It was a painful search, but I finally found the video. I had an odd deja vu experience skimming these sandbox escape headlines. Almost a decade ago experts were already exploring the realistic expectations/implications of having/needing to sandbox ‘a malevolent AI’.

That discussion was awfully optimistic though. Sandboxing has inherit limitations, but increased friction usually incentivizes little to none. I still think ‘malevolent humans’ physically/quietly attached to a network are far more scary though. AIs are haphazardly rowdy, pushy, and loud.

I was asked many times before. And no, I don’t actually use Reddit or Twitter/X! Last time I “perused” Twitter/X the “tweets” were seemingly inaccessible anyway. I hope Reddit follows suit or even better yet, becomes login only (pretty please).

Governments and other truly open/public entities should wise up and future proof themselves already. Maybe it’s already happening?

In fact, I quickly prototyped this theme mainly to show that it’s probably not that hard. “Agents” are kind of a thing now. There are no more excuses.

I tried to make that post land on April 1st to make it more confusing but got busy and messed up. For the sporadic peruser, every other day is April Fools day anyway. (:

Another systemd tip is the analyze timespan and calendar commands for timers. They’re good for planning out timer sequences or explaining them to the microsecond.

I think I told someone back when that the safest program might just be ‘hello world’ in the safest language of your choosing.

And that came out more like a joke, but there really was a note on how to achieve an almost perfect sandboxing score in systemd. The trick is to not relax.

There’s another old interesting thing called the estates of the realm. These are the loudest voices in/on the social sphere, roughly.

If somehow (anything’s possible nowadays) the fourth through sixth become as distrusted as the first and second, then know that for the record I was simply a commoner… just one of the people. ;)

On poetry is the funniest part as far as I’m concerned. Maybe that’s an amusing intro but a lot of scams today really do look like something out of the 14~17th century. Here’s a fun one: Would you put your full trust in John Keely? Another is Ming (14) or Minguo (20)?

I sometimes wonder what kind of 12-sided coin Keely would flip and how many would carry the suitcases. He was good. Ponzi and Keely have similar smirks but beware of cross dressing lawyers. I also made the mistake of scanning recent world news and surely all’s right with the world?

I saw a clip about the new Culdcept Begins and it made me think. My “new” 2/3DS is still in decent shape. These are the last three console games I remember playing:

• Ketzal’s Corridors (2012) (3DS) (Video)
• Culdcept Revolt (2017) (3DS) (Video)
• The Alliance Alive (2018) (3DS) (Video)

I’m lucky enough to still get free proof reading/linking. I’ve got random aggregators running about that I don’t read but others really like to.

Understanding what’s up at a glance today is no joke. Misleads are everywhere and journalists play on ultra hardcore mode. News link aggregators are so-so, but too linear, and lack dimension.

Parliaments have Hansard (uh from Thomas) for sussing out contradictions. But for the news, I’m thinking more like Reading Steiner… These timelines are wild bro.

Another funny common occurrence:

Question: Isn’t this AI generated?

Huh? Wait. What? This is AI? (Reads further; brain fully activated)

Response: I don’t know, I’m from these parts and this part can’t be right either. (Pulls out calculator and checks the elementary numbers) It’s AI! I knew it all along!