sliding window beats linear attention
4 experts across 3 network communities independently surfaced this.
4 experts
3 communities
1 sources clustered
Research & technical analysis
3 experts
Evidence, methods and technical implications.
“Huge thanks to my collaborators @RheaSukthanker, @CameronPashmina, and @Emy_Aze. Paper: https://arxiv.org/abs/2608.28444”
2 experts discussed this · 2 posts
Alexia Jolicoeur-Martineau: Simple beats complicated: We show that switching to a sliding-window attention mask with attention sinks (at no cost) beats linear attention post-training. Huge thanks to my collaborators Rhea Sukt…
Miguel Alonso Jr.: Simple beats complicated: We show that switching to a sliding-window attention mask with attention sinks (at no cost) beats linear attention post-training. Huge thanks to my collaborators Rhea Sukt…
James MacGlashan: Makes me wonder if you can remove the sinks if you use softmax-1?
Open the full discussion →
Established
AI field signal
Signal
9d ago
⚡ 26 h early
5 experts across 3 network communities independently surfaced this.
5 experts
3 communities
1 sources clustered
Research & technical analysis
2 experts
Evidence, methods and technical implications.
“GLM-5.3 is now open-weight. Their most capable model for agentic coding and cyber defense is now available to download, run, and customize. Weights: huggingface.co/zai-org/GLM-5.3 Tech blog: z.ai/blog/glm-5.3”
Developing
AI field signal
Analysis
2d ago
⚡ 6 h early
Claude Code capabilities assessment essay
3 experts across 2 network communities independently surfaced this.
3 experts
2 communities
1 sources clustered
“If you haven’t recently used Claude Code*, you might not understand where AI is at davidpreichert.substack.com/p/if-you-haven… #AI #agents”
“In particular, your imaginary friend or a character in a script cannot do any of these things: davidpreichert.substack.com/p/if-you-hav... davidpreichert.substack.com/p/ai-and-a-c... Does that necessarily mean AI agents can think, or reason, or are consciou…”
5 experts discussed this · 53 posts
david-p-reichert.bsky.social: Many don't see how much AI has changed recently. Coding agents aren't just chatbots. We can’t deal with the challenges of AI if we don’t understand it. Here's a post to provide some evidence (Claud…
david-p-reichert.bsky.social: This should be useful to technical people / researchers who haven’t recently used coding agents, but hopefully some insights for others interested in AI, too. Some key points below: ...
david-p-reichert.bsky.social: 1. There’s a huge gulf between a chatbot putting text into your brain and an agent using code to interface with its environment. 2. Code isn't “just text” or even maths. Code is a tool to observe a…
Open the full discussion →
Established
AI field signal
Signal
9d ago
⚡ 6 h early
2 directory members surfaced this signal.
2 experts
2 communities
1 sources clustered
“GLM-5.3 is now open-weight. Their most capable model for agentic coding and cyber defense is now available to download, run, and customize. Weights: huggingface.co/zai-org/GLM-5.3 Tech blog: z.ai/blog/glm-5.3”
Established
AI field signal
Resource
11d ago
⚡ 12 h early
ZCode GLM-5.2 coding harness release
4 experts across 2 network communities independently surfaced this.
4 experts
2 communities
1 sources clustered
Research & technical analysis
2 experts
Evidence, methods and technical implications.
“Blog: z.ai/blog/glm-5.3... Available now across all official platforms: Weights: huggingface.co/zai-org/GLM-... API: docs.z.ai/guides/llm/g... Coding Plan: z.ai/subscribe ZCode: zcode.z.ai/en Chat: chat.z.ai AutoClaw: autoclaw.z.ai”
Established
AI field signal
Development
7d ago
⚡ 4 h early
prime-agent v0.8.1 usability issues
3 experts are actively discussing the implications.
2 experts
1 community
1 sources clustered
“alright, i love Prime Agent, but it's been steadily becoming less stable i finally sat down and fixed it this morning. Rewrote a fairly large chunk of it so comms don't get bottlenecked, which eliminates all the global failures that used to happen it's pret…”
3 experts discussed this · 16 posts
Tim Kellogg: alright, i love Prime Agent, but it's been steadily becoming less stable i finally sat down and fixed it this morning. Rewrote a fairly large chunk of it so comms don't get bottlenecked, which elim…
Tim Kellogg: i can't send a real pull request because i'm not an authorized contributor they shut down open contributions because they got overwhelmed. but now I have to just hope that my code get seen by the r…
Nafnlaus 🇮🇸 🇺🇦: I'll say it once again: OSS projects this day in age need an AI project manager. *At least* to handle first-level work, even if major decisions ultimately get escalated to humans.
Open the full discussion →
DeepMind AI games research frontiers
2 experts are actively discussing the implications.
2 experts
1 community
1 sources clustered
“Google DeepMind article on 15 tears of Ai game research”
2 experts discussed this · 2 posts
Lynn Cherny: Google DeepMind article on 15 tears of Ai game research
Miguel Alonso Jr.: Google DeepMind article on 15 tears of Ai game research
austegard.com: 🤣😭😢😹😂😿😂😹😢😭🤣😭😢😹😂
Open the full discussion →
RL evaluation bottleneck analysis
2 experts are actively discussing the implications.
2 experts
2 communities
1 sources clustered
“If like me you're curious why most algorithmic papers in RL land with a thud, here's my argument as to why.”
2 experts discussed this · 4 posts
Eugene Vinitsky: If like me you're curious why most algorithmic papers in RL land with a thud, here's my argument as to why.
Eugene Vinitsky: As always my rule is write fast with minimal editing, so forgive the mistakes. They at least prove minimal LLM involvement.
Miguel Alonso Jr.: If like me you're curious why most algorithmic papers in RL land with a thud, here's my argument as to why.
Open the full discussion →
Ox Alpha model API launch
3 directory members surfaced this signal.
3 experts
1 community
1 sources clustered
“Track Ox Alpha's data live: https://t.co/noXptFpet3”
“Ox Alpha, a new stealth model on openrouter. People are saying it’s stronger than Sol & Fable. No benchmarks yet. Some are hypothesizing it’s Chinese. Everyone seems convinced that it’s not OpenAI or Anthropic openrouter.ai/stealth/ox-a...”
2 directory members surfaced this signal.
2 experts
1 community
1 sources clustered
“Over the last few months, I think the foundational capability you need for a "robot GPT" has become clear and has been demonstrated: with new in-context learning results from top startups, we see how general-purpose robots become generally useful tools: itc…”
model predictive meta-reasoning paper
1 directory member surfaced this signal.
1 expert
1 community
1 sources clustered
facility location interactive learning paper
1 directory member surfaced this signal.
1 expert
1 community
1 sources clustered
robot deliberate practice budget paper
1 directory member surfaced this signal.
1 expert
1 community
1 sources clustered
Established
AI field signal
Signal
13d ago
2 directory members surfaced this signal.
2 experts
1 community
1 sources clustered
“Some thoughts of anger vs anxiety for people in tech when it comes to what AI does to the industry in response to a comment I read that made if think. lucumr.pocoo.org/2026/8/24/an...”
Established
AI field signal
Signal
16d ago
1 directory member surfaced this signal.
1 expert
1 community
1 sources clustered
