kelp

Thanks for sharing this. I wasn't aware of this, going to keep on eye on it.

I don't know much about medical research, but I've looked into some of these trials. They are generally being done by startups and not the established players who sell the diabetes management stuff.

So think disruptive tech. A new player comes into the market with something that obsoletes the old tech, and maybe destroys a whole market. But that new player makes their own money off it.

So yeah, maybe Tandem, Dexcom and Medtronic don't have a big incentive. But someone else can make some real money if they can find a cure or novel treatment that obsoletes all those others things.

The alarm fatigue is no joke, and CGMs get less accurate the lower ones blood glucose goes, so false alarms on the low end are common.

My wife has a Tandem T:Slim and Dexcom G6 and she can't turn off the urgent low alarm. The alarms drive her totally crazy.

Though I'm happy to know at least the system will disable her basal rate if she's heading low.

My wife has t1d and when we were dating we backpacked the Lost Coast trail. She was pretty paranoid of bears, so made sure we didn't have any food in our tent and did a good job of hiding our bear canisters some place where any bears couldn't move them around if they got to them. She'd had a bad experience with bears pushing bear canisters off a waterfall.

So she had a low in the middle of the night, in the middle of nowhere. She was still verbal, but barely, and sent me to get her some snacks out of the bear canisters. However I didn't know where she'd hid them. After looking in the dark for a while I had to go back to the tent and try to get enough of an explanation out of her to find the hidden canisters. Fortunately she was abele to give me enough context that I was able to find the canisters and get her something to eat.

After that, I kept my own hidden stash of snacks in the tent and make sure she told me where the canisters were.

A commenter above guessed right. Though Leavenworth and maybe George were in our high school sports leagues.

This was an absolutely fascinating and chilling story.

The thing that struck me about it is the scam didn't work for a few reasons:

1. He typically had a practice of not using his main wallet for things like this.

2. He got wary and actually read the smart contracts.

This is a level of technical competence required that's going to mean most people have to offload this to a trusted intermediary. And then what's the point of all the decentralization ideology? Because we just re-invented banks.

You got it. :)

Only slightly related, but a story from back when pay phones were still a common thing.

My first or second year of high school I lived in a rural town in Washington state with very little to do. But it was also a tourist destination. In the summer me and my friends would just sit around downtown (this was 1 street with a 4 way stop) trying to figure out what to do with our days. One day the pay phone was randomly ringing and we picked it up.

It was two girls, near our age, who were also bored and had written down the number of that pay phone and were randomly calling it. They were from out of town (if they lived there we'd have known them, it was that small of a place) and staying at their grandfathers funky vacation home just a couple of miles away.

We ended up meeting up with them and got up to all sorts of trouble that summer. We were friends for years and years after. I'm still occasionally in touch with one of them, over 25 years later.

Just a completely random way to meet some new friends that I can't imagine happening anymore.

I kind of love this idea.

It reminds me of a redis use case we had at a former employer.

We had a cluster with a high double digit number of nodes that delivered a lot of data to various external APIs.

Some of those APIs required some metadata along with the payload. That metadata was large enough that it had made sense to cache it in Redis. But over time, with growth the cluster got large enough, and high volume enough that just fetching that data from Redis was saturating the 10Gbps Nic on the ElastiCache instance, creating a significant scaling bottleneck. (I don't remember if we moved up to the 25Gbps ones or not.)

But we could have just as easily done a local cache (on disk or something) for this metadata on each node and avoided the cost of the ElastiCache and all the scaling and operational issues. It would have also avoided the network round trips to Redis, and the whole thing probably would have just performed better.

He’s actually very Christian and I’m told attends services every week. I’m not sure what denomination though.

I remember staying up all night downloading Slackware 96 (version 3.0) floppy sets over my 28.8kbps modem. I think I used it for my first 6+ months of home Linux use.

In retrospect it was probably a good thing to learn on because Slackware was so simple but did so little for you. I remember having to read a long HOWTO from the Linux Documentation Project about how to get XFree86 working. Hand crafting /etc/X11/XF86Config, including experimenting to find modelines that worked for my CRT monitor. Learning how to get Apache setup, which was kind of pointless but fun since I was on dialup.

And setting up auto dial up with a PPP session whenever I used anything that needed to get online. That seemed like magic!

Learned a lot there before trying out RedHat (4.x, before Enterprise was a thing) and Debian 1.3 (Bo) and then staying a Debian user for a long time until mostly moving to Ubuntu for server things.

OpenBSD tends to really value high quality man pages.

However to submit a patch you have to email a diff to their mailing lists. They even prefer it inline vs as an attachment.

The GitHub repo linked in the post is just a read only mirror, not used by the OpenBSD project for development.

Yeah I didn’t mean settle for money. I mean the town agrees to follow the law.

Yeah and it's cured me of any idea of ever running for one of those offices. No way could I handle sitting in every one of those meetings for several years.

And also local and even to state politics are incredibly accessible.

It's quite easy to end up having a 1:1 conversation with a serious candidates for local office. Just 2 weeks ago I donated $500 to a candidate for CA state assembly and the next day I got a call from the candidate thanking me for the donation.

And it's pretty easy to volunteer for various organizations involved in local politics and get access to the current supervisor or state assembly member who is aligned with that organization.

And public comment is actually just like it's portrayed on Parks and Recreation.

I'm not a lawyer... but...

You sue the town for violating state law.

The town either settles, or you go to court, and if you win, the court compels them to comply with the law, and the town probably has to cover legal fees.

I always thought Google built Kubernetes as a commoditize your complement strategy: https://www.gwern.net/Complement

And the Weinsteins are both very smart. But being smart can mean you’re very good at making a compelling argument for something that’s wrong.

That’s what they do. Same with Jordan Peterson.

Yeah I don’t think it can do that, unfortunately.

Corollary 2 reminds me a bit of software like djbdns and qmail which was relatively simple and worked incredibly well for it's time.

I managed a number of production systems with both and it always worked beautifully.

That said, licensing confusion always made it difficult for anyone to repackage and distribute it (until 2007) and djb was not known for being easy to work with. So that probably also hurt adoption.

I feel like there could be an Explanation #4 which is something like: for many people lots of math is used so infrequently, it's hard to retain after you're done with school.

Maybe more drills would help that?

I did pretty well in high school math and got through the most advanced math classes my small town school district had to offer. Then I opted out of driving to the next town over to do a more advanced math class with something like the 5 other kids who qualified. I think it was pre-calculus or something.

So I ended up not doing math for a year, then I really struggled in my first college math classes, and had to retake one of them.

Since then I've basically just forgotten everything because it comes up so rarely. How often do I need to do fractions? almost never, and when I do, I can just look up a factions calculator. Never mind trig or calc.

Day do day, I think I need basic addition, subtraction, multiplication, and division. And I figure out percentages a lot. But I'm also doing all of that with a calculator.

I think financial cautions, like compounding interest are the most advanced things I run into regularly.

arm64

Otherwise it would be emulation not virtualization. Parallels on Apple Silicon only works with arm64 VMs.

But that's pretty much everything I care about. I have Windows 11, FreeBSD, OpenBSD, several Linux distributions. They all work fine.

And now I have FreeBSD 14 (13 won't boot and the issue is only fixed in -current) running in Parallels on my Apple Silicon MacBook Pro... works well.

This is hardly a production use case, but I've had a bare metal server at OVH for several years. It's currently running FreeBSD 13 with bhyve. I have about 4x VMs running on it, including OpenBSD and several different Linux flavors.

Things have just worked and I've really not had any issues with it in the years it's run. The VMs are stored on a mirrored zfs zpool with sparse volumes so they only take up space they actually need.

All that said, I find the bhyve config language and command line options to be pretty painful, so I've been only using it with this frontend: https://github.com/churchers/vm-bhyve

Other concerns I have are systemd dependencies and browser support.

I'm not as familiar with how it works in FreeBSD, but the OpenBSD Chromium maintainer asked Google to upstream their patches and the Chromium team refused. So OpenBSD has 2.6MB of patches to maintain to keep Chromium going. I just checked and it looks like FreeBSD has 4.1MB of patches in their Chromium port.

See: https://groups.google.com/a/chromium.org/g/chromium-dev/c/b5...

Gnome decided a while back that systemd dependencies were ok, so the BSDs (and Gentoo) have large patch sets to maintain to keep basic functionality going.

Those are just a couple of examples of large important projects that don't want to dedicate much if any effort to portability to BSDs or other non-Linux OSs.

Seems like that puts more and more pressure on a small number of volunteers to keep up porting efforts of some mainstream programs, otherwise *BSD becomes even more niche and limited.

Also wasn't it during the 5.0 release process when DragonFly BSD split off?

And, RE the point about "that's fixed in -current", I'm reminded if a talk from Netflix engineering about how they just run -current in production.

I think OpenBSD has support for some 802.11ac chipsets, but they don't operate at 802.11ac speeds.

For example see the CAVEATS section here: https://man.openbsd.org/iwm.4

I've read maybe 3 or 4 of his books, and it all tends to be pretty much secular.

He specifically argues against ideology, and argues for adapting buddhist teachings to the needs of another culture. IIRC, in The Art of Living he argues for an "American style of Buddhism" which is adapted to American culture.

I'm generally pretty allergic to religion, but I find his teachings to be really palatable and quit helpful.

So generally I think you'd find the Plum Village tradition pretty easy to swallow.

You might want to check out the documentary "Walk With Me" which came out in 2017 and gives a pretty good picture of what Thich What Hanh and Plum Villiage tradition is all about.

Another example right here: https://news.ycombinator.com/item?id=29980227

Intel is buying fab capacity from TSMC. Backing away from vertical integration to force their own fabs to compete on the open market.

It seems like echos of, and a less drastic version of the AMD/Global Foundries split.

And then Global Foundries couldn't keep up, cancelled their 7nm work, and AMD has been sending more and more business to TSMC.