/tech/ - Technology

101010

Posting mode: Reply

Check to confirm you're not a robot
Name
Email
Subject
Comment
Password
Drawing x size canvas
File(s)

Remember to follow the rules

Max file size: 350.00 MB

Max files: 5

Max message length: 4096

Manage Board | Moderate Thread

Return | Catalog | Bottom

Expand All Images


Random Thoughts Anonymous 03/05/2018 (Mon) 01:21:31 [Preview] No. 12441
It's a shame how inactive this board is. It seems only to be that way because there's about three other chans consolidating the userbase. In the interest of promoting activity, I'll try to make this thread a thread about technology related musings I have (not that I'm important or anything) that could hopefully be used as starting points for discussion.


Anonymous 03/05/2018 (Mon) 01:27:05 [Preview] No.12442 del
I have a bunch of things on my mind right now, some of which I'd like to keep private. But some topics include a new GNU/Linux distro, a new operating system made especially for RISC-V.

This new distro should be easy to use and configure. One of the flaws of current distros is either that they have concerns in their design (Ubuntu, Fedora, Debian, either being able to brick systems or having sucksomedicks), or being arcane and difficult to learn (Gentoo, and so on). This state of affairs makes it low motivation for Windows and Mac users to switch. So I think a new distro would give a lot of the control to the user while helping them make as informed a decision as possible.

The new operating system is necessary because GNU/Linux is very old and was made for a different time. What is needed is a system for the future, on hardware that is not inherently insecure, such as x86.


Anonymous 03/05/2018 (Mon) 14:36:31 [Preview] No.12443 del
>>12442
>>12441
I'll contribute a little.

I dislike most modern GNU/Linux distributions due to how bloated they are regarding RAM usage, my operating system shouldn't be using three-hundred megabytes of RAM when I barely turn on my computer, and more often than not, it is the graphical desktop (XFCE, GNOME, Etc.) that is responsible for this.

I could do away with graphical desktops completely and utilize my computer with only a command terminal and cut the ram usage down significantly. I don't care that my computer has eight gigabytes of RAM installed, it doesn't need to be this bloated.


Anonymous 03/05/2018 (Mon) 16:00:57 [Preview] No.12444 del
>It's a shame how inactive this board is.
Well this is to be expected. There is no reason for people to post here when content is exactly the same as on 8/tech/. We've tried improving things in meta threads, but nothing changed. There are still people posting merchants and conspiracy theories, saying everything is botnet woo hoo we're done woo they are behind this woo hoo hoo... I just don't have enough energy left to deal with this, so I just don't post here anymore. Occasionally, once per week, I come here to check if there is less bullshit, but recent update of board mascot made me visit even less often. Imagine you tell someone you know about this board and first thing they see is Terry. They naturally ask you about who this is and you say he is some programmer from the internet with schizophrenia, who likes to use the word "nigger" in every second sentence. Friend politely declines your advice to visit this site. And then people complain that there are no people posting here...


Anonymous 03/05/2018 (Mon) 21:01:26 [Preview] No.12445 del
>>12443
>I dislike most modern GNU/Linux distributions due to how bloated they are regarding RAM usage, my operating system shouldn't be using three-hundred megabytes of RAM when I barely turn on my computer, and more often than not, it is the graphical desktop (XFCE, GNOME, Etc.) that is responsible for this.
I see this as a symptom of lack of programming skill these days. GNOME has been downhill, as newer coders who don't know what they're doing step into vacated roles. The people who would be competent at this sort of thing are probably either in full time jobs or playing video games instead.

Fortunately, in the case of desktop environments, we do have, for now, some viable alternatives (that don't require using a window manager alone): LXQt, for example, is approaching a state of usability. I hear KDE is becoming less and less bloated (defined as memory usage), but I have no experience there. However, here, too, things could go downhill fast. Imagine what people would do if IceCat nor Waterfox existed after the transition to Firefox 57.

>I could do away with graphical desktops completely and utilize my computer with only a command terminal and cut the ram usage down significantly. I don't care that my computer has eight gigabytes of RAM installed, it doesn't need to be this bloated.
Yes, but you shouldn't have to sacrifice modern convenience to avoid this crud that shouldn't be there. At the moment, it can't be helped that there's a shortage of good programmers contributing to GNU/Linux---the languages used for this purpose are old, and complicated, so few youth now may be able to see the point. But there are a few solutions already coded and available for usage---it's just that it's really difficult to find them and implement them into your personal distribution. Distributions these days would have you believe that the only DE options are GNOME, KDE, or a TUI. But that's not true, and there are explanations about how you can do otherwise, but those aren't brought to the fore. A modern distro needs to teach people their options and empower them to make informed decisions, because not everyone will be comfortable sticking to terminals like you.


Anonymous 03/05/2018 (Mon) 21:10:58 [Preview] No.12446 del
(22.68 KB 175x216 avatar10757_2.gif.jpg)
>>12444
I can see we're you're coming from with Terry as mascot. I personally disagree with sticking with a figure like RMS, but there's no doubt that he's easier to tolerate for the less abnormal potential users.

I guess I can see how /pol/ crossover would be a repellent. On /g/, there's crossover, but it's the most convenient board. On 8/tech/, there appears to be much less crossover. So the two niches are filled, and that this site has fewer privacy concerns doesn't appear to be compelling enough for people to invest in this place.

I'm curious what you think this board needs to do differently in order for it to be more successful. Surely it can't just be a clone of some other board, even minus their flaws? It seems to me that the userbase just needs to contribute valuable enough information for this place to be a source worth checking. I'm only gradually beginning to gain more expertise that would make my posts meaningful.


Anonymous 03/06/2018 (Tue) 20:49:13 [Preview] No.12447 del
>>12446
Do you really want this place to become a user curated tech news aggregate? Then repost some news that you've never seen people on tech news sites that I like to read unironically. https://www.techdirt.com/blog/ http://techrights.org/?stories Just don't go crazy with the news reposts, one very good news article every 36 hours is good enough for this slow board.


Anonymous 03/07/2018 (Wed) 04:51:02 [Preview] No.12451 del
>>12441
>It's a shame how inactive this board is.
Not really. The userbase here got the board it deserved.

The board certainly had potential. The large file size could have been a great opportunity to share interesting /tech/ related videos and other files, for example. And being able to post files through Tor is great, since that functionality isn't available on 8/tech/.

Unfortunately, most of the few users here were not capable of creating good content for the board and, worse, were incapable of responding intelligently to good content when it was posted. Garbage in, garbage out. The userbase was garbage, so the board became garbage, and the couple of posters that were worth a damn left for greener, less retarded pastures.


Anonymous 03/07/2018 (Wed) 04:55:59 [Preview] No.12452 del
My GUI progression. It's been so long some of the names and details may be a bit off.

Started out on Redhat 3.0.3. Gnome or KVM were the two options being pushed. Tried both, went with Gnome.

After some upgrades, Gnome was transformed into some monstrous software system with its own installer. Tried it. It trashed my system on install. Today, I could have sorted it out on my own without so much trauma. Back then, I had to reinstall from scratch.

Tried KVM again. Holy fuck no.

You know what I really want? Something like the Windows for Workgroups desktop!

Boss: "THAT BETTER NOT BE FUCKING LINUX I SEE ON YOUR PC! Get rid of that shit."

Hello FVWM95. Thank you, God!

Boss: "THAT BETTER NOT BE oh. Carry on."

Later on I had the freedom to do whatever I wanted how I wanted, as long as shit got done. I tried out just about every DE, GUI, WM, whatever-the-fuck under the sun.

Gnome? KVM? Oh hell no.

Various others: nah.

Blackbox? Blackbox. End of line.

Later on I was made responsible for creating a public use internet kiosk. I went with Blackbox. After a major point upgrade, I discovered BB was dead in the water and was going to be dropped from the repository. Well, fuck. Back to trying out everything under the sun and ...

Fluxbox? Fluxbox. This works for the kiosk as well as my personal stuff. Whoot!

I've been sitting on the same style of setup for about a decade now. Fluxbox + fbpanel + idesk + xscreensaver. I like slim if I need a multi-user login manager. Otherwise, it's login to the command line, and startx.

I'm not trying to be coy, nor trollish. I'll grant that a graphical login manager is a must have. Also, idesk is beyond bizarre, and totally unmanageable as a desktop icon system. There are better replacements for idesk, but it does cut down desktop clutter like nothing else when you have to text-edit each shortcut you want. A good thing over the long run. Otherwise, I simply do not understand what people are looking for from a DE/GUI these days.


Anonymous 03/07/2018 (Wed) 23:18:00 [Preview] No.12453 del
>>12447
>Do you really want this place to become a user curated tech news aggregate?
Not at all what I had in mind.
What would be more interesting to me would be, frankly, along the lines of every thread being a guide. Genuine information exchange, rather than just sharing opinions. But that's just me, and I'm not at that point yet where I could be a valuable user in that kind of environment.

It does raise the question of what a chan should be, if not just anonymous Reddit.

>Unfortunately, most of the few users here were not capable of creating good content for the board and, worse, were incapable of responding intelligently to good content when it was posted.
I can see that.

>and the couple of posters that were worth a damn left for greener, less retarded pastures.
if only such places existed


Anonymous 03/07/2018 (Wed) 23:34:28 [Preview] No.12454 del
>>12452
I have never used GNOME, which seems to me to be a smart move on my part, but I have heard that it's bad and getting worse. I don't hear the same sorts of complaints with the other popular desktop environments like KDE and XFCE.

>Otherwise, I simply do not understand what people are looking for from a DE/GUI these days.
It makes transitioning from Windows and MacOS easier. I'd reckon most people who grew up with either have only seen the command line in heart-stopping moments, if at all. People also have different needs and varying levels of interest in personalization. Someone who wants something that just werks would likely pick KDE. Someone who wants their DE to provide nothing but the bare essentials and is willing to customize their experience might choose LXDE/LXQt.

What I'm saying is that people like yourself who have used Linux for a very long time should understand that Linux is overwhelming in many ways. These people say, "Just install Ubuntu," but then, those people get stuck with systemd, GNOME 3, and even may get their computer bricked because of a botched update. These people say the command line is not hard, but don't take into consideration that people only have a limited amount of time to work with; if they work 40 hours a week, the last thing they want is to give themselves headaches trying to learn something they shouldn't have to and didn't need to learn. These people who just call these migrants retards rather than seeing where they're coming from are the people who keep Linux from being widely adopted on the desktop---because their attitude is a brusque "Use it or move it."

It doesn't seem like it would be hard to create a distribution that teaches the user how to be free painlessly---it's just some current users have this self-righteous attitude that keeps them from acknowledging these complaints as flaws.


Anonymous 03/10/2018 (Sat) 09:42:25 [Preview] No.12456 del
(4.70 KB 250x247 1519070293541.jpg)
Last 24 hours looking at the Forth programming language. /g/ had a shill thread. I wanted to see if it could be an alternative systems programming language to C.

Forth has no obvious open examples on the internet of large-scale programs. Everything I saw written in it was just an isolated page of code, used as example or in comparison to other languages. Its paradigm is stack-based programming, which x86 largely does not use---it uses registers instead. Forth was being described as using "host OSs" which I gathered meant that it relies largely on its own system and does not work for other systems.

I can't say whether it is or is not a meme, having not programmed in it, but it's not obvious that it suits my interest of operating system programming, and it's concerning that, as has been pointed out, advocates of the language cannot or do not point to actual code to prove the viability of the program---they just say what has been claimed as Forth programs, or say that it will "change how you think about programming."


Anonymous 03/12/2018 (Mon) 19:14:13 [Preview] No.12468 del
the .xyz is down, that's a big hit on traffic for an otherwise slow but steady board.


Anonymous 03/12/2018 (Mon) 21:37:49 [Preview] No.12470 del
>>12468
You should be using the hidden service.
Also, post this kind of stuff on the meta-thread, please.


Anonymous 03/12/2018 (Mon) 22:34:34 [Preview] No.12472 del
>>12468
it would be a shame if people hadn't noted the alternative .net.


Anonymous 03/18/2018 (Sun) 04:38:30 [Preview] No.12492 del
So I had this plan. Go into technology as an entrepreneur, innovate, and make bank as well as the future.
I realized that it would not work in the US. Patent trolls accumulated all the ideas and sue anyone who seemingly comes into conflict with them. Only large and established companies like Microsoft and Google can stave them off and get things done.
But you can't work for those companies because not only will you never get into a position of real decision making, but your position will constantly be in danger due to hiring politics and economics. Of course, that's why entrepreneurship is the answer in the first place.

So the only rational choices are to stay put and do nothing, bow to venture capitalists and risk the integrity of your work, or leave for brighter places. I'm leaning toward the last.

Does anyone have an idea of places in the world with burgeoning technology sectors? Besides the usual suspects, and India, that is.


Anonymous 03/18/2018 (Sun) 05:27:51 [Preview] No.12493 del
>>12492
Seems like a wet dream than a plan. Work for a small company first. You would need some certificates and sometimes you would also need more than that like a degree. Once you've got more hands on experience, then move on to already existing small tech companies or make your own or work alone. If you're going to make something "new" then avoid China, India (still a horrible bureaucracy), US, and all of Europe under the EPO. Japan might be okay, their antitrust laws are sane, but you need to have lots of credentials to get a job there so it's not really worth it. Where's most of the money at? Webdev. Real boring but practical necessity for any company.


Anonymous 03/18/2018 (Sun) 10:04:30 [Preview] No.12494 del
>>12493
US is a sinking ship. I don't have anything close to a tech degree, and even CS graduates aren't guaranteed a job. There's no way I am getting a job soon, and there's no way I am going to advance in the career in a speedy amount of time, which I feel is necessary; nor would I be advancing in a reasonable amount of time, given the large amount of workforce that is already in the market, vying for promotion.

The US makes it very difficult for the not well-off and not well-connected to advance, and I am both. There's virtually no opportunity for me if I stay here.

I won't discuss my plans I have right now, but I feel like there's a way out for me if I just focus on accumulating enough wealth here first, in whatever way I can, then seek opportunity elsewhere.

Blogpost, sorry. My main point is that only full-grown adults have a shot at successful tech entrepreneurship in the US. Everyone else must wait decades for their chance, or else go where there are less restrictions on their activity and less barrier for entry. I don't think we have that much time left.


Anonymous 03/18/2018 (Sun) 13:13:10 [Preview] No.12495 del
>>12443
If you are interested; try this out:
https://github.com/martanne/dvtm
with https://github.com/martanne/abduco
It is a tiling window manager for the console.


Anonymous 03/20/2018 (Tue) 15:00:34 [Preview] No.12502 del
>>12494
As much as I hate to say it, many Western nations face the same dilemmas. And (most) Eastern nations are not too much better either. If you get a tech job expect it to be highly controlled by greedy vultures and/or government. Developing nations are where it is at I hear, because many are booming with new industries and are currently more old-fashion "Americana" than we are today as we face economic and social decline (similar to the USSR).

Back to tech skills: if you have those skills you could become a hero if you help invent some kind of new decentralized web-friendly platform where users can share content and information without being censored or controlled by authorities. With the FCC removing net neutrality and the US Congress now trying to remove CDA 230 protections we are about to be FUCKED BADLY (this includes image boards). We need more heros to provide their skills to create something we can use to bypass the control grid surrounding us.


Anonymous 03/21/2018 (Wed) 14:13:41 [Preview] No.12506 del
>>12502
Removing net neutrality isn't a bad thing


Anonymous 03/22/2018 (Thu) 05:28:24 [Preview] No.12510 del
>>12502
South America to me seems to be the best bet, aside from most of their governments being socialist andor backwards.

As for the technologies, I think what most people want already exists---like the distributed file sharing and other networks. The issue is that these rely on tech-literacy that most people lack. For example, people ask what the alternative is to Google. The answer is YaCy. The problem is that it takes a lot of memory to use on your main system, so it's better to have a server for it, which most people have no idea about. All of these things essentially require people to run their own servers to contribute.

What I'm more interested myself is the future of computing. Such as, figuring out finally a replacement for the C programming language, a new operating system, better general artificial intelligence systems, etc.

Once I get money and more knowledge, though, I definitely plan on contributing to the already existing tech.


Anonymous 04/08/2018 (Sun) 01:30:50 [Preview] No.12556 del
Haven't abandoned thread. Just have been learning.

I am currently pursuing Scheme. I like the LISP syntax so wanted to learn the language. Decided on Scheme as opposed to Common Lisp because the latter cares about Object-Orientation which is a dumb paradigm. Scheme focuses on the essentials.

That got me into the multiplicity of implementations, as well as the structure of Lambda Calculus itself. I still don't get that fully, but I lack much math background. I need to learn more. As for implementations, I've settled on Gambit and Guile for programming and extension, respectively. MIT/GNU Scheme is lesser than Gambit as far as I can see. I would also look into the Scheme Shell if it weren't tied to 32-bit computing.


Anonymous 04/08/2018 (Sun) 10:42:52 [Preview] No.12559 del
>>12556
How are you learning Scheme?
I've recently been reading SICP, which I feel is a fairly good introduction to functional programming, but I'm aware there are other sources (such as Little Schemer and the following books) that are better for learning the language itself.


Anonymous 04/08/2018 (Sun) 11:23:19 [Preview] No.12562 del
(3.51 MB 3840x1080 1521558228738.jpg)
>>12559
I ran into the problem of choosing an implementation and deciding whether or not I cared about whichever one I chose having support for Revised Report on Scheme versions 6 and 7. There was no obvious way beside starting from R5RS and reading afterward whether the extensions in the latter versions were justified in my view.

I have SICP on my drive ready to read, but I decided not to begin with that first because 1) It's a text book; 2) It assumes use of MIT/GNU Scheme; 3) It's a meme, whether or not it's good. I will read it likely after I read the actual Scheme standards and the particular Scheme specifications that I end up settling on. I'm on the 4th chapter of R5RS, only started reading yesterday. It's a short read, so it can't be anything more than 2 hours left.

After that, I'm going to look at actual programs written in the language and try and comprehend what's going on.


Anonymous 04/10/2018 (Tue) 07:31:58 [Preview] No.12577 del
Looking into optical computers. If we want personal AI, or AI in a human formfactor, silicon is almost certainly not feasible. The technology is some years away from the early adopter phase, I'd say mid to late 2019 at the earliest, but I know jack squat, and it's difficult to find up to date information on this subject.

Anyway, what this means is that it's mostly pointless to develop a replacement to GNU/Linux at this point in time. G/L will need to be ported to RISC-V fast and get some decent development for the system there, because this will be a slight window to get the average person, not just in the highly developed Western countries, but potentially across the world, into GNU/Linux. Tech companies are already investing heavily into architecture, but no one but Google with Fuschia OS has anything already developed to leverage the technology. It's inconceivable that not one of those heavy hitters is thinking of the consumer personal computer market in addition to whatever else they see potential in RISC-V. With Microsoft dissolving the Windows division, it's unlikely that there will be a port of Windows to RISC-V. Which, all of this means, is that companies are going to need to rely on UNIX and UNIX-likes to push out market-ready desktop operating systems.

Of course there's still this stigma surrounding Linux that it's not functional for everyday just werks desktop computing tasks. If I'm correct about the above, then I think we'll either see investment in the already existing corporate distributions like Ubuntu or Red Hat, or each company is going to hack their own distribution together---they might even collaborate and come up with a standard. The benefit of this move would be that the companies could more easily acquire support from device manufacturers for drivers and software developers for ports (such as games or heavy-duty applications for professionals like CADs). Yet the state of Linux isn't looking too good, given that something as basic as the Desktop Environment is still buggy and fractured. All of this is pure speculation, so dwelling on it is not very productive. What is important is that GNU/Linux will be here to stay, but it won't be able to extend itself much further unless this opportunity is seized.

Meanwhile, we have the wait for optical computers, which will be magnitudes faster than electrical computers, and so more powerful/efficient/cost-effective. As soon as this comes out, developers should be rushing to get the foundation down for the next generation of "personal" computers.


Image Boards Anonymous 04/13/2018 (Fri) 06:44:24 [Preview] No.12610 del
(103.66 KB 1067x1600 uli.jpg)
(1.00 MB 2592x1944 red room 2.jpg)
I read a thread asking what would be next after imageboards if they no longer have a place in the future, on 8chan/tech/ I think. I''ve been thinking about that question.

4chan is somewhat primitive in comparison to the newer imitations. 8chan and Endchan and the like allow for not only multiple file uploads per post, but also multiple file types, beyond merely images and videos. At that point, you have a filesharing platform that is only an "imageboard" in legacy. From there, we can abstract to the nature of a 'chan'.

What features make a chan? The list is surprisingly long if one gets fine-grained with the details.

>file attachments
>anonymized by default
>topic/theme segregation
>centralized hosting and moderation
>ephemerality of hosting
>lack of profiles and microblogs excluding this thread
>linear, chronological ordering of posts

And on you could go. Compare this to Reddit, and ask yourself whether you would be looking at a regression in either direction, or if they are both valid solutions to different problem sets, with their own separate lineages of progression and degeneration. I lean toward the latter stance; that position reframes the original question, "What will replace imageboards?". It entails the question, "What are chans for?". And that is a very big and complicated topic of debate.

But for the sake of this conversation, how about the answer of "Quick, no strings attached, discussion and file sharing"? That necessitates profiles not being a requirement, having the ability to share many and multiple types of files at once. Reducing user liability for what is posted is where wanting to increase anonymization comes in. Ephemerality of content, some say, is an essential feature of chans, but to me it seems to be more of a practical matter of how much data really can the host afford to keep. So while there is a baseline of data that needs to be kept for the site to be functional, there doesn't seem to be a limit or maximum amount of data that can be hosted.

What would people do with a chan? Discuss and share files. Drop in, drop out, whenever one feels like. Meetups maybe, even some regular socializing would be possible---unless strict anonymity was enforced (anti-contactfagging and anti-faceposting). But at the end of the day, it would be about saying and sharing what one wants without fear of reprisal. From a user standpoint, it's also a source of thought, ideas, and digital content. People also treat chans like Reddit, with a heavy emphasis on linking and responses to outside content, but that is more a feature that is possible and not inherent to its nature. And of course, the chans are also sources of original content.

With all of this in mind, it doesn't seem like there would need to be too much evolution of chans right now to produce better platforms. Going off the last point in the previous paragraph, there's room to include collaborative content creation, like an online GIMP like Google Docs or something for making memes. Chan hosts would need to encrypt user identifiers to the extent that even they wouldn't know who is who, if total anonymity became an ideal. Embedding linked websites right into the thread is also a wild idea, but again, would be more suited to Reddit-likes. Maybe there's something to do about the content segregation---how boards, and even threads, can keep users away from the website in its totality. Would the trend be toward broader board topics, like a single technology board andor science board, or toward more specific topics, like programming, security, hardware, and so on? This, I do not know.

Anyway, I'm exhausted talking about this for now, but would like to hear thoughts if people have them. I likely will expand later.


Distributed Sites Anonymous 04/13/2018 (Fri) 07:05:00 [Preview] No.12611 del
(974.46 KB 1200x900 red room 3.jpg)
(2.30 MB 3000x2250 red room.jpg)
On a different, but related note. I was thinking about "website" models: centralized, decentralized, and distributed. A lot of people want distributed sites, but they have a few problems:
>hosting illegal or immoral content
>incentivizing hosting
>reliability of hosting, mostly from data storage standpoint, but also from a privacy standpoint

Some of these problems are easier to address than others. For one, privacy can be enforced in the software specification: maybe it encrypts all personal identification before the user connects to services, so no one can know who the user "really" is. Easy said at least. As for data storage, as long as one host is up, assuming as much data as possible has been synchronized between hosts, then all legitimate data will be hosted.

What about illegal content? I think the software can allow for hosts to maintain their own blacklists of users. This would require that the software not implement random identifiers for users each time they connect, but allow for users to be tracked (this would nevertheless not unveil their anonymity: it would just associate the user's activity with their single anonymized identifier, like an encrypted IP address in theory). Hosts could ban/blacklist users so that they would no longer automatically host content that they post; then there's also filtering, which would prevent them from seeing posts from certain users at all---but those are two distinct features, though users might want to avail themselves of both rather than one or the other at a time.

The last problem I thought of was incentivizing hosting, rather than punishing non-hosting. If it's a distributed system, it could force each user to be a host for content, but that load could get really heavy really quick, so it's better not to do that, lest someone's personal computer get stalled by the demand. So how to get users to opt-in by contributing their own server? That's what I was looking up before I made my posts tonight, and figured it would be good material for conversation.

Does anyone have any thoughts on this: how do you get people to willingly host content in a distributed system? I have heard that private tracker torrenting sites have a ratio system or something; and then there's some sites that have taken to giving out their own cryptocurrency to contributors. Any other methods for making this work?


Optical Computing vs Quantum Computing Anonymous 04/16/2018 (Mon) 02:34:21 [Preview] No.12648 del
LIGHT: How optical processing can solve some of the world’s most complex problems.
https://www.labs.hpe.com/next-next/light
>Beausoleil spent about a decade working on quantum computing. That’s when he experimented with chips made out of diamonds. His team found that it was too difficult to create enough diamond chips that had exactly the same qualities, making it impossible to manufacture them in a repeatable and predictable way. This particular approach to building a quantum computer wouldn’t scale.
>Beausoleil concluded that photonic technology offered a faster and more practical route to success than building an actual quantum computer. “When somebody does manage to create a true quantum system with entanglement, that’s going to be awesome,” Beausoleil says. “Right now, we’re leaving that one on the table and just trying to take advantage of coherence.”
>[...]
>Optical computing is an emerging field with experimental components. The challenge is to raise the yield of working components to a commercially viable level. The Labs approach is to create a round of equipment, test it to identify flaws, then determine necessary changes in the chip design or manufacturing method.
>[...]
>Labs researchers are also exploring new applications for their optical circuits. For example, they’ve designed a system called an energy minimization computer that changes the state of an optical circuit to find the configuration that consumes the least amount of energy. This concept applies perfectly to solving NP-hard problems like the traveling salesman.
>[...]
>This technology won’t replace general-purpose electronic computers, because lots of problems aren’t NP-hard problems. But a photonic system-on-a-chip could be used as an accelerator running alongside a CPU in a conventional computer. Emerging computing platforms like The Machine, a next-generation system under development at Hewlett Packard Labs, offer even more intriguing possibilities. The Machine will hold huge amounts of data in memory and allow users to plug in different processors as needed, depending on the type of computation they want to perform.
>This technology won’t replace general-purpose electronic computers...
I still haven't learned why this would be the case.


Optical Computing vs Quantum Computing Anonymous 04/16/2018 (Mon) 02:35:12 [Preview] No.12649 del
LIGHT: How optical processing can solve some of the world’s most complex problems.
https://www.labs.hpe.com/next-next/light
>Beausoleil spent about a decade working on quantum computing. That’s when he experimented with chips made out of diamonds. His team found that it was too difficult to create enough diamond chips that had exactly the same qualities, making it impossible to manufacture them in a repeatable and predictable way. This particular approach to building a quantum computer wouldn’t scale.
>Beausoleil concluded that photonic technology offered a faster and more practical route to success than building an actual quantum computer. “When somebody does manage to create a true quantum system with entanglement, that’s going to be awesome,” Beausoleil says. “Right now, we’re leaving that one on the table and just trying to take advantage of coherence.”
>[...]
>Optical computing is an emerging field with experimental components. The challenge is to raise the yield of working components to a commercially viable level. The Labs approach is to create a round of equipment, test it to identify flaws, then determine necessary changes in the chip design or manufacturing method.
>[...]
>Labs researchers are also exploring new applications for their optical circuits. For example, they’ve designed a system called an energy minimization computer that changes the state of an optical circuit to find the configuration that consumes the least amount of energy. This concept applies perfectly to solving NP-hard problems like the traveling salesman.
>[...]
>This technology won’t replace general-purpose electronic computers, because lots of problems aren’t NP-hard problems. But a photonic system-on-a-chip could be used as an accelerator running alongside a CPU in a conventional computer. Emerging computing platforms like The Machine, a next-generation system under development at Hewlett Packard Labs, offer even more intriguing possibilities. The Machine will hold huge amounts of data in memory and allow users to plug in different processors as needed, depending on the type of computation they want to perform.
>This technology won’t replace general-purpose electronic computers...
I still haven't learned why this would be the case.


Anonymous 04/16/2018 (Mon) 02:37:45 [Preview] No.12650 del


Anonymous 04/18/2018 (Wed) 21:16:08 [Preview] No.12653 del
>>12649
That sounds really cool. I've always liked the idea of having a completely modular computer, more so than what we currently have; the ability to have multiple processors for different tasks in a standard level computer would be incredible. It will never happen, because most users don't care for solving NP-Hard problems, though.

I spoke with a guy who had a PhD in photonics about 6 months ago. Although that conversation was mainly about carbon fibre, the bandwidth increases alone seem like they would be able to replace normal computers easily. Imagine being able to compute on a terabyte of data in seconds. We're hitting the limits of silicon, and quantum computers require massive amounts of energy to keep the conditions correct. Photonics looks like the only way forward.


Anonymous 04/18/2018 (Wed) 21:34:52 [Preview] No.12654 del
>>12653
To add to my first paragraph (which is worded like shit), my main ideas would be having a modular CPU. So, the APU, cache etc. and the controller for all of the chips would be separate, rather on one die. There would be a much higher latency, especially for the cache, but that could potentially be negated by using a high-bandwidth connection and by using large packets. But it would be much easier to upgrade in response to exploits. Meltdown/Spectre? Just buy a new cache module that mitigates the exploit. Intel ME spying on you? Buy an open source controller that you know can't be hacked remotely.
It would make cooling a bit of a nightmare, though.

While I'm at it, why can't we have open source CPU designs at all, like we do with 40xx and 74xx chips, where different manufacturers make functionally identical products? If the basic chips are simpler (as described in my first paragraph), it wouldn't be too difficult to accomplish. The truly paranoid people could even produce their own modules that would work in a standard motherboard then.


Anonymous 04/20/2018 (Fri) 00:05:06 [Preview] No.12656 del
(523.20 KB 1920x1080 1520474434370.jpg)
>>12654
>While I'm at it, why can't we have open source CPU designs at all, like we do with 40xx and 74xx chips, where different manufacturers make functionally identical products?
I believe this is the goal of RISC-V, a silicon project which is just getting rolling this year.
Qualcomm, Samsung, nVidia, Western Digital, have all invested in it, so it's possible that there'll be more competition for consumers in the area of CPUs; there's also Sifive and Lo-risc which I believe are companies more committed to actual open-sourcing.

Normal people however can't make their own CPU in their garage---it takes highly advanced equipment, controlled space, and enormous amounts of money.

>>12653
>We're hitting the limits of silicon, and quantum computers require massive amounts of energy to keep the conditions correct.
>Photonics looks like the only way forward.
An argument I don't find too bad from quantum enthusiasts is that 60 years ago, electric computers took up entire rooms, and now they fit into the palms of hands.
Obviously there's limitations that restrict the potential mobility of quantum computers no matter what, but it's possible that a consumer could get a fridge-sized quantum computer in their own homes some decades from now.

But that is far off. If you want better personal computing now, I see nothing more promising than optical computers.

>the ability to have multiple processors for different tasks in a standard level computer would be incredible. It will never happen, because most users don't care for solving NP-Hard problems, though.
It won't be available for "most consumers" in the first place. Windows is over and Macintosh will never get off silicon. These computers are going to run OS that won't be geared toward non-professionals; and so they won't be the target market initially.



Top | Return | Catalog | Post a reply