[ overboard / sfw / alt / cytube] [ leftypol / b / WRK / hobby / tech / edu / ga / ent / 777 / posad / i / a / R9K / dead ] [ meta ]

/tech/ - Technology

"Technology reveals the active relation of man to nature"
Password (For file deletion.)


IRC Chat





| Catalog | Home

File: 1608526406313-0.png (131.31 KB, 989x889, cootact.png)

File: 1608526406313-1.png (351.07 KB, 1165x603, Screenshot_2020-12-09 Yond….png)


>Yonder is an A.I. software company that discovers the hidden groups who control and amplify online narratives, so companies can navigate an unpredictable, ever-evolving internet with confidence. While we can’t fix the whole internet, we can help with this part of it. Our mission is to create a more authentic internet, where everyone can experience a true sense of belonging.



File: 1608526406694.pdf (650.48 KB, Yonder_ABCs_Online_Viralit….pdf)

They give info to companies on how to increase their influence and reach.
>their promotional pdf attached
I wonder if we can get our hands on some of these reports they make for companies.


Sounds liberal as fuck and I bet it's only going to target """alt-right""" and lefty stuff.

File: 1608526399670.jpg (487.95 KB, 744x1218, 1607043116285.jpg)


>The research paper in question deals with possible ethical issues of large language models, a field of research being pursued by OpenAI, Google and others. Gebru said she doesn’t know why Google had concerns about the paper, which she said was approved by her manager and submitted to others at Google for comment.
>The paper called out the dangers of using large language models to train algorithms that could, for example, write tweets, answer trivia and translate poetry, according to a copy of the document. The models are essentially trained by analyzing language from the internet, which doesn’t reflect large swaths of the global population not yet online, according to the paper. Gebru highlights the risk that the models will only reflect the worldview of people who have been privileged enough to be a part of the training data.
Kind of a weird thing to get fired over, especially when it's your damn job to do exactly this.
11 posts omitted. Click reply to view.


That wasn't it at all. Reading comprehension: F, see me after class.

What she is saying is that AI is being trained using language data from the internet. Only the wealthiest people on the planet have access to the internet, therefore AI will be trained by speech/text patterns of the wealthiest humans. And she's right. Last thing we need is an imperialist, anti-communist AIs.


Also, people don't act on the internet like they act in real life.


is it tho? considering social network is modern equivalent to crack cocaine psyop, I'd think researchers would have access to rather diverse data including poorer part of population no?

I might actually read her paper if it's openaccess


File: 1608526405035.pdf (1.19 MB, sp (1).pdf)

Apparently this is it? I can't find the actually published version. (If it was published yet.)


>Do this or I quit
>We accept your resignation
How the fuck was she fired?

File: 1608526389387.jpg (52.88 KB, 487x503, 7644392049ced531c00976a892….jpg)


how come it seems like no progress has been made? we keep hearing about research but it seems like we never get to see any results. what gives?
what even is quantum computing anyway?
5 posts and 1 image reply omitted. Click reply to view.


>superposition, or the ability to operate on many qubits at the same time instead of consecutively on bits as it is in computers
so what does that mean from a programming point of view? in what specific way does that effect the time complexity of an algorithm?


Imagine a rat trying to find a way out of a maze. The classical way, the rat has to go all the way down a path, and only after hitting a dead end can it then try checking another path. A quantum rat can go down every path simultaneously, solving the maze much faster



>>5894 is a pretty good explanation tbh. it's also why right now google, ibm, et al. are measuring their successes in building a quantum computer by its quantum volume, a number that takes into account both amount of qubits and its effective error rate, because every qubit that is added and can operate relatively error free theoretically increases the capability to operate on data exponentially by allowing larger data structures. then all you need is an actual algorithm that can take advantage of this capability in a way that makes it more efficient than classical computers.

as >>5894 said pathfinding is one example of a problem that might benefit from this, so in complexity terms if we take A* (which worst case can get polynomial cost in bounded tree) and compare it to say a quantum binary tree search then, assuming there are enough qubits with a negligible error rate, you would get at least an exponential speed up.

pathfinding of course is fairly unique here because its memory bounded first and foremost, and every attempt at a quantum pathfinding algorithm i know of currently uses a qubit per pathing decision for example, so on large enough paths this really does exhaust capabilities of current quantum computers if you take into account that the "best" quantum computers (google, ibm) at the moment have around 50 qubits. ibm is "planning for a 1000 qubit computer by 2023" but right now this is just big words.

but to generalise it: if you are working on an algorithm that requires finding the correct state in a large state space, then quantum will in general offer (usually) an exponential kind of speed up there. and this is before we get to unique solutions that use entanglement like https://en.wikipedia.org/wiki/Shor%27s_algorithm


Can't wait for quantum computers to decrypt all logged tor traffic that passed through CIA-owned nodes.


Haskell and Ocaml are the only programming languages I use.

File: 1608526401023.jpeg (78.27 KB, 800x600, image.jpeg)


>hides your face from coppers
>no more facial recognition


white people hate wearimg masks, they'll get rid of them soon when the vaccines come out


True, this is eternally based


File: 1608526403582.jpg (27.54 KB, 570x570, il_570xN.2249988798_l0ga.jpg)


I'm glad I no longer get weird looks from visitors


The pandemic happened a month after my country approved a law that was made to stop protesters from covering their faces. Pretty funny.

File: 1608526323426.jpg (33.06 KB, 444x250, home.jpg)


Do we need a leftist Luke Smith?
11 posts omitted. Click reply to view.


hahahahahah the absolute state of leftypol


he just does whatever /g/ thinks is cool. nothing about him is intelligent or groundbreaking.


Go listen to some of the podcasts he's done, he's some kind of Tradchrist-monarchist freak living on land he bought using a gift from a rich family member and memeing about technology he doesn't really understand (I'll make my dotfiles reference the regular user account as SU so I never have to enter a password, this is a good idea lmao) to idiots that think being cool is hipster software + edgy jokes

Literally a soyboi living off of daddy's money and youtube bucks larping as ted, could you be more pathetic?


Potentially. What does the left Luke Smith (LLS) do differently?


also flirting with antisemitism

File: 1608526396657.jpg (48.79 KB, 359x363, 1597826941514.jpg)


>There are 160 apps on my phone. What they’re actually doing, I don’t know. But I decided to find out.


The worst thing about phones is that the only way to win is to not play at all.




>There are 160 apps on my phone


Closest I've gotten is LineageOS without gapps

File: 1608526384506.jpg (35.01 KB, 460x345, 1606371923237.jpg)


Amazon is using union-busting Pinkerton spies to track warehouse workers and labor movements at the company, according to a new report

Holy lol.
1 post omitted. Click reply to view.




well well well, its cowboy time


HN is such a disgusting community.
>Oh, guess I'll stop buying from Amazon now
>Umm it's voluntary exchange


For a place called Hackernews they're filled with lolbertarians with a cucked mentality.


What's alternative to it?

File: 1608526394982.jpg (22 KB, 500x331, Vulpes_vulpes_laying_in_sn….jpg)


Mozilla shoots itself in the foot again to please its overlords at Google.

>I came across this video recently, in which Mozilla add-on developers talk about the implementation of the new "DeclarativeNetRequest" feature that Google intends to use in order to replace the current "block WebRequest" used in Manifest V2 by pretty much any content blocker worth of it's name. The video is only 3 weeks old: https://www.youtube.com/watch?v=tpDFS-GUytg

>To summarize the important bit of the video, Mozilla is planning to adopt the new functionality without replacing the current one, essentially to ensure compatibility between add-ons in Chrome and Firefox while also maintaining the content blocker functionality.
>However they are very careful to always say, wether in this video or in the corresponding blog posts, that they will support blocking WebRequest "for now", or "we don't have any IMMEDIATE plans to remove content blocking capabilities" which concerns me, because they are definitely going to implement DNR in the future, even if only at browser start up at first.
>The fact that they don't give us a definitive answer regarding content blocking features is a big concern. Surfing the web without UBO is…unthinkable at this point.
4 posts and 1 image reply omitted. Click reply to view.


Opera GX gang represent


Firefox is controlled opposition.


File: 1608526397822.png (66.35 KB, 1083x410, browser-user-share.png)

Just imagine if they had done this shit while they were the underdog to Internet Explorer.


they havent done anything yet


False, they've done A LOT of dumb shit over the years.

File: 1608526391752.jpg (36.73 KB, 720x551, mousecleaning.jpg)




Xenocara is pretty nice, no need to run as root.



File: 1608526391252.jpg (211.83 KB, 1080x897, 1606859999825.jpg)


3 posts and 2 image replies omitted. Click reply to view.


Also, the patent submission date is pure schizo magnet.


File: 1608526396135-0.png (366.88 KB, 656x678, IMG_2531.PNG)

File: 1608526396135-1.png (478.61 KB, 667x752, IMG_2530.PNG)

I love technology


Real life tends to be worse than fiction.

>Personality Traits Can Now Be Predicted Using Smartphone Data




See also >>5933


The headline made me think they could do all that just by you carrying a smartphone but it's an analysis of heavy smartphone use. Not that that's much better.

Delete Post [ ]
[ overboard / sfw / alt / cytube] [ leftypol / b / WRK / hobby / tech / edu / ga / ent / 777 / posad / i / a / R9K / dead ] [ meta ]
[ 1 / 2 / 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 / 13 / 14 / 15 / 16 / 17 / 18 / 19 / 20 / 21 / 22 / 23 / 24 / 25 / 26 / 27 / 28 / 29 / 30 / 31 / 32 / 33 / 34 / 35 ]
| Catalog | Home