/tech/ - Technology

>>13351
<3. Avoid Two-Way Engagement
<Reviewing an unsolicited patch from a contributor in a sanctioned region should generally be fine, but actively engaging them to better understand their issue, diagnose the problem, or help improve a patch or modify code would likely cross the line. If the contributor is linked to a sanctioned entity or region, in general, it is best to keep communications strictly one-way. If a patch is received and you improve it and submit it upstream, that should be fine, but going back and forth in communications with the SDN developer likely would not.
<4. Avoid Contributions Enabling SDNs
<Accepting unsolicited patches that fix general issues in your open source project should be okay. However, if the changes directly benefit a restricted party’s products or services, it could be a problem. For example, if a developer from AcmeSDN (and AcmeSDN is an SDN subject to OFAC sanctions) contributes a driver that enables the AcmeSDN processor to work in your software, that contribution would likely be an issue. Think carefully not just about the source code, but the impact of these unsolicited patches.
<5. Avoid Indirect Contributions
<Sanctioned entities might try to contribute indirectly through third parties or developers acting "individually." Developers should understand other contributors' affiliations and raise any concerns with their community and legal counsel. For example, if in the prior example, AcmeSDN paid a developer in a country not subject to sanctions to make the driver contribution enabling AcmeSDN’s processor, that would still likely be an issue. A common pattern is that an SDN’s developer is blocked from making a contribution, but then a very similar (or the same) patch is submitted to the project from another account or email address. It could be an anonymous email account. Just because the contributor has been obfuscated does not change an assessment of the situation.
So basically the sanctions are placing an enormous burden on libre projects to both a) keep extreme track of the identity and national origin of contributors in order to avoid the wrongPost too long. Click here to view the full text.

>>13351
been reading this, it seems that the "OhFuck" sanctions, mean that everybody can use each others code, but devs aren't allowed to talk directly to each other anymore, they have to talk past each other. Bunch of confusing shit.

>>13353
I don't see why this is such a big deal. Can't we just move to an open source repo or something?

>>13353
>keep extreme track of the identity and national origin of contributors in order to avoid the wrong ones, making software projects about people instead of the software
Yeah that's the difference between technical people and non-technical people.
Technical people absolutely despise this kind of "personality drama", the reaction to this will be: there are no people, there's only code

I'm so angry at how little attention this controversy is getting in other places of libre software enthusiasm.

>>13348
>what else is there to say?
Apparently deep-seek caused a big sell-off in tech stocks. There's a lot of speculation on that. Mine is that it proved that the big players aren't the only game in town, and that new competition can emerge, and that the trend probably goes towards commodity AI rather than siloed monopoly rent AI.

Deepseek being open source is also nice.

With it being a pure reinforcement learning design it doesn't need fine tuning and that's probably why it only cost like 6 million to make.

If you want to run the big boy model with 600+ billion parameters you need like 405 gigabyte of memory preferably speedy video memory. Graphics cards need an entire order of magnitude increase in video memory capacity. The 8-16 gigs on consumer cards and the 24-48ish gigs on profesh cards doesn't cut it anymore. Maybe this is finally whats going to cause mainstream PCBdesing to incorporate optical data-lanes.

The Hated One made a good and brief vid about big tech lobbies trying to kill Open source AI because bigtech can't compete with the open stuff on cost efficiency. His take is to go bug your political representatives to not let big tech hord all the AI-tech for it self.

My question is, could the tech monopolies really block Opensource AI ?

Can't the computer wizards just go to some other country and ask for opensource friendly regulations. That country could import a massive tech-boom for free. Possibly even get better AI. It's not like this is huge immobile industrial technology.

https://farside.link/invidious/watch?v=5NUD7rdbCm8

>>13304
> face ATTENDANCE
Bio-metrics is a cyclical fad.

The very technology that enables you to detect a bio-metric feature also allows you to make fake duplicate, that circumvents it. It's conceptually flawed.

https://github.com/hanweikung/face_anon_simple
Simple face exchange for anonymity

>>13019
https://archive.ph/20240422172830/https://www.wired.com/story/section-702-reauthorization-expansion/

>Legal experts—including a rare few attorneys who’ve argued cases before the FISA court in the past—say the new ECSP text ensnares owners of facilities housing equipment used to store and carry data, as well as commercial landlords and virtually anyone with access to communications equipment in those spaces. The text, they argue, may be interpreted by the government as granting it authority to compel the assistance of “delivery personnel, cleaning contractors, and utility providers,” among others.
How they can reconcile forcing normal people to become spies, is hard to fathom. Ethically and strategically, it sure is egregious, however abusing people like that, fosters malicious compliance.
>Criticism of the 702 program largely stems from revelations of abuse in a declassified court filing from 2022, which describes rampant misuse of the 702 database by the FBI. Investigators at the bureau have been caught unlawfully scouring 702 data for information on American protesters, journalists, and political donors.
This type of hard persecution is what kills the legitimacy.

>What's next for the internet?
it's hard to predict technological trends, a new counter strategy to surveillance extremism might be increasing data noise, if enough useless data clogs up the system, that might mitigate the harm it does.

Good post by Kevin Gosztola on the recent history of domestic spying with a foreign pretext:
https://thedissenter.org/biden-expands-ranks-government-spies/

Seems like one of the aims of this bill is to go after Gaza genocide protestors by pretending they're Hamas.

>>13024
>Seems like one of the aims of this bill is to go after Gaza genocide protestors by pretending they're Hamas.
Peak Zionism ended a while ago, the political pendulum had already begun swinging in the other direction. The Zionists going full retard with the mass-murdering are accelerating that process. And it's not just mass-politics that are changing, the militarized Zionism project is more trouble than it's worth in material terms as well. Soon everything tied to it will become politically tainted.

https://noyb.eu/en/ag-cjeu-facebook-must-minimize-personal-data-ads-eu
Here's an interesting case about minimizing personal data collection and retention.

This ties in well with what i think the near future privacy politics will become, where people consider data related to people as analogous to some kind radioactive waste material.

There's two major apps that we could use moving forward.
Blue Clover: https://nnuudev.github.io/BlueClover/
Kuroba-Ex: https://github.com/K1rakishou/Kuroba-Experimental

We were very close with Kuroba-Ex but the developer is a bit of a dick and asshole. Thus why I moved to Blue Clover.

Sauce: https://github.com/K1rakishou/Kuroba-Experimental/issues/780

>>12758
based thank you

Since leftypol has been added to kurobaEx, would it be difficult to fork it and add leftychan?

>>13296
Why fork, why not just send in a compatibility for both patch?

>>13296
You'll need two things:

- to find the pull request or commit where leftypol was added
- to be able to compile and test the app

For the first one you can look on github and for the second one you need to figure out how they build that app, are they using Android Studio? Idk anything about android development (or ios for that matter) but I'd download the source on github and open it in Android Studio and see what happens.

I'm not sure what you mean concretely by "scrambling the logic of a CPU". Do you mean to say that a key would be needed to load data into registers? Do you mean to say that a key is needed to know the instruction set? Are you referring to fully homomorphic encryption?

The other problem is that suppose this organization of yours can make "scrambled CPUs". How can the rest of us trust that organization? What's to stop some interested party from compromising it?

>>13274
There is a philosophical question whether microchips can be something that has verifiable operations or whether it's doomed to be a black box. And here the answer is, yes it's possible to build a clean stack that is verifiable all the way down.

If you want to go as hardcore as the philosophical thought experiment you need to have an organization that starts from scratch with a clean-slate processor, that can be a super basic design and therefore it's plausible to do this in reality. Whether a nearly omnipotent opponent exists that requires this level of commitment, is another matter. The motivation for doing this would be more an exercise in scientific rigor and precision.

On the realistic side of things you would scramble a cpu design if you have to produce it in a fab that you don't want to trust. That is from the perspective of a cpu designer.

From the perspective of an end-user: the scramble interface could be exposed to end-users too, they could generate their own de-scrambler.

>>13236
Drew is a scheming bastard that wants Stallman and most of the FSF gone so him and his friends can install themselves. You don't want people like Drew leading FSF because these scheming ambitious types don't give a shit about ideals or principles and will 100% sell out.

Drew already has a bad track record. He used to be a huge fascist then suddenly flipped to being a progressive trans ally. Not a trustworthy man.

Also he's a huge hypocrite accusing Stallman of being a pedophile while being an avid browser of /u/ and /c/.

>>13239
I wonder if this is one of those cases where an accusation is a confession. The most self-righteous always seem to have skeletons in their closet.

>>13239
I do not care about attention seekers with nothing interesting to say.

I do not care about hating on them. I do not care about their opinions.

>>13239
>he bookmarked /c/
>1 of the 200 threads on /c/ is a loli thread
>HE IS A PEDOPHILE
This is just as retarded and dishonest as the attacks on RMS you're complaining about.

We've been discussing it here:
>>13230

Thankfully the culprit was exposed quickly and isn't garnering a lot of sympathy. It looks like this is going to fizzle out more quickly this time, especially after the controversy just ignited over the US attacking worldwide free software collaboration.

>>13224
God don't even fucking bring the Cloudcancer faggotry up, it makes me so mad. These faggots are fucking enemies of the people, doing their best to make Firefox unusable

https://xcancel.com/Sn_darkmeta/status/1845502888480579860
A group claiming to be Palestine solidarity activists is claiming credit for the hacks and DDoS. Apparently we're supposed to have sympathy for victims in Palestine after they just attacked the good guys.

Here's a real juicy bit of their write-up:
>On June 1st, 2020, four major publishing houses— Hachette Book Group, Penguin Random House, HarperCollins, and John Wiley— filed a lawsuit against the Internet Archive in the U.S. District Court for the Southern District of New York, claiming that the Internet Archive’s regulated digital lending practices constitute copyright infringement. On March 25th, 2023, the court ruled in favor of the publishers. The negotiated ruling issued on August 11, 2023, prohibited the Internet Archive from lending books for which electronic copies are sold digitally.

>On August 11th, 2023, the major music companies Universal Music Group, Sony Music, and Concord (along with their subsidiaries Capitol Records, Arista Records, and CMGI Recorded Music Assets) also filed a lawsuit against the Internet Archive in the same federal court in New York regarding the archive's Great 78 project, demanding $621 million in damages for alleged copyright infringement.

>The archive was supposed to be a reference for information, but the site has started to resemble piracy sites. Frankly, we are astonished, as the picture has become clear: The archive is officially responsible for this circus they created to escape from lawsuits and financial crimes. Do you want a job with us, Brewster Kahle?

In other words, this entity wants us to be mad at Internet Archive for violating copyright law. They want us to think IA actually deserved to have a bunch of scumbag publishers go after them. A hacktivist group siding with copyright law? First one I've ever heard of.

>>13227
>A group claiming to be Palestine solidarity activists is claiming credit for the hacks and DDoS.
Yeah they probably are neither a Palestine solidarity group nor the people who done the hacks and ddos.

>In other words, this entity wants us to be mad at Internet Archive for violating copyright law.
Publishers want to get rid of public libraries, that's all there is too it.

IA probably didn't violate copyright law, by any reasonable interpretation. (I do understand the irony of putting the word reasonable in the same sentence as copyright in present year)
And it's just madness to accuse people who run websites to be engaged in mercenary naval warfare.

There is a technical aspect. Libraries do have to pay to lend books. But during covid many people were cut off from libraries on account of "lock-downs" (horrible terminology) and could not use the book lending services. However publishers did not offer a refund to libraries, that means they took license money but nothing was offered in return.

There is another argument, the extremist expansion of copyright was not exactly done with democratic consent, powerful lobbies wrote those laws and then it was foisted onto society. To what extend copyright law falls into the category of unjust might-makes-right and justice respectively is debatable. Unjust laws are not valid, it's not a crime to violate unjust laws but it is a crime to enforce unjust laws. A recent example of an unjust law would be banning birth-controle medications, because it violated the right to self-determination and it is political interference into medical treatments. Basically medical treatments ought to be decided by patients and their doctors, without any third party interference.

If they kill Internet Archive that will create damages in opportunity cost for billions of people. The copyright pushers already have infringed on personal property rights of billions of people by infecting digital goods with malware that they call "digital rights management". So what they are doing likely isn't above board either.
Post too long. Click here to view the full text.

>>13227
>"Palestine soldarity group"
A Zio-op, non plus ultra

archive.org is back, btw
https://blog.archive.org/2024/10/21/internet-archive-services-update-2024-10-21/

>In recovering from recent cyberattacks on October 8, the Internet Archive has resumed the Wayback Machine (starting October 13) and Archive-It (October 17), and as of today (October 21), has begun offering provisional availability of archive.org in a read-only manner. Features like uploading, borrowing, reviewing items, interlibrary loan, and other services are not yet available.

>Please note that these services will have limited availability as we continue maintenance.

Same here! Text is based!

That's so 1990.

https://eregime.jcink.net/
Good textboard

bumpy