NewsAI & Machine Learning
Small language models now run on a laptop
Models of a few billion parameters run on ordinary machines, cutting the cost of using AI for smaller organisations.
Small language models have improved quickly over the past year. Work that used to need a large server now runs, in good part, on an ordinary laptop.
The effect shows up directly in cost. Document search, summarising, answering customer questions — these can now be built without a large monthly API bill.
For small and mid-sized companies in Bangladesh this matters. Dependence on the internet drops, data stays on your own server, and costs become predictable.
For anyone wanting to work in this area the starting path is the same: Python, API integration and data handling. Model deployment is comparatively easy after that.