Skip to content
AI NEWS JOURNAL
AI, IN PERSPECTIVE

Search AI news

Search headlines, summaries and articles to connect the dots.

Newest first

UniExit: A Unified Latency-Aware Early-Exit Runtime for Heterogeneous Edge AI

arXiv:2611.09159v1 Announce Type: new Abstract: Edge applications increasingly combine CNN-based perception with edge-deployable Transformer-based language models, but deploying such heterogeneous workloads on resource-constrained devices is challenging because inference latency varies with both input difficulty and runtime hardware state. Early-exit inference reduces computation by terminating at intermediate stages, yet existing methods are often model-specific and rely on static latency profiles that become stale under dynamic edge conditions. This paper presents UniExit, a unified latency-aware early-exit runtime for heterogeneous edge AI. UniExit abstracts CNN blocks and Transformer layers as ordered stages and applies a common latency-aware routing policy while retaining model-specific quality mechanisms. It predicts the cost of reaching the next exit from runtime state and inference progress to make budget-aware continue-or-exit decisions. UniExit combines runtime-state-aware latency prediction, cache-based prediction reuse, trend-aware fallback, and CNN-specific exit stabilization to provide low-overhead and robust online control. We implement UniExit on multiple edge and edge-cloud platforms and evaluate it with representative CNN and Transformer-based language-model workloads. Across single-workload, concurrent heterogeneous, and edge-cloud settings, UniExit consistently improves deadline satisfaction and latency-accuracy trade-offs over existing progressive inference baselines, while reducing redundant exit-module computation and task-level energy consumption.

What Are LLM Agents Actually Learning on Low-Budget Molecular Optimization Benchmarks?

arXiv:2611.08522v1 Announce Type: new Abstract: Recent work reports large gains for LLM agents on low-budget molecular optimization. We measure the contributions of online optimization, retrievable task information, and permissive oracle landscapes. On PMO at 50 oracle calls, a GPT-4o mini agent's sum AUC falls from 14.44 to 4.85 after task names and curated reference compounds are hidden. An identity-blinded condition that states only the oracle form (e.g., similarity to a hidden reference) scores 5.39, while a deterministic retrieval-only baseline with no LLM reaches 12.16 (84% of the full agent). A full 23-task 2x2 ablation and a three-seed replication on eight tasks support this interpretation. On a protein-targeted benchmark with an AutoDock Vina composite oracle, a single-edit random walk is a strong lower bound; across three seeds it obtains composite 0.636 +/- 0.008 and docking improvement -1.11 +/- 0.04 kcal/mol. As a positive control, a retrieval-free population agent exceeds REINVENT at a matched 50-call AUC budget, with its strongest result at 100 calls. These results motivate two inexpensive sanity checks: retrieval-safe PMO prompts and a random-walk floor for docking-based optimization.

b12207

A new fast path for d_state=64 has been added to the SSM_SCAN kernel in ggml-cuda, improving generation throughput for Zamba2-7B-Instruct from 11 to 15.1 tok/s—approximately a 35% increase—by eliminating CPU fallbacks for this state size. The change was verified with real-world Zamba2 hyperparameters and maintains correctness, though a separate 28% of scan calls still fall back to CPU due to an undiagnosed issue.

b12208

The ggml library has been updated to make tied selections in the radix Top-K algorithm deterministic on CUDA and Metal backends by replacing atomic compaction loops with a deterministic scheme using simdgroup scans and coalesced memory accesses. This change ensures byte-identical outputs across runs, improves performance by 20-45% on M5 Ultra hardware, and enables deterministic testing for Metal and ROCm. Redundant CUDA kernel resets were removed, and Vulkan/Metal changes were reduced or skipped.

Show HN: Coldframe, a honeypot that poses as a supermarket fridge controller

I built Coldframe, a honeypot that pretends to be a supermarket refrigeration controller. Inspired by the events on industrial fridge installations in the past weeks, I wanted to see what people would try against a convincing fake.To learn more about this I built a honeypot that resembled perfectly the Danfoss AK-SM 800A and Dixell XWEB interfaces. Both of them are recognized on Shodan as real ones (an not as honeypots). It’s one Python file using only the standard library, and it builds on eppser/danfoss-ak-sm-decoy and public research by Synacktiv, Claroty and Ironhackers.Things that were fun to get right:- Matching real units. I used Shodan records of real devices to match header order, page titles, certificate fields and TLS versions. The Danfoss persona sends nginx headers in the same order as a real AK-SM 800A. The Dixell persona sends no Server header and pads its page to exactly 8,202 bytes.- Recognising exploit attempts. It recognises CVE-2025-41450, the Danfoss “code of the day” login bypass, and records whether the submitted code is correct for today. It also tags attempts targeting three other Danfoss CVEs and the XWEB vulnerabilities.- Letting them in. Default logins open a fake store dashboard with defrost buttons, so you see the second move, not just the first. Nothing behind it is real.Two instances have been live for a day. So far they’ve only seen generic scanners and Shodan’s crawler, which is called refrigerator.census.shodan.io.Feedback welcome, especially from anyone who has seen these devices in the field! Comments URL: https://news.ycombinator.com/item?id=50406292 Points: 15 # Comments: 1

Show HN: Neural Network Formal Verification

Article URL: https://github.com/giovanni-norbedo/awesome-nn-formal-verification Comments URL: https://news.ycombinator.com/item?id=50412710 Points: 1 # Comments: 0

Show HN: Throng, a project first, terminal 2nd IDE for command line people

I've been beavering away at this since July, and its finally getting to a point where its stabilising.It started Windows only, but I recently converted over to Linux, so I've been ironing out the kinks in the last week. You can install it via apt or rpm, or .exe or .msi on Windows.Many more features to come, including native agent support and more preview features (images, PDF, document formats). Claude works pretty well in it, but I've not tried many other agents yet.I rarely leave throng now - 99% of my work happens inside it. I stopped using Visual Studio around 10 months ago (in favour of VSCode), and now I don't even use VSCode. Comments URL: https://news.ycombinator.com/item?id=50412226 Points: 1 # Comments: 0

Cisco Catalyst C1300-24XTS Review A Neat 24-port 10GbE Switch

In our Cisco Catalyst C1300-24XTS review, we take the switch apart to see how it works and test it under full load to see if it holds up The post Cisco Catalyst C1300-24XTS Review A Neat 24-port 10GbE Switch appeared first on ServeTheHome.

How Greenlight Financial built an autonomous SOC solution on Amazon Bedrock AgentCore

Greenlight Financial resolves more than 90% of up to 4,000 monthly security and risk cases without human intervention using Luna, a tiered multi-agent system on Amazon Bedrock AgentCore. This post covers the architecture, episodic memory for continuous learning, and the expansion from SOC operations into financial risk triage.

SageMaker Inference now supports Multi-Instance GPU to maximize GPU utilization for model endpoints

Amazon SageMaker Inference now supports GPU partitioning with NVIDIA Multi-Instance GPU (MIG), which divides one GPU into hardware-isolated partitions so you can run a different model in each or take a right-sized fraction of a newer accelerator. This post shows how two optional API fields place one model per partition, and what we measured against a whole older GPU.

Serverless MCP server for AI agent data lake access on Amazon ECS Fargate

Learn how to build a serverless Model Context Protocol (MCP) server on Amazon ECS with AWS Fargate that gives AI agents secure, metadata-driven access to an enterprise data lake. Exposing a new dataset requires only a JSON file, with no per-dataset integration code.