What changed
New this week
Everything USASI added in the 7 days from Oct 13, 2026 to Oct 19, 2026, the date this page was built. Each entry is dated and links to its page. RSS feed
Added to the catalog
- T3 Code · Open model or toolT3 Code, MIT-licensed, maintained by T3 Tools, Inc.
- Ping Labs · OrganizationPing Labs (T3 Tools, Inc.), maker of T3 Code and T3 Chat
- U.S. Federal Government · OrganizationParent catalog record for U.S. federal AI job listings from USAJOBS.gov
Other additions: site features and people profiles
As the changelog records them (profiles carry no separate “added” date).
- Hugging Face dataset usasihq/usasi-ai-jobs and Space explorer; daily refresh can push the snapshot
- AI on GitHub (/github/), with a page per project; methodology updated
- AI on Hugging Face (/huggingface/ and /huggingface/datasets/), with a page per model and dataset; methodology updated
- AI Hiring Index at /jobs/index/: weekly open-job counts by employer and state, week-over-week change, and WARN Act pairing
- Profiles: Jesse Genet, X Freeze, Kim Isenberg, Mesh Times, argo, Ben Davis, maria
- Voices of SI at /voices/; 107 creator profiles under /people/; Andrej Karpathy also listed as a creator
- Federal AI jobs at /jobs/at/federal/, linked from /jobs/, sourced from USAJOBS.gov
- Jobs by role at /jobs/role/forward-deployed-engineer/; later roles are new config entries
- Fourteen more role pages at /jobs/role/<role>/
- Similar jobs on every job page
- Employer jobs pages at /jobs/at/<employer>/, linked from /jobs/ and each organization page
- A page per current listing with the employer's description, apply link, and job posting data for search engines
- Per-document pages at /policy/<slug>/; site search now opens these pages for policy results
- People directory at /people/ and profile pages at /people/<slug>/; organization pages list linked people
- News items for events from July 1 to October 12, 2026; news filters and month headings on /news/
- Pages at /find/memory/ for 8, 12, 16, and 24 GB of VRAM and 8, 16, 24, 32, and 64 GB Apple silicon Macs, linked from the finder and the local AI hub
- The State of American AI at /report/, with the October 2026 edition at /report/2026-10/
- Jobs pages by U.S. state (/jobs/in/<state>/), city, and topic
- Twelve finder configurations with publisher-documented memory figures, spanning chat, documents, code, and transcription
This week’s changelog entries (20)
- AI jobs on Hugging Face: a searchable snapshot and explorer (Oct 18, 2026)
- AI on GitHub: the most-starred open-source AI projects, each with its own page (Oct 18, 2026)
- AI on Hugging Face: the most-liked open models and datasets, each with its own page (Oct 18, 2026)
- Hugging Face jobs explorer is a free static Space (Oct 18, 2026)
- Jobs: weekly AI Hiring Index (Oct 18, 2026)
- T3 Code and Ping Labs in the catalog; seven new profiles (Oct 18, 2026)
- Remote, hybrid, and on-site labels from employers; all of NVIDIA's listings; four universities (Oct 17, 2026)
- Voices of SI: 107 tech creators and communicators (Oct 17, 2026)
- Federal AI jobs from USAJOBS.gov (Oct 16, 2026)
- Jobs by role, starting with Forward Deployed Engineer (Oct 16, 2026)
- Jobs from NVIDIA, Micron, Salesforce, Intel, and six more employers on Workday (Oct 16, 2026)
- Jobs from ServiceNow, Lawrence Livermore National Laboratory, Hugging Face, Nomic AI, and FAR.AI (Oct 16, 2026)
- Jobs from SpaceX and Anduril (Oct 16, 2026)
- Jobs: search first, filters by category, level, and pay, 15 role pages, and five more employers (Oct 16, 2026)
- A Jobs page for every employer and for every current listing (Oct 14, 2026)
- A page for every document in the policy tracker, starting with Executive Order 14434 (Oct 13, 2026)
- A People directory with a page for every profile, and a larger, filterable news section (Oct 13, 2026)
- Local AI by memory: what fits 8 to 24 GB of VRAM or an Apple silicon Mac (Oct 13, 2026)
- The State of American AI, a monthly report, and Jobs pages by state, city, and topic (Oct 13, 2026)
- Twelve more documented local setups for the finder and the memory pages (Oct 13, 2026)
News
- vLLM 0.32.0 released with a structured decisions endpoint and wider low-precision KV cache support
- GitHub makes open source license compliance checks generally available for Enterprise Cloud
- GitHub makes AI Scan for pull requests and agentic autofix for code scanning alerts generally available
- NIST CAISSI, NSA, and CISA publish voluntary incident response practices for super intelligence training and evaluation
- NVIDIA releases Nemotron 3.5 Super VL and Nemotron 3.5 Bolt 74B-A9B open models under OpenMDW-1.1
- NVIDIA releases Nemotron Labs Dory1.0-4B, a looped language model, under OpenMDW-1.1
- Reflection AI releases the Beam model weights on Hugging Face under Apache 2.0
- Ai2 releases HiRO-ACEv2, an updated climate emulator and 3 km downscaling system, under Apache 2.0
- Anthropic gives a METR team employee-like access and describes publication terms for outside evaluators
- IBM releases Granite Switch 4.1 models combining chat, decisions, and speech-to-text under Apache 2.0
- Meta releases Muse Spark Open, an open-weight version of Muse Spark, under the Muse Model License
- Anthropic describes Claude election safeguards ahead of the 2026 U.S. midterms
- Anthropic opens a Singapore office and signs an MOU with Singapore's digital ministry
- Google makes Gemini 4 Argon available to developers and to Google AI Pro and Ultra subscribers
- Liquid AI releases LFM2.5-24B-A2B, an open-weight mixture-of-experts model for on-device agents
- SEC declares the Agility Robotics and Churchill Capital Corp XI Form S-4 effective; vote set for November 2
- Agent Plugins 1.0.0 published as an open specification for packaging agent skills and MCP servers
- Agility Robotics unveils Digit 5, a humanoid robot designed to work near people
- Ai2 releases BenchMIRT, a method, code, and data for auditing LLM benchmarks prompt by prompt
- Ai2's Bolmo byte-level language model research is published in Nature, with new checkpoints
- Amazon releases ALoDLM-1.7B and ALoDLM-8B diffusion language models for non-commercial research
- AMD agrees to acquire Taalas, a developer of AI inference silicon
- AMD releases Instella-MoE, a 16B mixture-of-experts language model, under a research-only license
- AMD releases ROCm Core SDK 10.0.0 with updated framework support and breaking AMD SMI changes
- Anduril acquires Popoto Modem, an underwater acoustic communications company
- Anthropic releases Claude Fable 5.1 and, for vetted organizations, Claude Mythos 5.1
- Anthropic releases Claude Opus 5
- Anthropic releases Claude Opus 5.5, the first model in its Claude 5.5 family
- Apple publishes LensVLM-9B, a research vision-language model that reads compressed images of text
- Arc Institute releases PIE, a model for predicting cell responses to perturbations
- ARC Prize names the winners of the first ARC-AGI-3 milestone prize for open-source solutions
- Arcee AI announces Genesis-Science-1, an open science model being built with the Department of Energy
- Biohub's esm 3.4.0 package adds full support for ESMC and ESMFold2
- CAIS and Scale AI release HLE-Diamond, a refined 1,000-question subset of Humanity's Last Exam
- Cartesia releases Ink-2, a speech-to-text model for real-time voice agents
- Cartesia releases Sonic-3.6 text-to-speech with Odia, Urdu, and Hindi-English code-switching
- Character.AI begins a voluntary Chapter 11 process to pursue a court-supervised sale of the company
- Cisco releases Antares-350M and Antares-1B, open-weight models for locating vulnerable code
- Cline introduces Cline Desktop, an early desktop app for running Cline agent sessions
- Cloudflare open-sources Cloudflare OS, a platform for agent workspaces and internal apps
- Cognition releases SWE-2, a coding model post-trained from Kimi K3, in Devin
- Cognition signs a memorandum of understanding with the Department of Energy to join the Genesis Mission
- Common Crawl releases experimental web graph embeddings for web hosts
- DSPy 3.4.1 makes InstructionProposer the default for GEPA and adds Python 3.15 support
- Faiss 1.15.0 adds a Metal IVF-PQ index, EDEN quantizer indexes, and a cuVS IVF-SQ backend
- FAR.AI launches the AI Security Leaderboard and a Minimal Standard for Safeguards
- FastVideo releases FastH3 Preview v1, an open-weight 4-step distillation of MiniMax H3
- Figure announces Helix 2.5, a humanoid control model pretrained on its Index dataset
- GitHub MCP Server 2.0.0 adds structured outputs and output schemas for newer MCP clients
- GitHub's Spec Kit reaches version 1.0.0
- Google DeepMind introduces Gemini Robotics 2, ER 2, and On-Device 2 models for robots
- Google DeepMind introduces SynthID Bio and open-sources code for watermarking AI-designed proteins
- Google DeepMind launches AlphaGenome Atlas, precomputed predictions for human DNA variants
- Google DeepMind releases AlphaProtein Novo code and non-commercial generator weights
- Google introduces WeatherNext 3, an hourly global weather model that uses live satellite data
- Google releases Gemini 3.8 Flash and, for vetted defenders, Gemini 3.8 Flash Cyber
- Harvey open-sources a synthetic law firm knowledge expansion of its Legal Agent Benchmark
- Hugging Face adds reinforcement learning environments to the Hub as tagged dataset repositories
- IBM releases Granite Speech 5.0 TurboCTC, two compact English speech recognition models
- IBM releases the Granite Time Series PatchTST-FM-r2 forecasting model under Apache 2.0 or OpenMDW
- Inception releases Mercury 2.5, a diffusion language model, and previews Mercury Voice
- Intel releases OpenVINO 2026.4 with multi-token prediction decoding in OpenVINO GenAI
- Isaac Lab 3.0 Early Access published with multi-backend physics and kit-less workflows
- Lemonade 2026.39.1 switches to calendar-based version numbers and adds image editing
- Linux Foundation announces the operational launch of the x402 Foundation for agent payments
- Liquid AI releases LFM2.5-2.6B, an open-weight on-device model trained for agentic workloads
- Liquid AI releases LFM2.5-VL-3B, an open-weight vision-language model for on-device use
- llm-d 0.10.0 focuses on operational hardening and deprecates its own CUDA and AWS images
- LM Studio introduces Bionic, a separate agent app for open models
- Marin begins pretraining Marin 535B-A23B, a mixture-of-experts model
- Meta launches Muse Image and previews Muse Video, its first Superintelligence Labs media models
- Meta publishes MetaEncoder-30B, a gated Apache 2.0 multimodal encoder for decision tasks
- Meta releases Muse Spark 1.1 and opens the Meta Model API in public preview
- Meta releases Muse Wonder and Muse Wonder Full through its Model API and Muse Code
- METR publishes a summary of its predeployment evaluation of Claude Opus 5.5
- METR's president testifies to a Senate subcommittee on AI agent incidents
- Microsoft AI publishes a draft Code of Conduct for MAI models and opens a public consultation
- Microsoft AI releases the MAI-Thinking-1 reasoning model in public preview on Microsoft Foundry
- Microsoft commits to the Genesis Mission with a SPARK program office and four initial projects
- Microsoft releases VibeVoice-ASR-Streaming, an MIT-licensed streaming speech recognition model
- Midjourney announces its first acquisition, Co–Star
- Midjourney launches its V8.2 image model
- MLCommons publishes MLPerf Storage v3.0 results with KV cache and vector database tests
- MLCommons releases MLPerf Client v2.0 with image generation and agentic AI tests
- MLCommons releases MLPerf Endpoints v0.7, a benchmark for buyers of hosted inference
- MLPerf Training adds an LLM post-training benchmark, starting with round v6.1
- Model Context Protocol publishes its 2026-07-28 specification, removing protocol-level sessions
- Modular 26.6 and Mojo 1.1 open the Mojo compiler to outside contributions
- Modular releases Mojo 1.0 alongside MAX 26.5
- Mozilla.ai introduces Otari, an open-source LLM control plane under Apache 2.0
- NIST signs an agreement with DOE's Office of Science to join the Genesis Mission
- NIST's CAISI publishes an assessment of Z.ai's GLM-5.3 cyber capabilities
- Nous Research releases Hermes Agent 0.20.0 with conversational voice and A2A v1.0 support
- NVIDIA adds CUDA-Q Logical, a fault-tolerant quantum orchestration layer, to open-source CUDA-Q
- NVIDIA and Cisco announce plans for USFP, a pluggable optics form factor, and an open specification
- NVIDIA and inaugural partners launch the Open Secure AI Alliance for open AI security tools
- NVIDIA contributes Vera Rubin rack and compute tray designs to the Open Compute Project
- NVIDIA Dynamo 1.5.0 adds custom worker-selection policies and deprecates KVBM
- NVIDIA launches Open Agent Safety Platform, pairing open-source OpenShell with a Sentry reference design
- NVIDIA releases Cosmos3-Edge, a 4B omnimodal world model, under OpenMDW-1.1
- NVIDIA releases CUDA Python 1.0 with semantic versioning commitments
- NVIDIA releases NeMo Speech 3.0, the first major release after the NeMo repository split
- NVIDIA releases Nemotron 3 Embed multilingual text embedding models in 1B and 8B sizes
- NVIDIA releases Nemotron 3.5 Lightning, an open 30B mixture-of-experts model for agents
- Ollama 0.35.0 adds decision models through a /v1/systemone endpoint
- Ollama 0.40.0 runs supported models on MLX by default on Apple Silicon
- ONNX Runtime 1.30.0 adds Go bindings and expands CUDA and WebGPU generative AI support
- Open Silicon Photonics for AI Systems initiative becomes an Open Compute Project workstream
- Open WebUI 0.11.0 ships a redesigned interface and optional sub-agents
- OpenAI publishes MCP Extensions, a specification and SDKs for ChatGPT plugins, under Apache 2.0
- OpenAI releases GPT-6 Astra in its API
- OpenAI releases GPT-6 Sol and GPT-6 Luna in its API
- OpenAI releases the GPT-5.6 model family, GPT-5.6 Sol, Terra, and Luna, in its API
- OpenClaw 2026.8.1 adds conversation search, remote sessions, and masked credential requests
- OpenHands 1.27.0 adds OpenCode and Pi as built-in ACP agents
- Perplexity releases pplx-decider open-weight decision models under Apache 2.0
- Perplexity releases pplx-embed-v2-late multimodal late-interaction embedding models under MIT
- Poolside releases Laguna S 2.1, a 118B-parameter open-weight model with a 1M-token context
- Poolside releases Laguna XS 2.1, a 33B-parameter open-weight coding model, under OpenMDW-1.1
- Prime Intellect releases prime-rl 0.8.0 with multi-agent reinforcement learning support
- PyTorch 2.13.0 released with FlexAttention on Apple Silicon and the torchcomms backend
- PyTorch 2.14.0 released with torch.switch, NVGEMM in Inductor, and a preview NCCL backend rewrite
- Ray 2.59.0 makes Ray Data LLM and Ray Serve LLM generally available
- Red Hat AI 3.5 becomes generally available with EvalHub and wider llm-d support
- Reflection AI announces Beam, a 501B-parameter model whose weights it plans to release
- Reka releases the Reka Inverse Dynamics Model for labeling video with camera motion under Apache 2.0
- Runway releases Ruby, a model that converts SDR video to HDR
- Salesforce and NVIDIA announce Koa, a CRM reasoning model post-trained from NVIDIA Nemotron 3 Super
- Salesforce signs a definitive agreement to acquire Listen Labs; the deal has not yet closed
- Scale AI releases DistressBench, a benchmark of crisis support in language models
- Sentence Transformers 6.0.0 adds MultiVectorEncoder for ColBERT-style late-interaction models
- SGLang 0.5.21 adds a Decisions API and runtime prefill-decode role switching
- Sierra open-sources Hyper-τ-bench, an evaluation of models that build customer-service agents
- Skild AI introduces S1, a robotics model that learns tasks from a video demonstration
- Strands Agents releases Strands harness, an Apache 2.0 general-purpose agent harness
- The Dioxus team joins Cognition, which says it will keep supporting Dioxus open source work
- The Firetiger team joins Cursor to work on agents for software in production
- The NVIDIA-founded Open Secure AI Alliance moves to the Linux Foundation
- Thinking Machines Lab releases Inkling-Small, a smaller open-weight Inkling model
- Thinking Machines Lab releases Inkling, its first open-weight model, under Apache 2.0
- tinygrad 0.14.0 released with tool calling in tinygrad.llm and parallel kernel compilation
- Together AI publishes Tev1-4B-experimental, a decision model whose weights license is not yet final
- UK AISI and NIST's CAISI publish a preliminary assessment of Kimi K3's cyber capabilities
- Unsloth introduces Unsloth Desktop, an open-source app to run and train models locally
- Unsloth release adds Browser use, Qwen-Image-2.1-Turbo, and text-diffusion LLM training
- Vercel announces it is acquiring Better Auth; the authentication library stays MIT-licensed
- Waymo announces a framework for incident management exercises for automated vehicle fleets
- World Labs announces that robotics company SceniX is joining it
- World Labs introduces Atlas, a world model offered through early access
- WRITER releases Palmyra X6, post-trained on GLM 5.2, alongside WRITER Agent upgrades
- xAI makes Grok 4.5 available on its API
- xAI makes Grok 4.7 available on its API
- Zyphra releases ZUNA1.1, an open-weight EEG foundation model, under Apache 2.0
New on AI on GitHub
None this week. Tracking started with the Oct 18, 2026 snapshot: projects that enter the ranking after it are listed here with the date of the fetch that added them. The snapshot is refreshed about once a week.
New on AI on Hugging Face
Models
None this week. Tracking started with the Oct 18, 2026 snapshot: models that enter the ranking after it are listed here with the date of the fetch that added them. The snapshot is refreshed about once a week.
Datasets
None this week. Tracking started with the Oct 18, 2026 snapshot: datasets that enter the ranking after it are listed here with the date of the fetch that added them. The snapshot is refreshed about once a week.
New job listings, by employer
13,021 current listings were first seen in employers' job feeds this week, at 80 employers. “First seen” is when USASI first read the listing, so an employer whose job board was first collected this week counts all of its listings.
Support Us
Help keep USASI useful.
Optional. No USASI account required. Payment takes place on the linked provider’s website (Buy Me a Coffee).
About supporting this project