Free Tool That Runs AI on Your Laptop Just Got a Fix

There's a piece of free, open-source software called llama.cpp that does something quietly remarkable: it lets you run an AI chatbot on your own laptop or phone, with no monthly subscription and no conversation ever leaving your machine. It's built by volunteers and maintained in public, and roughly 130,000 people have bookmarked it — a huge number in the world of free software. This week it pushed out update b11332, which is not a new feature but a repair job.

The repair targets what developers call "recurrent memory" — essentially the AI's short-term memory, the part that helps it keep track of what you said earlier in a long back-and-forth. An incorrect internal check in the code could cause the program to misbehave or stop working. This update fixes that. Think of it like a car recall for a faulty sensor: the car still drives, but one thing was giving bad readings, and now it doesn't.

Who should care? Honestly, most people won't notice. If you open ChatGPT, Gemini or Claude in a browser, this release has zero effect on your day. But if you're one of the growing number of people running AI locally — because you want privacy, offline access, or to avoid paying per message — this is worth updating, especially if you've had long conversations cut off unexpectedly. The update ships ready-made versions for Windows, Mac, Linux, Android and even certain Snapdragon phone chips.

The bigger story is what open-source AI maintenance actually looks like: dozens of tiny, unglamorous fixes with names like b11332 that outsiders can't decipher. The catch is that there are no new capabilities here, and installing it still takes a bit of technical comfort compared with opening a website.