#Deployment/Engineering

AIMiniMax has upgraded its Agent, renamed Mavis, and introduced Agent Teams that run multiple role-based Agents in parallel on desktop. The team uses Leader, Worker, and Verifier roles so complex tasks can be split, checked, and reported at key checkpoints, and it merges TokenPlan and Agent Plan into one subscription with credits shared between Agent and API. The post also discusses the added token, handoff, and retry costs of multi-Agent work, and says the Agent will be open-sourced alongside MiniMax M3.

Why it matters: The post explains why multi-Agent helps long tasks and where its verification, token, and aggregation costs come from, useful for judging when a team setup beats a single Agent.

AIMiniMax says the M2 series could not generate the rare token "嘉祺" in names like Ma Jiaqi, and its investigation traced the cause to post-training. The company found the token was learned in pretraining, but low coverage of rare tokens in post-training data caused lm_head vectors to drift. Adding synthetic full-vocabulary repetition data restored generation for these tokens and reduced Japanese-to-Russian mixing from 47% to 1%.

Why it matters: The post traces a specific token failure through tokenizer, embedding, and lm_head checks, showing a reusable way to diagnose post-training generation problems.

AIFunAudioLLM has released Fun-ASR-Nano-2512-hf, a Hugging Face Transformers-compatible version of its end-to-end speech recognition model, which supports Chinese, English, and Japanese. The Chinese coverage includes 7 dialect groups and 26 regional accents, and a separate Fun-ASR-MLT-Nano-2512 checkpoint handles 31-language recognition. Developers can run the model natively in Transformers 5.17.0 without custom model code or trust_remote_code=True.

AIDeepSeek has made its discount on DeepSeek-V4-Pro permanent, after previously extending it until May 31, 2026, 15:59 UTC. The company is encouraging developers to build with the model.

AITri Dao says that after a mathematical rewrite, all transformer operations can be expressed as a series of GEMMs with epilogues. Given a few optimized primitives, LLMs and novice humans can write near speed-of-light kernels for transformer ops. The related CODA work fuses memory-bound surrounding ops into the matmul epilogue, and LLMs can also write CODA kernels approaching speed-of-light.

AICursor is doubling usage for new users invited on the Teams plan for the next month. The offer applies only to users invited by a Teams plan member during that period.

AICognition's Devin AI software engineer can now build, run, and test code natively in its own Windows virtual machine, including migrating .NET Framework apps to .NET Core. The Windows capability is in beta for Enterprise Cloud and Dedicated Deployment customers, with the same SOC 2 Type II and ISO 27001 controls as the Linux version. Citi and Mercedes-Benz are named as existing Devin users.

AIPaddleOCR 3.5 now supports Hugging Face Transformers as an inference backend, letting users run PP-OCRv5 and PaddleOCR-VL 1.5 models directly within the Transformers ecosystem. Users can select it with engine="transformers" while keeping the same PaddleOCR pipeline, which the post says eases integration for RAG and Document AI applications.

AIAwni Hannun highlights a workflow to write PyTorch models, export them, and run them with MLX on Apple silicon. The linked PyTorch post says ExecuTorch now includes an MLX delegate that runs PyTorch models on Apple silicon GPUs, supporting LLMs, speech-to-text, and MoE models.

AIAt Google I/O, Google introduced Gemini 3.5 Flash, which the author says has become part of the daily research cycle. The author says a team of agents in Antigravity 2.0 recreated the original AlphaZero paper and built a playable web version from two prompts, coding the reinforcement learning pipeline in JAX/Flax and training a ResNet model via self-play on multi-TPU pods.

AIEugene Yan shares Cloudflare's description of a vulnerability discovery harness that runs eight stages, from reconnaissance to report writing. The pipeline uses about 50 concurrent agents to hunt for bugs, independent agents to try to disprove findings, and a trace step to confirm whether attacker input reaches each bug. Reachable findings feed back into new hunt tasks before a report is written against a predefined schema.

AICognition has released Auto-Triage in Devin Automations, which lets Devin respond to Slack messages, Linear events, GitHub activity, schedules, and webhooks. Devin can investigate with connected observability tools and the codebase, then post a summary, tag an owner, or open a PR. Devin runs in network-sandboxed environments with added protections against prompt injection and data exfiltration, and a limited-time offer gives $200 in credits for a first automation.

Why it matters: The post shows how an agent handles alerts and bug reports from existing team channels, a practical pattern for teams weighing automated incident response.

AISebastian Raschka reviews recent open-weight LLM architecture changes aimed at reducing long-context memory and compute costs. He covers KV sharing and per-layer embeddings in Gemma 4, per-layer query-head budgeting in Laguna XS.2, Compressed Convolutional Attention in ZAYA1-8B, and mHC with CSA/HCA compressed attention in DeepSeek V4. The article reports that DeepSeek V4-Pro uses 27% of single-token inference FLOPs and 10% of the KV cache size of DeepSeek V3.2 at a 1M-token context.

AIShanghai AI Laboratory's Intern Large Models introduces Intern-S2-Preview, a 35B scientific multimodal foundation model, and says it matches the trillion-scale Intern-S1-Pro on core scientific tasks. The post says it is the first open-source model with material crystal structure generation and strong general capabilities, with shared-weight MTP plus KL loss improving acceptance rate and speed. It is already supported by vLLM and SGLang, with weights on Hugging Face and ModelScope.

AICognition's Devin can now spin up an Android Virtual Device, letting it build, run, and test Android applications directly on its own machine. The new emulator support gives Devin an Android equivalent of computer and browser use, so it can open apps, inspect behavior, reproduce issues, and verify changes. The feature is available now for teams using Devin.

AIThinking Machines, founded to advance human-AI collaboration, says its first bet is interactivity built into the model rather than added as scaffolding around a turn-based core. The company argues that how people work with AI matters as much as how intelligent the model is, and that interactivity should scale with intelligence. The post links to a blog detailing these interaction models.

AIThinking Machines Lab announced a research preview of interaction models that take in audio, video, and text continuously and respond in real time without external turn-detection harnesses. The model, TML-Interaction-Small, is a 276B-parameter MoE with 12B active parameters, paired with an asynchronous background model for sustained reasoning and tool use. The post reports competitive intelligence scores and lower turn-taking latency against GPT-realtime and Gemini Live models, along with new interactivity benchmarks where baseline models largely failed.

Why it matters: The post explains a time-aligned, full-duplex design and benchmarks against turn-based models, showing how interaction and background reasoning can be split across two cooperating models.

AICognition reports that deploying its Devin agent on hardware-in-the-loop and software-in-the-loop workflows cut failure triage time and multiplied test generation at automotive customers. One team reclaimed 2K–4K engineering hours monthly across about 4,000 tickets, while RV Tech rose from 1–2 to 10–15 generated tests per day. Devin also helps convert bottlenecked HIL tests into SIL equivalents to catch failures earlier.

AIBerkeley AI Research describes adaptive parallel reasoning, in which a reasoning model decides when to split independent subtasks, how many concurrent threads to spawn, and how to coordinate them. The approach targets the latency, context-rot, and cost problems of long sequential reasoning, which can require millions of tokens and tens of minutes for complex tasks. Existing methods such as self-consistency, Tree of Thoughts, ParaThinker, and Hogwild! Inference fix the parallel structure outside the model, which wastes compute on simple problems.

AIOpenAI reports that its automated system found accidental chain-of-thought grading in RL runs for several released models, including GPT-5.4 Thinking and GPT-5.4 mini. Its analysis found no clear reduction in CoT monitorability, though the company says subtler effects cannot be ruled out. OpenAI says it still avoids grading CoTs during RL and has fixed the affected reward pathways.

Why it matters: The post shows how accidental chain-of-thought grading was detected and tested, giving a concrete method for checking monitorability risks in RL training.

AICognition has opened its Asia-Pacific headquarters in Singapore to expand its autonomous software engineering platform, Devin, across the region. The company says OCBC saw up to 30% improvement in code and test case generation, and its system integration test first-pass rate rose from below 50% to over 80% after deployment. Cognition is building its Singapore team across engineering, go-to-market, and partnerships, with Richard Spence leading APAC.

AIOpenAI's Fidji Simo argues that biological data infrastructure is the missing link needed for real-world impact, even though it lacks flashy announcements. She credits CZI (Chan Zuckerberg Initiative) for its work, pointing to the Virtual Biology Initiative linked in her post.

AIDeepSeek has extended the 75% discount on its DeepSeek-V4-Pro API until May 31, 2026, 15:59 UTC, up from the previously announced May 5 deadline. The post also notes integration updates, including setting Claude Code to deepseek-v4-pro[1m] for a 1M context window.

AINVIDIA's developer account announced a livestream titled "Scaling NemoClaw: Roadmap, OpenClaw Collaboration, and Real-World Integration" as part of Nemotron Labs. The post provides no further details on the roadmap, collaboration terms, or integration specifics.

AIAndy Jassy announced that OpenAI's models will become available directly to customers on Amazon Bedrock in the coming weeks. The availability will be alongside the upcoming Stateful Runtime Environment, giving builders more model choice. More details are expected at the AWS event in San Francisco tomorrow.

AIXiaomi's MiMo-V2.5 is a native omnimodal model that understands text, image, video, and audio within one architecture. It is a sparse MoE with 310B total and 15B activated parameters, and supports up to 1M tokens of context. The repository also notes a config.json and tokenizer_config.json update that users who downloaded before commit 4da2748 should re-pull.

Why it matters: The repository documents a 310B-parameter omnimodal MoE with a hybrid attention design, useful for comparing long-context efficiency against other open multimodal models.

AIMistral AI has released mistralai/Mistral-Medium-3.5-128B-EAGLE, an EAGLE draft model for speculative decoding with the 128B dense Mistral Medium 3.5. The companion model, which the source says replaces Mistral Medium 3.1 and Magistral in Le Chat and Devstral 2 in Vibe, has a 256k context window, handles text and image input with text output, and is served with vLLM or SGLang using three speculative tokens. The model is released under a Modified MIT License that allows commercial use with exceptions for companies with large revenue.

AICognition has released Devin CLI, a local coding agent that runs in the shell with access to the codebase, tools, and environment. Users can choose among frontier models including Opus 4.7, GPT-5.5, and SWE-1.6, and hand sessions off to cloud agents with their own computer, which keep working after the laptop is closed.

AIDeepSeek has reduced the price of input cache hits across its entire API series to one-tenth of the original price, effective immediately. The existing 75% discount promotion for DeepSeek-V4-Pro remains active until May 5, 2026, 15:59 (UTC).

AIDeepSeek is offering a 75% discount on the DeepSeek-V4-Pro API until May 5, 2026, 15:59 UTC. For Claude Code, users can set the model to deepseek-v4-pro[1m] to unlock a 1M context window, while OpenCode requires v1.14.24 or later and OpenClaw requires v2026.4.24 or later.

AIMckay Wrigley says his coding split moved from 80/20 Claude/GPT to 80/20 GPT/Claude within three months, and he trusts GPT 5.5 for engineering. He still finds Claude better for non-coding agent work, calls Opus 4.7 underwhelming, and attributes Anthropic's issues to compute constraints.

AINVIDIA AI Developer announced a live "Ask the Experts" broadcast on running Gemma 4 on DGX Spark. The post provides only a broadcast link and no further technical details.

AIMeta has committed to tens of millions of AWS Graviton cores, according to Andy Jassy, Amazon's CEO. Jassy argues agentic AI is becoming as much a CPU story as a GPU story, since multi-step orchestration and real-time reasoning are CPU-intensive. He says the Graviton5 instances deliver up to 33% lower latency between cores.

AIThe DeepSeek API now supports V4-Pro and V4-Flash through both the OpenAI ChatCompletions and Anthropic interfaces. Developers keep the same base_url and set the model parameter to deepseek-v4-pro or deepseek-v4-flash. The legacy names deepseek-chat and deepseek-reasoner will be discontinued on 2026-07-24, and until then they map to the non-thinking and thinking modes of deepseek-v4-flash, respectively.

Why it matters: The source gives exact model names, an unchanged base URL, and a July 2026 discontinuation date, so developers can plan their migration from legacy names.

AINVIDIA and Google DeepMind experts will host a live session on Friday, April 24 at 11:00 AM PDT showcasing Gemma 4 running on DGX Spark. The demos cover vision translation, long-context document Q&A, and real-time code generation, with audience questions welcome.

AIFactory has released an Automated QA skill that drives an app as a real user would, filling forms, typing into terminals, and calling endpoints, then posts a structured report with screenshots, terminal snapshots, and API traces as a single updating comment on each pull request. Teams can run it on every push or make it an optional CI check triggered by a PR label, comment command, or manual dispatch, and developers can run /qa locally in any Droid session. Automated QA is available today in all Factory plans.

AICognition argues that enterprises building cloud agents face three problems: shared container kernels, the inability to persist agent state across async gaps, and the scale of orchestration, governance, and integrations. The post says VM-level isolation with hypervisor-level snapshots was needed for Devin, and that organizations must also rebuild engineering processes around agent execution.

AIAnthropic says three changes to Claude Code, the Claude Agent SDK, and Claude Cowork caused recent quality complaints, and the API was not affected. The fixes were resolved by April 20 (v2.1.116), and the company is resetting usage limits for all subscribers as of April 23.

Why it matters: The postmortem traces three separate changes to specific dates and versions, showing how a bug in context management can look like broad degradation to users.

AIAt Google Cloud Next this week, Google introduced its 8th-generation TPUs, split into two variants: 8t for massive-scale training and 8i for low-latency inference. Google presents the launch as a milestone in its accelerator roadmap, aimed at optimizing the full AI stack. The post links to a blog with further details on the systems architecture.

AITri Dao points to Wentao Guo's explanation of a mathematical rewrite of the MoE backward pass, which reduces activation memory and speeds up training, especially for fine-grained MoE. He also notes the work leverages new Blackwell features, including 2CTA MMA and CLC, to build fast MoE kernels.