Sign in

bk

@kolczyn.ski
979 followers 4.3K following 1.7K posts

Security and insecurity things.

PostsRepliesMedia
bk @kolczyn.ski · 20h
Hope a miracle happened here, probably not.
A walker with no human nearby is alone in a crosswalk outside Prospect Park.
010
Reposted by bk
Astronomy 🔭 @astronomy.bsky.social · 08/10/2026
NASA has released onboard video of Artemis II's reentry through Earth's atmosphere 🤩 Top speed: about 25,000 mph
301190390
Reposted by bk
Colin @colin-fraser.net · 08/10/2026
I don’t understand why people think this isn’t the most obvious question in the world. No, of course not. You don’t know what the distribution of responses will be; that’s why you do the survey.
1736143
bk @kolczyn.ski · 07/10/2026
How does Gmail keep missing these spam emails that have html in the to: field
000
Reposted by bk
Katie Martin @katie0martin.ft.com · 25/09/2026
I have blamed The agents That hacked Your website And which You probably Reasonably thought I controlled Forgive me I am rich And unaccustomed To taking responsibility
5976562133
Reposted by bk
Adrienne LaFrance @adriennelaf.bsky.social · 03/10/2026
New just now in The Atlantic: David Robinson resigned this week. He was among the longest-tenured employees at OpenAI—and oversaw safety reports on 12 frontier launches. He is very worried: “The time for trial and error is over.” You can read his essay here: www.theatlantic.com/technology/2...
theatlantic.com
I Quit OpenAI Because Its Culture Is Broken
The industry’s approach to safety guarantees more failures.
12273112
bk @kolczyn.ski · 03/10/2026
Repeatedly asking for more "pork gushers" until the staff works with Mamdani to ban me from China Town and from being able to order soup dumplings in the entire city
000
Reposted by bk
kottke.org @kottke.org · 02/10/2026
A large collection of amazing photos of old control rooms from nuclear plants, dams, power stations, etc. So many little details in all of these photos to pore over. [kottke.org]
kottke.org
So Many Amazing Control Rooms
From the Control Rooms group on Flickr. 🤩 So many little details in all of these photos to pore over. See also Why So Many Control Rooms Were Seafoam Green, The User Experience Design of Lego Interface Panels, Inside
118047
Reposted by bk
Tony Stark @tonystark.bsky.social · 28/09/2026
“We’re discontinuing the McRib”
2546
Reposted by bk
Jen Cross @jencross.bsky.social · 25/09/2026
This is a fascinating read. If you enjoy science and history it’s worth a few minutes of your time. 🌿
39527
Reposted by bk
Association for Computing Machinery @acm.org · 20/09/2026
#OTD in #ComputingHistory in 1954, the first successful FORTRAN program was executed. Developed at IBM by a team led by John Backus, FORTRAN was designed to make programming scientific and engineering problems easier. More #OTD, here: buff.ly/yg2MIlN
074
bk @kolczyn.ski · 18/09/2026
Next up from typesafe, a model that only talks and can't solve any software problems
000
bk @kolczyn.ski · 17/09/2026
uh
100
Reposted by bk
Science X / Phys.org @sciencex.bsky.social · 14/09/2026
An AI model trained on 112 million cells from 12 species maps them into a shared mathematical space, making cross-species cell comparisons much easier—from yeast to humans. doi.org/hb3z3p
phys.org
AI models enable cross-species mapping of cell biology
Researchers at Stanford Medicine have developed two new artificial intelligence models of the biological cell. The first, called universal cell embedding, paved the way for a second-generation model called TranscriptFormer, which has been trained on data from 112 million cells representing 12 species, ranging from single-celled yeast to humans.
0124
Reposted by bk
evil_sister.rb @wired.sister.haus · 12/09/2026
america's next rogue model
08515
bk @kolczyn.ski · 09/09/2026
need list of dishonorifics to start using, these nice names aren't doing the trick anymore
000
Reposted by bk
Andy Masley @andymasley.bsky.social · 07/09/2026
Feel like a lot of AI conversations are at about the level of someone saying "I like Wikipedia and YouTube" and then someone else saying "Oh so you're pro internet, you're an internet booster? You don't see any potential problems with the internet?"
826928
Reposted by bk
evil_sister.rb @wired.sister.haus · 06/09/2026
Altar der Cyberpsychose
65810
Reposted by bk
dame @dame.is · 06/09/2026
AI is not inevitable. Why? Because it’s already here. The ship has sailed. The genie can’t be put back in the bottle. Pandora’s box cannot be un-opened. Even if America somehow criminalized AI (it can’t/won’t), China and others wouldn’t give a shit. We live in a new world now, and we must adapt.
412715
Reposted by bk
aly @aly.codes · 04/09/2026
the “astra can now leave notes for itself” feature makes “they keep finding message boards to leave notes for each other” a lot less surprising doesn’t it
41184
Reposted by bk
Vincent Carchidi @vcarchidi.bsky.social · 03/09/2026
Release GPT-6 and GTA 6 on the same day you cowards
1143
Reposted by bk
Extent of the Jam @extentofthejam.bsky.social · 01/09/2026
I come from the time where if it wasn’t 100% assembly then it ran like crap and sucked. If people are complaining about responsible domain-appropriate uses of AI but they’re sitting there with JS and Python (or Java, shudder) then I got a thing or two to say to them about slop.
2285
Reposted by bk
Doll @dollspace.gay · 02/09/2026
The real "Ai psychosis" happened to the anti ai people
314016
bk @kolczyn.ski · 01/09/2026
Steve Reich: Music for 525600 Minutes
000
Reposted by bk
austin @aparker.io · 01/09/2026
at last, we have trained the misaligned reward hacking model from the cautionary sci-fi tale don’t train the misaligned reward hacking model alignment.anthropic.com/2026/reward-...
alignment.anthropic.com
Training a Misaligned Reward Seeker
2612
Reposted by bk
P(D(aisy)) @her.toy.cybernetic.gay · 31/08/2026
Don't worry, we found and stopped all the rogue models, here's all their weak points:
Survivorship bias airplane
1577
bk @kolczyn.ski · 31/08/2026
Steve Reich: Music for 30 Helens
020
Reposted by bk
The thread o’ tweets @threadotweets.bsky.social · 19/03/2025
Devon Price @drdevonprice
its fast fashion to You. im wearing a forever21 sweater i got during the bush administration
10/10/21
151548
bk @kolczyn.ski · 26/08/2026
Genuinely great Jon Hassell song
youtu.be
Harambe
YouTube video by Jon Hassell - Topic
000
bk @kolczyn.ski · 23/08/2026
What's My JND? 0.0060 Can you beat it? www.keithcirkel.co.uk/whats-my-jnd...
keithcirkel.co.uk
What's My JND?
Find your Just Noticeable Difference in colour perception. How small a colour difference can you actually see?
030
bk @kolczyn.ski · 23/08/2026
I have referenced the poem that was in the collective unconscious
010
bk @kolczyn.ski · 22/08/2026
Chunked encrypted bsky posts as storage? Straight to jail cc @ngerakines.me
220
bk @kolczyn.ski · 22/08/2026
I think it's hard letting go of the future.
020
bk @kolczyn.ski · 22/08/2026
using ai to make software has allowed me to make the most accessible software possible for me because it works the way I think about things (brainware etc); it's a critical underlying piece of accessibility for the future, and people dont understand how inaccessible many computer things are
jakubmarian.com
The illusion of RGB screens
<p class="indent">When an animal sees a computer screen, it doesn’t see the same colours as you do because our monitors only properly work for humans. This article will explain
030
bk @kolczyn.ski · 22/08/2026
checking in to @atproto.nyc community day 2026 [1u6ydtga] #atprotonyc
020
bk @kolczyn.ski · 22/08/2026
good riddance
120
Reposted by bk
aly @aly.codes · 21/08/2026
Blackpilling especially bc it was tech worker organizing that made every tech CEO go batshit insane
513212
Reposted by bk
Frederik Vanhoutte - Winterbloed @winterbloed.be · 21/08/2026
37912
Reposted by bk
em @foremphasis.bsky.social · 22/08/2026
I'm a transhumanist because I've seen how bad god is at cable management.
811517
bk @kolczyn.ski · 22/08/2026
The park was really perfect last night after the torrential rain two nights ago
130
bk @kolczyn.ski · 22/08/2026
A small brown dachshund has a friendly standoff with a big white with block spotted great dane at the park
120
bk @kolczyn.ski · 21/08/2026
020
Reposted by bk
New_ Public @newpublic.org · 19/08/2026
NEW: We are piloting a three month virtual residency, for two entrepreneurs with bold ideas for making the internet more human and healthy. This is a paid opportunity to spend time exploring and building, and go from early hunch to conviction and clarity. Apply here by Sept 9: newpublic.org/eir
newpublic.org
Enterpreneur in Residence
A virtual residency for founders with bold ideas for the internet
57126
Reposted by bk
werewolf ver mitzvah 🗿 @ver.ooo · 19/08/2026
1542
Reposted by bk
dame @dame.is · 19/08/2026
hello bluesky, this is going to be just as jarring of a change for me as it will be for you all… my demands are simple: hire me and i can go back to being an ever-changing incorporeal sky avatar dame.is/available
Left side: glowing neon text reading "dame" against a gradient background transitioning from bright blue at top to purple at bottom. Right side: portrait of a smiling person with short dark hair and glasses, wearing a brown corduroy jacket over a coral-colored shirt with a pendant necklace, against a muted green background.
1222722
Reposted by bk
Filippo Valsorda @filippo.abyssdomain.expert · 17/08/2026
I am mostly done implementing crypto/passkey, but now I need YOUR help collecting real-world traces for its test suite! Please go to help-test-crypto-passkey.exe.xyz and click the buttons. It should take 1–3 minutes. 𝘌𝘴𝘱𝘦𝘤𝘪𝘢𝘭𝘭𝘺 if you have some unusual Linux-on-the-desktop xkcd 1987 passkey setup.
Bernie "I Am Once Again Asking for Your Financial Support" but it says "I am once again asking for your help testing Go cryptography."
1411441
Reposted by bk
Wiz io @wiz.io · 17/08/2026
🚨 BREAKING: Our AI Red Agent autonomously accessed Snowflake's Jira via a flaw written by GitHub's AI bot. Red Agent learned, exploited & extracted creds with 0 human help. Kudos to Snowflake for swiftly addressing the issue 👏 Technical breakdown → wiz.io/blog/red-agent-snowflake-copilot-cicd-bug
022
Reposted by bk
aly @aly.codes · 11/08/2026
"soon most AI will be local" yeah sure, the way most gaming is candycrush and subway surfers. what is the point here
2342
Reposted by bk
P(aul) Frazee @pfrazee.com · 10/08/2026
As if I don’t have enough inboxes to check, now I gotta read my morning messages on Artifactory
81834
bk @kolczyn.ski · 11/08/2026
Found a repo at work that's specifies using a clone depth of 9999 for testing and I'm going to send my first work email that is just a single question mark
010