#Architecture

Thore Graepel, a core member of the AlphaGo team, has left Google DeepMind and is now arguing publicly that today's large language models lack the structured reasoning machinery that made AlphaGo's famous Move 37 possible - and that fixing this requires a fundamental architectural rethink.

A new paper from Intel researchers introduces BITCOS, a distribution-aware storage layout for ternary LLMs that exploits the real-world prevalence of zero weights to push effective bit-widths well below the conventional 1.585-bit floor - delivering up to 1.28× throughput gains in production serving.

A new runtime technique called Speculative Macro Commit (SMC) uses a fast drafter model to pre-execute multi-step tool-call chains in parallel, slashing wall-clock latency for tool-using LLM agents by up to 44.9% compared to sequential execution.

A new EMNLP 2026 paper shows that swapping flat text or pairwise triples for hypergraph-structured state serialization yields consistent gains in how well small language models learn to simulate text environments - with the biggest wins appearing under distribution shift.

A paper accepted at ARES 2026 proves that LLMs can pass every full-precision safety audit while harboring backdoors that activate only after INT8 or 4-bit quantization - exposing a critical gap between where models are validated and where they are deployed.

A new arXiv paper proposes Meta-Persona Anchoring combined with Filtered Temperature Scaling to combat the well-documented tendency of LLMs to converge on near-identical responses - even when temperature is cranked up to maximum.

A new arXiv paper introduces the CSE benchmark to study how LLM instruction-following collapses sharply - not gradually - as the number of simultaneous constraints grows, revealing a phase-transition structure with direct consequences for production AI systems.

LinkedIn published a paper detailing a production agentic support system that continuously rewrites its own prompts and retrieval strategy - with A/B test results showing a 30.6-point jump in routing accuracy and a 9-point gain in self-serve resolution.

Multi-agent LLM systems fail at rates that are poorly understood - and the research community still lacks a principled way to measure why at the orchestration layer. An early community proposal called OrchestraBench aims to change that, introducing metrics for cascade resistance, failure recovery, and task decomposition quality.

A deep dive into the AWS-backed stack powering the rebuilt Cruiser Aviation platform: from Next.js 16 App Router conventions and Aurora Serverless v2 to a dual-country billing abstraction and a pure-functional FIFO hours ledger.

A flight school's production SaaS was rebuilt from scratch on AWS in 11 days after a year of compounding technical debt - this is the story of the autopsy, the decision to rewrite rather than refactor, and the planning, schema, ETL, auth migration, and cutover that made it work.

How an internal LinkedIn content-automation platform deleted its Railway worker, Redis, and BullMQ entirely - replacing them with Vercel Cron and a Postgres-as-queue pattern - shipping ~2,400 fewer lines, ~85% fewer scheduled invocations, and eliminating a whole class of silent distributed-systems bugs.

Four products - marketing site, AI sales concierge, CRM, and customer portal - for a Swiss aircraft dealership, shipped as one Next.js app on one Postgres. A study in breadth without sprawl.