I've just compiled the latest 6.18 kernel, released this morning ("Mon Aug 3 11:22:13 2026 +0200" on the 6.18.42 commit message), and there's a 6.6, 6.12 and 7.1 as well. To my layman's eye there appear to be enough "leaks" and "corrupts" in the changelog to warrant security concerns. With the way Linux kernel CVEs work there won't be any yet, the team needs to now look at this new set of releases to judge each commit.

So, expect another Debian kernel update in the next week or so.

Don't conflate "forces a burial site to last 100,000 years" with the half-life.

I'd assume the majority of the elements' half-lives are much shorter than 100,000 years, so that they will have dropped in population enough after 100,000 years for the waste to be considered 'safe'.

Wikipedia: Radioactive waste gives a run down of the isotopes of concern and their half-lives.

That and the recent (the patch was released Xmas day, which alerted blackhats to it) exploit for MongoDB.

Source, a security researched on Mastodon who did some work on reporting the situation to Ubisoft: https://cyberplace.social/@Gos... https://cyberplace.social/@Gos...

As a point of reference, the current ISS cooling system can reject 75kW of heat, from a combined area of 84.864m^2. Those radiators are only part of the system, the External Active Thermal Control System, and they are continuously rotated to optimise heat loss from them.

Ref: https://en.wikipedia.org/wiki/... and the cited https://www.nasa.gov/pdf/47348...

So that ballparks 2GW requiring ~2250m^2 of radiators. Remember, it needs mechanisms for rotating, and to be placed so that they're mostly not radiating at each other, plus all of the plumbing and pumps for the transfer of heat to thrm.

I forgot to say... I was made aware of this article by utilising the /. RSS feed.

Having used Google Reader for some years before the shutdown, I migrated to hosting my own Tiny Tiny RSS instance on my home server. It works. The developer can be ... a little hostile ... to people not thinking before they report issues, but otherwise it works well.

Now, if only sites that offer RSS feeds would always think to exclude them from any bot detection rate limits.... Even YouTube was guilty of not thinking about this, at least a couple of years ago when I had to reduce the poll rate of some channel feeds from there.

The historical landers took a different path/trajectory down to the Moon. They came to an almost-halt laterally, and then descended straight down, giving them a good static view of their landing area from very high up. The problem with this is you use up a lot of extra fuel due to gravity losses. That then has an impact on getting less payload (mass) down to the surface.

The modern landers are trying to be more efficient in fuel usage, so are doing something akin to a "suicide burn" where they adjust their orbit so it will impact the surface, and then only just cancel out lateral and vertical motion in time. This entails facing some amount of 'sideways' until very shortly before landing, so that the engines are still cancelling out lateral motion. That in turn means a downward facing altitude sensor, often LIDAR based (lower power requirements than the RADAR that historical landers used) can't even see the landing area until the lateral velocity is mostly cancelled out and the bottom of the lander is actually facing the surface. This in turn means there's an issue with lunar dust being kicked up by the landing engines, just as the lander is trying to judge its distance to the surface to decelerate to a near-stop in time (but not too soon).

I'm getting most of this from Scott Manley's video on this topic: https://www.youtube.com/watch?... . Specific discussion on this topic starts at https://youtu.be/ISZTTEtHcTg?t... . Discussion of the landing system, including how little mass made it down on a Surveyor: https://youtu.be/ISZTTEtHcTg?t... .

From TFA:

If this cooperation agreement isn't in place, the only alternative left for law enforcement is to issue a European Investigation Order (EIO), but responses for these can take up to 120 days, which isn't ideal when you want to catch a drug dealer who's only in the country for a weekend.

So, instead of campaigning to get those responses speeded up they want to do away with the entirety of PET.

There's something to be said for not charging "too fast" as it will mean the battery/phone getting hotter whilst that happens. Pixels have a feature where it'll learn your nightly habits and aim to be 100% charged something like 30 minutes before you'll wake up, trickle charging the whole time.

I suggest you go to TFURL that I linked and read the details of the experiment, rather than us going back and forth with me pasting relevant parts. For instance all participants were still receiving the other support they normally would have. That's both a confounding factor for the money side of the study and also a possible explanation for the gains by all groups.

Again, there were differences in the groups that received more money, just not in all areas.

Now, the actual project's own reporting does echo this:

The number of participants who reported stably housed doubled in all payment groups from enrollment to the ten-month mark. This jump was statistically significant amongst all groups.

But, you appear to be ignoring:

And participants in those two groups reported an increase in full-time work, while the control group reported decreased full-time employment.

Specifically Group A ($1k pcm) 18% to 23%, Group B (Lump Sum) 24% to 37% and Group C ($50 pcm) 26% to 21% (a decrease).

And do I actually have to point out that $50 pcm is not nothing. For some of these people that alone could have made a difference. I'm guessing the researchers judged that they needed to offer all particpants some remuneration to help gurantee their continued involvement, hence not having a "paid nothing extra" control group.

Also the tier of Prime Video you'd be on with a free trial is the one with embedded ads. Drove me nuts watching a few things recently.

I tried DDG searches periodically throughout the Bing outage.

Initially they all returned a flat out error message, after a timeout. Literally no results. No way to "try again with ".

Later on there was a message about having problems, with hints about how to have DDG just go query Google (or some others I forget). For Google it was to start the query "!g ". I believe it then literally just performed a redirect to that other search engine.

So, yes, it does seem that in order to return any results DDG was 100% dependent on Bing on that day. Obviously they might be rethinking that now.

Can you tell I was typing that reply on my phone at the end of a long day ?

Despite those statements DDG search was entirely hamstrung when Bing search was down. So, even if the statements are true in the sense that other sources are used to augment results, and possibly used solely for certain classes of queries, of Bing is returning "bad" results then DDG if likely to do so as well.