All AI news

AIMike Krieger, Anthropic's account owner, used an interactive Claude artifact to visualize the Iron Tangle level from Dungeon Crawler Carl, saying it helped him finally understand the layout. The post shares a link to the artifact, with no further details about its features.

AIJetBrains describes building Air Context, a RAG pipeline that gives LLM agents semantic code search over real repositories instead of grep. The first installment covers parsing, chunking, and vectorization, arguing that fixed-size line chunks split related code and that structure-aware chunking using language grammar produces better retrieval units.

AIZ.ai reports that GLM-5.3 helped build and optimize the inference infrastructure for GLM-5.3-Flash. The system went from first successful run to production readiness in under two weeks, with end-to-end throughput tripling over the initial baseline. The team credited dense feedback from local correctness tests, execution traces, microbenchmarks, and end-to-end measurements for enabling targeted hypothesis testing.

AIMosaic, Ponder Studio, and Revyl used early access to Google's Gemini Flash models to test agentic video understanding on long footage. Mosaic reports a 97% cut in median token usage and nearly double the ability to handle complex edits, while Ponder Studio reports a 0.967 F1 score and about 72% lower token costs for B-roll selection. Revyl says the approach improved mobile UI bug-catching accuracy by 65%. The capability is available now for video uploads and YouTube videos via the Gemini API.

AIAnthropic is hosting office hours for teams interested in using Claude Tag for on-call work, and it is asking Team or Enterprise plan users to share triage feedback. Claude Tag can start investigating when a Slack alert fires by pulling metrics, diffing deploys, and checking flags to propose a likely cause and fix. Sign-up is through a Google Calendar booking link.

AITogether AI says teams can migrate from closed to open-source models in weeks to months, not years, using a managed service. The guide recommends defining the use case, shortlisting candidates with relevant benchmarks, and replaying existing traffic against them to test accuracy, cost, and latency against current targets.

AIPart 2 of the zero-trust agents series moves security checks from agent code to the Gemini Enterprise Agent Platform runtime. Model Armor screens prompts and responses, Semantic Governance Policies judge proposed tool calls against intent and business rules, and Agent Anomaly Detection flags multi-turn drainage that single-turn checks miss. The same Customer Support and Returns Agent from Part 1 is used, with the companion demo open-sourced on GitHub.

Why it matters: The post walks through a concrete refund agent under four attacks, showing how screening, intent judgment, and anomaly detection each catch what the others miss.

AIThe vLLM team trained a DSpark speculative decoding draft model for Kimi K3, a 2.8T-parameter model, using the Speculators library on GB300 NVL72 hardware. They added a MooncakeHiddenStatesConnector to stream hidden states from disaggregated vLLM inference nodes to training nodes across multiple machines. The released speculator raises single-stream interactivity from about 110 to about 435 tokens per second per user on math reasoning, with up to about 3.5x higher output throughput under concurrent load.

Why it matters: The post shows how hidden-state extraction and Mooncake transfers let a 2.8T-parameter model's speculator be trained across multiple nodes, a reusable pattern for similar setups.

AIA hands-on guide by @pandemicsyn walks through writing evals that detect when an AI agent claims to have completed a task it never did. Working through a demo agent that fails on purpose, the author refines the checks until they can distinguish real work from mere claims of work. The post includes a coding agent skill that can guide readers through the exercise.

AIAugment Code reports that size-adjusted output per active developer rose from 12.3 to 55.7 between November 2025 and July 2026, while median time to merge fell from 11.2 to 3.1 hours. The post says the company added specialized agents wherever work was piling up, across planning, review, verification, feedback, and incident response, and kept engineers responsible for product decisions, architecture, and production risk.

Why it matters: The post pairs internal productivity and quality metrics with the order in which agents were added, showing how review and verification bottlenecks shaped a software delivery pipeline.

AIGoogle Developers Blog describes autofinetune, a project applying autonomous agent loops to LLM post-training with Tunix, Gemma, and Cloud TPUs. In an SFT case study on FunctionGemma, an agent ran 20 automated experiments on a Cloud TPU v5e-1 to adjust LoRA settings, optimizers, and learning rates. In a GRPO case study on Gemma 3 1B for GSM8K math reasoning, the agent ran 40 experiments on a Cloud TPU v6e-1 and improved total reward by about 10%.

AIGiles Thomas extended the GPT-2-style code from Sebastian Raschka's "Build a Large Language Model (from Scratch)" into a 6-expert, 2-active mixture-of-experts model and trained it from scratch over 8 days. Raschka praised the project as interesting LLM work done on a single GPU.

AIHugging Face announced a broadcast titled "Training Agents 4: From reward functions to environments" on X. The post gives no further details on its content or speakers.

AIRadixArk links to a blog post on DeepSeek V4.1 at lmsys.org. The post text provided contains only the URL, so no further details on the model or its performance can be reported.

AIFireworks AI describes a four-stage path for teams moving from renting closed frontier models to training their own, starting with API use and prompt, context, and harness engineering. The post uses the UIPad computer-use dataset to show that Kimi K3 ties GPT 5.6 Sol overall at 87.7 but wins three of four categories while costing about half as much, suggesting routing. After roughly three hours of training on the training split, the tuned Kimi K3 outperforms GPT 5.6 Sol on the held-out test set.

AISherwin Wu, an OpenAI executive, says employees now use Astra to vibe-code custom forms, deploy them on ChatGPT, and share the links instead of Google Forms. The post describes this as an emerging internal use case around ChatGPT sites.

AIGranola says users can add personal jargon in their own settings. The post provides no further detail on how this works or what it affects.

AIMistral AI helped a European energy operator migrate 40,000 lines of Fortran 77 to C++ for a reservoir simulator with no test suite. The post explains a parity harness that checks numerical agreement between the two codebases, and a workflow where agents coder, tester, and reviewer migrate modules under human review. Its authors note the approach covered the self-contained first sprint of 40,000 of 300,000 lines and that dependent systems would bring additional challenges.

AIAmy, a 58-year-old with no coding background, built Wally, an AI voice aid app using Manus, for her brother Jamie after cancer surgery at Brigham and Women's Hospital left him unable to speak. The app combines one-tap urgent and everyday phrases with golf news, live tournament scores, and family messages, all placed on his phone's home screen.

AIGoogle Developers Blog argues that teams building AI coding agents should replace end-to-end benchmark scores with behavioral evaluations that test discrete, observable actions. Examples include asking clarifying questions on underspecified prompts, running a local validator before marking a build change complete, and consulting live search for current information. The post recommends fast, deterministic unit-style checks, outcome-based LLM-as-a-judge checks for complex tasks, and batch runs that track aggregate pass rates over time.

AIInferact says months of vLLM tuning for agentic workloads, validated on SemiAnalysis's AgentX benchmark, let open-source models reach up to 130K tokens per GPU-second. The company claims this is 106 times cheaper than Opus 5 API pricing. The work is described as part of a vLLM blog post covering architecture, framework, and runtime optimizations.

AIGoogle AI Studio is encouraging developers to start building apps with Gemini 3.8 Flash, linking to a guide on ai.dev. The post offers no benchmarks, pricing, or feature details beyond that invitation.

AITinker has published a cookbook recipe for fine-tuning a model to predict event probabilities from reference information. The recipe starts with a dataset from ProphetArena and can be adapted to custom data, with code available on GitHub.

AIAndrew Ng presented an AI Engineering Skills Map for using coding agents such as Claude Code, Codex, Cursor, OpenCode, and Pi. The workflow he describes covers planning, execution, and deployment with monitoring, and he identifies five key skills: directing the workflow, enabling agent autonomy, reviewing the work, customizing the agent and its environment, and coding agent foundations. The source says these skills matter more as the agents evolve quickly.

AIGoogle's Gemini Enterprise developer experience team tested agent governance workflows without internal shortcuts and fixed friction points across its agent governance products. Fixes included documentation stating that enabling the Identity-Aware Proxy API is a hard requirement, auto-allowing essential Google-managed platform APIs in the Agent Gateway, and adding Private Service Connect and Cloud DNS setup guidance for Semantic Governance. The team also published ready-made Logs Explorer queries for monitoring Agent Gateways and Content Security.

AIDatabricks used Unity AI Gateway tracing and Genie One to find seven small MCP-server bugs and eliminate an estimated $1.2M in annual wasted AI spend and lost productivity within an hour. The bugs drove about $499K per year in wasted tokens, roughly 12,000 engineering hours per year in agent wait time, and 1,409 tool errors in a single 24-hour window. Matei Zaharia argues that analyzing tracing data for AI workloads will become a routine form of operational data analysis across companies, much like finance and security.

AIRadixArk published a blog post on post-training with Miles aimed at understanding and generating the multimodal world. The post's main text is only a link, so no further technical details, figures, or benchmarks can be confirmed from the source.

AIPrime Intellect reports that rebuilding RL weight transfer for GLM-5.2 on NIXL and ModelExpress cut sync time from 86.1 seconds with NCCL to 3.9 seconds in its fastest setting. The method traces vLLM's loader to find each tensor's runtime layout, then reads only the needed source bytes over RDMA and replays the rest locally. Most remaining latency comes from vLLM's pause consensus, which the team reduced by syncing every wave instead of every 32.

AIMeta Engineering describes an AI agent for a compliance domain that stores expert knowledge in structured, auditable files and separates it from reasoning procedures called recipes. Expert feedback is diagnosed, compiled into verified text edits, tested against regression suites, and reviewed by humans, all without retraining the underlying model. Meta reports that domain experts rated outputs useful almost all the time and that assessment time fell from days to minutes.