🧠 The AGI Obsession That’s Stalling Real Engineering
A thoughtful critique of Silicon Valley’s AGI mythology—calling for pragmatic innovation grounded in engineering discipline.
Tom Phillips argues that the industry’s fixation on artificial general intelligence has derailed practical engineering. Referencing Karen Hao’s ‘Empire of AI,’ he critiques the cult-like drive for massive LLM scaling that ignores environmental and ethical costs. Instead, he advocates for efficient, focused AI systems solving tangible problems.
🔗 Read more 🔗
🚗 Two Years of ML vs One Month of Prompting — Lessons from Honda
A striking example of how generative models are reshaping industrial ML—trading data pipelines for creative prompt optimization.
Honda’s analytics team found that large language models could replicate results from two years of supervised ML work in just a month of prompt tuning. By switching from SQL and XGBoost to prompt engineering, they achieved similar accuracy at lower cost, revealing a shift in ML bottlenecks from data collection to effective prompting.
🔗 Read more 🔗
🐍 portable_python: Self-Contained Python for Linux
🔗 Read more 🔗
💧 The AI Water Issue Is Fake
🔗 Read more 🔗
💻 How to Deploy LLMs Locally
A detailed yet approachable walkthrough for developers seeking hands-on experience with local LLM infrastructure.
A comprehensive guide to running large language models locally. It covers benefits like transparency and cost control, along with setup advice on hardware, VRAM, quantization, and CPU offloading. Frameworks such as llama.cpp and fastllm enable efficient, open-weight deployments for experimentation and customization.
🔗 Read more 🔗