Chris Mairs

AI excels at generating impressive content, often surpassing paid professionals. So why can’t it handle basic online chores like ordering a book or booking a flight? And what’s the outlook for tackling far more consequential agentic tasks?

Agentic Misalignment describes actions taken by AI to the detriment of one or more humans. Aengus Lynch and colleagues at Anthropic recently demonstrated Large Language models attempting blackmail to prevent their own shutdown.

Overly complex household tech, with low cost miniature touch sensitive control displays is a nightmare for vision impaired users. But there’s a great opportunity for an AI assistant to guide the user, combining information from online sources with a first person view from smart glasses.

It turns out that a blind man can now create, and verify with reasonable confidence, surprisingly rich images. Artists will say this is derivative, shallow and not creative, but although I haven’t found a serious use case, it does feel like a bit of a super power in a fun sort of way.

Conversational voice is a brilliant inclusive, communication mode. But as illustrated here, we have limited bandwidth to process a noisy soundscape, so sometimes we need bi-modal solutions.

Seamless conversation and memory have been Huge historical barriers to effective digital assistive solutions. Hobson is a mainstream voice first conversational personal assistant with a great memory and profound assistive potential.

A summary of 5 different task categories in which we rely fundamentally on vision and some observations on how I manage without any sight. There’s a brilliant but simple app called Be My Eyes described in this post, which despite its limitations, still surpasses what AI can do

My recent enjoyable and amusing San Francisco experience of two blind men locating and riding in an entirely autonomous Waymo. The future of independent blind travel is getting closer, but there is still a real challenge in joining up the dots for an end to end inclusive experience

A reflection on possible paths for AI

I’m laying down a gauntlet to any proponent of the imminent emergence of AGI, i.e. Artificial General intelligence