I think this Gemini behaviour is really confusing because at the moment we expect software outputs to be deterministic. We've internalised that there's no point to press retry if it doesn't work the first time, but now suddenly that's actually essential because the tool is intrinsically flaky

Using Gemini to control Google Home works better than the old assistant on a good day, but every 20 requests or so the inherent LLM stochasticity kicks in and it insists it has no idea how to control my lights

As a developer working in the public sector I can tell you I'm *begging* my bosses to let me use AI tooling unhindered. These takes are really hard to understand

Regression to the mean again. Here we pick an extreme bin of one variable (value-to-sales) at time x and plot what happens over subsequent time points to a correlated variable (returns).

It's everywhere when you learn to recognise it

Lots of security theater in my UK mobile operator's new auth but this one is funny

Yeah sure your fridge may not need an app but this bag of (delicious) basmati rice on the other hand

So close but so far with Gemini on Android at the moment. Next release I guess

Regression to the mean strikes again

Premier Inn is the UK's biggest hotel chain but something is wrong with this calendar

American payment systems are just so funny to an outside observer. Yes of course you pay for the car from the future by posting a slip of paper with your signature on it to the offices of the auto maker

Some spicy recommendations on the 7-year-old's Disney profile