HN user
Site Reliability Engineer
https://pwnguin.net
Plus, electricity costs money while the sun is free.
Paradoxically, wholesale electric costs are negative during the day here in California[1], especially in the Salinas region. In theory you could get paid to tumble dry clothes.
It's true that markets are more _adversarial_. But there's still a lot of trouble with distribution shifts even in server metrics. As an example, our SRE team got paged a few times in the past month for traffic drops due to the World Cup. This stresses the nowcasting alert in several dimensions:
- there's no seasonal pattern to the matches, they happen sorta randomly.
- they drive increased query traffic in the hour or so before the game
- then during the game usage drops, sometimes to below "normal" depending on time of day and who's playing
So... now the accuracy of your forecasting tool depends on correctly predicting when world cup matches happen, and also who wins them!
edit: and this is just one recent example. others involve severe weather, national gameshows, earthquakes, and when you celebrate christmas.
Apparently not given how many comments in opposition I got.
The difference is that vacuuming seven times as often can compensate for a less effective process. Folding 7 times doesn't eliminate wrinkles.
You're telling me we've finally got robots that can do our laundry?
Did you watch it actually fold that shirt? Super slow and bad at it.
The linked paper suggests brain sizes were only measured once at visit 5.
After the third !, the probability of a fourth probably skyrockets =)
There's a difference between how you describe using "honestly" and how claude seems to prefer tokens like "honest" and "load-bearing." An example from some coworkers attempting to replace product managers with Claude.
Deliberately avoid a heavyweight "alert governance" process; the lightest recurring check that keeps FP-rate honest is the right dose.
And one for load bearing:
Five open questions still stand; the load-bearing two are the runbook-AC contradiction (ratify "high-priority set only") and pinning the "high-priority set" definition + SLO source-of-truth before Milestone 3 (small-sample noise on a low-traffic fleet).
One hopes that "you cant ship any more features until your test suite stops failing due to timeouts" would motivate them.
I think the request here is not about sounding like Majel Barrett but in keeping the output extremely terse and unobtrusive.
There's been a few studys showing that novices love LLM output that's long, but experts hate it. As an example, I've been tasked with using some agentic PM tool to write specs, and it keeps generating these huge page long outputs with "HBR voice" bolded summaries of paragraph long bulletpoints. I.e.:
Right-size hard, and watch the one open-ended edge. Endorse the DRI's simplifications wholesale: drop the runbook-per-alert mandate (keep 1–2 diagnostic-only runbooks for the high-priority set), and ride durability on the existing weekly incident + monthly operational reviews — no new governance. The single scope-creep risk is the coverage strand (gaps are defined by absence); bound it to gaps evidenced by real, already-missed customer-facing outages, not a proactive gap hunt. Curing ownership gaps (e.g. foo-bar, no clear owner) is finite in-scope work.
There's dozens of these every iteration. I can't imagine trying to deal with that via voice, I would just zone out after the second sentence.
I was just wondering if there was a postgresql specific version of https://use-the-index-luke.com i could send a coworker who seems oblivious to the perf penalties of missing indicies. Certainly a good start!
Sure but why pussyfoot around the issue? We should be actively encouraging each other to punish misbehaving companies. It's the right thing to do.
Probably because doing what you suggest would not look great in court:
"Dear Jeff Bezos,
Here is my signed confession letter of intellectual property theft.
Yours Truly, Mike Taylor"
but what if i learned to type Colemak?
Paging Vernor Vinge to the white courtesy phone.
Would 1.0 have fixed the wide variance in scoring?
TIL someone took over the now defunct jcmit dataset[1] (archive[2]). I expected his dataset to die off when his website did, but I guess someone found the data dump on archive.org and revived it. Which raises a question: how will this dataset fare five years from now?
[1] https://www.jcmit.com/mem2010.htm [2]: https://web.archive.org/web/20250716092935/https://jcmit.net...
Project Euler is like this a lot -- it's presented as a coding challenge but half the problems can be optimized to
return 5 # because mathI'm mostly just surprised the graph starts at 5 seconds for a mean value for all datapoints. I would have assumed it starts much closer to 1s. Which just makes the poll responses even crazier. Who is picking B when you have 25% more capacity than you need?
But I suppose the question is underspecified. How does the load balancer know which systems are busy? What happens to a request if the load balancer routes a request to a busy server?
I mean, it's the rest of the subway line thats the problem -- how many ancient sites do you tunnel through to reach the next stop?
Is it? I'm pretty sure oncologists will say "you have stage 2 breast cancer," but I wasn't in the room at the time.
Who among us has not launched a nuke in Civilization just for the spectacle?
My last promotion from level 3 to level 4 was a 50% increase in pay basically, with no change in responsibilities or scope.
performance reviews are scheduled separately from 1:1s
The performative annual review meetings can be separate, sure. But managers should be discussing with their directs in 1:1s sufficiently that no criticism or praise contained within is heard for the first time.
Thats fine, though if you do that forever you'll probably harm your promotion chances. Which, if your goal is just to get by, sounds fine?
But generally speaking, it's a chance for you to speak about your work to the person writing your performance review, and get feedback, which may be in short supply otherwise for various reasons.
AI Exponential
How much of the policy prescription changes if the exponential is actually just a series of sigmoids[1]?
Because they have so many 1:1 recurring meetings scheduled each week that they don’t have time for anything else.
Dude, a a weekly 1:1 should be 30 minutes long. And managers should have at most 10 directs, so 5 hours total out of a 40 hour work week. Something has gone haywire and it's not the 1:1 thats the problem.
I was requested to set up 1:1s not only with my team, but with each of the other teams we interfaced with, team leads on those teams, designers, stakeholders, interns, product managers who wanted to interface with us, the security team, and an endless list of other people. ... All the managers were just shuffling from one 1:1 to the next. Many never had time to deal with issues from the 1:1s because they were so busy moving on to the next 1:1.
Yes, managers go to meetings but they're not all 1:1s and if they are, the problem isn't too many middle managers, it's not enough of them. But what you describe does not sound like a 1:1. At most it's a cross-functional meeting, and should have multiple people from both sides.
The worst were the managers who had silly agendas for every 1:1, like my manager who blocked out the first 10 minutes for us to talk about our weekends with each other in a performative manner, 5 minutes per person. I could be dealing with an urgent issue in prod and he’d get angry if I tried to rush past the forced chit chat about our weekend to get back to business.
It sounds like someone got halfway through the ManagerTools guidance on 1:1s and decided they could improvise a better solution and failed. The purpose of 1:1s is to build and keep relationships, and they encourage this chitchat as relationship building, but the key thing is that the direct goes first and gets to talk about _what they want to talk about_. If you want to talk about work that's great! The best way to build a relationship is working towards a common goal, and work is pretty much the only expected common goal anyways. And if your manager _wants_ to talk about their weekend, they can, but the recommendation is to always let the direct set the first 10m of the agenda -- if a manager wants time on a direct's calendar they can always ask for more, but the reverse is much harder.
Took me a minute to parse the headline -- Sanctioned as as in "imposed penalty" (ie "sanctions"), not as in dictionary definition #2 "official permission or approval".
Perhaps because "US territories" are a thing, perhaps because it's way more newsworthy if LE bans the US, or perhaps im just a dummie.
Are you familiar with the Search Engine podcast?
Wish I could recall the podcast I listened to a few years ago that was telling the history of robo-dialers and caller ID spoofing. The general gist was that AT&T was making money off it from 1-900 operators so they weren't eager to self-regulate. So even though ending spam calling is a bipartisan issue, feet were dragged on the implementation.
If anyone's eager to do podcast archaeology, IIRC one of the angles was investigating dead government agency phone numbers, and some lady entrepreneur in the 80s. Might have been Reply All, but the market regulation angle makes me think Planet Money.
of course, politicians exempt themselves from the spam call category. Political speech is the most important speech!
By who? The only US people who seem to like the tarrifs are the ones front running trades on their announcement.