The '90s called, and want their Amazon back. Amazon in 2025 competes in various fronts.

If the laid off engineers worked in the "Tat-Bazaar" competitors exist beyond Target and Walmart and include the websites Costco, BJ's, and for Amazon Engineers located in LatAm, Sites like Mercado Libre.

If the fired engineers worked in Amazon Prime Video, the doors are wide open at Netfilx, Apple TV (plus or not plus, I do not remember), Disney+, HBO Max, Peackcok/Universal+, and plenty of other streamers worldwide.

In Cloud, engineers fired could go seek employment in competitors like Microsoft Cloud, oracle cloud, Google Cloud, IBM Cloud, OVH, Hetzner, and plenty of companies offering OpenStack clouds. Heck, since netflix (and many other large companies) runs on top of amazon clud, maybe they want some engineers with intimate knowledge of how things work inside the "black box"

Were those engineers working at AI efforts at Amazon? OpenAI, anthropic, Google, Grok, meta and many more are there waiting.

Were said engineers working with gadgets like alexa? plenty of consumer electronics companies looking for good engineers

I see plenty of competition for Amazon engineers, if all these companies were also not reducing and re-aligning their engineering efforts.

And please, I/we do not buy the argument that the Ilumminati are sitting on the boards of all the large companies, in tall leather chairs, with cigars and cognac, plotting the end of the peasants class.

Again, in this particular case, is not lack of competition, but lack of unionization.

I do not think there are no competitors. There are plenty of competitors, but said competitors have ALSO been firing engineers left, right and centre. And that "synchronized firing of engineers" can not be stoped by anti-trust laws, but it can be slowed down by unionization.

1.) DELL and HP-ink* DID NOT kill codec support retroactively. Unlike Synology (also mentioned in the article), who removed the codecs RETROACTIVELY FOR ALL MODELS PAST AND PRESENT, probably because they could not be arsed to keep different versions of their DSM 7.3 OS with or without the codecs depending on the HW model, and also, because the oldest models could not for the life of them keep doing the transcoding with all the extra stuff that was dumped on them form DSM 5 all the way to DSM 7.3.

2.) The thing is, if you had a DELL or HP-ink computer model XYZ-rev1 from early 2025 you had the codec, and then all of the sudden, when you buy a second computer model XYZ-rev02 from late 2025 with pretty much the same hardware, sudenly, you do not have said codecs, even if the hardware encoder/decoder is still there.This baffled some customers, and offered untold click-bait-rage potential for tech news outlets everywhere.

3.) DELL and HP-ink want to either save the price of the olive** by not paying the 24cents for the codec royalties going forward, or upsell you to a higher priced fuller featured laptop. Either way, is money in their pocket.

4.) Cue people developing scripts to re-inject support by automaticaly downloading and extracting drivers and miscelaneous files (like .inf files) from older computers from DELL and HP-ink website (or failing that, downloading them from the generic driver packs from Intel and AMD iGPU drivers), complementing them with FFMPEG 8, and making some registry and other tweaks in 3... 2... 1...

5.) Again, this is not retroactive to older models, if the laptop you bought a few quarters back had the codecs active then, it still has the codecs active now. Is models within the same series, with very similar hardware and model numbers the ones that now ship without said codecs that are causing confusion.

* A few years ago HP split into HPe (e for enterprise) for servers, networking and datacenter stuff, and HP Inc for PCs and Laptops and Printers. HP-ink is a pun.

** https://www.forbes.com/sites/m...

Please let me clarify that I was tanking about Krafton and their AI innitiative specifically.

Seen many a comment, so I'll respond generically:

Your top performers are separated intwo groups, those who are genuinely excited to work with AI, and those who do not, with a very small group of "don't cares". Guess who stays and who leaves.

In the mid-range and low tier something similar occurs. Also, in those two populations the group of "don't cares" is much bigger.

This will self-select out the people that will either actively or passive-agressively oppose AI to leave on their own volition. And avoid mass layoffs which may attract scrutiny in SouthKorea specifically, and more generaly, worldwide.

Also, some people that were mid-tier before AI, may move up or down the tier list as AI is introduced. Good in both cases (the one who moved up became more productive and we keep, the ones who moved down are primed for the next round of layoffs).

Then you focus on the low tier workers post AI. Doing a "justifiable firing" case for those should be easy. If you achieve the desired employment levels, then mission acomplished. Otherwise:

Then comes the work-mobbing/slow firing of the mid-tier workers that were not smart enough to go on their own volition until you acquire satisfactoy employment levels.

You do not need to achieve full nationwide blanked coverage in one of the Gs, before you start to deploy the next Gs, if there are customers who want the service, and are willing to pay reasonable rates for it.

And customers are not only you and me with our smartphones, is also consumers with wireless broadband, customers and companies with latency sensitive workloads (and 4G brought a reduction in latency, 5G brough even morem and 6G even more), companies which need network slicing, or companies which need energy savings @ a fixed Mbps (each digital G has lowered the pJoules per bit down significantly). Also, things like self driving cars, including cargo transportation platoons, companies which need network slicing, etc.

These are just a few examples of capabilities that will be enhanced in 6G.

I agree with you that Huawei was in the technical lead in 5G, and that lead will extend in 6G. but that lead is not sooo big. Depending on the specific area, I'd meassure that lead as less than 36 months tops, and that streches it.

Except for telcos that do the 6G rollouts in the 2029~2031 timeframe (which are few and far between), not really relevant...

The main issue is the cost advantage. With Huawei and ZTE being significantly less expensive than Nokia or E//.

In RF/BSS only Samsung is an alternative, cost-wise, from a big company. There are other alternatives, but at that point, you are assuming a risk if you are a small telco, or playing kingmaker if you are a large one.

As I said, In Servers (for the NFV core) Germany (and europe) have Jackshit domestic alternatives. In DCN (Data Comm Network), they only have Nokia, and in optical, they have Nokia or Adtran (IIRC, E// killed/ejected their DWDM dept).

So, less options == more price AND less flexibility.

Full disclosure: Was in Huawei's payroll in my country in the late '00s, and worked with them as an independent contractor in the mid '10s. Have kept up to date in the area, and still have contacts inside.

Agree 100% with you. Voluntary severance packages should be offered in a no questions asked first come first served basis.

If manglement* or HR can say no to a request, it means that trust is broken. Manglement KNOWS that the employee wants to leave, and the employees know that they know. This leads to the employee embarking in a mad race to find a new job, for fear of being fired, and manglement trying to replace the employee ASAP (for fear of them leaving). I'v seen it first hand, with a very gifted storage Sysadmin. It was not pretty.

While I despise AI initiatives in their current form, this is the way to do it.

You declare your AI intentions and lofty goals, then give the employees a decent (or, in this case substantial) voluntary resignation package.

No bad blood, and if you need to re-hire these people in the future, no burned bridges.

I hope more companies idd things like this.

JM2C
YMMV

I mean, both are Europeans, and Nokia has a slight touch of Germany

Also, they shall not be using Cisco, Juniper or Arista Routers (too USoAn). Lucky for them, Nokia has a router line inherited mostly from Alcatel.

But, things get realy nasty once we continue, they shall not be using Samsung RF or DWDM equipment(too Korean), even though it is the leader of low priced fiber, and in 5G RF (and probably in 6G too).

And for the NFVs (Network Function Virtualization) at the core of the network, they can not use HP/Dell servers (too USoAn again), no Lenovo Servers (too chinese), no ASUS/ACER/Supermicro servers (too Taiwanese),

And finally, no NEC and/or Stratus Technologies ft servers (too Japanese & USoAn) for the cloud controllers and critical servers...

Good luck with that 6G network.

Because it fell to the floor a short while after being unveiled, so, if the teacher in the "humanoid robot classroom" were reading the grades outloud would say: "Adol, F"

Nickels also cost more to make than the coin's worth, and the dollar bill deteriorates so fast, that a coin replacing it is the best option.

the 2 dollar bill can take it place in stipers thongs, as the minimum note.

Also, for coin operated vending machine convenience, ramping up minting of half dollar coins would be prudent.

Today, we are gonna learn how to bypass al security safeties with simple hand tools. We will need a breaker bar, a torque wrench and our safety glasses

I was the NOC sysadmin, later, manager for Value added services.

The closest I was ti compiling was for verifone terminals before the telco job. The compiling on the PC was fast enough. Dumping the code into a POS to test because the simulator was shit was the real bummer.

That and the frequent defrag/optimization of the machine.

I would start working when I walked through the door, Since my machine was only mine, I'd turn it on on mondays, endure the 20min boot + Opening of apps (+ Memory dumping process*), and turn it off on fridays, ah, good times

* After booting and opening all your "workworse apps", you would call a script that would request 85% to 90% of the total RAM of the machine, forcing everything to SWAP. Afterwards, slowly, things would come back from swap, but only the really usefull stuff, all the flaff (codepaths seldomly used, if at all) stayed on the swap. Made a huge difference on Win2000 and XP, less so on latter editions, as the memory manager was slowly refined.

cue some enterprising developer hoping yo make a name for him/her-self making a FOSS to extract the old Icons from online installers and re-implementing them under coarse and fine grained control of the user in 3... 2... 1...