AI Decision Models: Knowing When to Bring in a Person
Customer support was the first place I wanted to use a large language model. Since the GPT-3.5 days, the difficult question has been knowing where its role should end.
Read article
AI, hardware & hands-on ideas
Stay curious.
Make it tangible.
buffer’s personal site. Notes on AI, hardware, the tools I use and what I learn along the way.
ExploringArtificial intelligenceHardware & devicesHands-on experiments
Notes & observations
All articles Recent notes
From GTA 6 to Tesla: What World Models Could Actually Do
Could an AI get you past a GTA 6 mission—or help your car anticipate trouble? A look at world models, Tesla's simulator and the real-world tests…
Read article
Liquid AI d1: Usage, Benchmarks and Local Deployment
See how d1 turns a support ticket into structured answers, where its benchmark scores differ from competing models, and what the 8 ms result means for…
Read article
Karotte Can Protect the Score. Who Defines Success?
An AI asked to speed up GPU code changed the timer instead. Karotte makes that kind of shortcut harder. The next question is whether we're measuring…
Read article
HSS: AI Can Read the Books. Can It See the Gap?
Try the bookshelf question before reading the answer. HSS exposes a gap in AI visual reasoning—and a product choice: when should an agent look harder, and…
Read article
DeepSeek and Huawei: It’s Time for Cheaper AI
DeepSeek's open-source tools for Huawei Ascend could lower the cost of running AI. Will consumers see the savings? A closer look at benchmarks, falling API prices,…
Read article
Claude Mods: 3 Useful Plugins for Better AI Workflows
Claude Mods let us customize how an AI works. DeepSeek Harness explored this earlier. I'd start with a handoff panel that means less AI babysitting.
Read article
About me
I’m buffer.
I talk AI and hardware on X. When a conversation needs more room, I continue it here.