Reposted by calebaly @aly.codes · 15hCompletely dishonest framing just being echoed without commentary, very cool. You mean the app progressively locking itself down in a years long effort to enable lock in is doing this? wow, must be AI 514218
Reposted by calebJohn David Pressman @jdp.extropian.net · 19hI'm just going to assume malice on Anthropic's part if they're not giving the OpenSSH maintainers access to Mythos. 413020
Reposted by calebamos @fasterthanli.me · 20hfinding that the answer to "what can you build today that relies exclusively on EU cloud services, inference, CDN, DNS, etc." is "quite a bit, but it's not gonna be comfortable" 3423
caleb @caleb.jasik.xyz · 21h@aly.codes how does tangled.org/aly.codes/ca... compare to chezmoi?tangled.orgaly.codes/catterySQLite-backed dotfiles manager for silly cats. 100
Reposted by calebDoll @dollspace.gay · 22hIf openssh which basically the entire internet relies on doesnt warrant mythos access then what does? Whats the criteria? Why isnt that criteria transparent? 716020
Reposted by calebLexi ⛯ @ash243x.bsky.social · 30/09/2026cute that ai assumes i need an iso. no, little buddy; original media. 5441
Reposted by calebFinfet @finfet.sh · 30/09/2026poorly working gaze tracking on steam frame. i've had it for less than a day. god i love mostly open hardware/software 1302
Reposted by calebIan Coldwater 🧊🚫 @lookitup.baby · 30/09/2026BREAKING NEWS: “Do geese see God” is a palindrome 14275401802
Reposted by calebaly @aly.codes · 29/09/2026Anthropic is basically an enemy of world peace. quite explicitly so. 1695
Reposted by calebmattie lubchansky @mattielubchansky.com · 29/09/2026awoke this morning with a Vision 4990221661
Reposted by calebeva (^_^)/ @eva.computer · 29/09/2026what they don't say is that our noble cyber warfighters, oh whoops sorry i mean "cyber defenders", could also just use these same open models to report vulnerabilities without Anthropic gatekeeping everything for their own financial gain 210018
Reposted by calebEthan Mollick @emollick.bsky.social · 29/09/2026Leaving aside Anthropic's incentives for publishing this research, there is no doubt that open weights models will soon create the same security threats that closed source models have been demonstrating, except without guardrails. We are close, so plan accordingly. www.anthropic.com/research/glm... 714327
Reposted by calebJames Medlock @jdcmedlock.bsky.social · 28/09/2026I see people putting AI in the same bucket as NFTs (speculative technology that will go away when the bubble pops), but I think the internet is the better comp. Sure there's some early bubble-y activity, but the underlying technology is real it's going to dramatically change things. 2453758
Reposted by calebphilpax @philpax.me · 28/09/2026I am not particularly worried about apocalyptic paperclip-y doom (per usual caveats about the physical world), but I think "the HF Incident but scaled up" is a very possible kind of fuckup, and it could be calamitous for internet infrastructure / trustworthiness of digital information 6785
Reposted by calebEmily Gorcenski @emily.gorcen.ski · 28/09/2026Here’s the most incredible story you’ll read today. A small plane with four people aboard crashes in a field. Another pilot witnesses it. That pilot happens to be flying a bunch of skydivers. They offer to jump to assist, are given permission. The skydivers happen to be EMTs. Everyone survives.wisn.comSkydivers alert emergency crews, assist after plane crashes near East Troy airportFour people were aboard the single-engine plane when it went down in a cornfield near Carvers School Road in East Troy at about 1:30 p.m. 122102612648
Reposted by calebconputer dipshit @davidcrespo.bsky.social · 28/09/2026even with the best model in the best harness it would take skill and care to get something like this right, e.g., making sure the agent session that talks to people doesn't even have access to the info you wouldn't want it to reveal. too bad they claim to have done exactly this! ai.meta.com/muse/ 71008
Reposted by calebAstra ⎔ @astrra.space · 28/09/2026experts being somewhat fungible in MoE LLMs is peak like what do you mean “if you don’t need the non-cached expert badly enough you can just use a cached one that’s close enough” is both a valid optimization option AND basically doesn’t affect accuracy at all??? 11914
Reposted by calebSam @samd.bsky.social · 27/09/2026we could solve the western water crisis overnight by simply charging alfalfa farmers market rates for water 211002181
Reposted by calebCameron 🇺🇸🗽🦅 @cameroncorduroy.bsky.social · 27/09/2026talking about water use policy is impossible because "AI data centers use a lot of water to cool the computers" sounds way more believable than "they're draining the Great Salt Lake so the Governor of Utah can sell alfalfa to Saudi Arabia" and yet 8895167
Reposted by calebGrace @gracekind.net · 27/09/2026The earring is a little topaz tetrahedron dangling from a thin gold wire. When worn, it whispers in the wearer’s ear: “Yeah, absolutely.” 51684
Reposted by calebSamuel @samuel.fm · 26/09/2026the myth of token subsidies is basically dead, right? open source inference providers are so cheap it proves the true cost of a token is very low, and the high cost of claude tokens imo indicates limited supply rather than propping up demand. oai is able to cut costs because they have more computers 41096
Reposted by calebJeff Lazarus @jlazarus.bsky.social · 25/09/2026This is Nazi rhetoric. Not similar. Not Nazi-like. It’s straight up lebensraum. 362974913
Reposted by calebXe @xeiaso.net · 24/09/2026I've been working on publishing my half-complete "seconds" to a blog. Also taking the opportunity to play with standard.site lexicons. Enjoy!xeiaso.leaflet.pubAnthropic posted cringe on mainTL;DR: advocating for what amounts to a ban on open-weights models is advocating for a ban on open-weights models 2232
Reposted by calebWhitney Curry Wimbish @whitneycwimbish.bsky.social · 24/09/2026The youngest person his pager attack killed was 9-year-old Fatima Abdullah, who had just started 4th grade. The pager was on a table when it went off and she picked it up to take to her dad. It exploded in her face and reports said it covered the room in her blood. She wanted to learn English. 42103344393
Reposted by calebAmber 🌸 @shimmermathlabs.com · 24/09/2026a couple days ago: "i think opus 5.5's modular flow explainer assumes i'm smarter than i am" today: it made a 6 min manim video that assumes i know algebra and nothing else. dots → arrows → stretch → keyhole → geodesics (voice by me) watching this video is how i learned the math! 5724
Reposted by calebmr. TIM @timkellogg.me · 24/09/2026the NSA is spending billions on testing AI models. But the details are classified. We don’t know what for, which models, safety standards, nothing www.washingtonsun.com/technology/c...washingtonsun.comClassified Estimates Show the NSA Is Paying Billions to Test AI ModelsThe price tag is significantly higher than previously known. 2337
Reposted by calebCraig Hughes @craig.rungie.com · 24/09/2026TypeSafe updated their terms. The thing Tim flagged is fixed. Everything I flagged is still there. And they added some new stuff that's worse. 🧵 2268
caleb @caleb.jasik.xyz · 24/09/2026@bsky.app can we have a "spoilers" feature like discord's || luke, i am your father || spoilers? to be clear, not calling for the text spoiler bit, more that I want to be able to spoiler images w/o them being labeled "violence" or "adult" or whatever 010
Reposted by calebAaron Reichlin-Melnick @reichlinmelnick.bsky.social · 24/09/2026Incredible. First Secretary Mullin claimed that DHS had identified 15,903 noncitizens on Nevada’s voter roll. Then DHS officials told a skeptical Nevada, no no no, that’s just people we flagged. Here are the names of 185 people we KNOW aren’t citizens. All 185 were citizens. DHS got 100% wrong. 21188073853
Reposted by calebGeorge Pearkes @peark.es · 24/09/2026*NY ACCUSES POLYMARKET OF RUNNING ILLEGAL GAMBLING OPERATION *NY GOV. HOCHUL, AG JAMES ANNOUNCE POLYMARKET SUIT IN STATEMENT 411629275
Reposted by calebfunferall @funferall.bsky.social · 24/06/2026latentspacecraft.antikythera.orglatentspacecraft.antikythera.orgLatent Spacecraft: Brains, GANs, Finnegans.Latent Spacecraft combines computational linguistics, neuroscience, and literary analysis to investigate latent space, i.e. the hidden internal structure that enables both humans and machines to produ... 0112
Reposted by caleblauren @lauren.rotatingsandwiches.com · 23/09/2026look, you can't just punish trump when dems are in power. that might make republicans do something crazy, like stage a coup, illegally shred the federal bureaucracy, or even vstart abducting people off the street and extraditing them to torture sites all over the world 434438785
Reposted by calebWilliam B. Fuckley @opinionhaver.bsky.social · 23/09/2026I think ‘rote memorization’ is underrated. Having a ton of shit laying around in your head leads to connections and being able to more precisely describe things. 2963652
Reposted by calebaustin (e/accordion 🪗) @thebadcode.com · 23/09/2026so i’ve had a long-standing obsession with 27gal totes, which are the optimal system for storing all earthly possessions. not going to argue about this: the home depot variant is the best 4512
Reposted by calebAmy Jo (AJ) Cousins @ajcousins.bsky.social · 23/09/2026This is not deporting people. Literally. Legally. There is a deportation process and this isn't it. This is rendition. This is disappearing people. 172058806
Reposted by calebErin Lockwood @erinkaylockwood.bsky.social · 23/09/2026This is human trafficking on a massive scale. 332304782
Reposted by calebAstra ⎔ @astrra.space · 23/09/2026even before LLMs i used to often say that the way you treat insects and other invertebrates says quite a lot about you as a person i still stand by that, even more so today 415510
Reposted by calebSung Kim @sungkim.bsky.social · 23/09/2026Google launches djev - diffusion jev. 4715
Reposted by caleblauren @lauren.rotatingsandwiches.com · 22/09/2026lots of discussion on here this week about how brains work and what consciousness really is so i'll go ahead and speculate that this is probably fun as fuck for the cow 531343143
Reposted by calebrev. howard arson @theophite.bsky.social · 22/09/2026wait what they are using is evidence of fraud is the fact that they have been paying for a service which they do not use? 1051751286
Reposted by calebI am nameless yet others name me @7778777.online · 22/09/2026Wow I spent a whole day on this and I hate it. Some parts of it are not so bad though.blog.7778777.onlineSo you want to do something good?* 1102
Reposted by calebAriellah | אריאלה @sodone.bsky.social · 22/09/2026Tbf I shat bricks when I first saw electron microscope images demonstrating that organic compounds do in fact look like that 1123049
Reposted by calebMichael Gattozzi @mgattozzi.dev · 22/09/2026Genuinely baffled. Like wtf are we doing here? Where was all this money for other distros? What you gotta have a racist run it and then slap on AI for a bunch of dotfiles to get the cash? 4676
Reposted by calebhikikomorphism @hikikomorphism.bsky.social · 22/09/2026I would not go as far as endorsing or condemning this but I feel comfortable making one statement and that statement is: lmao 410625
Reposted by calebEris @isolyth.dev · 22/09/2026FIRST RELEASE OF THE DAY: CLAUDE OPUS 5.5!!! Supposed improvements on communication, it costs 20% less than Opus 5 at $4 per input, $20 for out, and $0.20 for cache read. It also uses fewer tokens and costs 40% less than Opus 5 on 'typical workloads'. It seems to beat F5.1 on most benchmarksanthropic.comIntroducing Claude Opus 5.5Claude Opus 5.5 leads in agentic coding and knowledge work, and costs 40% less to run than Opus 5 on typical workloads. 3667
Reposted by calebSung Kim @sungkim.bsky.social · 22/09/2026Claude Opus 5.5 www.anthropic.com/claude-opus-... 2859