Based on a recent breakthrough known as the Semantic Decoder this past May:
https://www.popsci.com/technology/ai-semantic-decoder/?amp
Concerns regarding privacy aside, a technology like this could, in the near future, allow for general purpose brain-computer interface without requiring invasive surgery of the kind currently envisioned by Neuralink. Further progress with LLM's would also allow an easy way for an AI to interpret specific thought commands.
Resolution criteria is as follows:
The brain-computer interface must be noninvasive i.e. does NOT require surgery to be used.
It must be portable.
It must be cheap. For now, I'll use the current average price of a smartphone as a benchmark, adjusted for inflation.
It must be general purpose, meaning that one could use it to link your brain to almost any computer, smartphone, or IoT device, and therefore allow for direct control over it (we'll limit it to contemporary devices or near future IoT devices, as I am uncertain that one can use it to control computers built more than 20 years ago).
It must have off the shelf availability, meaning that one can reasonably expect to purchase it at any "Big Box" store (such as Best Buy or equivalent), or at the very least, there are dedicated chains which can sell the device (such as an Apple Store, a Verizon store, etc.). Online availability isn't enough, except in the event that all big box retailers decide to be fully virtual by 2030.