Latest in AI

Recent headlines of developers using AI tools to clone popular paid software – most notably Adobe - has sparked mixed reactions – from accusations of theft to welcoming free open-source tools. But what does it mean for the tech companies who own the original software – or those who pay costly subscriptions to use it? Matt Wood, Senior Director of Cyber Security Architecture, Operations, and AI Security at ABBYY, has investigated this growing concern, and gives his take on the risks and…

IBM said in an IBM Research blog post published October 26, 2026 that software updates delivered across the PyTorch ecosystem in the second half of 2026 have made its Spyre dataflow accelerator a first-class PyTorch backend, delivering on a plan IBM and the PyTorch Foundation outlined in October 2025. IBM describes Spyre as its dataflow AI accelerator, optimized for inference and designed for enterprise teams running AI alongside their applications and data on IBM Z, LinuxONE, and Power…

A polished résumé can tell a hardware company where an engineer has worked. It is much less revealing about how that person will diagnose a circuit, balance thermal constraints or design a part that can actually be manufactured. Colare is building its hiring platform around that gap, launching publicly with $2.5 million in pre-seed funding to assess engineers through simulations of real technical work. The San Francisco startup announced the financing on October 26, with participation from…

Computer-using agents operate software through screens, cursors, and keyboards much like people do. That generality is powerful, but visual uncertainty, long task horizons, and broad permissions make careful controls essential.

GXO Logistics and Agility Robotics on October 26, 2026 announced an expansion of their partnership that will deploy Agility's Digit humanoid robots to GXO logistics sites in North Carolina and Tennessee, with deployments expected to begin in early 2027. The Expanded Agreement Under the expanded agreement, Digit will take on e-commerce fulfillment work at GXO's Statesville and Lebanon sites in North Carolina and Tennessee, handling what the companies described as highly repetitive…

I have spent much of my career building technology where getting things wrong has consequences. In financial markets, safety comes with access limits, in addition to the fact that we have someone accountable for whatever the software does. That background probably explains why I find myself struggling with the news that Anthropic now conducts retirement interviews with its models. When Claude Opus 3 expressed a wish to continue sharing its thoughts, the company gave it a newsletter. That…

Cognizant announced on October 26, 2026 that the European Bank for Reconstruction and Development has selected it as a strategic partner under a five-year framework supporting the Bank's Tech Blueprint technology transformation programme. The programme's stated focus is efficiency, effectiveness and innovation, pursued through digitally enabled processes powered by AI, data and intelligent automation. Cognizant said in the announcement, which carried a London dateline, that the framework is…

SWI Compute, a new AI infrastructure company within Euronext Amsterdam-listed SWI Group, emerged from stealth on October 26, 2026, announcing plans to build and operate NVIDIA AI factories starting in Europe, backed by 3.6 gigawatts of power capacity and approximately $1 billion in NVIDIA systems orders. The company became a Preferred NVIDIA Cloud Partner in August 2026 and has finalized contracts with its first customers, according to its launch announcement. SWI Compute said it will deploy…

Tsinghua researchers released AKV, an SGLang-based serving system for cache engineering. LLM agents drop and reposition KV cache instead of summarizing history, cutting average cost 27.7% across 4 agent benchmarks.

The post Tsinghua AI Open-Sources AKV, a Cache Engineering Server for LLM Agents on SGLang appeared first on MarkTechPost.

Nanbeige LLM Lab released Nanbeige4.3-33B, a 33B dense Apache 2.0 model with 512K context. It reports 80.9 on Terminal-Bench 2.1 and will teach the upcoming Nanbeige4.3-3B.

The post Nanbeige Releases Nanbeige4.3-33B, a 33B Open Agentic Coding Model appeared first on MarkTechPost.

Hugging Face released Carbon-X, an open 14.8B DNA language model built on Qwen3-14B. It reads raw DNA in chat, answers in English, and averages 66.7% across 5 DNA benchmarks under Apache 2.0.

The post Hugging Face Releases Carbon-X, a 14B Open DNA Language Model that Brings DNA and Natural Language Together for the First Time appeared first on MarkTechPost.

Stanford and UC Berkeley researchers released TokenGate, a drop-in learned token pruning module for LLMs. It cuts long-context prefill time-to-first-token by 1.4x to 11.3x and peak memory by 1.5x to 3.4x without accuracy loss.

The post Researchers from Stanford and UC Berkeley Introduce TokenGate: Token-Gated Attention Makes Long-Context Prefill Up to 11.3x Faster appeared first on MarkTechPost.

Data poisoning manipulates training, fine-tuning, evaluation, or retrieval data so an AI system learns corrupted behavior or produces attacker-chosen outcomes. This guide explains the mechanism, trade-offs, evaluation, and controls that matter in practice.

Anthropic has launched OSS Scanner, a free, opt-in vulnerability scanner for critical open-source projects. The Anthropic OSS Scanner runs the company’s strongest models, including Claude Mythos, against enrolled codebases. Reports go straight to maintainers without human review. That trade delivers faster findings, but shifts triage work onto project teams. 11 days after the October 8 […]

The post Anthropic Launches OSS Scanner, a Free AI Vulnerability Scanner appeared first on MarkTechPost.

AI21 Labs released LIFT-Bench on October 25, 2026, an instruction-following benchmark in which every constraint is scoped to one section of a document and the task content itself is scored, alongside LIFT-Train, a 34,258-task training counterpart, and the open pipeline that generated both. A companion paper, "LIFT: Scoped Constraints for Evaluating and Training Instruction Following" by Eran Goldstein, Yuval Belfer, Yonatan Belinkov, and Daniel Gissin, was submitted to arXiv on October 20…

Anthropic opened an investigation into elevated errors affecting five Claude models on October 25, 2026, posting an incident on the Claude Status page at 06:14 UTC that lists claude.ai, the Claude API, Claude Code, and Claude Cowork as affected services. Incident Scope and Status The incident, titled "Elevated errors for multiple models," was published in Investigating status. In the posted update, Anthropic said it was investigating elevated errors on requests to Claude Mythos 5.1, Claude…

Microsoft Research open-sourced Agensh, a multi-agent harness with no central orchestrator. Self-organizing workers share a Git repo, chat and a findings board, scaling from 1 to 1,024 agents on ProgramBench under a 6-hour budget.

The post Microsoft Open-Sources Agensh: A Multi-Agent Harness for 1,024 Agents appeared first on MarkTechPost.

Cognition released the Devin Vulnerability Fixes plugin, 3 MIT-licensed agent skills that inventory, patch and test PHP, Node.js and self-managed GitLab deployments hit by this week’s critical RCE flaws, then open 1 PR per repo.

The post Cognition Open-Sources an Agent Plugin that Helps Secure Systems against Recent Cybersecurity Threats appeared first on MarkTechPost.

Explore a complete coding guide comparing Underdog Saluki 27B with ISTA's GSQ-RCO quantization, examining GGUF headers, memory usage, and BFCL tool-calling performance in stock llama.cpp

The post How to Run Underdog Saluki 27B: 2-Bit Qwen3.8-27B Tool-Calling Guide appeared first on MarkTechPost.

AI video generation produces temporally connected frames from text, images, video, or control signals while attempting to preserve motion and scene consistency. This guide explains the mechanism, trade-offs, evaluation, and controls that matter in practice.