three.
Of course everything has a coprocessor in the era of SoC.
The demarcation between what is the CPU, what is hardware is getting ever more vague. Take a look at modern RISC multiply/divide methods. The custom silicon that is considered CPU will be more like SIMD for things like neural net maths.
You keep your POWER and MIPS ISAs they don't scale so well from the 512bytes ram microcontroller all the way upto 512exabyes ram datacenter storage array. They also bring with them their own unique baggage that the world does not need today. They were great in their day, but if you could start over would you simply remake them again? Or take their best ideas and make something better?
I agree with IP concerning RISC-V transistor layout has a way to go to be upto ARM, but...
I don't agree tool chains are an issue, gcc/LLVM improvements of the past 30 years can all be leveraged. The ecos systems concerning tool chains are easily ported and reused.
Indeed the same is true of the other kinds of transistor IP what exists inside a modern SoC implementation that too can all be leveraged.
Besides the real tool chain improvements for a CPU target are to leverage a particular implementation of IP see Intel/AMD and M0/coretex/arm64. See also recently intel no longer maintains their own compiler, they use LLVM now.
Many applications for RISC-V are in the area of MCUs and overlap Microchips products lines, SoCs and multifunctional controller chips. The applications need reliability over optimisation, robustness over extreme low power, low cost over performance, low density technologies that can be fabricated on older process nodes. All this stuff is largely already here.
While SiFive maybe working on a version to compete with ARM64 in the datacentre with virtualization et al. SiFives holy grail, not mine. What has taken intel/amd/arm 20 years to develop is not going to take RISC-V 20 years to catch up to where we are today.
Yes they (RISC-V/SiFive) may always seem to be behind but I don't think they need to get at the front to be wildly successful.
So what?
Now someone look up if DNSSEC is also in use?
At least they have TLS 1.3 lets take a moment to applaud that.
As bad as Virgin Media.
Hey we are in 2021 guys. All the other major UK ISPs support IPv6 now, also mobile operators too, as this is a UK centric news site I wonder if it were enabled would more than 50% of the UK traffic be IPv6?
Maybe the problem is more the other way, charges for IPv4 are a money maker, especially to organisations that have had allocations for years that were granted for free.
So adoption of IPv6 will make that money spinner come to an end sooner. All entities want to milk that one for as long as possible.
So as it should be.
If you are making a statement (anonymous coward) you should put your name on it, so that liabel laws can be up held in the era of digital misinformation.
If you are voting that should be you own personal and private choice, so that you are less subject to the coercive situation that exists in some countries voting systems due to political/military enforcement of the populations' way of thinking.
No but as I commented in a previous thereg story on this topic.
If it were the UK, Consumer protection laws should override certain items in their contract terms.
1) the vendor should be allowed to breakdown costs at the point of sale, this would indicate just how little the vendor receives, apples terms don't allow this.
2) the vendor should be allowed to offer the customer a choice to buy via an alternative mechanism at the point of sale
3) this means that the customer is given choices for the same product that may have different total costs to the consumer, but apple terms also ensure their platform has best pricing so you cannot charge more for delivery on an apple platform
Apple should breakdown the various costs to allow vendors to all carte the items, app store listing, app approval, app marketing, app distribution, app support, app payment.
Apps once reviewed get (counter) signed by apple, digital signature, to allow distribution and loading from any source.
Epic are big enough to be able to manage everything themselves.
If vetting apps is a significant cost then all vendors should be forced to pay for this process, per update if needed. Not have the small % of paying apps fund the rest.
It's a good job the PC turned out to be multi functional computer, I don't see why the mobile phone is something special.
Anyhow all walled gardens eventually collapse, give it another 10y.
Yes this is a major problem, that the UK payment card industry also has.
My view is that it should be required at the point of sale to itemize transaction costs. So it is clear to the consumer how much the vendor is actually getting, how much is being lost to a transaction platform.
This then leads into a vendor providing the consumer with options on how to reduce overall costs to the consumer, by using a different transaction platform. It is also my view that this should legally be allowed (via consumer laws) regardless of apple/google/visa/Mastercard contract that attempt to inhibitor this information and consumer choice.
Obviously for Epic this means allowing to option to buy from Epic store.
Where this ethos affects UK/EU payment cards is that recently (jan 2018) it became illegal to charge less to the consumer based on the different rates of credit/debit/bank transfer charges and breakdown those additional costs to consumer by way of surcharges.
1) all such charges should be made transparent to the consumer, so itemization at point of sale should be allowed and in some situations made compulsory.
2) vendors should be allowed to legally offer alternative options at point of sale to reduce transaction charges.
Lets not forget Microsoft tried for this with the Windows 8 era
Now we are left with an office product that is extremely difficult to buy without their online services 365 attached to it.
A product that once you have a copy compatible with Windows 10 (that looks likely to not be made obsolete anytime soon) doesn't have any significant new features added to office over any 6y view.
I'm all for spending 150 GBP for a copy and after 3y since activation 5 GBP per year for security updates. Plenty of profit at those price points, but alas not a product that exists (at those price points anyway).
Of course streams has versions. 'rpm -qa' will work like before, simply save it to a file.
It's just that there maybe 3 versions of a patch as they tinker with the final version in CentOS Streams distro and the RHEL only gets the 1 version, the last one.
The question is, will the version numbers used be exactly in sync?
This begs the question what if I only install a CentOS Steams update when the RHEL version number for the same package exists, do I now get back to CentOS like before and effectively skip the 2 tinkering with package updates.
I'm going to presume it would have passed your cursory defender scan test. Hey it sounds like a sofisticated gang with access to modify the original O/S loading images. Hmm I'd bet they know someone might inspect the system for such concerns.
So as a countermeasure I'd not have the bad actor software reveal itself within the first 100 hours of powered up operation. Even when active stay somewhat dormant as the goal would be to get as many into victims hands before the supply chain is rumbled.
No what they needed to do was build their own image and blap whatever was supplied. This would be after an inspection of BIOS or other system elements that may present similar security concerns.
Does the proposed CentOS Stream package version number mirror the RHEL package version numbers?
So if you don't want to be patch beta tester (re the upstream versus downstream of RHEL aspect), what happens if you always delay your CentOS Stream updates until they match the RHEL update?
If version numbers never match someone can build index to compare source code inputs to translate.
Will be interesting if Rocky is operational by Q3 2021 existing CentOS 8 installations can just run a script to reconfigure yum.conf to point all package locations at Rocky distro servers without needing to reinstall. Since all ABIs, package names and versions should be compatible.
To me Fedora is meant to be in the Red Hat upstream distro space. Why don't they have a Fedora respin that mimics exactly the technical goals they are looking to achieve with CentOS Stream. In Fedora there is reign to add/remove entire suites, methods, subsystems. It is not clear at this point if that is a CentOS stream objective, I find that unlikely.
I can only guess this is a short sighted commercial goal.
The issue is that I can see Rocky attracting enough sponsorship to cover respin, build and distribution costs, all because Red Hat are unwilling to offer a low enough subscription model price to cover what CentOS is and what that market segment wants.
It is a FREE product he ordered (firebase) it got upgraded without his explicit consent. It has nothing to do with limits, the quantity of resources Google gives away in a free product is upto them.
But the obvious contract he agreed to and his purchase order was for the FREE product.
I assume the real issue is the small print of the contract, what would allow Google to getaway with billing like that.
My recent experiences with Azure and Oracle Cloud Infrastructure is that the services are added to a free tier plan and upgrade is explicit, maybe to prevent adding a non free service accidently, you have to remake the profile on top of a normal profile that is chargeable.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
