@cyberquail.bsky.social
Reposted
Came across this amusing and comprehensive history of early Swedish computing: www.youtube.com/watch?v=UXSB...
The Birth and Glory of Swedish Computers
YouTube video by Asianometry
www.youtube.com
March 6, 2025 at 9:21 PM
Came across this amusing and comprehensive history of early Swedish computing: www.youtube.com/watch?v=UXSB...
Reposted
Microsoft's AI Toolkit VS Code extension
Run DeepSeek R1 7B & 14B distilled models for Copilot+ PCs on VS Code for your Copilot+ PCs.
Video: www.youtube.com/watch?v=GotH...
Run DeepSeek R1 7B & 14B distilled models for Copilot+ PCs on VS Code for your Copilot+ PCs.
Video: www.youtube.com/watch?v=GotH...
Run Distilled DeepSeek R1 models on Copilot+ PCs
YouTube video by Windows Developer
www.youtube.com
March 4, 2025 at 3:42 AM
Microsoft's AI Toolkit VS Code extension
Run DeepSeek R1 7B & 14B distilled models for Copilot+ PCs on VS Code for your Copilot+ PCs.
Video: www.youtube.com/watch?v=GotH...
Run DeepSeek R1 7B & 14B distilled models for Copilot+ PCs on VS Code for your Copilot+ PCs.
Video: www.youtube.com/watch?v=GotH...
Reposted
I've lived democratic backsliding & authoritarianism across multiple countries. I've watched repression creep across staff planning sessions & editorial rooms, from both the left & right.
The #1 way it happens? Self-censorship.
Before your opponents ever censor you, self-professed allies do.
The #1 way it happens? Self-censorship.
Before your opponents ever censor you, self-professed allies do.
Hell yeah. Hold the line. This is what we need people to do.
Slow things down, don’t comply in advance so the courts have time to stop them. It’s starting to work.
Slow things down, don’t comply in advance so the courts have time to stop them. It’s starting to work.
Received a letter from POTUS today purporting to remove me as Commissioner and Chair of the FEC. There's a legal way to replace FEC commissioners-this isn't it. I've been so fortunate to serve the American people and stir up some good trouble along the way. That's not changing anytime soon.
February 7, 2025 at 11:00 PM
I've lived democratic backsliding & authoritarianism across multiple countries. I've watched repression creep across staff planning sessions & editorial rooms, from both the left & right.
The #1 way it happens? Self-censorship.
Before your opponents ever censor you, self-professed allies do.
The #1 way it happens? Self-censorship.
Before your opponents ever censor you, self-professed allies do.
Reposted
An uncensored version of R1 is released 🔥
“R1 1776 is a DeepSeek-R1 reasoning model that has been post-trained by Perplexity AI to remove CCP censorship. The model provides unbiased, accurate, and factual information while maintaining high reasoning capabilities.”
huggingface.co/perplexity-a...
“R1 1776 is a DeepSeek-R1 reasoning model that has been post-trained by Perplexity AI to remove CCP censorship. The model provides unbiased, accurate, and factual information while maintaining high reasoning capabilities.”
huggingface.co/perplexity-a...
perplexity-ai/r1-1776 · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
February 19, 2025 at 3:22 AM
An uncensored version of R1 is released 🔥
“R1 1776 is a DeepSeek-R1 reasoning model that has been post-trained by Perplexity AI to remove CCP censorship. The model provides unbiased, accurate, and factual information while maintaining high reasoning capabilities.”
huggingface.co/perplexity-a...
“R1 1776 is a DeepSeek-R1 reasoning model that has been post-trained by Perplexity AI to remove CCP censorship. The model provides unbiased, accurate, and factual information while maintaining high reasoning capabilities.”
huggingface.co/perplexity-a...
Reposted
A deep dive into understanding long CoT reasoning in LLMs!
They believe that the model size is the primary factor limiting the emergence of the behavior. However, scaling up with model size for long CoT with open-source RL Infrastructure still faces challenges.
Please read their post for details.
They believe that the model size is the primary factor limiting the emergence of the behavior. However, scaling up with model size for long CoT with open-source RL Infrastructure still faces challenges.
Please read their post for details.
February 6, 2025 at 3:40 AM
A deep dive into understanding long CoT reasoning in LLMs!
They believe that the model size is the primary factor limiting the emergence of the behavior. However, scaling up with model size for long CoT with open-source RL Infrastructure still faces challenges.
Please read their post for details.
They believe that the model size is the primary factor limiting the emergence of the behavior. However, scaling up with model size for long CoT with open-source RL Infrastructure still faces challenges.
Please read their post for details.
Reposted
Why do LLMs trained on over 90% English text perform so well in non-English languages?
They find that they learn to share highly abstract grammatical concept representations, even across unrelated languages!
They find that they learn to share highly abstract grammatical concept representations, even across unrelated languages!
February 6, 2025 at 7:19 AM
Why do LLMs trained on over 90% English text perform so well in non-English languages?
They find that they learn to share highly abstract grammatical concept representations, even across unrelated languages!
They find that they learn to share highly abstract grammatical concept representations, even across unrelated languages!
Reposted