If you read what Stallman writes about hardware, he imagines a future when we will all be able to print our own processors based on free hardware designs.
How close is this future? To my mind, it seems to distant to care about. Like interplanetary travel distant.
Because, really, what's the point of freeing up computer related technologies when there will never, ever be a point when computer technologies will not be controlled, ultimately, by massive corporations?
>>53826711
>there will never, ever be a point when computer technologies will not be controlled, ultimately, by massive corporations
Not when it becomes possible to 3D print your own processors at home. Even if that is 500 years away, it's possible because (forgive that Marxist lingo for lack of better words) it shifts the means of production to the end user.
>>53826711
The immediate task is to develop a way of scanning cpus and investigating random samples don't contain back doors.
you can make 8bit computer...
so 100 years
>>53826711
You can already do it, if you have enough funds to do so, RiscV is a thing nowadays
>>53826711
>it shifts the means of production to the end user
what about the printer itself? are we supposing that just anyone can access it like a public good? and the raw materials? even if robots do the mining, who controls those robots?
really, with pretty much any modern technology, it's impossible to get the ball rolling yourself because of the overheads. it takes an army of people any way you look at it.
>>53826711
pic related
>>53826711
>print our own processors based on free hardware designs
Does that fucking retard even know how processors are made? Are you going to make a fucking 10 transistor processor with a 3d printer? Fabs cost billions of fucking dollars and this fat fucker thinks 3d printers will be able to compete?
>>53826711
dafuck is Fassino doing sitting next to gentoo?
>>53827240
Protip : this is how much code monkeys know about hardware
>>53827303
>130 nm trannies are viable
1980 called. They want their technology back. Don't bother replying, this thread will be gone by the time your web browser manages to render it.
>>53827344
You fucked up the part when you reply to the person you intended to, reddit
>>53827240
First of all, calm yourself. It's not what he said, it's what I said.
Here's what he says:
>We can envision a future in which our personal fabricators can make chips, and our robots can assemble and solder them together with transformers, switches, keys, displays, fans and so on. In that future we will all make our own computers (and fabricators and robots), and we will all be able to take advantage of modified designs made by those who know hardware. The arguments for rejecting nonfree software will then apply to nonfree hardware designs too.
>That future is years away, at least. In the meantime, there is no need to reject hardware with nonfree designs on principle.
We're talking about the distant future, so it's open-ended to say the least. The specifics and the costs of the technology involved will all be different. But the complexity probably will be something similar. That sheer complexity is what makes me doubt it will ever happen.
>>53827240
>>53827303
Do you guys really know how "fabs" work?
It's all too easy to hold "it's really complicated" over other people when you don't know about it yourself.
It is difficult for several reasons.
One challange is that making ASICs is too expensive and requires tools that is not likely to come to everyones homes.
A solution would be to use a FPGA where the production can still be made by specialists.
These are a lot slower than high end cpu's but people would be in more control.
But then we have to look at the current state of FPGA's.
Every model is programmed in a different way and it is usually using proprietary methods.
The sane response is to just use a CPU and trust it works as it is documented.
Buying a CPU / GPU is fairly cheap when they are mass produced. so the idea that we should be able to create them on our own is neither necessary nor easily obtainable.
>>53827302
Probably some bullshit seminar about freedom tm.
BTW, where I live the communists have convinced the comune to switch to opensores, it's going to be a shitshow.
What would be the point?
Think about machines you have in your house.
>toaster
>television
>kettle
>computer
>etc
How often do you use them? At least a few times a week.
How often do you need a new processor? a few times a week? no, once every few years.
Unless we're talking about a replicator that can make anything from food to positronic brains, it makes no sense.
I'm all for anarchism, but what does he expect?
Small communities working together, bartering, writing free software and laying 100s of kilometers of fibreoptic cable they printed out at home so they won't have to rely on the tyranny of ISPs to access the internet?
>>53827710
Pretty much. Also parrot fucking.
>>53827240
Maybe we invent materials which are feasible for 3d printing processors. Keep in mind, that computation in a more abstract way is just a calculation based on logical gates. 2000 years ago computation was all about, if water runs through a gate or not and the people then probably couldn't imagine having millions of those logical gates in a device the size of your hand.
>>53826711
A few years ago I saw a national lab demo a 3D printer that could print an entire laptop.
Well actually it was a soldiers helmet with integrated computer functionality, but all the core components were there, and the curved CPU geometry was interesting as the part curved with the helmet. The transistor size was around 15 microns, but considering was printed from raw feed stock and input data it was very impressive.
Took two months to print, in one of the most advance research labs in the world. However the methods they used were very simple and it would not surprise me to see this tech becoming wide spread in 20 or 30 years for small low run companies or the filthy rich individual.
However if you just let me use a more common method of manufacturing (as printing is relatively new as can't shrink well) then a we could pool resources and batch run custom CPU sets with 1 micro transistor size assuming order size of at least 1000, which is definitely small enough to topples the industry giants with turn around time of a year or two. That disruptive tech is already here more or less, less then 5 year if the market wants it.
>>53827792
Interesting, do you have a link?
Is it like inkjet printing but with semiconducting inks?
>>53827890
>inkjet printing but with semiconducting inks?
Isn't this the only possible way any of this engineering is feasible?
Design and test the thing completely on the computer and then literally print it out atom by atom without worrying about parts?
>>53827890
No, it was one of those guest speaker event from back when I was in college.
If I remember correctly it was basically just extruding feed though syringes attached to robot arms. Then they would bath and temper to get final properties.
The battery was the part that had the most issues given all the potential energy (even when uncharged) it had was created reactions before it could be sealed. They did print a battery as proof of concept, but didn't use it as anything off the shelf was way better.
The real barrier to the build was calibrating the flow and feed rates of the different stock, each had there own values and optimal needle diameters. I was just blown away that many ranged from 50 to 5 microns depending on the stock type. Once they got those setting they didn't have much trouble repeating the setup, so once people publish those values then regular folk don't have to deal with all that complex fluid dynamics and stuff as the algebraic approximation woks well enough. Even if they did publish, we could just empirically generate the needed values through lots of trial and error.
>>53828072
Actually they compromised. Pre-doped silicon nano crystals suspended in a gel, printed in layer by type and point fused with laser passes. Much high quality then the typical circuit inks, but still not as good as full mono-crystals or atom assemblers. But surprisingly close given how much faster and easier it was compared to atomic assemblers as these were effectively very very large nano assemblers.
Trick is not trying to force it, but setup conditions were the parts assemble themselves, let physics do the hard part for you.
>>53827456
Yes, now fuck off. There's no fucking way anyone will ever be able to produce silicon ingots or have precise enough instrumentation to lithographically fabricate processors (that don't have 130+ nm transistors) in their homes.
>>53828393
Okay, I'll just stick with Windows then.
>>53828393
Actually, I could do.
Well let me clarify, I couldn't purify the feed stock to the grade needed (that refinement setup would add 30 years to the build time), but could buy raw powders at grade and work from there. Silicon ingot are simple, I have made some small ones as part of my grad studies. They just take huge amount of time and electricity, but the tolerance for temp controls can be done with an arduino, high end thermal couples and a empirical generated thermal model of the furnace being used.
the lithographically fabricate process has gotten much easier now that FIBs are dropping into the sub $300,000 price range.
sure my chips would only be comparable to the chips ~1991 or ~1993 depending on any number of things, but with custom architecture and optimized software an early 1990 chip can still give the new stuff a good run for the money.
I mean we are only looking at a factory setup cost of 3 to 5 million plus another 2 to 3 million in material and labor. Once setup the site cost would fall fast.
So just give me 10 million,10 years and a good crew of people to work with and let's show them what we can do!
>>53827552
sudo apt-get install solitaire
>>53828835
Can you personally recommend any texts where I could learn about this process?
>>53828893
Not really, I might have some basic text on ingot formation somewhere but you probably find the same stuff online.
You just melt the silicon in a inert gas environment. Lower a seed crystal in and slowly pull it up and down dunking it to grow the crystal, the slower the better. So you want your furnace to have a very small but significant range so the part you are making doesn't melt but the molten puddle you are dunking it in does.
It like making a lollipop, if your kitchen was thousands of degrees and an asphyxiation hazard among other minor changes.
As for the second part, I am not as familiar with that but the other people I worked with explained enough that I am sure I could figure it out, as they could do it.
>>53829178
Thanks, man.
I'll do it just like you said.
>>53829178
Oh, I forgot.
I dunked it to make things easier, but you get better results if you just slowly pull so rings are less likely to form. But I had trouble with the pull rates, which is why my ingot had rings. Still useful, just couldn't be used for larger chip sizes.
>>53826784
>scanning cpus
it's being done. xray son
>>53831033
So where are the resources reporting on the xray scans and the analysis done on them?
>>53827552
Yes everyone should switch from closed source, means that closed source will die eventually
>>53831118
As nice as that would be, I'm pessimistic about that outcome happening. As long people don't value their own freedom in computing, as long as people think in terms of "open source/closed source", it is not going to happen.
>>53831114
I'm pretty sure Jacob spoke about it here. It may be more for determinisic hardware verification to make sure things aren't intercepted and sabotaged in transit
https://www.youtube.com/watch?v=L-LQ7SBnic0
>>53831114
here's more. just search.
https://www.caleupe.com/writing/jacob-appelbaum-a-technical-action-plan-in-the-face-of-surveillance/
it only costs like 200k to produce a line of motherboards. any white yuppie group-buy can do this.
>>53831316
>a motherboard is a CPU
Good job, /g/.
>>53826767
intellectual property then becomes even more important to holding onto record profits and thus people will get death sentence for piracy. give me marxist expression any day over this dire future.
