In the Monday announcement, OpenAI claims the recently updated version of GPT-5 responds with “desirable responses” to mental health issues roughly 65% more than the previous version. On an evaluation testing AI responses around suicidal conversations, OpenAI says its new GPT-5 model is 91% compliant with the company’s desired behaviors, compared to 77% for the previous GPT‑5 model.

I don’t particularly like OpenAI, and i know they wouldn’t release the affected persons numbers (not quoted, but discussed ib the linked article) if percentages were not improving, but cudos to whomever is there tracking this data and lobbying internally to become more transparent about it.

Root system growing around and into another trunk. Two sticks as one, fused by nature. I say it counts.

Thunder

Funny seeing this as the first thing in Top as I am missing my flight connection due to it. ¯_(ツ)_/¯

My LG does picture frame looking pictures it came with (never connected to internet).

They seemed to have thought the brightest white of the sun would be the best choice.

Nighttime does not exist if you turn on the TV and the Shield is sleeping.

I would bet Newsmax leadership is actually fine with it, but realizes their viewer base would immediately lose trust in them, which is not fine (for them).

The reason this reminds you of Relativity is because that’s what it is with modifications.

Credit where credit is due: https://github.com/relativty/Relativty

He basically built (and advanced) Relativity’s open source headset.

Graphene could sandbox the integrity check, just like they do with the Play Store.

Set of cron jobs that check services, then send a Matrix message if there’s an issue.

For the cron jobs, I pipe stderr to another script that watches those and does the same.

If all fails, and internet is unavailable and the router crashes, a Pi will toggle a relay, cutting and resupplying power.