stderr @stderr.dk · 4hstatic.klipy.comMichael Scott's Iconic 'How the Turntables' QuoteAlt: Michael Scott from The Office: "Well, well, well. How the turntables..." 010 0
Reposted by stderrElizabeth Gibney @lizziegibney.bsky.social · 15/10/2026There will be metrologists partying in Paris tonight 🕺🧪 I've been following this for more than a decade and know how hard they've campaigned to make it happen. Sad day for me though, as no more reporting on the leap second. So long my pesky friend! 4317 0
Reposted by stderrRichard Van Noorden @richvn.bsky.social · 15/10/2026Governments have voted to abolish the leap second immediately and replace it with a leap hour. by @lizziegibney.bsky.social | @nature.com www.nature.com/articles/d41...nature.comWorld’s timekeepers upend how they adjust time to avoid catastrophic ‘negative second’Governments have voted to abolish the leap second immediately and replace it with a leap hour. 151 0
Reposted by stderrFree Software Foundation @fsf.org · 13/10/2026Without Ada Lovelace, widely recognized as the first computer programmer, we might not have #FreeSoftware today. Thank you for your contributions, Ada! #AdaLovelaceDay #FSF 083 0
Reposted by stderrSarah Drasner @sarahedo.bsky.social · 13/10/2026Let's talk a little more about Margaret Hamilton because she deserves it In 1961, Margaret Hamilton was tackling massive systems, having previously wrangled room-sized computers that blared like a foghorn whenever they crashed. 4484143 0
Reposted by stderrThe C Programming Language @c-official.bsky.social · 12/10/2026In C23 variables have pronouns and misgendering is undefined behavior. And no you can't just always use they, smartass. It literally won't compile 731884 0
stderr @stderr.dk · 11/10/2026To whom it may concern @merriam-webster.com Can I suggest a new phrase? cait video [noun] [plural] cait videos : AI slop video featuring one or more cats (i.e. cAIts). Examples of "cait video" in a sentence: Your cait video is the worst waste of water I have ever seen. Say no to cait videos. 000 0
Reposted by stderrbryan newbold @bnewbold.net · 07/10/2026ah damn, RIP Margaret Hamilton. she took safety-critical software engineering seriously, and got people to the moon and back. news.mit.edu/2026/margare... 102144294733 0
stderr @stderr.dk · 02/10/2026*Duran Duran starts playing in the background* No-no-Notorious... 000 0
stderr @stderr.dk · 24/09/2026What is this "sleep", you're all talking about? Please tell me more. It sounds interesting. 001 0
Reposted by stderrChristen Rexing, PhD, MPH @christenrexing.bsky.social · 23/09/2026Happy Wednesday! 241789280 0
stderr @stderr.dk · 22/09/2026I can not accept any more followers at this time. Sorry for the inconvenience. I hope you'll understand. 100 0
Reposted by stderrstderr @stderr.dk · 19/09/2026Does anyone here have two or more #WesternDigital #WD8005FRYZ disks? I got 3 disks where the serial number starts with AM23 and 1 where the serial number starts with AN0S. The last one is much slower (5+%) than the other three and I would like to know, if others are seeing the same problem as me. 132 0
stderr @stderr.dk · 19/09/2026Does anyone here have two or more #WesternDigital #WD8005FRYZ disks? I got 3 disks where the serial number starts with AM23 and 1 where the serial number starts with AN0S. The last one is much slower (5+%) than the other three and I would like to know, if others are seeing the same problem as me. 132 0
Reposted by stderrIan Coldwater 🧊🚫 @lookitup.baby · 19/09/2026static.klipy.comSlow InternetAlt: I Cast Extremely Slow Internet 1927771 0
stderr @stderr.dk · 19/09/2026Just to show how much 5+% is... Ordered by fastest drive to slowest drive, I get: A 0.125% increase from number 1 to 2. A 0.544% increase from number 2 to 3. A 5.314% increase from number 3 to 4. I don't really care about a 0.125% or 0.544% increase. I do care a lot about a 5.314% increase. 030 0
stderr @stderr.dk · 19/09/2026I was going to use the 4 drives in a RAID-6, which means that each write to a data block will change the P and Q syndromes for the given stripe, i.e. each write to the RAID will involve 3 drives. With 4 drives in the RAID, that's a 75% risk of hitting the slow drive when writing. 120 0
stderr @stderr.dk · 15/09/2026"AI actor Tilly Norwood doesn’t need sleep, a living wage or a say in its roles" It also doesn't need to be "interviewed". It's not real. You might as well "interview" a spreadsheet. This is PR from Xicoia and Particle6 Group thinly disguised as news. STOP IT! 010 0
stderr @stderr.dk · 10/09/2026Next on CNN: In an exclusive interview, Elizabeth Wagmeister asks a Magic 8 Ball about Tilly Norwood's chances of winning an Oscar.static.klipy.comMagic Eight Ball: Don't Count On ItAlt: Magic Eight Ball: Don't Count On It 010 0
Reposted by stderrto wound the autumnal kitty @blurry-line.bsky.social · 07/09/2026A megaphone is 10¹² times the size of a microphone 25719 0
stderr @stderr.dk · 03/09/2026Sounds a bit like Futurama where, instead of moving the Planet Express ship faster than the speed of light, they "just" use the dark matter engines to move the rest of the universe in the opposite direction.static.klipy.comFry's Skeptical SquintAlt: Fry's Skeptical Squint 1200 0
Reposted by stderrFreya Holmér @freya.bsky.social · 01/09/2026there's a pretty simple connection here 1. humans become good at what they practice doing 2. outsourcing to an AI or even other people, means you have stopped doing the thing 3. if you stop doing the thing, you will stop improving, and, long term, get worse at it (cont.) 161306326 0
stderr @stderr.dk · 31/08/2026To make it easier for school children (and me) to remember the names of all 50 states, they are now called "State 1" to "State 52". There are no "State 44" or "State 46". Don't ask about it. Thank you for your attention to this matter. 140 0
stderr @stderr.dk · 31/08/2026Hey Clankers. Solve the halting problem. Don't make any mistakes. Don't give any output until you're completely done. 000 0
Reposted by stderrBrian Honan @brianhonan.bsky.social · 29/08/2026Happy Skynet day to all those who celebrate 6252191723 0
stderr @stderr.dk · 28/08/2026Hey @tracketpacer.com . I didn't see my boss yesterday, but today he visited my office and I got a chance to ask him about hiring you. I'm sorry to say that he unfortunately doesn't think we can justify hiring more "networking people" at the moment. 😞 I hope you find a great job soon. Good luck. 120 0
stderr @stderr.dk · 27/08/2026I'm glad you liked it. I must say however that I'm not sure if you got the right "stderr". There's a quite a number of us, it seems. If you're sure you got the right one, can you (without doxxxing any of us) give me a hint about how I know you? 110 0
stderr @stderr.dk · 27/08/2026WTF, @wsj.com ?! This is libel! That marmot was with me the whole time serving soup at the local homeless shelter! 1122 0
stderr @stderr.dk · 26/08/2026And 0.999999940395355224609375 is stored as sign = 0 ((-1)^0 == 1), exponent = 01111110 (2^-1 == 1/2), mantissa = 11111111111111111111111 (1 + 8388607/8388608 == just under 2). 000 0
stderr @stderr.dk · 26/08/2026The exponent has to be multiplied by (-1)^sign * (1 + mantissa/(2^23)) to get the actual value of the float. For example 0.0000000000000000000000000000000000000117549435 is stored as sign = 0 ((-1)^0 = 1), exponent = 00000001 (2^-126), mantissa = 00000000000000000000000 (1 + 0/8388608 == 1). 100 0
stderr @stderr.dk · 26/08/2026I'm not sure what you mean by "bits 3-8 are the 0-1 exponent range". For 0 < x < 1, the 8 exponent bits, as stored in the float, has a value between 00000001 (1) and 01111110 (126) (with the exponent bias), which translates to an exponent between 2^(1-127) = 2 ^ -126 and 2^(126-127) = 2 ^ -1. 110 0
stderr @stderr.dk · 26/08/2026If you can live with "uniform distribution" in a range from 2^n <= x < 2^(n+1), the problem is much easier. Set the sign bit to 0, the exponent to 127 - n, and the mantissa to the 23 most-significant bits of your 32-bit unsigned int. Unfortunately, that is, of course, very rarely what we want... 110 0
stderr @stderr.dk · 26/08/2026And that's breaks the "uniform distribution". While the ranges gets smaller and smaller the close the numbers gets to zero, the count of numbers in each range stays the same, 2^23, so the distance between numbers in a range has to become smaller (halved) as we move to a range closer to zero. 110 0
stderr @stderr.dk · 26/08/2026And likewise for numbers less than 1: 2^23 equally distance numbers in 1/2 <= x < 1 2^23 equally distance numbers in 1/4 <= x < 1/2 2^23 equally distance numbers in 1/8 <= x < 1/4 etc. 110 0