All AI news

AIXiaomi MiMo reports that tool-call repetition in MiMo-V2.6 reached over 0.05% of responses across agent harnesses, causing stalled agents and wasted context. The team traced the cause to an RL flooding penalty set at 32 calls per turn, which missed smaller excess behavior, and replaced the approach with a specialized teacher distilled via MOPD. Repetition rates for both Pro and Flash dropped substantially, at roughly $90,000 versus an estimated $2.31 million for the alternative fix.

Why it matters: The post traces an agent failure to a reward blind spot and compares the costs of two fixes, offering a transferable debugging method for RL-trained tool-calling models.

AISebastian Raschka's fifth Reasoning from Scratch video explains log-probability scoring and self-refinement for LLMs. It covers token probabilities, PyTorch implementation, numerical stability, and a self-refinement loop evaluated on MATH-500, with the log-probability concept linked to cross-entropy loss in pre-training and distillation.

AISGLang's /v1/score endpoint returns scores for exact requested labels such as Yes/No or A/B/C, and its multi-item scoring (MIS) computes shared context once while keeping candidates isolated. On Qwen3-8B, 16-candidate p95 latency dropped from 54.1 ms with Generate to 20.6 ms with MIS. On Qwen3-0.6B, MIS p95 stayed under about 100 ms as load rose, versus seconds for Generate and SIS.

AICanvases in the GitHub Copilot app are customizable interfaces that you and the agent share, such as kanban boards, dashboards, or checklists. You create one by running /create-canvas and describing the workflow, what you can do in the interface, and what the agent can do. Changes made by either you or the agent appear immediately in the shared canvas, and completed canvases can be saved as reusable extensions.

AIGoogle Cloud has launched a managed reinforcement learning fine-tuning service (RLFT) that lets customers adapt Gemini models using a reward function they define instead of labeled answers. Users supply prompts and a reward function, while Google handles the RL infrastructure and proprietary model internals. The guide advises exhausting prompting and supervised fine-tuning first, and notes that RLFT suits tasks that are easy to score but hard to demonstrate.

AIUsing the Neuron Kernel Interface, Reactor and Amazon's Neuron Science team built a kernel-centric path to real-time autoregressive diffusion video generation on Trainium. They addressed dynamic shapes, memory access patterns, and cache management, which are hard for generic compilers, and developed techniques intended to generalize across models.

AIThis time I'm stunned: a product promo video made directly with Opus 5.5, with no AI video, AI music, or AI image generation models connected from start to finish, outputting both 16:9 and 9:16 versions in one go... The full prompt is posted directly below 👇 (1/3)

AILlamaIndex argues that extraction confidence scores are useful only when they help decide what can be automated and what needs human review. Using ExtractBench, the post compares extraction systems after confidence filtering, reporting that LlamaParse Agentic Plus reached 66.48% recall on expected fields at a 97% precision target. The post covers confidence cutoffs, precision versus recall, score coverage, score granularity, and human review volume.

AIMicrosoft's Foundry Agent Service preview lets developers attach a named, ordered egress policy to a hosted agent, allowing only approved destination hostnames. The walkthrough uses an invoice agent, an Audit-mode RAI policy with a Deny default, and Allow rules for two finance and vendor hosts, configured outside the agent code. Network egress controls are preview features, not GA, with no preview SLA, and are not intended for production use.

AILovable describes how its Chats feature lets a workspace-level chat agent hand work to project builder agents and receive progress back. The design records each agent's history as an append-only, forkable trajectory, and passes messages through durable inboxes that activations wake. Agents can suspend at iteration boundaries and resume on freshly deployed nodes without killing long-running runs.

Why it matters: The post details how trajectories, inboxes, and activations let agents share work and resume after deploys, useful for designing comparable agent systems.

AIKling AI's blog identifies six limitations of current AI video generation, including temporal consistency, character consistency across shots, unrealistic physics, long-form generation, fine details and text, and prompt control. It recommends workarounds such as reference images, shorter single-action clips, storyboards, and adding text or logos in post. The article says Kling VIDEO 3.0 and VIDEO 3.0 Omni offer reference-based subject consistency to help reduce these problems.

AIThis guide compares six watermark removal tools, including Kling AI, HitPaw Watermark Remover, Picsart, Adobe Photoshop, Fotor, and Cleanup.pictures, based on mark type and editing control. Kling AI's IMAGE 3.0 uses natural-language prompts and annotated images to rebuild marked areas in context, while IMAGE 3.0 Omni adds refinement with native 2K/4K output.

AIMemory and supported Skills can live there too. It needs host setup, permissions, and matching OpenClaw versions.

AIThe agent checks earlier actions before continuing interrupted helpers; chats you stopped stay stopped.

AIBuilding a model of the program, targeting a tricky state machine or race-prone part of the code 2. Finding counter-examples in the model. These are suspected bugs 3. Reproducing the bugs 4. Fixing the bugs in the code It's not that the whole codebase is formally verified (yet!..), more that the hairiest parts of the code are modeled, checked for counter-examples, and fixed.

AIAnthropic says it made claude.ai 3x faster in two weeks, and the same speedup applies to the Desktop app. A linked blog post explains how the team used Claude to measure, debug, and improve performance, and includes prompts and methods.

AICursor's Eric Zakariasson shared a prompt for improving an LLM agent harness to lower token cost per completed task without losing quality. The prompt covers the system prompt, tool definitions, cache layout, tool results, compaction, and subagents, and reports that one team's round of these changes cut overall token cost about 7%.

Why it matters: The prompt gives a concrete checklist for cutting agent token cost per completed task, with tested figures on cache layout, tool offloading, and compaction.

AI…techniques as described in the blog post.

AIGitHub rebuilt the pull request view in the GitHub Copilot app to keep review fast on very large diffs, testing it on an open source pull request with 2,200 files, over a million changed lines, and more than 400 inline review comments. The core difficulty is that review comment heights can only be measured at render time, which breaks the fixed-geometry virtualization used for code-only diffs. GitHub split the document height into a deterministic code domain and a separately measured domain for comment blocks.

AIif you're building agents, i highly recommend reading this (or giving it to your agent and ask it to implement the findings)

AIMicrosoft Research reports that running physical AI inference on onboard GPUs can limit robot performance and battery life, while offloading inference to edge or cloud GPUs improved results in mobile manipulation tests. In its evaluation, smaller onboard GPUs slowed mapping and planning by up to 383% compared with an A100, and large onboard GPUs such as Jetson Thor drained robot batteries by up to 160%.

Why it matters: The study measures how offloading robot inference to edge or cloud GPUs changes task success, battery life, and model size, offering evidence for infrastructure design.

AIPeople kept calling the wash sequence “real.” It came from a reusable UGC product-review template in Awesome Seedance. The open-source repo has 463 sourced cases, 264 cross-model retest runs, and 25 templates you can adapt:

AIAwesome Seedance is now open source: • 463 cases traced to their original posts • 264 cross-model retest runs with public verdicts • 25 copy-ready templates + 60 installable AI-video Skills Pick a clip, copy a template, swap the brackets, and generate. GitHub:

AIThe author compiled 436 AI videos made with Seedance into a prompt collection and a Skill, now available on GitHub as awesome-seedance. The post asks GitHub users to star the repository to help it reach a thousand stars in a day, and notes the X post has about 400,000 reads.

AIThe Together AI blog shows how to fine-tune a Qwen3.5 4B base model into a classification model using about 38,000 examples sampled from six Hugging Face datasets, at a training cost of roughly $17.0. The tutorial covers cloning the tev1 repository, normalizing data with provided scripts, launching a Together AI fine-tuning job that takes about 25 minutes, and deploying the result to a dedicated H100 endpoint.

AIBoris Cherny used Opus 5.5 to formally verify the Claude Agent SDK with Lean, and short prompts produced 16 PRs fixing bugs and race conditions. He also combines Lean and TLA+ to find issues in data flow, concurrency, and state management, and says Claude is strong in both languages even though he does not know them well.

AIA prompt shared by Alex Albert asks Claude to recreate San Francisco's Market Street as it stood on April 17, 1906, before the earthquake, using Blender. It requires building a source file from Sanborn fire insurance maps, the Miles Brothers film, period photos, and USGS topography, with reusable Blender Python generators for facades, street lamps, and vehicles, ending in a 10-second video up the street.

AIIts better 3D modeling and vision mean you can build an entire world from a single prompt. Historically accurate San Francisco Market street in 1906, pre-earthquake

AIUnsloth says the 7B Qwen-Image-2.1 text-to-image and editing model can run locally on 12GB VRAM using its GGUF builds. It also states that the model performs on par with Nano Banana 2.0, and that Dynamic FP8 can run on 6GB of VRAM via offloading for higher quality. The image lists int8 at 7.26 GB with mean LPIPS 0.064 and fp8 at 7.12 GB with mean LPIPS 0.112, and says int8 is the default.

Why it matters: The post gives concrete local-run settings, VRAM figures, and GGUF and FP8 options, which helps readers judge whether the model fits their hardware.

AIReally cool to see MiniCPM5-2B being used in a practical multi-agent workflow like this — especially with the workers actually handling matching, short payments, duplicate references, and disputes through tool calls. Appreciate all the work you put into testing and documenting this. Such a nice case for the community 🙌