There does not yet exist a (strong) chess engine that tries to force its opponent to win against it (who is assumed to have the same objective)

You might enjoy hatetris (tetris, not chess, but it optimises against you in a way that felt similar)

I measured my office too, with an Adafruit SCD-30 sensor, it also got to 1500ppm faster than I expected. And it took a long time (12+ minutes with fully open windows) to get it down to an acceptable level again. Certainly compelled me to do that more often.

Surprisingly, i couldn't find any calculator or theoretical approach for estimating this (given room of a certain size, how long does gas need to equilibriate with outside atmospheric composition to within some tolerance, through a hole of certain size)

I wonder if there is a way to integrate this woth home assistant for historical data collection, then use it with other sensors to build a prediction model or some usedul analysis. A bit less principled approach but voild be interesting!

Here a Youtube playlist of some people playing it: https://youtube.com/playlist?list=PLMrpfY5oU1DY79EQTQ_aD0-Ub... (using cards submitted by their viewers)

Do you regularly test your AI on the https://github.com/google-deepmind/formal-conjectures collection?

The Gnome desktop environment usability degradation in recent releases (stuff like drag and dropping files and folders between the desktop and the file explorer not working anymore, or not being able to create new empty files with a right click by default without having to create custom templates, being unable to pin apps to the launcher without messing with files, and more) was so horrendous that it felt like actual sabotage was being committed. Who in their right mind would decide to make the UI actively worse?!

I wonder if/when DeepMind will try to tackle the problem of finding potential room temperature, ambient pressure superconductors (and if it can be approached similarly to protein folding)

From a couple of weeks ago:

>Google DeepMind will establish its first “automated science laboratory” in the UK...

>The lab will focus on using AI tools to develop new materials for superconductors, solar cells and semiconductors.

(https://www.ft.com/content/b20f382b-ef05-4ea1-8933-df907d30c...)

Not enough training data and not enough theory for synthetic data, sadly.

It's not a choice if it's an underhanded default opt-in without knowledge, understanding or explicit consent

And it used to reset to "on" after random firmware updates.

I would prefer that the setting not exist because the functionality doesn't exist.

Coming up around 2041 (hopefully) will be the https://habitableworldsobservatory.org - which will be the first telescope sensitive enough to detect Earth-like exoplanets around Sun-like stars! Check out the "Simulated Observation of the Solar System" video toward the bottom of that page, coolest thing I've seen in a while!

It's still good for asking questions, I'm not aware of many more other places where friendly subject matter experts hang around (IRC, Mathstodon, Reddit?).

It would be better if closing questions would cost 1000 reputation. That's one advantage AI has over it - it will at least try to answer your question every time and not just randomly shut you down for its own (wrong) reasons.

I don't understand why lawyers haven't gotten on this train yet. The number of possible class action lawsuits must be unbelievable