Are you sure the Agent is not called M3gan
You know, M3gan, like this M3gan:
https://en.wikipedia.org/wiki/M3GAN
111 publicly visible posts • joined 31 Jan 2024
The people developing linux kernel patches at microsoft probably use Windows only to fill the (macro ladden) expense report and time tracking excel files that are feed intro microsoft Dynamics ERP.
Otherwise, they use Linux IDEs to develop for the Linux kernel in linux, just like every other linux dev out there... maybe using WSL2, but linux, nonetheless
Microsoft has a problem, in the short term, once Win10 in out of support, they will only have to care about X86-64 8th gen or older ;-)
Redhat also has a problem now, but as soon as RHEL 9 is out of support, they only have to care about X-86-64 v3 ...
Ubuntu only cares about X-86-64 even if the debian base still handles X86-32
Something similar happoens in the ARM camp, hardly any distro of Renown runs ARM v7 anymore.
the only way for these behemoths to solve the Architecture complexity puzzle is to leave behind older versions of said architectures
I've had good experiences with DD-WRT. Haven't tried OpenWRT yet, as by the time I wanted to flash the only router in my stable that supports it (casualy, a TP-Link), it fell prey to the 4/32 rule
https://openwrt.org/supported_devices/432_warning
Having said that, the only way to be sure that your router will have alternative firmware from the get go, is to buy it second hand, or buying one which already has it, or has a close derivate.
I went with a GL.inet GL-MT1600, which has a derivative of OpenWRT, but has plenty of vanilla builds available. Additional are a coverage in my flat is provided by a pair of wired second hand Linksys E900s with DD-WRT acting as APs. So, no need for mesh, not that I beliueve in mesh in a harsh RF multidwelling environment were all the neighborgs are morons, with red brick walls instead of gypsium planks.
I'll stick with stock for a while, and then change to Stock OpenWRT at some point in the future.
Wish me luck.
If your SAP instance has to interact with other stuff that is aleady (or that you want) at AWS, it may benefit you to have the HPE instance there. Also, these companies can strongly negotiate with amazon. Which in turn hetsa sweetheart deal from HPE. So, while not cheaper than on prem, it will be competitively priced.
For this analysis, please forget desktop linux. Follow the money, and the money is in server or embedded.
Problem was/is that there is no coordination between the following groups:
LTS kernel Mantainers
CIP SLTS
AOSP (google) i.e. android
And most SERVER/ ENTERPRISE distros
When LTS kernels went from 2 years to six years, and multiple ones a year, I (and many others) said the LTS kernel mantainers were overextending, we were downvoted to oblivion, but, ultimately, proven right.
One shoudd have one LTS per year (or even better, one every two years), but in agreement that as many stakeholders as possible will use it.
What ended up happening was that the LTS maintainers designated the LTS kernel as they saw fit, and expected everyone else to follow them, instead of negotiating to get as broad support as possible for a prospective LTS kernel.
While Google does not care one way or the other about using LTS on android, their sheer scale means they do not need to. If you do not negotiate with them, they will not go out of their way to use the LTS the maintainers see fit.
We are clear that RHEL will never use the LTS, as that makes life easier for the Copycats.
But there is hope that the relevant copycats may offer their distros in two flavours:
RHEL bug for bug "reverse engineered" kernel and LTS Kernel (the article states that)
But again, the LTS maintainers do not negotiate with them, rather, they have to go out of their way to use the LTS.
So, part of the failure was with the Kernel LTS maintainers DIC-tating, instead of coordinating LTS emitions.
Anyhow, a good reason to use Ubuntu Server.
If you are a personal user, and want to be on the Up-and-Up, the path of least resistance is to get the ESU until 2026, then 0patch until 2028, and then re-evaluate your options.
You may want to be on the Up-and-Up either because your moral compass is strong, or because your work demands you to be on the Up-and-Up (say, you work in industries regulated by the HIIPA,PCI, GDPR, etc)
There are options to get 3 years ESU, or LTSC2019 for us normal folk, without going to the high seas, or bending the license so much it becomes a pretzel, but those are fortuitous or very convoluted.
And better not talk about LTSC IoT 2021
All my clients connect over WiFi, but I need extra ports for the NAS and the (wired) satellite APs around the house.
So, not for me.
Also, far as I can see, the device has 3 detachable antenas, but neither PCB antenas nor internal type antenas for diversity
Given that one of the main points of WiFi 6 is MIMO, I find that lack of antenas, suprisingg, to say the least.
IMHO< a good first effort that requires polish. Will wait for the next iteration
JM2¢ YMMV
PS Went with a GL.iNet (which also ships from the factory with a fork of OpenWRT), only time will tell if I did right or wrong.
OpenWRT does not work on Broadcom Routers, but DD-WRT does, because DD-WRT has certain agreemenst with broadcom that OpenWRT does not.
https://openwrt.org/meta/infobox/broadcom_wifi
Similarly, there are routers that OpenWRT supports that DD-WRT does not (Like my own TP-Link PoS). And then, there is the Tomato Family. The open router firmware is an iceberg, that many a HW vendor does not want to navigate...
Also, there are routers with propiertary firmware, based on things like Windriver OS and other RTOSs that require SIGNIFICANTLY less HW grunt to run.
In the elden times, and more so with the 4/32 warning (https://openwrt.org/supported_devices/432_warning), many an old router will not be supported by Open FW going forward.
Finally, many of these vendors will use the Toolkit provided by the SoC provider (redundancy intended) for their propiertary OS. This allows them to esaily deploy a propiertary/"diferentiated" solution, with "knowledgeable" support from the SoC provider. And if there is a security vuln, we the users rip the HW vendor to shreds, but behind the scenes, the HW maker rips the SoC vendor to shreds, and if push comes to shove, sometimes is the SoC vendor will bear part or all the £€¢¥$ cost of the security vuln.
Weird, I'm going in the other direction. I can somewhat afford the HW tax, and stuff "just works". But the direction of their PC hardware is not to my liking (expandability and repairability wise).
I'll take my current (intel Macs) stable as far as it will go, and then (and only then, not now, then) will decide what to do.
Who knows what the future will bring...
Mint or Zorin Core are good choices and are rocking a somewhat capable older PC.
CrunchbangPP, AntiX or the "reborn" DamSmallLinux if your hardware is truly ancient (Core2Duo or older) or severley underpowered.
good luck in your linux adventures...
Writen from my late 2014 macbook air, with a bunch of linux VMs, including TAILS and Kali
Agree 100% with you.
And with the existance of 0patch, is even easier to try to wing it.
Having said that, sadly, many small businesses (and some individuals in independent/contractor roles) do not know full well what their obligations are under HIPPA, PCI, GDPR, CCPR and other similar laws.
If they try to "wing it", they may get into a wordl of hurt, like failing an audit, losing their certifications, or having to pay massive fines out of pocket ("cyber-threat" insurance will not cover it if the SW was not up to snuff) if there is an incident/breach...
So, perilious times ahead for SMB and independents...
According to my research, the processor itself can run Win11 fine (HVCI, MBEC and all the goodies), but the integrated graphics can not (no DX12).
Probably lenovo would have to tap either innosilicon or mooore thread for a discrete laptop class GPU with DX12 and WDDM 2.x drivers.
Having said that, china is de-emphasizing Windows, for their homegrown linux desktop,so there is that.
¿Est-cé que vous avez besoin de un dessin?
Is not the same having simple tea and coffe (and caffeinated soft and energy drinks) in the office, as is having a small battaillon of baristas in the compannies canttens making sofisticated coffees and teas with lots of ingredients for free...
Drinks with caffeine (coffee, certain teas, guarana, certian fizzy drinks and energy drinks) are not only a perk, but by their very nature (with moderation) make workers "slightly" more productive.
A company would be dumb not to offer them...
Having said that, there is a huge difference between a free brewed coffee in a Mr. coffee with cow milk and a free "cafe espresso abarth with soy milk, extra foam, pumpkin spice and cinnamon", if you know what I mean...
«chip vendor Transmeta hired him. It got the company the low-level expert knowledge they needed to build their Crusoe VLIW chips, which ran x86-32 code by emulating it. »
Transmeta also ran AMD64. Actually, AMD hired them to simulate the architecture before the actual silicon was ready, for the benefit of SW/OS developers
https://www.theregister.com/2000/11/07/transmeta_helping_out_amd_ms/
If you read other news, what intel is proposing is making a processor that boots right into 64Bit Compatibility mode with 32 bit support. No option to boot in 16bit real mode, remove every single trace of 16bit support (both 16bit real mode and 16bit protected mode) from the silicon and also remove a lot of unused 32bit stuff that is not used anymore (like the Virtual x86 mode that Windowx 3.x and (x used to run DOS programs).
Well behaved 32 bit SW (Like Win32.exe and up, and most of 32 bit *nix SW) shall be fine but UNDER A 64Bit OS, that will be the only type of OS those chips will run...
Most likely Intel and/or AMD will no develop fixes like this for processors older than 8th gen, or Zen2 (respectively).
Ditto for drivers for other parts of their SoCs (nee processors).
Basically, when Microsoft meet with Intel and AMD and said ¿For which processors could you guarantee driver and microcode updates for the 10 year life of a "Win11" OS?, the answer was what we know now.
All the 7th gen and Zen+ exceptions, were oneoffs negotiated between the HW maker and intel/AMD
And I am guessing something similar will happen withy Win12, the Processor gen supported will be decided by the processor makers themselves, by denying support (Microcode and driver updates) to older gen processors.
Your statement is false.
Intel emulated dx9 only.
There is NATIVE support for OpenGL, Vulkan, DX12and crucially: DX11 too.
I also play olden games, but you are lucky I was not the project manager for alchemist. I've focused on DX12 and Vulkan ALONE, and handled OpenGL via Zink, and the DX9/11 via BOTH microsoft's emulators (DX9on12 and DX11on12) AND DXVK (on a per game basis, wichever works best).
After this debacle, Win11 will be sacrificed in the altar of Win12, to produce what Microsoft always wanted, a Windows Client ecosystem where all client machines are on pretty much the same version of Windows.
Win10 has not one, but many expiration dates.
Plain Win10 22H2 Oct next year
Win10 LTSC 21H2 Oct 2026
Plain Win10 22H2 + ESU Oct 2026/27/28
Win10 LTSC 19H1 ~ late 2029 ~ Early 2023
Win10 IoT 21H2 ~Early 2032
Many orgs, speciall the larger ones, can jump from one to the other with relative ease.
This gives ample time for Win11 to run its course near the end of life for Win10. Perhaps it will completely run out of support a few month before Win10, or most likely, a few months after Win10.
Most Likely, Win11 WILL NOT have an ESU (just like Vista and 8.x did not have ESUs). For this to come to fruition, Microsoft will have to anounce the EOL of Win11, and launch Win12 in the 2026~2028 timeframe.
But, I have to be transparent. Since the Win Client and Win Server codebases are the same, Microsoft can increase the support span of the client easily by using the server stuff, since server Lifecycles are longer.
But, even with that caveat, come 2032, whether you are still in Win10, or in Win11, if you want to stay in Windows, your only option will be Win12 (of course, Linux, android and xBSD are options too).
And since by then the TPU and Secureboot requirements will be more than 10 years old (and mandated for OEMs since 2018), and since the 40 TOPS NPU requirement will be 8 Y/O by then, you can expect that, except for niche cases, all reasonable client machines in 2032 will be compatible with Win12
As for other Win12 client Requirements, expect something similar to Win12 + Copilot. PErhaps a few tweaks to main memory (to be able to keep a couple of small ML/AI models in RAM), Video driver level from WDDM 2.0 to 2.4 (for added stability) and from HWL 11 to HWL 12.x (for more functionality and performance), and AVX-2 (or AVX2 + AVX-VNNI and/or AVX-IFMA) or worst case scenario, AVX-10 instead of the current SSE4.2.
And of course, the NPU may demand slightly more than 40TOPS. but then again, most chips with NPUs that make the cut now now EXCEEED 40TOPS. Oh, and SSDs will be mandatory.
So, for pretty much anyone, their best bet is to hang on to their current machine for dear life (with any flavour of Win10 or Win11), and wait until the Win12 minimum requirements are anounced.
Think of benching as a new name for purgatory.
As the article clearly says, while "benched" you are still employed, and have a certain time (a few weeks) to find a project where a significant # of your hours are billable. Failure to do so means Redundancy/firing/pink-slip.
So, benching is the start of the process, and also the name of the specific way Kyndryl is doing it.
Also, Kyndryl will probably offshore (to places like India and the Philiphines), but also nearshore to places like Mexico, costa rica and other places in LatAm.
Try to run the same app (say, libreoffice) in windows, mac and linux, and tell me if the OS makes a difference or not.
try to program a non-trivial application in windows, mac and linux, and tell me where it is easier.
the os can have virtues, both for users and programmers.
companies would be foolish not to extoll said virtues.
In theory, the microkernel is FOSS, is handled by the "OpenAtom Foundation", a chinese non-profit that manages it and other FOSS projects in china. Said foundation has a cooperation agreement with the Eclipse Foundation.
Take that as you may.
The good thing for Huawei, and many of the other memebers of the foundation, is that they control the hardware (for example, Huawei makes their own processors), so, for them is the same effort to do a driver for HarmonyOS as it is to do a driver for Windows or Linux
And, the OpenAtom foundation is leaning heavily onto Risc-V
Agaim take this as you may.
I interacted with COMPANIES (Huawei and Nokia, technical training, as a contractor) that used MS Office.
Premade PowerPoint presententations that I had to modify (for example, translate, reshufle, merge or expand) did not render properly in libre. And I was not paid to correct formatting. And any kind of animation/ timing stuff wouldnot work.
Excel forms for expenses reinbursement, time keeping and grading of participants (using whatever formulae/macros behind the scenes) did not work properly in libre.
Word documents that I had to modify, or example supplementary documents and/or exams ((by translating, correcting, amplifying or merging) did not work/render correctly in libre, and I was not paid to correct formating.
easier, faster, cheaper, and less responsability/liability if something broke if I used MS Office , so that is what I did.
When I was teaching at the university before that, then yes, every freacking presentation, exam, paper and grade spreadsheet done by me, was done with (at the time) OpenOffice...
Anecdotal, I do know, but using libre to interact with the msoffice world is not the bed of roses some zealots claim it to be, except in the most trivial cases
THe corpos are my age.
They remmeber when they fell in the trap of buying all those 486SX, and NextGen machines, and Cyrix 486DLCs (a 386 pin compatible processor with a 486 instruction set) to save a few bucks, and then got burned by needing the math cocprocessor down the line.
They will not make the same mistake with AI PCs.
AI will not drive adoption of new PCs, instead, when the company decides that a PC has to be replaced, the replacement PC will be an AI PC just because...
... Just because the corpos do not want to miss on the killer device side AI Application (it does not matter if it materializes or not), also, because it is almost certain that some form of NPU with a certain TOPS will be a hard requierement for Wiin12, therefore, an AI PC is their best bet for a computer that can be upgreaded to the next OS. Buying a Non AI PC in 2024 or 2025 and beyond is professional suicide in corpo world.
Having said that, your best bet in corpo world is to hang on to what you have in a safe manner (Supported Win11 or Win10 with ESU) and delay as much as possible new HW purchases until prices go down and there is more clarity about Win12
My not so humble predicition: Win11 EOL will be anounced around 2026 and Will not have an ESU. Win 12 will be announced around the same timeframe. That way, the end of support for all incarnations of Win10 (Win10 IoT 2021 will be supported until Jan 2032) and Win11 (EOS Anounced 1 year in advance, and LTSC duration 5 years only) will be closely matched, and world + dog in the microsoft Sphere will be finally in one only Windows... Win12
PS: In my native language, Spanish, Win12 sounds VERY similar to Windoze, I chuckle every time. And do not get me started about 5G
You are right.
But Pat, Lisa, Michael and Enrique know it.
That's why, unlike the the new machines of my youth, that were made to entice the people that bought their machines a year or 2 ago, the machines of nowadays are made to entice people that bought their machines 5 or more years ago.
If you are happy with the peformance* of your machine, the new machines are not aimed at you.
Performance in the broader sense, and mostly for corpos. ¿Is the security performance/features good enough? ¿Are the centralized management features of your machine fleet good enough? ¿Is the OS and SW supported enough so you can pass HIPPA/PCI certification du jur? ¿Your isurance company is satisfied with said level of support?
Those are the reasons to upgrade a machine nowadays, more so for a Corpo.
And these to events werew by corpos for corpos
Probably the TOPS requirement will be mandatory, maybe a little bit higher (maybe not). And therefore, the processor cutoff will probably start at lunar lake and AMD AI 300 processors.
Minimum RAM will increase to accomodate the AI models you want to run device side.
Minimum (boot drive) space will increase.
SSD type boot drive will be mandatory.
The WDDM level of your display and HW feature level will increase as well. (currently is WDDM 2.0 and HWL 11)
Other than that, nothing really.
Other than the TOPS, most OEM Win11 PCs sold since Win11 entered the market meet the requirements, and, long before Win12 appears, HW makers will be forced by Microsoft's OEM conditions to sell PCs with those features onboard anyway, to Keep the OEM licenses... It happened before with Win10. OEMs were forced to sell PCs with secureboot and TPM since 2018, a few good years before Win11 entered the chat.
OS, Software license and short term HW savings Vs. People retraining costs + Bussiness disruption as procecess adapt.
Decitions, decitions...
My point being, Linux adoption has upsides and downsides. There is no right or wrong answer. Each and every organization has to make their own analysis, and decide accordingly.
The most hyped up thing about AI PCs is that the model runs locally, and the data remains on the PC, giving privacy and latency advantages.
Is like the math coprocessors of yore, better have and don't need, than needanddon't have.
If a corporation NEEDS to replace a PC (and remember that the article is referring to TWO corporate events) the replacement PC may as well be an AI PC.
Having said that, no sane corporation will replace a PC that does not need replacing just because it lacks AI
In case a worker needs AI, but the PC does not need replacing, the PC will not be replaced, but rather re-deployed. A long way to say that AI is not leading replacement, but rather, lack of performance, natural attrition, (staged) end of suppor of Win10 and formal replacement policies are...
I agree 100% with you, but just a small correction, aversion every six months is yesteryear news. Nowadays there is only one new versionper year (21H2, 22H2, 23H2, 24H2, etc) and a "fatter" update six months after that (21H1, 22H1, ...) that does not get much more than a "tolkien" consideration in therms of additional support for microsoft to save face.
I've said it many times and in many places. The last instances of the win10 codebase to drop support will be IoT 21H2 and server 22H2, both in early 2032...
expect support for win11 to lose ALL support close to that timeframe, and win12 to appear a few years early, in order for microsoft further simplify the support matrix.
more so now that they have to support more than one microarchitecture again...
Microsoft charitably sings a version of GRUB so that you can (dual) boot tpm enabled machines with linux.
enough to keep the FTC and EU of their tail.
they simply got tired of the monetary cost of supporting so many processor/chipset/gpu combos, as well as so many OS variations.
the last win10 codebase stuff (win10 IoT 2021 and winserver 2022) will go out of support in early 2032. Expect that ALL support for win11 will end close to those dates, and that win12 will debut a few years befor that date...
And it will be defined by companies. Around oct 2025 there will be a ramp-up of adoption, from corporations that do not want to pay for the ESU.
Then, for the next 3 years, smaller ramps from companies in the ESU threadmill, that either finished the transition, or do not want to keep paying ESU, then, around jan 2030, another ramp up from companies on LTSC 2019 that go out of support, then, finally, around Jan 2032, the last holdouts using IoT 2021...
expect win 11 (like vista and 8 before them) NOT to have an ESU, so, the latest laggards will have to go from 10 to 12 in one fell swoop.
I hope, for their sakes, that any astronaut scheduled to board the starliner tells nasa, in no uncvertain terms, but very diplomatically, they are NOT entering the capsule, and deal with it.
is not like there is a police force or security guards onboard the station. Also, these peole are very capable, and can find jobs elsewhere literaly anywhere in the world.
I, personaly, subscribe for hybrid.
two or three days in the office, and 3 or two at home.
personally I'd take 3@ the office and 2 at home, but that's me (mostly to set the example). To each (team member) his/her own.
Sometimes y'all do need the face to face time to generate rapport intra and inter team (among other things).
JM2C
YMMV
In principle, Microsoft could (and should) twist AMD's (figurative) nipples and make them fix every chip AMD themselves deemed adequate for Win11.
Part of the dealio for a chip to qualify for Win11 was a compromise from AMD (and Intel) to supply drivers support and security fixes for said products.
That's how we eneded at Win11 launch with Intel's 8th gen onwards and AMD's 2nd gen Ryzen...
So, AMD, get patching, before big bad microsoft comes for you...
The second supplier contract was given to AMD, Siemens, NEC, Harris, and many others. It was actually standard practice at the time.
What I do not get is Why IBM did not go with the 8086 in the first place. Yes, in principle making a systme with an '88 is cheaper, when the system is small, say, the controls for a traffic light. But when the system is a full fledged general purpose PC, the difference is negligible...
Well, a wasted opportunity.
Aonther thing I do not get (and another wasted opportunity), is that Intel told IBM of the impending 87186. If intel were wise, they would have used in the OG XT chips compatible with what they knew would end up inside the 187, even if not buying them from intel (rememeber, second supplier was not something you forced the chipmaker to do, it was par for the course at that era).
This meant that the PCjr and the klatter X86s were more expensive than the 80186 and 80188 clones of the era.