Tootfinder

The rewrite of the Sisyphus server continues, I've got all of the GET endpoints and about 60% of all the endpoints finished overall. Some initial testing on the `/workers` and `/queue` endpoints show some worrying results from the old version and some great results from the new version. The `/workers` endpoint started throwing errors at about 800 RPS on the Python/Flask version (the old one), the rewrite is showing around 4800 RPS (Golang/Gin). The `/queue` endpoint doesn't show a…

Finally some progress. But it's baffling.
The very first step of the de-embed filter is to execute FFTs of dinFwd to vectorTempBuf3 and dinRev to vectorTempBuf1.
It appears that somehow the FFTs trade places in the failure scenario, with the frequency domain version of dinFwd being written to vectorTempBuf1 and dinRev going to vectorTempBuf3.
I don't even know where to start debugging this, it smells like something going wrong in the depths of vkFFT with some state …

This lunatic… this idiot son of modern-day barbarians is proclaiming that his morally-vacuous plastic-trinket empire is going to commit genocide and what’s the rest of the world doing? Anything? Where are the condemnations? Where are the sanctions? Where’s the mass mobilisation against this Fourth Reich?

99% of stakeholder "ideas" are just copying a feature they saw on a competitor's product. And the features they notice most are the ones that yell at their users the loudest.
Now all software (not just consumer apps, but also metrics dashboards) yells at you, and trains you to attend to its yelling. Strategy becomes dominated by the overriding need to make the good number go up.

The circular logic of our metrics
We design what is familiar. The design patterns we adopt are the ones yelling at us the loudest.

Because @…’s Flipboard site won’t scroll for me (?!?), here’s the paragraph with the punch:
❝It’s worth noting that over the decades, Pixar has explored topics such as, but [not*] limited to: becoming a widower, suddenly becoming a single or adoptive parent, environmental collapse, existential dread, motherhood, and job burnout. The first Incredibles has a subplot where Elastigirl thinks Mr. Incredible is having an affair, Toy Story 3 has the toys accept death by furnace, and Elemental is about interracial relationships. The last two movies Docter directed were Soul, which is about a Black man finding joy in his life again, and Inside Out—maybe the most therapy-ass movie (and series) in the studio’s entire 40-year tenure—is about messy preteen emotions that lead to its main character running away from home.❞
* The word “not” is missing in the original, surely an editorial mistake?
https://flipboard.com/@gizmodo/i09-ifm0rl1hz/-/a-M010weUPTnu1lZIfiOE3oQ:a:1876139665-/0

Human Traces II🚶
人类踪迹 II🚶
📷 Pentax MX
🎞️ Harman RED 125 (FF)
If you like my work, buy me a coffee or a roll of film from PayPal #filmphotography

The largest federal prison in the United States holds roughly 4,000 inmates.
The new warehouses DHS is purchasing in Phoenix and Philadelphia are designed to hold 8,000 people.
𝗧𝗵𝗲𝘆 𝗮𝗿𝗲 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗰𝗶𝘁𝗶𝗲𝘀 𝗳𝗼𝗿 𝗶𝗻𝗰𝗮𝗿𝗰𝗲𝗿𝗮𝘁𝗶𝗻𝗴 𝘁𝗵𝗲𝗶𝗿 𝗲𝗻𝗲𝗺𝗶𝗲𝘀.
Operation Liberty is ready to deploy investigators to these sites immediately.
We have the vans.
We have the personnel.
We have the strategy.
The only thing we are missing is the funding to scale.
$100,000 allow…

I just gave to the Save America Movement!
Join the fight »

Hm. Google Gemini has humans reading sampled chats for "quality assurance." And you can't opt out. How nice...
I wonder, how ethical would it be to flood Gemini with chats that say mean and nasty things about the human reviewers, for them to see? Or to encourage them to find a better job?
Not that I recommend being cruel to anyone, ever. But... guerilla tactics can be effective.

I compared the privacy of ChatGPT, Gemini, Claude and Perplexity — here’s the one you should trust most with your personal info
I read the fine print so you don't have to

An eventful day at the home office so far... just spent the last 20 minutes trying to clean up the mess of pushing bag of besan (chick pea) flour off the top shelf in the pantry, which burst on impact, maximising dispersion on the coffee making stuff & the rest of the pantry. Much blue air was created. But before that had a magical few minutes when a pīwakawaka flew into the house, happily flew around a bit, & happily flew back out again.

The 100baseT1 decode benchmark is dropping below real time when I have a lot of traffic on the link. Do not want.
Sooo now to figure out how to either double the speed of the decode or steal some more time from something else along the critical path like the edge detector or CDR filters that I've spent a ton of time optimizing.
It looks like the workload scheduler might be a weak point here: the greedy algorithm makes the eye pattern run as soon as data is available, which ma…