Cline SDK is live. 74.2% on Terminal Bench with Opus 4.7. Ahead of gpt-5.3-codex and Gemini 3.1 Pro. Open source.
LLMS
-
Comparative analysis of GQA, MLA, and Gated DeltaNet model architectures
By
–
Yes! It would be interesting whether this happens in GQA:SWA models only though or also in MLA or Gated DeltaNet models, for example.
-
Curiosity about Gemini’s arrival in the local apps race
By
–
Really curious to know when Gemini will join the Cowork & Codex race to build a local application that is not reserved only for developers. Antigravity hasn't published updates on X for a month, and remains very software-focused.
-
TACO: A New Training-Free Framework for Terminal AI Agents
By
–
What if your terminal agent could learn to ignore the noise and keep only what matters? Researchers from University of Manchester, HKUST, and Beihang University present TACO — a plug-and-play, training-free framework that automatically discovers and refines compression rules
-
Learning LLM Architectures Through Scratch Implementation and Model Evaluation
By
–
A little talk on what we can learn from implementing LLM architectures from scratch in Python and PyTorch. And how I approach new open-weight models, compare them against reference implementations etc:
-
AI Agent Architecture for Persistent State and Context Management
By
–
If you build anything that runs for longer than one session, this is an architecture to test. The agent holds its state, the workspace holds its context, and you stop re-explaining from scratch every time. Star the repo and get started https://
github.com/holaboss-ai/ho
laOS/releases
… @Holabossai -
holaOS Beta 0.1 Release Introduces Agent Computer Management Layer
By
–
holaOS has shipped its beta 0.1 release, adding a management layer on top of its Agent Computer foundation: a Dashboard, Sub Agents, and Multi Workspaces. Each workstream gets its own isolated context, memory, and agent.
— 🚨 AI News | TestingCatalog (@testingcatalog) 13 mai 2026
No need to have a reset between sessions. The whole… pic.twitter.com/a7hLGBegADholaOS has shipped its beta 0.1 release, adding a management layer on top of its Agent Computer foundation: a Dashboard, Sub Agents, and Multi Workspaces. Each workstream gets its own isolated context, memory, and agent. No need to have a reset between sessions. The whole
-
δ-mem: An efficient online memory mechanism for LLMs
By
–
// δ-mem: Efficient Online Memory for LLMs // One of the more elegant memory mechanisms I've seen this month. Most long-term memory work either inflates context or retrains the model. This paper shows a tiny external state, coupled directly into the attention computation, can
-
Grok AI’s Deep Understanding of Personal Data: Scary Implications and Agent Capabilities
By
–
This will be an interesting problem. Go to Grok. Ask it "please simulate a conversation between Michael Mignano and Robert Scoble about the future of personal data." And it will. Very well. It knows me VERY DEEPLY. That is scary. But on the other hand, now my Agents know me