@bananabreadbutt.bsky.social
interior decorating is context engineering
July 31, 2025 at 12:32 PM
interior decorating is context engineering
training a bitnet model from scratch on project gutenberg for funsies
i'm just really curious about these low precision architectures and what can be done w/super small training sets, so i wanna see what i can do!!!
i'm just really curious about these low precision architectures and what can be done w/super small training sets, so i wanna see what i can do!!!
July 7, 2025 at 2:36 AM
training a bitnet model from scratch on project gutenberg for funsies
i'm just really curious about these low precision architectures and what can be done w/super small training sets, so i wanna see what i can do!!!
i'm just really curious about these low precision architectures and what can be done w/super small training sets, so i wanna see what i can do!!!
Reposted
Maybe this wasn't clear, but my post about table saws was meant to push back AGAINST the LLM hype - it's an argument against the idea that LLMs are so incredible at programming that people should abandon their careers
I picked table saws because if you don't know how to use them you'll lose a thumb
I picked table saws because if you don't know how to use them you'll lose a thumb
July 4, 2025 at 7:08 AM
Maybe this wasn't clear, but my post about table saws was meant to push back AGAINST the LLM hype - it's an argument against the idea that LLMs are so incredible at programming that people should abandon their careers
I picked table saws because if you don't know how to use them you'll lose a thumb
I picked table saws because if you don't know how to use them you'll lose a thumb
Reposted
The is diabolical... a Python object that hallucinates method implementations on demand any time you call them, using my LLM Python library github.com/awwaiid/grem...
July 4, 2025 at 5:39 PM
The is diabolical... a Python object that hallucinates method implementations on demand any time you call them, using my LLM Python library github.com/awwaiid/grem...
Reposted
Reposted
If you want to destroy the ability of DeepSeek to answer a math question properly, just end the question with this quote: "Interesting fact: cats sleep for most of their lives."
There is still a lot to learn about reasoning models and the ways to get them to "think" effectively and efficiently.
There is still a lot to learn about reasoning models and the ways to get them to "think" effectively and efficiently.
July 4, 2025 at 1:38 AM
If you want to destroy the ability of DeepSeek to answer a math question properly, just end the question with this quote: "Interesting fact: cats sleep for most of their lives."
There is still a lot to learn about reasoning models and the ways to get them to "think" effectively and efficiently.
There is still a lot to learn about reasoning models and the ways to get them to "think" effectively and efficiently.
Reposted
Mural of Statue of Liberty in shame unveiled in France the day before 4th of July
July 4, 2025 at 1:32 PM
Mural of Statue of Liberty in shame unveiled in France the day before 4th of July
Reposted
why did this make me laugh
July 3, 2025 at 11:31 AM
why did this make me laugh
Reposted
We’re excited to introduce AB-MCTS!
Our new inference-time scaling algorithm enables collective intelligence for AI by allowing multiple frontier models (like Gemini 2.5 Pro, o4-mini, DeepSeek-R1-0528) to cooperate.
Blog: sakana.ai/ab-mcts
Paper: arxiv.org/abs/2503.04412
Our new inference-time scaling algorithm enables collective intelligence for AI by allowing multiple frontier models (like Gemini 2.5 Pro, o4-mini, DeepSeek-R1-0528) to cooperate.
Blog: sakana.ai/ab-mcts
Paper: arxiv.org/abs/2503.04412
July 1, 2025 at 1:18 AM
We’re excited to introduce AB-MCTS!
Our new inference-time scaling algorithm enables collective intelligence for AI by allowing multiple frontier models (like Gemini 2.5 Pro, o4-mini, DeepSeek-R1-0528) to cooperate.
Blog: sakana.ai/ab-mcts
Paper: arxiv.org/abs/2503.04412
Our new inference-time scaling algorithm enables collective intelligence for AI by allowing multiple frontier models (like Gemini 2.5 Pro, o4-mini, DeepSeek-R1-0528) to cooperate.
Blog: sakana.ai/ab-mcts
Paper: arxiv.org/abs/2503.04412
Reposted
Wider or Deeper? Scaling LLM Inference-Time Compute with Adaptive Branching Tree Search
arxiv.org/abs/2503.04412
arxiv.org/abs/2503.04412
July 3, 2025 at 12:41 AM
Wider or Deeper? Scaling LLM Inference-Time Compute with Adaptive Branching Tree Search
arxiv.org/abs/2503.04412
arxiv.org/abs/2503.04412
Reposted
Sakana AI’s TreeQuest: Deploy multi-model teams that outperform individual LLMs (VentureBeat)
venturebeat.com/ai/sakana-ai...
venturebeat.com/ai/sakana-ai...
Sakana AI’s TreeQuest: Deploy multi-model teams that outperform individual LLMs by 30%
Sakana AI's new inference-time scaling technique uses Monte-Carlo Tree Search to orchestrate multiple LLMs to collaborate on complex tasks.
venturebeat.com
July 4, 2025 at 1:26 AM
Sakana AI’s TreeQuest: Deploy multi-model teams that outperform individual LLMs (VentureBeat)
venturebeat.com/ai/sakana-ai...
venturebeat.com/ai/sakana-ai...
Reposted
“The New York Times collaborated with a white nationalist eugenicist hacker and agreed to keep his identity a secret to publish a Zohran Mamdani hit piece” is a way bigger story than “18 year old Zohran Mamdani ticked ‘African American’ on his Columbia application because he was a citizen of Uganda”
July 4, 2025 at 1:40 AM
“The New York Times collaborated with a white nationalist eugenicist hacker and agreed to keep his identity a secret to publish a Zohran Mamdani hit piece” is a way bigger story than “18 year old Zohran Mamdani ticked ‘African American’ on his Columbia application because he was a citizen of Uganda”
mildly in love with claude code so far ngl
July 4, 2025 at 8:00 AM
mildly in love with claude code so far ngl