porridgeraisin

7 ms·

91.
by porridgeraisin 2mo ago
At a basic level, the steps involved in setting up a fab are extremely serial in nature. While there are ways in which each step has been cost-cut in the past, amdahl's law itself limits it to a minimum of 2-3 years. If you want good y
92.
OpenAI's ad business hits $1B annualized revenue run rate(reuters.com)
1 points by porridgeraisin 2mo ago|0 comments
93.
SK Hynix CEO sees memory chip shortage lasting until 2030(sammyfans.com)
67 points by porridgeraisin 2mo ago|82 comments
94.
OpenAI acquires Mac Minis, Mac Studios for AI training(cryptobriefing.com)
4 points by porridgeraisin 2mo ago|0 comments
95.
by porridgeraisin 2mo ago
No, the above attack writes that function into bashrc, meaning the next time the user runs sudo themselves, you harvest their password.
96.
by porridgeraisin 2mo ago
I mean, I saw this on twitter, and thought ok maybe its a nice exploit. But really? its the usual docker root thing? I wouldn't even consider that a vulnerability tbh, every personal laptop I had I add myself to docker group. Yes, you
97.
by porridgeraisin 2mo ago
Yep, but I mostly get SMS OTPs. Whatsapp is all the other messages
98.
by porridgeraisin 2mo ago
I use whatsapp primarily, so no tethering needed, but SMS of course for one-time passwords and the like. I didn't want to open my phone for that. So I had a coding agent write an android background app for me that listens for SMS and f
99.
Uber Solved DB Overload: From Static Rate-Limiting to Intelligent Load Mgmt(uber.com)
1 points by porridgeraisin 2mo ago|0 comments
100.
Teamwork: When AI Becomes a Research Partner – Google Antigravity Blog(antigravity.google)
4 points by porridgeraisin 2mo ago|0 comments
101.
I have made an OS in CUDA(twitter.com)
3 points by porridgeraisin 2mo ago|0 comments
102.
by porridgeraisin 2mo ago
This has happened in many packages contributed to by this person. Mkdocs e.g
103.
by porridgeraisin 2mo ago
I think people are overestimating the speed at which the transition will happen. I think it will happen much slower than is often portrayed online. This slow transition period will also help answer the "where will they go" questio
104.
by porridgeraisin 2mo ago
A lot of it will get automated the same way very many industries got automated. A lot of physical labour got automated once a primitive for it was created. Similarly, we have now a primitive for automating knowledge work. In the next few ye
105.
by porridgeraisin 2mo ago
> Anything more complex is overentineering. For general workflows, I agree. Doing some DAG system is just a waste of time if you don't have a grounded verifier for every node. If you are going to human verify, why make a dag of suba
106.
by porridgeraisin 2mo ago
You train an addend to the weights rather than the weights themselves. You limit the dimensions[1] available in this addend to a small fraction of the dimensions of the weights, making training cheaper. When you have a few example chats you
107.
by porridgeraisin 2mo ago
> Framework 12 ... days without charging I don't see how? it is a 50Wh battery. Even at 6W that is about 8h of usage? On what OS + workload on x86 are you seeing just 6W usage? Or is it that each day your usage is only 2 hours?
108.
by porridgeraisin 2mo ago
Mountaineering. For indians, you also have the religious aspect because the immigration center you go to to visit Mt Kailash is there. That immigration building was the white one that got completely shredded in the popular video going aroun
109.
by porridgeraisin 2mo ago
Those takes are wrong with or without this particular processor.
110.
IBM Unveils Next Generation Dual-Architecture Processor for IBM Z and LinuxONE(newsroom.ibm.com)
175 points by porridgeraisin 2mo ago|131 comments
111.
Ilya Sutskever's SSI and Nvidia Announce Long-Term Strategic Partnership(nvidianews.nvidia.com)
6 points by porridgeraisin 2mo ago|0 comments
112.
AWS to acquire DuckLabs, the Amsterdam-based company behind DuckDB(aboutamazon.com)
3 points by porridgeraisin 2mo ago|1 comments
113.
by porridgeraisin 2mo ago
Well ya, that efficiency is why it's split. There is also the other idea where you run your attention layer on the GPU/TPU/Trainium and the FFN on the SRAM accelerator. Because KV cache is more difficult on cerebras etc, whil
114.
by porridgeraisin 2mo ago
It is being used as part of a combined system. For example AWS is pushing for Trainium + WSE 3. The WSE 3 does the decode and the Trainium does the prefill. Even in nvidia land rubin + LPU does a similar thing. It has its downsides of cours
115.
by porridgeraisin 2mo ago
These are not replacing GPUs, they are entirely complementary. It's the same with cerebras, groq etc, they are all complementary to the GPU.
116.
by porridgeraisin 2mo ago
Articles are mostly a prompt for people
117.
Walgit leans on S3 primitives but has classic distsys bugs(twitter.com)
4 points by porridgeraisin 2mo ago|0 comments
118.
by porridgeraisin 2mo ago
Tbh the way you're treated depends on the hotness of the task. A couple of years ago, people I know got paid OK for relatively simple programming and logic RLHF tasks. But very soon it turned dark because that sort of data was required
119.
by porridgeraisin 2mo ago
A lot of new parents in my family are simply not giving any screen to kids until they are much older. The only exception is using it to play some brainrot to get them to eat if it's turning out to be a particularly difficult meal. Most
120.
by porridgeraisin 2mo ago
Do they claim the opposite?

More ›