[ 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: 1608526417636.jpg ( 112.17 KB , 1130x678 , singularity.jpg )


Is a singularity even plausible? *asking leftists*

Like my personal opinion no. because if we enter a singularity what's to say that we didn't have the intelligence to create our own information I:E if we manage to create bots smarter than us then why would we let them fuck up the world. we would probably use them to teach us what they know and then burn them once we have the information.

humans know bots don't have feelings feeling and consciousness are only experienced by living things bot can only retain information and use that information to make more information then we grab that information. That's what we do now with calculators after all

what do you think?


Well of course it's possible, just not on the near future. A better question is if it is desirable. Also keep in mind humans have a tendency towards self-destruction, just look at what's happened in the latest centuries.

And who's to say a sufficiently intelligent AI can't develop feelings of their own? They'll most probably be completely alien to us, though.


If the Singularity means an artificial intelligence that improves it self faster and faster, that is not possible, because it violates the law of diminishing returns. If it just means that humanity could build vast artificial intelligences that run most of society like in the Ian Banks Culture novels that is possible.

File: 1608526402175.jpg ( 67.95 KB , 800x420 , 34234-3.jpg )


I can understand why one would use python for practical purposes : it is the most 'battery included', 'scripting' programming language that requires least amount of fuss to get the work done 'initially'. And yes lots of important scientific research framework depends on python so you don't have much choices on few domain specific problems.

actually that's a spook. I am willing to bet a lot of money even simple boiler plate web applications developers would have wished they are writing in java the moment exceptions start flying.

also those salty researchers who just want to stick it up to their programmer friends with clever jupyter notebook demos would be sorry when they realize numpy could have been much more usable and less human error prone should they chose a language with, you know? BLOODY TYPE AWARE SYSTEM that can allow computer program to deduce which operation feeds what operands without mere mortals fumbling through yet another stack overflow question or never ending reference manual pages???

like seriously, I would furrackign volunteer to port legacy app I am tasked to expand to BCHS or even bloody perl script or fucking racket scgi on my off time than torture myself with this bullshit language of jokes

python & its perpetrators should be gulaged for sabotaging type system and wasting workers time while inflating exploitation rate for bourgeois class
12 posts and 2 image replies omitted. Click reply to view.


Scientists/cybsec guys/data science guys whatever dont write big software projects, they use tools to solve immediate problems and python is great for that because its concise and simple to write. Also perl you mention in first post as alternative is much less maintanable and error prone


B-But Python is popular and I want to feel hip and special about the programming languages I use!!


Then learn lisp or Haskell, python is too pragmatic to be hip


may I ask which country you are wage cucking at? I would be surprised if you did not see all the entry level python job openings nearby


File: 1608526415703.png ( 31.41 KB , 200x155 , R.png )

R-using scientist here, I honestly don't see the point to Python. When you really need efficient routines that are going to be massively iterated you simply write them in C++. In fact all the heavy duty R stuff is coded in C++ on the back end.

File: 1608525985240.jpg ( 64.31 KB , 1000x615 , GHOST-GUNNER-2-CNC-Gun-Mil….jpg )


Is it something practicle for insurgent groups or is it a giant meme?
10 posts and 3 image replies omitted. Click reply to view.


This depends on the design. Guns need to deal with some variation in the amount of pressure each cartridge develops. A design meant for small-scale manufacture should have enough tolerances and maybe a way to tweak spring pressures to allow for some variety in materials. I saw one gun design a while back that specific exactly how to heat treat the steel for the springs inside it, to get it just right.




File: 1608526386754-0.jpg ( 356.51 KB , 1500x1000 , DiCenzo_WIRED_ISIS_Weapons….jpg )

File: 1608526386754-1.pdf ( 1.35 MB , Frontline_Perspectives_Ins….pdf )

File: 1608526386754-2.pdf ( 1.29 MB , Islamic-State-Recoilless-L….pdf )

not a meme. plug n play heavy weapons are the future.


Interesting. I did a bit of amateur rocketry years ago, this is way more sophisticated. Small arms will only get you so far.


These file dumps are super interesting. Really puts into perspective how simple things can be made.

I ran across a youtube video some years back where a guy hand made a rifling button and the press to push it though a home made barrel blank.

As somebody with a nicely equipped shop, I think it is fair to ask the question: at what point do you become the man?

File: 1608526405307.jpeg ( 141.74 KB , 640x634 , 3EBEE9D6-B75B-4757-A771-0….jpeg )


Any of you guys do this? I set up BOINC on some old laptop I had and started using it to search for pulsars and shit. Also, [email protected] became the most powerful “supercomputer”.
2 posts omitted. Click reply to view.


i mean yeah seti was stupid, but still, distributed computing makes it easy for scientists to get computing power quickly and cheaply, it was able to expand massively with COVID and became the most powerful computer in the world as i said, its very flexible.


Actually wasn't much of a waste of electricity since there was no low-power modes or thermal throttling back in the day. Better to use idle cycles for something than nothing.

Now we have Bitcoin, which is a scale of wastage worse than everything before it combined.


>Better to use idle cycles for something
Completely unrelated but I hate retards that keep their computers turned on all day and night just for nerd cred, not even running a home server or whatever.


who the fuck just leaves it on?


A lot of people on technology imageboards who want to seem cool showing a high uptime on their rice or something

File: 1608526393615.jpg ( 146.24 KB , 565x730 , 1606876145673.jpg )


Do you use a CRT monitor?
4 posts omitted. Click reply to view.


I used to. i Had two CRT monitors, one was a 19' one for small stuff and the other is a 24' widescreen trinitron CRT, that i never got to working properly. The screen would drop like a water fall, maybe i got ripped off, or a capacitor needs changing. CRTs are nice if you have the space.


i don't play smash


Also they can quite literally explode.


Remember to sit 3 meters away from the screen, it's good for your health.

File: 1608526407258.png ( 225.21 KB , 560x315 , sybil.png )


Japan Will Fund AI Technology to Pair Singles Based on 'Emotional Quotient'

This is Silicon Valley tier, in that it doesn't actually address the real issue.


they’re one step away from state appointed gfs


And I thought hookup apps were dehumanizing enough.


File: 1608526408466.jpg ( 181.99 KB , 720x712 , mikucccp.jpg )

if this leads to personalized vocaloid gfs running GNU/Linux CRITICAL SUPPORT


you really saw japan in the headline and had to post this shit huh


Based on the description this is not even an expert system just some weak ass similarity metric?

The stated goal is to make people breed, not to make them happy.

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.

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 / 36 ]
| Catalog | Home