Search: “engineering”

Yesterday I shared our open-source Clash Royale simulator and its recurrent PPO agent here. A training loop is easier to understand when you can watch it, so we put a small interactive version online:

arXiv:2609.31140v1 Announce Type: new Abstract: Most Vision-Language Models (VLMs) are built by extending pretrained Large Language Models (LLMs) with visual modules and multimodal alignment. However,

arXiv:2609.31568v1 Announce Type: new Abstract: AI tutoring could markedly improve learning outcomes for students in developing regions such as Vietnam, yet the two obvious paths both fall short. Clou

arXiv:2506.04734v3 Announce Type: replace Abstract: Reasoning models represented by the Deepseek-R1-Distill series have been widely adopted by the open-source community due to their strong performance

arXiv:2609.31186v1 Announce Type: new Abstract: Artificial intelligence is advancing rapidly, with increasingly capable systems taking larger roles in reasoning, decision-making, scientific discovery,

arXiv:2609.30662v1 Announce Type: new Abstract: Large language models (LLMs) can plan, use tools, write code, and execute long-horizon workflows, yet strong local competence does not guarantee project

https://preview.redd.it/23cumykhabsh1.png?width=944&format=png&auto=webp&s=0b1d1c437fc845169f521b31793e2b1325dab0dd Reminder that you can put Qwen 3.8 27B as a subagent and its a workhorse. Pic: using

arXiv:2609.30334v1 Announce Type: cross Abstract: AI development agents are increasingly used to support and partially automate software architecture tasks. To explore how practitioners perceive this

I agree. Learn the fundamentals, folks! If you plan to be a solid builder using agents, this is a really important point. There is no way around it. I have recently started digging deep into evals, in

I know there is nothing new with what I am saying but I recently started with hermes agent (it’s been a while I wanted to but did not have the time). Qwen3.8fn 6bpw exl3 (from turboderp) on a 6x3090 (

Banger paper from Microsoft and colleagues. We talk about human taste being important in this AI era. But for recursive self-improvement, an agent's taste also matters. The big question is: How often

blank cells in tables can mess up agent decision making many document OCR tools struggle with reading multiple blank cells in a document and cause values to be shifted from the original column and row

arXiv:2609.30233v1 Announce Type: cross Abstract: Task and motion planning (TAMP) problems remain difficult even with full observability and object-centric states because discrete decisions are tightl

How do you actually build an effective harness with Claude? We had Thariq (@trq212) from Anthropic at Navigate 2026 to talk about 'Unhobbling Claude', the difficulties and processes on how to build ag

arXiv:2609.28836v1 Announce Type: cross Abstract: While various multimodal methods combining imaging and tabular data for Alzheimer's disease (AD) diagnosis were proposed, they are often limited in ge

Pay attention to System One models if you are building custom harnesses and agentic workflows. Great to see DSPy integrate Jev and now support System One models for calibrating agentic workflows. I'm

arXiv:2609.30177v1 Announce Type: new Abstract: Query understanding (QU) plays a critical role in production search systems, translating raw user queries into search execution plans that drive downstr

1. We will keep accelerating. Our AI efforts are only 3 years old, vs 6 and 10 years old for Anthropic and OpenAI. If our second derivative remains strong, SpaceX will reach pole position in about 6 m

arXiv:2609.26828v1 Announce Type: cross Abstract: NAND-backed storage offers the capacity needed to scale LLM prefix caching, but its block I/O path incurs CPU cache contention and host-DRAM staging i

building in the agent space? this is your room at SF @Techweek_. think modern gatsby: black tie, red carpet, cocktails, live jazz & art deco glamour. 300+ founders, founding operators & investors acro

The availability of resources for AI workloads can be challenging across the industry, especially accelerators. This can slow your AI workload deployment if it’s built around a specific type of accele

Latin America’s small and medium-sized businesses are the heartbeat of the region's economy — accounting for more than 60% of total employment in the region, according to United Nations estimates. And

arXiv:2609.25804v2 Announce Type: replace Abstract: LLM agents increasingly work on long-horizon tasks, and the decisions they make along the way, such as which hypothesis to test or which implementat

arXiv:2609.27321v1 Announce Type: new Abstract: Language-model agents increasingly face long-horizon tasks with evolving state, interdependent decisions, and delayed outcomes. Scaling their training r

arXiv:2608.28021v2 Announce Type: replace-cross Abstract: Large language models increasingly author Infrastructure-as-Code (IaC), where one insecure default is provisioned straight into production. Pr

I’m the developer of TensorSharp. Inspired by vLLM PR #57250, I added and optimized a native /v1/systemone endpoint for Jev-style structured decisions using DiffusionGemma GGUF. This uses DiffusionGem

arXiv:2609.24838v1 Announce Type: new Abstract: Medical agents increasingly combine general reasoning models with specialized clinical tools, yet their capabilities remain largely fixed by what clinic

arXiv:2609.25498v1 Announce Type: cross Abstract: Deploying Large Language Models for runtime operational triage incurs prohibitive latency (>100-500 ms), high VRAM requirements (>4-8 GB), and excessi

arXiv:2609.22917v1 Announce Type: new Abstract: Non-technical stakeholders frequently cannot write the SQL needed to extract insights from operational databases. We built and evaluated a Text-to-SQL a

As part of our efforts to pace the frontier, we’re committed to supporting independent assessments with deep levels of access across training, evaluation, and deployment. That access should enable thi

Cool paper showing how effective tuning a system prompt for an agent can be. Recommended paper if you tune agent harnesses. This paper presents EvolveTrade, which treats a trading agent's system promp

I am trying out a few fine-tunes of Qwen 3.5 9B @ IQ4_XS @ 131K context and trying to go mostly local (free beats cheap, after all). However, it is much slower than at, say 16K context, even when I am

arXiv:2609.22142v1 Announce Type: cross Abstract: Kubernetes scheduler plugins that score candidate nodes are, in production, hand-tuned heuristics. We ask whether a learned scoring function - trained

arXiv:2609.21112v1 Announce Type: new Abstract: Training a visuomotor policy calls for abundant demonstrations that closely match the target environment, yet collecting them anew remains expensive. Ex

Grok 4.7 places @SpaceXAI as third, after Anthropic & OpenAI, for agentic coding. When factoring in that Grok is significantly faster & lower cost, it’s a great choice for your everyday workhorse. Gro

Hey everyone, A few classmates and I got tired of the usual browser AI workflow: copying text from a documentation page, opening a separate tab, pasting it into a chatbot, and switching back. We also

Jev and the System One Model: RLCD, intelligence/, reliable AI, & the end of chat-first AI https://www.latent.space/p/jev @typesafeai CEO @CompleteSkeptic explains why AI can solve extraordinarily har

Meet Wally. Our inference stack for open frontier models, built to be the fastest place to run them. Performance snapshot: GLM-5.3 Flash: 380 tok/s. GLM-5.3 Max: 790 tok/s Qwen3.8-27B: 485 tok/s DeepS

arXiv:2609.21493v1 Announce Type: new Abstract: Multimodal large language models, or MLLMs, perform well at visual understanding and structured generation, yet these capabilities do not establish whet

The Information: Sources: OpenAI is developing features to counter SpaceX's Grok Bot “teammates”, and has discussed creating a personal AI assistant to compete with Meta's Muse — OpenAI pioneered AI t

I shipped something I've been building for the last few weeks : phantom-kv , a refusal-removal system for large language models that doesn't touch a single weight. Instead of editing the model, it loa

I forked llama.cpp to implement Declarative Attention (arXiv:2609.02737, Google DeepMind and KAIST AI). The model declares in its own output which context chunks it needs <focus magic_chunks='N'>), an

Impressive paper showing the impact of a good harness. Improves Qwen3-8B from 41.2% to 91.8% on long-horizon robot tasks without any change to the model. The gain comes from the harness. GAVEL keeps a

Sharing my successful experience here. My setup: RTX 5060 Ti 16GB (x16 PCIe slot) RTX 3060 12GB (x4 PCIe slot) RAM: 32GB CPU: Ryzen 5 5600 OS: Linux So I have a total of 28GB. Average tk/s: 50, and th

Note: This writeup was put together with the help of AI. So I've got dual AMD MI50 32GB cards. If you know these cards, you know the story - AMD dropped official ROCm support for gfx906 after ROCm 5.7

I have always posted about budget builds on here, and often asked how we are going to run the next big models. Often Plenty of downvotes too or folks telling me that it's not running if I'm getting 5t