Hah, that was fast! Thank you. They must have had preview access. It didn't bode well that SimonW [0] had to explicitly tell GPT-5 to use python to get a table sorted correctly (but awesome that in can use python as a tool without any plumbing). It appears we are not quite to AGI yet.
I missed the edit window before noticing the completely fair flagging of this post, so I'm going to try again, this time, following the rules I have now read more recently and carefully. Sorry folks, second post, new norms. This is what I wish I'd posted:
I don't think their current team thinks about the product website very much. I think this has become even more true since the Broadcom acquisition, but even before that, they still had - and have! - a banner about being "here for you in these unusual times." It's from early COVID times.
As of this writing, their latest blog post is about _adding PayPal as a payment option._
It's from December, before the Broadcom deal closed. I think "add paypal" and "eliminate non-enterprise paid tiers" are different enough as product directions to illustrate the impact of Broadcom's thinking on the product.
For anyone coming to this conversation later, there's a site at talk.storytime.solutions dedicated to talking about Pivotal Tracker alternatives.
Gmail generally works fine. Outlook works too most of the time.
"Outlook Enterprise" is a mess that refuses email for no good reason. Sometimes it's because Microsoft's DNS resolvers are broken (and can't validate SPF/DKIM), sometimes it's because the mail server rewrites message headers and then tries to validate the signature (which fails, obviously).
Set up an embassy. Register your domain for Outlook, but don't really update the DNS (just add Outlook to SPF and DKIM to pass the validation, but don't change the MX). Then tell your MTA to send through Outlook servers when the destination is there (detecting this is a bit tricky), otherwise route normally.
I haven't really implemented this in production, but it worked for me one time as a proof-of-concept when I had an issue with disappearing mail - my message went through that time. Later it worked without any tricks, so I haven't bothered.
Sorry, nope. It was my own idea (to best of my memory, I haven't seen this anywhere else, so I think it was an original one), I've tried it out, it worked, but I never finalized it, nor wrote anything about it.
It wasn't anything complicated, though. I've just did the documented steps to set up Outlook with my own domain (not sure if that's a free option, I have MS365 subscription for Office apps), except that I made no changes that would disrupt my existing mail system - I've added to SPF and DKIM instead of setting/replacing them, and I haven't touched any MX records at all. Then I've just grabbed Outlook's SMTP details and sent a test email to my other test Outlook account via SMTP and it got delivered with my email address, which gave me a confirmation that my idea had actually worked. I haven't really updated my MTA to do the routing thing, as I was about to replace it anyway (I did since then, replaced Postfix with Maddy).
Maybe your domain isn’t on blocklists, but what about your IP?
Assuming you don’t send spam, the question of whether or not your IP is on blocklists is primarily a function of both how long you’ve had your IP address, and how well-behaved its neighboring IPs are.
For example I just tried checking[^1] the public IPv4 address of a VPS I’ve been managing for about a year. It’s never sent or received any email for at least as long as I’ve been using it, but it’s showing up on two blocklists![^2]
Surprisingly, my home IP address (which is a dynamic IP, in a pool of other residential IPs) is only on two blocklists[^3] as well. I would have expected more, because in my experience IPs known to be residential are almost always blocklisted, just as a matter of fact!
Of course this doesn’t check the main blocklists used by Microsoft and Gmail. I’d expect my home IP to be on those (because I’d expect the entire pool to be), but maybe my VPS might not be!
Anyway, the point I’m trying to make is that whether or not the battle has been “lost,” it’s definitely stacked against anyone who doesn’t start out with essentially a known-good, static IP address that you can control the reverse DNS record for.
You could do absolutely everything else right, but if you can’t get ahold of an IP address from a reputable provider that isn’t known for spammers using their service, you’ll probably have a lot of trouble with delivery of outbound mail. And that’s not a battle that I want to fight right now…
Yeah... this tends to be the issue. Also, I wouldn't even bother trying to get removed from the UCEProtect blacklists, it's literally just extortion. (luckily I use a small hosting provider so they're not even on the UCEProtect lists)
You're naively ignoring the simplest solution: smarthost through a provider with a good reputation.
You still get to control your incoming email, your filtering, you get logs of everything, you control your email at rest, and you'll still get good logs for outgoing, but deliverability simply is no longer an issue.
So, what other objections do you have for email self-hosters?
Then either you didn’t configure your server correctly or you were trying to run a server on an IP address that’s part of a blacklisted netblock (e.g. residential).
I’ve had a mail server in colo for over a decade, and I even recently had to change IP addresses on that server, and I’ve had zero deliverability issues. Set up SPF, DKIM, and reverse DNS, and obviously don’t do anything stupid like send spam or leave an open relay, and you should be fine.
> Then either you didn’t configure your server correctly or you were trying to run a server on an IP address that’s part of a blacklisted netblock (e.g. residential).
This is frequently the case but not always. Sometimes you don't have any server issues, and originating IP is totally fine, but your messages are 250-accepted then somehow just disappear into the void without reaching the recipient mailbox (not even the "spam" folder).
Fortunately, it's rare (in my experience), but super annoying when this happens, because with FAANGs there's absolutely no way to reach out for any technical support (unless you know someone who works there and they can help you).
Same. Rclone is wonderful because it supports a ton of different backends which makes it super easy to mirror. It's also got some great features like crypt where you can encrypt everything locally, thus sending the data all as ciphertext.
It's still sea-level. The transition altitude just changes the altimeter setting from one that matches the current air pressure to a standard pressure setting.
Planes report pressure altitude via their transponders. 20-30 feet up and down is very normal for an autopilot.
GPS altitude is used for vertical guidance for certain types of GPS approaches (i.e. "LPV" approaches[1]) and requires the airplane's avionics to be equipped with a WAAS[2] receiver that provides accurate altitude information.
You might be underestimating the amount of traffic Google sends to publishers through Google News. Anecdotally, I get Android notifications from CBC, Global, etc. through Google News daily and do sometimes click on them.
We had such a remediation done recently, in Alberta close to the rocky mountains.
Costs were around $2,000 CAD to install a sub-slab depressurization system (i.e. a fan that pulls air from below the house and vents it away from the house).
Radon values dropped from around 500 Bq/m3 to less than 20 Bq/m3.
Those systems must be new. The classic solution is to install drain pipes before pouring the slab, so one can imagine how difficult it would be using 1980's construction techniques to retroactively add piping below a finished house.
We also had to dig trenches to lay natural gas lines, but we have a way to do those with horizontal boring techniques (of course then people who didn't know what they were doing put them straight through sewer lines, causing backups, visits from the Roto Rooter man, and subsequent explosions due to dumping natural gas straight into the sewer main).
Is it safe to assume they're using something like that with perforated pipes to exhaust radon?
I had a side-job installing active, subslab depressurization systems in the 90s, and they were quite well established at the time. The EPA has quite clear and cogent guidance on radon, paraphrased as follows:
Before going with active depressurization, start by installing a sealed sump cover, caulking the basement wall-to-floor joint, and caulking all the cracks in the basement walls, in that order. If that doesn't get the number down where you want it, drill the subslab access hole and install the ventilation piping to the outdoors, but don't install the fan in the middle. Only of those fail to get adequate results, install the fan.
It's really simple and quite cheap. I don't think we ever did a job that was over $1000. Costs of running the fans were pennies a month, and Fantech still sells the classic FR-100 fan all these years later, though there are even quieter options now.
So make the interior of the house the path of most resistance, and see if it will passively vent itself, and if not bore a horizontal tunnel halfway under the house and install a fan to do the job?
No! There is no horizontal boring, at least not on any system I ever touched or heard of. I don't know where that idea came from.
There's typically enough gravel under the slab and around the foundation that there's plenty of soil-gas transport without doing anything more. So the install is just a simple vertical hole through the corner of the slab somewhere. Maybe you scoop out a few handfuls of soil before sticking the pipe in the hole, but there's no horizontal boring. Here's a very typical one:
Not shown is the crawlspace portion of that system, which would use a horizontal perforated pipe _laying on the surface of the crawlspace floor_ and covered with a plastic membrane that's taped to the walls:
>Those systems must be new. The classic solution is to install drain pipes before pouring the slab, so one can imagine how difficult it would be using 1980's construction techniques to retroactively add piping below a finished house.
For instance, the first subway systems in London were installed via cut and cover. Dig a trench, build a roof. Usually those trenches were in the street, or a former canal, not under a building.
Using modern tunnel boring equipment, Seattle's attempt to move the downtown section of 99 into a tunnel ended up stalled for a year because the ground shifted and pinched the boring machine in place. Oopsie.
And of course if you're worried about gasses, puncturing a solid piece of concrete is just going to make that problem worse.
Because of the question of liability: If it was lab made and accidentally released, was it due to recklessness or criminal negligence? Is someone guilty of involuntary mass-manslaughter? Or if this was state-sponsored research, could they be found liable for the damage caused?
https://x.com/fchollet/status/1953511631054680085