Do you know what places people who are interested in research congregate at? Every forum, meet up or journal gets overwhelmed by bullshit with a year of being good.
3,929 karma · joined February 12, 2024
It's a lazy blog post that should be thrown out after a minute of thought by anyone in the field.
If op is trying to catch up to frontier models locally on a budget 1/5th of what you can get in the west then I can see why she would feel the way she does about Ai.
My heart goes out to all the oppressed programmers in the EU.
>The bourgeoisie cannot exist without constantly revolutionising the instruments of production, and thereby the relations of production, and with them the whole relations of society. Conservation of the old modes of production in unaltered form, was, on the contrary, the first condition of existence for all earlier industrial classes. Constant revolutionising of production, uninterrupted disturbance of all social conditions, everlasting uncertainty and agitation distinguish the bourgeois epoch from all earlier ones. All fixed, fast-frozen relations, with their train of ancient and venerable prejudices and opinions, are swept away, all new-formed ones become antiquated before they can ossify. All that is solid melts into air, all that is holy is profaned, and man is at last compelled to face with sober senses his real conditions of life, and his relations with his kind.
LLM training—and the massive (potential) copyright infringement that everyone is engaging in to train these models—is the latest contradiction in capitalism. For the first time in my lifetime, this contradiction isn’t harming the workers but is instead affecting a segment of the capitalist class.
Not since the abolition of slavery has one class of capitalists required the destruction of another to modernize the means of production.
We are in for an interesting decade.
Deepseek R1 is something that you can run in a garage on hardware that the average software engineer can buy with a months salary and when it came out last month it was better than _every_ other model.
Whenever I talk to people who rave about SICP I get the feeling they never made it to chapter 4 and beyond.
> In this chapter we will describe processes in terms of the step-by-step operation of a traditional computer. Such a computer, or register machine, sequentially executes instructions that manipulate the contents of a fixed set of storage elements called registers. A typical register-machine instruction applies a primitive operation to the contents of some registers and assigns the result to another register. Our descriptions of processes executed by register machines will look very much like ``machine-language'' programs for traditional computers.
https://mitp-content-server.mit.edu/books/content/sectbyfn/b...
The transistor level simulation happens in 3.3.4 "A Simulator for Digital Circuits".
That is a arbitrary object detection problem, which is in general harder to solve than just using an ips.
You can absolutely have open source machine code.
The issue is and always has been that you need to have access to the same level of abstraction as the people writing the source code. The GPL specifically bans transpilers as a way to get around this.
In ML there is _no_ level of abstraction other than the raw weights. Everything else is support machinery no different to an compiler, and os, or a physical computer to run the code on.
Linux isn't closed source because they don't ship a C compiler with their code. Why should llama models be any different?
What do you think an open source matrix should look like?
Mr Musk, we can't afford a shitpost gap between communist and capitalist AIs!
And it's not until the end of the last chapter that you build a stack machine in about 40 pages from memory.
Because vertical scaling is now large enough that I can run all of twitter/amazon on one single large server. And if I'm wrong now, in a decade I won't be.
Compute power grows exponentially, but business requirements do not.
Not everything needs to be made for a chimp.
Rust developers are promising that rust in the kernel should have no impact on c in the kernel.
If rust in the kernel has no impact in c then you should be able to remove the rust from the kernel and move it to it's own project, since it will have no impact on the c either way.
Black letters on white background, xheight of between 10 to 30 px, tiff format, mono column layout, etc., etc., etc..
People get terrible results because they treat it like a phone app and drop a barely legible colored jpg of a bent page and wonder why it's garbage.
If Rust truly is zero overhead on the C side of things then they should be just able to fork Linux indefinitely without any worry about what upstream does. Just fast forward every change and you're golden.
Then they can write every driver they can imagine to their hearts' content.
The fact they aren't doing this tells me that the promise that this won't impact C people at all isn't much of one.
If you're asking them to accept your code then yes, you' are asking them to support your project forever.
If you weren't sending them patches then they couldn't block you. Since you are they can.
Again, it's not their job to support this great idea you have that means they have to change how they've done everything for the last 30 years.
It's yours.
If you need someone to do free work but you can't convince them then you either get their boss to tell them to do it, you take over their job, or - the one thing that no one under 30 ever seems to do - fork and do the work yourself without anyone stopping you.
Crazy how much drama happens in open projects like openAI.
I know which of the two languages was easier and more pleasant to hire for - which should be impossible as I kept getting told no one uses forth.