The UX for setting up child accounts makes me certain none of the people involved have ever tried doing the full process themselves.

It took me an hour to set up a Minecraft account and just spent an hour to get a Fortnite crew sub.
#gamingparent #ux #extrahell

There are also what I consider good ones, like this. I think it's good because while it's a niche question about a specific thing, it does not ask for trivia but rather for something I believe a expert in the field could answer without looking it up in a textbook.

Part of the data is open and you can query it with a huggingface account.

And based on my random sampling, yeah I can understand why the current models score so low.

A lot of the questions in my random sample seem to be for trivia.
huggingface.co/datasets/cai...

At first glance, it looks good. Crowd sourced hard questions which the models currently fails.

The example questions on the website look like a good balance of facts-a-expert would know and things an expert could deduce.

But marketing is marketing. Let's look at the raw data.

I totally did not accidentally pick up the second book (about eels), continue at the bookmark and read three (very interesting chapters about Sigmund Freud and Johannes Schmidt), while thinking "this eel/psychology diversion is really long, I wonder when we get back to ADHD"

What the heck #brain ?

Rate my #homeassistant - dashboard

It's getting there, slowly. Running on my old trusty iPad2. Getting that to work took a bit of trickery as the HS dashboard required modern JS, but #novnc to the rescue. I'll write it up soon.

Changed to winter tires on our "second car", real workhorse of bike will continue to serve us throughout winter.

#winterbiking #longtailbike #tern

I made myself a twitter archive from my export, and while going through the old tweets I found this fun little thread.

I attempted to make the chapter on QR codes from Mikko Hyppönen's book fit into a QR code.

#qr #twitterarchive

nyman.re/twitterarchi...
nyman.re/twitterarchi...

Finally got around to bridging my bluesky account into the fediverse so I can cross reference my bluesky posts there, as easy as following @ap.brid.gy

a interconnected internet, makes me nostalgic

The other finding in that paper (which was only looking at people aged 53 to 83) was that people who drink daily outperformed those who never drink, and that correlation was stronger than the morning vs evening person...

So not sure what to think about it :-D

For example, the research only looked at the cognitive performance of people aged 53-86, which feels like kind of a big deal to leave out.
And when checking the paper, unless I misunderstand something, there is actually a stronger correlation than the sleep one.
[4/n]

Winds are changing, hopefully for the better.
#riptwitter #ripreddit