Daily AI trends · Thursday, 15 October 2026
What was rising in AI on 15 October 2026
The fastest-rising AI project was Arekejoker/qwen21-nsfw-pose-skill (Agents). The top GitHub repo gained 659 stars in 24 hours (Arekejoker/qwen21-nsfw-pose-skill). LFM2.5-24B-A2B was breaking out across 3 sources.
- 77 items trending
- 8 new models
- 35 discussions
- 1,443 GitHub stars gained by tracked AI repos
Last recorded Thu, 15 Oct 2026 23:30:52 UTC.
Top movers
- Arekejoker/qwen21-nsfw-pose-skillAgents · momentum 94 · up 26% in a day · ★ 2,003 · +659 in 24h · Created 3 days agoQwen-Image 2.1 NSFW/姿势出图 skill:作者级提示词配方 + NSFW LoRA + 136 姿势 VNCCS 人偶控制;适用 Claude Code / Codex / Cursor / OpenClaw
- onehumanbeing/esp32-iphone-agentAgents · momentum 91 · ★ 2,423 · ~580 in 24h · Created 4 days agoLet an AI model use a real iPhone with a $7 ESP32: AirPlay eyes, USB HID hands, any OpenAI-compatible vision model. No Mac, no developer account, no jailbreak.
- google/embeddinggemma-2Vision · momentum 90 · up 7% in a day · 144K downloads · Released 31 days agoVision on Hugging Face · open source · ♥ 2.4K
- jialinyyzz/humanizerInfrastructure · momentum 89 · up 15% in a day · 75K downloads · Released 35 days agoModel on Hugging Face · open source · ♥ 1.7K
- Qwen/Qwen-Image-2.1-TurboVision · momentum 89 · 15K downloads · Released 7 days agoVision on Hugging Face · open weights · ♥ 873
- openbmb/MiniCPM-V-4.7-35B-A3BVision · momentum 86 · 9.3K downloads · Released 9 days agoVision on Hugging Face · open source · ♥ 791
- LiquidAI/LFM2.5-24B-A2BLLMs · momentum 86 · 6.3K downloads · Released 4 days agoModel on Hugging Face · open weights · ♥ 387
- abenzerps/Qwen-Image-2.1-Uncensored-QuantizedInfrastructure · momentum 85 · 2.4M downloads · Released 25 days agoModel on Hugging Face · open weights · ♥ 4.4K
- AtomicChat/Qwen-Image-2.1-Turbo-Uncensored-GGUFCoding · momentum 85 · 296K downloads · Released 6 days agoModel on Hugging Face · open source · ♥ 489
- Paper/CLASHarness: Benchmarking Multi-Agent Conflict in Computer Use via a Controllable and Verifiable Generation PipelineResearch · momentum 85 · ▲ 25 upvotes · Hugging Face PapersVision-language models (VLMs) are increasingly deployed in multi-agent systems to automate complex computer-use workflows, yet existing benchmarks mostly assume consistent, cooperative environments.
Breaking out
- LFM2.5-24B-A2BNamed on 3 sources: Hugging Face: LiquidAI/LFM2.5-24B-A2B; Hacker News: LFM2.5-24B-A2B: Built for fast on-device agents; Hugging Face blog: Run fast, private agents on-device with LFM2.5-24B-A2B
Fastest-growing AI repos on GitHub
- Arekejoker/qwen21-nsfw-pose-skill+659 stars in 24h · 2,003 stars · 414 forks
- onehumanbeing/esp32-iphone-agent+580 stars in 24h (estimated) · 2,423 stars · 322 forks
- rociiu/talorys+204 stars in 24h (estimated) · 1,134 stars · 101 forks
Trending models on Hugging Face
- google/embeddinggemma-2Vision · Open source · 144,280 downloads · 2,395 likes
- jialinyyzz/humanizerInfrastructure · Open source · 75,253 downloads · 1,729 likes
- Qwen/Qwen-Image-2.1-TurboVision · Open weights · 15,315 downloads · 873 likes
- openbmb/MiniCPM-V-4.7-35B-A3BVision · Open source · 9,278 downloads · 791 likes
- LiquidAI/LFM2.5-24B-A2BLLMs · Open weights · 6,286 downloads · 387 likes
- abenzerps/Qwen-Image-2.1-Uncensored-QuantizedInfrastructure · Open weights · 2,415,278 downloads · 4,382 likes
Called it
- Comfy-Org/Qwen-Image-2.1Tracked at 5,376,977 downloads on 01 Oct 2026 17:48 UTC, passed 10,000,000 on 15 Oct 2026 10:01 UTC
- bherbruck/solidcraftTracked at 964 stars on 12 Oct 2026 18:30 UTC, passed 2,500 on 15 Oct 2026 07:01 UTC
- wheresryan22/anatomyTracked at 1,020 stars on 11 Oct 2026 11:30 UTC, passed 10,000 on 15 Oct 2026 00:30 UTC
- yihui-dev/awesome-opus5-5-videosTracked at 1,221 stars on 01 Oct 2026 17:48 UTC, passed 5,000 on 14 Oct 2026 17:30 UTC
- CopilotKit/openmuseTracked at 3,527 stars on 01 Oct 2026 17:48 UTC, passed 5,000 on 14 Oct 2026 16:30 UTC
- Jakeschincariol/replica-skillTracked at 938 stars on 08 Oct 2026 09:30 UTC, passed 2,500 on 14 Oct 2026 11:01 UTC
- Qwen/Qwen-Image-2.1-TurboTracked at 0 downloads on 09 Oct 2026 15:01 UTC, passed 10,000 on 14 Oct 2026 10:01 UTC
- nerkyor/Qwen3.8-27B-Coder390-EfficientThink-Opus5.5-GPT6Astra-Grok4.7-DSV4Pro-K3-SFT-RLOO-MTP-DFlash2Tracked at 8,873 downloads on 09 Oct 2026 18:01 UTC, passed 100,000 on 14 Oct 2026 10:01 UTC
Most-discussed threads
- Muse Spark OpenHacker News · 560 points · 319 comments
- Wayland desktop for Linux with skeuomorphic designHacker News · 466 points · 184 comments
- Why We're Banning AIHacker News · 217 points · 172 comments
- Krea 3 – An image model built for editing at scaleHacker News · 227 points · 76 comments
- Call Center Jobs Are Declining, but 14,000 Calls Show Humans Still Outperform AIHacker News · 35 points · 47 comments
- Simons Foundation launches $15M effort to help mathematicians navigate AI eraHacker News · 218 points · 176 comments
AI headlines
- Bluesky expands Attie, its AI tool for building custom feeds, with a "social vibe coding" feature to create sites, games, quizzes, and more on the AT ProtocolTechmeme · News
- Sources: some Anthropic investors discussed a ~$1.5T valuation for its IPO, down from ~$2T they had expected, after disclosures showed slowing Q3 revenue growthTechmeme · News
- A researcher says modern neural networks are far more interpretable than believed, with most neurons understandable in models up to trillions of parametersTechmeme · News
- Granite Switch 4.1: Chat, Decisions, and Speech from One EndpointHugging Face · Lab blog
- Perplexity open sources SPACE, its AI agent sandbox, and says partners including CoreWeave, Akamai, and Crusoe agreed to offer SPACE to customersTechmeme · News
- Bringing GPT-Rosalind to academic research labsOpenAI · Lab blog
- Google says it has 400+ publications in its News AI pilot, which tests features like overviews; participants include Financial Times, CNN, and The Texas TribuneTechmeme · News
- Volvo says it will use Nvidia's Hyperion platform to launch a Level 2++ hands-free, point-to-point automated driving system by 2030Techmeme · News
- Memo: Accenture has told many employees, some of whom had been spending as much as $1,000 monthly, that it will now only spend $29 a month per person on ClaudeTechmeme · News
- Anthropic says it is giving employee-like access to METR as part of its pledge to partner with evaluators for assessment of its models and safety practicesTechmeme · News
- Be There for the Beginning: Celebrate the Launch of NVIDIA RTX SparkNVIDIA · Lab blog
- Tell us: have you used AI to help with online dating? The Guardian · News
- Accelerating the search for antimicrobial peptides with Google AntigravityGoogle DeepMind · Lab blog
- Open From Data Center to Desktop: NVIDIA Expands Nemotron Family With Highest Efficiency Super VL and Bolt ModelsNVIDIA · Lab blog
- Tech market sell-off deepens amid fears over US inflation and AI debtThe Guardian · News
- Dial Up the Cloud: NVIDIA DLSS 4.5 Controls Arrive on GeForce NOWNVIDIA · Lab blog
Get this in your inbox at 7am
One free email each morning with the AI repos, models and stories that are picking up speed. Your own time zone, no spam.