Signals: local AI in the wild.
The shift to open-weight and on-device AI is not our opinion alone. Here is where practitioners, labs, and builders are working it out in public. We link to the source, we do not reproduce it. Follow a link to read it where it was published.
Pinned
Full archive
- Running Qwen 3.8 Flash Next on a $1000 Local AI Build with Strata
- Building AI for Reliable Execution: Lessons From Industrial Robotics
- ThursdAI - Oct 8 - OpenAI drops 722 math papers, Haiku 5.5 hits 10 cents & more
- ttok 1.0
- I expect rapid progress but not towards general superintelligence
- Step 5 Preview (Tested): This MODEL BLEW MY BRAINS AWAY!
- Reflection AI Announces Beam 501B Open-Weight MoE Model
- I Gave Two AI Supercomputers a Real Job
- Chatting with Alex Ziskind about the Future with AI
- Google Launches EmbeddingGemma 2 Open Model
- We're using GLM-5.3 Flash instead of frontier models on a massive production codebase
- Mistral Is BACK - Mistral Large 4 First Test (Le Chonk!)
- DeepSeek Harness Just Got a Web Agent
- Ollama v0.40.0
- The Cyber Risk Discourse is Broken
- llm-openai-decisions 0.1a0
- Quoting Felix Rieseberg
- Reflection AI Is About to Release a US Open-Weight Model to Take On DeepSeek and Qwen
- You can now try Aleph Alpha's Kolibri 78B for free online here.
- Qwen3.8 27B addition in words
- Strata takes the promise of "MoE models just need a total amount of VRAM+RAM" and makes it a reality
- Strata is seriously impressive, running Qwen 3.8 Flash Next on hermes at 512k context.
- ThursdAI - Oct 1 - OpenAI joins the assistant race, CoreWeave drops serverless GPUs & more
- Ollama v0.35.1
- Phone-Sized Open Models Match Opus Coding Benchmarks
- The DGX Spark 64GB is a Local AI Warning Sign
- Open models and the future of Physical AI with NVIDIA
- The First Thing AI Found Wasn't an Overcharge
- M5 Ultra vs 2 DGX Sparks… The Number You're Not Looking At
- Building an AI Workstation for Someone Else
- Gemini 4 Argon Explained
- Quoting Matthew Green
- Why Dwarkesh is Wrong about Computer Use + How OpenAI shipped its Jev competitor in 1 Week
- Ollama v0.35.0
- Quoting Anthropic Frontier Red Team
- Photo Scrubber : local face blur & metadata removal
- Language Models for Text Classification: From Bag-of-Words to Jev
- [Release] GSQ-RCO GGUFs for Qwen3.8-Flash-Next, plus a 50% expert-pruned Coder build at ~1.89 bpw
- Claude Code's Next Era : Thariq Shihipar, Anthropic
- GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price
- 98% of Qwen - Bonsai 2
- Claude Sonnet 5.5 Is INSANE - Seriously, This Model Is Ridiculous!
- Don't Buy a Local AI GPU Until You See These Trends
- Xiaomi Mimo V2.6 Is INSANE? - Pro & Flash FULLY Tested!
- OpenRouter: from Seed to Stripe : with OpenRouter's Alex Atallah & AMP's Anjney Midha
- DeepSeek is CRAZY
- Ollama v0.34.4
- New MacOS 27 for clusters
- Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war
- MiMo v2.6: Xiaomi Just Built the Best Open Model
- Ollama v0.34.3
- M5 Ultra… Apple Wasn't Messing Around
- Claude Opus 5.5 Is INSANE - Hands-On With the BEST Model Yet!
- Debating RSI, the US-China Gap, and Jaggedness with JS Denain of Epoch AI
- The current balance of power in open models
- Qwen Image 2.1 + Hermes Agent Fully Local AI Image Generation
- Codex's NEW A2A Communication & Crazy NEW UPDATES: You don't know about these SECRET UPDATES!
- The M6 Mac mini Had Me Checking My Numbers
- Jev introduces a new shape of LLM - System One, aka Decision Models
- I Waited Two Years for This Mini PC… Then I Tested It
- ThursdAI - Sep 17 - TypeSafe's Jev is a ChatGPT moment for decisions, Pace the Frontier splits the labs & more
- An AI Supercomputer On Your Desktop: Checking Out The MSI XpertStation WS300
- Is Bonsai 2 27B REALLY 98% of Qwen 3.8 27B fp16?
- My AI Agent Lives OUTSIDE My Computer
- proto-v0.3.0
- Ollama v0.34.2
- The Only Local AI GPUs Worth Buying Right Now (Ranked)
- Doom in Real Time
- 8 RTX Pro 6000's Wasn't What I Expected
- I ran Qwen 3.8 27B locally for 30 days, here are the results
- Ollama v0.34.1
- Cognition SWE-2 First Test - Is THIS a Better Kimi K3?
- Humanity's Last Invention : Richard Socher of Recursive
- I Ran A 27B Model On A Hand-Sized PC… Didn't Expect This
- DeepSeek Just Made Long Context Cheap
- I made an app to help me optimize llama parameters on my hardware
- DeepSeek Fails the Rubik's Cube Test
- OpenAI solves Navier-Stokes, Meta's Muse a free AI agent that's really good, DeepSeek V4.1 shrinks KV cache, and one doomer post causes OpenAI to consider pausing training + more AI news
- Open-Source AI & Open Models Reading List
- I Dropped Multi-GPU Temps by -15°C with ONE Mod
- Deepseek did it again...
- Ollama v0.34.0
- DeepSeek V4.1 Flash: Stronger, Faster, More Accessible
- 8 DGX Spark Cluster with this Switch
- vLLM v0.29.0
- Latest open artifacts (#24): Motif-3, GLM-5.3, Hy4-preview and open model licenses
- Why I'm back to Qwen 3.8 27B and Hermes Agent - Local AI Model Fatigue
- NVIDIA didn't want me to do this
- All That VRAM Needs a Bigger Brain
- OpenClaw Power, MacBook Simplicity: Five Days With Grok Bot
- MSI EdgeXpert (DGX Spark) Unboxing
- Understanding ChatGPT Work
- Voodoo Quant beats Unsloth Dynamic 2.0 KLD by 95% in Qwen3.5 0.8B and 2B
- Qwenthropic
- I didn't give up - extGemma4-40_5B returned
- KEEP AI LOCAL! Explaining Agentic AI and The Loop
- GLM-5.2 is the step change for open agents
- Banning Open Source AI Would Be A Mistake
- Frontier post-training recipe review with Finbarr Timbers
Picked by an on-device, open-weight AI pipeline running on our own hardware. An item lands here when a trusted source carries it, or when several independent sources are echoing the same story. One link per story, never a reproduction, and we prune by hand.