HN user
Tangentially reminds me of what happened to the South Korean gov data center [1] where a no-backup ~900TB data center got erased due to a battery fire.
Withno external backups piecing together all the lost functions must havebeen hair raising, and more forensic archeolgy than data recovery.
Last i heard i think they had restored a quarter of the lost services/data.
Wouldnt the weight of the old chassis be a impedement to mods , older models tended to be more robustly buildt. Or what cut-off date for mods would be appropriate (last 10 years?) as beyond a certain cutoff would not be cost effective and most likely would be a hobby project done for the love of it.
Well there is a close equivalent today where a similar situation is playing out.With the argument being made that the dev cost aside - the pricing is being tied to what the potential future treatments would have cost.
Save a child from crippling illness and a lifetime of pain with a single dhot: Jonas Salk 0$ ,Doug Ingram 3.2M$ [1]
[1] https://web.archive.org/web/20260118223436/https://www.cbsne...
Basically a new iron curtain didving the world into digiatl blocks.The era of open internet/science is on its last legs with the potential forr bifurcation into incompatible ecosystems high , the onger the exchange is disrupted. As recently as this month the USgov has donce a Wolf Amendment style declaration for the Scientific collaboration NSF while shifting its purview under the military. To add to that its trying to rope as many countries into its Pax Silica idea intentionally to exclude China while simultaneosly coercing its 'allies' into using its nerfed offerings [1]
So maybe some isolated switzrland/singapore type locales would exist for US/EUusers to be able to dip their toes across the curtain legally without reprucursions.
[1] https://nitter.net/RnaudBertrand/status/2069574934972797089
Apologies if it came across as accusatory , not the intent. A lot of the people within the ecosystem however view it as such a s the primary beneficiary of corps and individuals an afterthough if at all.
For the BSDs they still have the problem see FreeBSD getting it bits reused by Sony/Apple etc with very littke to show back in terms of code. They do give grants of support however if they solved a code issue that the open code struggles with. It will never see the light of day as their chnages are propreitery as the licensing allown.
I think GPL and copyleft in general is getting less and less relevant as time goes on. Looking at GPL specifically it relies on scarcity. The reason companies would agree to the terms of the GPL
Therein is the great misunderstanding , the GPL was never written for 'companies' , it was and still is for the User. You, Me a $MEGACORP , sentinel islander - it does not matter the rights are granted to all equally to reuse/modify/offer for sale as long as the contributoins come back to the commons.
What is happening now is akin to the 'enclosure system' in early Britain when the commons which had been for the benefit of all were fenced off and the peasants thrown off the land to seek wages in the newly industialising system.
When no one is contributing to the GPL commons the options become more restricted. If one isnt a corp that can write their own library or a 10X coder that can bash it out on their own , leaves the users looking at proprietary solutions or restricted offerings with two tier licences.
So in a way yeah most coder/engineers have developed an antagonistic relation to the GPL commons , which is leading to its decline in some sectors.However if/when the share of GPL drops to a level where the adverse effects can no longer be ignored , there will probably be attempts to rollback the clock.
I believe in microcontrollers its already pretty ubiquitous , see their utilisation by WesternDigital with their SwerV core thats already shipping since 2019. At speeds and complexity comparable to desktop/server cores from Intel/AMD they are still lagging in perf though improving as more cores get deployed. Also to add into the mix the whole geopolitics with non-US players hedging. So potential is there will just depend on what will be the base case like Windows was for Intel.
Eeh even ther its a stretch , when people talk about Africa - they should really specify where exactly. PLaces like SouthAfrica [1] already have a robust Fiber network with accelerated buildout of FTTH. Ditto for most of Eastern Africa countries which have FTTH to most of the major cities and subururbs with accelerated buildouts ongoing. Unless its a conflict area most regions are getting wired up pretty fast to enhancce business connectivity - the speeds and bandwith for starlink make noe economic sense once a developing pop are factored in.The only major push for many countries approvals is basically strong armed and shaken down by the US admin on behalf of Musk[2].
[1]https://ctcommunications.co.za/blog/south-africa-fiber-rollo...
[2]https://tech.yahoo.com/science/articles/us-pushes-nations-fa...
I think one of the issues is premature optimization/minimzation(?) . All the wearables so far that feature some human interactive display are essentially cramming the mobile concept into different device classes. Take the AR glasses , they are essentially a computing unit output display , input (joystiicks/sensors) all being crammed into one unit with a powerpack to drive it all. I believe the desktop concept, or in this case 'Walkman' concept would be better fit for wearables .Where one can separate the processing from the display/output and power each separately.Having a cable from your hip pack to the headunit is not so strange/difficult as previous gen of walk/discmans can attest. However the current ruling desing is wireless everything , super slim devices and all unified with the power supply.
Would actually be a good biz model for the Colo facilities that keep shutting down as everyone moves to the big cloud providers.Now if they can get their hands on enough GPUs and RAM.
I believe until the hardware designs catch up to be more commodized ala cryto mining evolution from GPUs to ASICS for specfic algos. Designs (like Google TPUs equivalent) would also need to evolve to be more memory dense to be able to handle them. Untill then it seems will be system time shares for the larger models , probably with a bring your own model and pay as you go.
As far as Europe is concerned they have recently signed up to the 'Pax Silica'[1] and willingly givrn the LLM space over to the US incumbents with buildtin legislation banning Chinese models and coperation with them. So EU will be a renter of the LLMs that the US allows them to use. In the long run OpenSource will dominate as it did in the DB(MySQL/Postgres)/ServerOS(Linux/BSDs) versus Proprietery rent seeking alts like Oracle and Microsoft et al. Would be interesting to see what the global startups using Qwen/DS/Kimi etc within the EU-US space navigate the cutting edge OpenSource LLMs vs seeking/getting a permission slip from the US gov.
How does this play aout in the current realm of ID verifiation laws sweeping across most of the EU and US. As usual is using 'protect the children' : UK,AuS, FR for social media and Operating Systems while also pushing 'national security' : this current iteration. End result seems to be total end of online anonymity while the data slurpers and data brokers continue plying their trade uninterrupted.
Concur, if it was an established builder with decades of know-how and experience doing a new buildout system maybe would be plausible to expect a reduction in the timeline for delivery. For comparison , the Chinese and Koreans who ae about the only players that have been consistently building/delivering reactors at scale over the last 30 odd years take a minimum of 9-10yrs. Expecting a commercially viable fleet buildout in half that time with less human capital and institutional knowhow is wildly optimistc. Maybe they subcontract out to the Koreans and get the running start that way?
I see a slightly different parallel here - they are basiclaly building a framework that takes the US adn friends back to the early 90's where cryptology was considerd a munition and all export products were nerfed. Just like then those that wanted to collaborate with the rest of the world found a way (printing /tshirts etc) similarly now those labs within the US sphere have that decision to make.Unfortunately its the Darios and Sam who are pretty much in favour of regulatory capture environment.With no other frontier labs in the US committed to OSS , and CHN models banned - the devs are pretty much hosed. I doubt the rest of the AI labs in China et al will follow suit in hobbling their own models.As its seen more as a commodity not a wondertool with a moat. At the end of the day the ask is going with a Oracle/Microsoft over a GNU/Linux type of environment.
Seems in all thing tech at the moment the US legal system is accelearting a great split and erectinga digital iron curtain, from AI models to the more mundane like TLS certs. Its been standard for a while for many Linux distros based in the US to toe the party line - like RedHat having notices pretty similar to this one by LE. Seems any meaningful Open Projects will have to choose what path they want to take, be like RISC-V and relocate or LE and others and enforce the divide.
So basically it will become more or less similar to the structure for SQlite and Fossil by Dr.Richar Hipp et al , basically seems most projects that have the requisite manpower/maturity will end up at that kind of structure. In the long run may be interesting from a chain of trust (human as well as code) and interop as any dev from these projects (guilds?) would already have some trust build in.
Famously back in the day Grindr , which had a plot point in the Silicon Valley series . Probably more obscure ones that havent been heard of outside software in the Hard tech space like MotorSich (Ukranian) was being courted by Chinese investment got blocked due to US pressure. And very recently the whole TikTok fiasco.
One of those names that forces a double take when seen disconnected from context:
'Lean or purple drank is a polysubstance drink used as a recreational drug. It is prepared by mixing prescription-grade cough or cold syrup containing an opioid drug '
proving that one of the hardest problem in CS - 'naming things' still keeps on keeping on.
I think the article has a giant blind spot as far as China is concerned , considering they have already a mature enough memory ecosystem via YMTC that Apple was considering sourcing from them. As well as continued expansion in the DRAM and HBM Fabs [1]. It feels like the memory cartel once again trying to incentivise their various govt to cough up some more tax breaks/funding to cushion the AI buildout bet that they made and the bubble seeming about to pop. In any case if they leave the consumer market underserved it should be no surprise if before that 2030 prediction we are all on cheaper YMTC memory modules.
[1]https://www.tomshardware.com/tech-industry/semiconductors/ym...
Funnily enough there is some level of control that can be exerted by the US gov via the distros (at least the major ones - see legalese restrictions on Redhat/Ubuntu etc when you want to download , stating the various US gov laws/sanctions that they follow) and also via the kernel - i think some time back Russian kernel maintainers were removed.
So Open source it may be , however there are still pressure points that can be used. I believe this is one of the main reasons RISCV foundation moved to Europe.
Yep i think thats why even POWER isnt sold standalone but as part of the Z/i series packages as a unit.
They will probably market the ARM inclusion similarly - as something that the package provides.
As far as POWER i think only Raptor[1] does direct marketingof the power(hehe) and capabilities
Im thinking maybe as a compliment to x86 offerings and eventual displacement as a primary offering , i do not see them ditching POWER.
The architecture might be non-standard and not very widespread however for what it does and workloads that are suited to it. I dont think any ARM design comes close , maybe Fujitsu's A64FX.
How does this compare to chisel [1] , i never could get around the whole scala tooling - seemed a bit over the top. Though i guess it is a bit more mature and probably more enterprisey
To expand on that - there are many body systems that depend almost entirely on the 1g glodilocks zone. Lymphatic systems movt , venous blood returning deox blood to the heart and even some digestive processes. Keenly dependent on a g value that allows proper muscle tone/function to the systems at play. Too little or too much and and human life becomes non-viable. Throw in the effect of ping ponging between microgravity and 1g and the issues multiply.
Whats the stdlib situation for swift in comparison to newish languages like go or rust. I know its not batteries included lke python - and doesnt have a massive dev ecosystem of helper libs seeming to be mostly tied to macOS/iOS operating system API/ABI.
Yep and the LLM tools are giving flasbacks to the Frontpage/DreanWeaver to geocities ipeline for building the sites.
Still early innings but i bet this plays out the same way - not everyone will have the time sink to vibecode all the software workflows they require.Maintainance iwse and security wise holes will still remain for the personaly non tech user. Devs and orgs will probably limit the usage to a helper sidecar rather than the hyped 100% LLM generated apps. Reminds me about the hype
Not to mention that p3 on its own was prettymuch functional and p2 quite stable and the major issue was migrating/porting all the legacy over to p3 .Hence bridges like six and 2-to-3 that at least attempted to smooth the transition over by allowing bot to coexist for a time.
With wayland they seem not to be even entertaing this optionality - with wayland itself being not yet feature complete to standalone.And the attempts to bridge like xwayland coming way after the fact and pushing a oneway path with no coexisting situation.
As a result introducinga whole lot of friction and surprises in UI functionality. So yeah at a time when the presentation layer should be a boring afterthough, it is too timeconsuming in part of a Linux setup and daily usage.
Yeah seems likethe business interests have overridden the adpotion needs. Knowing the IETF process is molasses slow , they still have not made moves to close that gap.For open source at least a implementation RFC that interested parties could work with - none avaiable.
They want to sell a technically brillant protocol that is single vendor propriety/patent restricted.
Their bet should have been of open protocol and captalizing on fist mover advantage to drive their business side witha large partner like ericsson/cisco etc.
Of course theres also the soveriegnity angle knowing what went on with another swiss company CryptoAG.
Or more interestingly with the low-earth sat/data network. Seeing as projectssuch as starlink are basically mil in nature with a side of barely profitable civilian use. The whole data centers in space makes more sense. These are not for running cat blogs and video streaming , which is waht they are/will be marketed as. Realworld application will always be a command and control node spanning the globe for the mil use. And as its rolloed out globally can basically provide jammingfree links for the autonomous commands from space.