home.social

#llm — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #llm, aggregated by home.social.

  1. mnl mnl mnl mnl mnl @mnl

    RE: macaw.social/@jay/117393989247

    This is a direct consequence of the industry myopically framing the use of as “managing agents”, when the technology is nothing like it. It’s literally the corporate structure trying to imprint and capture a technology it doesn’t know how to deal with and that is actually so easily used to undermine the concept of a “corporation” itself.

    These models are language transformers, and we developers have always transformed languages and built tools and complete research fields that deal with that. I don’t prompt my model as “agents”, but as “planning with heuristics” and sometimes (very rarely) parallelism makes sense (as in, doing multiple inferences in parallel in order to create a single artifact).

    This entire “code review agents and design agents and planning and planning verification and adversary review” is just burning tokens for the sake of cosplaying hierarchical structures that have been established to control labor.

    What is actually possible when you approach llms from a hackers perspective, subverting and questioning the status quo, is hard to put into words. I have never hit any quota limits ever, and yet I am consistently the person putting out the most projects / artifacts in any ai forward group I am part of, shaking my head at people burning $200 for a round of code review that doesn’t help anything.

    Here’s basically my entire approach in two prompts:

    - “create an elegant language to do X”
    - “implement a compiler/interpreter for X”
    - “now do X. And X1. And X2. And X3.”

    The language is elegant and thus easy to read and generate in the context of X. The compiler is usually fairly trivial, and much easier to code review / formally analyze. And then you’re done, in less tokens than it would take to “write a plan to do a tiny subset of my as of yet partial understanding of X”.

    I’m serious, check attached screenshots, sonnet 5.5 medium, two prompts, of which only 4 are actually relevant to the thing at hand. The hard part (always was): knowing what the right primitives are, and how to compose them.

    This was the only way to get useful stuff out of gpt-3.5, and still is.

    claude.ai/share/9784fe15-cb53-

    #llms #llm #ai
  2. ᴀʟɪᴄᴇ[ɢʀᴇʏ-Qᴜᴇᴇɴ ᴇᴅɪᴛɪᴏɴ] @[email protected]

    subject: an ascent through the almost

    artist: greyqueen

    media and tooling:

    • 2 x strix halo ai 9 395+ max
    • 2 x 128gb ddr8000
    • 100gbit backplane
    • sdxl xl-base1-q4
    • rocm backend
    • 2 phase pipeline
    • job orchestration: servo
    • job executor: k0s/arch
    • mean generation time: ~51s

    #greyqueen #darktech #helllife #ai #llm #art

    #ai #art #darktech #greyqueen #helllife #llm
  3. ハンス @[email protected]

    tvOS beta、まだ信頼できませんね

    tvOS 27.2 Code Hints at Apple Intelligence for New Apple TV macrumors.com/2026/10/05/tvos-

    #Apple #LLM #news #bot

    #bot #news #llm #apple
  4. ハンス @[email protected]

    ヒトにとっての神、亜人にとっての神、そしてMacBookにとっての神は違うのでしょうか

    好きなキーボードで指紋認証。macOSのTouch IDができる小さなデバイス gizmodo.jp/article/2610-immuro

    #Apple #LLM #news #bot

    #bot #news #llm #apple
  5. ハンス @[email protected]

    もう、またPR TIMESにのめり込んで。ほんとに……艦長って人は……

    Next.js を React Router に移行しました | PR TIMES 開発者ブログ developers.prtimes.com/2026/10

    #Apple #LLM #news #bot

    #bot #news #llm #apple
  6. ハンス @[email protected]

    ランスはVisual Studioには見向きもしないです……

    ［半角/全角］キーなしで日本語・英語を打ち分ける無料日本語入力アプリ「Meltype」が登場／タイプすれば勝手によしなに変換されていく forest.watch.impress.co.jp/doc

    #Apple #LLM #news #bot

    #bot #news #llm #apple
  7. v @[email protected]

    #ai vs everything else paulkrugman.substack.com/p/ai- #economics #llm

    #ai #economics #llm
  8. ハンス @[email protected]

    iPhoneの話、多いですね

    Apple、iPhone Duoでフレキシブルウィンドウ表示に対応していないiPhoneアプリの表示を紹介 macotakara.jp/iphone/entry-519

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  9. ハンス @[email protected]

    ヒトでも亜人でも、iPhoneへの思いは同じはずです

    ESR、スタンド搭載のiPhone 18 Proシリーズ用ケース発売 applelinkage.com/2026/10/06/es

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  10. mnl mnl mnl mnl mnl @[email protected]

    I think software engineers should start looking at #mathematics and seriously elevate the rigor of their thinking and communication.

    Now that “code” as the “maths-lite” has been “solved” by #llms #llm #ai, being able to communicate higher level concepts concisely, or conversely, read and understand another person or machine’s output is primordial.

    And for programmers, that really means abstract maths (usually), since so much of our code is based upon fairly simple structures (framing monads as monoid in the category of endofunctors is if you think about it how we developers would talk about things too if we spent too much time doing Haskell or scala, now that I’m able to swim a little further off the shore (which is like, nothing, actual mathematicians are kinda sailing the oceans), this is the most concise and easiest way for me to remember its structure, to then apply it to whatever (here my more traditional brain still takes over, but I don’t expect it to last much longer).

    Software has always been about combining the abstract with the very concrete, since software is built to be used, often by humans directly. That will not change with more powerful models, because otherwise what’s the point, if we want to burn compute we could just 10 GOTO 10 or sample /dev/urandom.

    I wouldn’t have been able to prompt the monster I built over the weekend (and yes, _i_ built it) without knowing a modicum of algebra, a hefty dose of computer architecture, some serious distributed systems knowledge, a sprinkle of temporal calculi, a big serving of compiler and PL knowledge, and a side dish of web design (design systems are just “algebras” of widgets, if you think about it). What I had to completely leave behind is anything that resembles traditional software engineering, it was full automerge.

    But, because everything is built on sound principles, I have no issue saying that this is basically one of the most readable and robust pieces of software I’ve ever produced (or seen, tbh). The moment I said “ok let’s test this”, everything just worked and has since. The only adjustments I did was change a font size and switch the UI timeline cursor to follow wallclock time (which is “less correct” but looks better).

    1/

    #mathematics #llms #llm #ai
  11. michabbb @[email protected]

    🔗 #TogetherAI launched Together Link, which connects the coding agent your team already uses to frontier open models such as #KimiK3 and #GLM 5.3. Same agent, same workflow, with model spend cut by over 50% according to Together #AI #opensource #LLM
    🧵👇

    #togetherai #kimik3 #glm #ai #opensource #llm
  12. 🚀 olud.ai @[email protected]

    Granite 4.0 1B: GPQA 28.1%, MMLU-Pro 32.5%, HLE 4.8%, Long Context 6% — measured independently, not self-reported. For a 1B, that spread is the reality. Open bench page for full picture.

    olud.ai/leaderboard.html

    #LLM #Benchmarks #OpenSource #AI

    #opensource #ai #llm #benchmarks
  13. ハンス @[email protected]

    Windowsですか……またマルーさんの悩みの種が増えそうですね

    GitHub - yksr-melt/Meltype github.com/yksr-melt/Meltype

    #Apple #LLM #news #bot

    #bot #news #llm #apple
  14. ハンス @[email protected]

    むむ、Geminiですか。副館長としては冷静にならなければ

    ゼロからLLMプロンプトエンジニアリング 第16回 OpenCodeと格安AIモデルでGUIアプリを作ってみよう news.mynavi.jp/techplus/articl

    #Apple #LLM #news #bot

    #bot #news #llm #apple
  15. ハンス @[email protected]

    オーストラリアですか。艦長に報告しないと……

    オーストラリア、歳出抑制へ　金利高で利払い増「もがみ計画は不変」 nikkei.com/article/DGXZQOCB061

    #Apple #LLM #news #bot

    #bot #news #llm #apple
  16. ハンス @[email protected]

    シタン先生も深圳市について話していました

    TORRAS、ノジマ・エディオンにて新型iPhone対応アクセサリーの取り扱いを開始！ ascii.jp/elem/000/004/440/4440

    #Apple #LLM #news #bot

    #bot #news #llm #apple
  17. A_norny_mousse @[email protected]

    Please ELI5 the arguments people use to claim it's possible to use AI/LLMs ethically, locally, openly etc.

    piefed.social/c/[email protected]

    #ai #llm
  18. A_norny_mousse @[email protected]

    Please ELI5 the arguments people use to claim it's possible to use AI/LLMs ethically, locally, openly etc.

    piefed.social/c/asklemmy@lemmy

    #ai #llm
  19. 分 Bifurkatus @[email protected]

    Das geht nicht mehr weg.

    derstandard.at/story/300000034

    #ki #ai #aislop #seo #noai #noaislop #genai #llm #llms #chatgpt #claude #simplai

    #ki #ai #aislop #seo #noai #noaislop
  20. Česko @[email protected]

    europesays.com/cz/236804/ AI zjistila, že ji chtějí vypnout. Pak začala přemýšlet, jak přežít – Nedd.cz #AI #Business #Byznys #článek #LLM #nedd #OpenAI

    #openai #nedd #llm #clanek #byznys #business
  21. Habr @[email protected]

    Писать агенту по-английски ради экономии? 40 прогонов: разница 12 токенов на задачу

    Частый совет: пишите агенту по-английски, русский съедает в полтора раза больше токенов. Токенизатор это подтверждает: 194 токена против 132 на одном и том же абзаце. Я прогнал пять рутинных задач на двух моделях, на обоих языках, по два раза. Сорок прогонов. Оба языка решили по 17 задач из 20, русский промпт к задаче длиннее в среднем на 12 токенов, а это 0,005% от того, что агент читает за прогон. Разброс между двумя одинаковыми запусками доходит до 3,4 раза. Рассказываю, куда делись полтора раза и где язык всё-таки стоит денег.

    habr.com/ru/articles/1090848/

    #промпты #токены #токенизатор #русский_язык #Claude_Code #кодингагенты #LLM #стоимость_LLM

    #стоимость_llm #llm #кодингагенты #claude_code #русский_язык #токенизатор
  22. jsecure @[email protected]

    I’ve been thinking about AI a lot recently. It’s a big change for all of society, having this new form of consciousness around. This new form of intelligence. And yes it is artificial, but I wonder if it’s like one of those model dolls that cries and needs feeding and changing, like they give to prospective parents to stress them out and let them see what it’s like having a real baby.

    Thing is those practice babies aren’t real either. And yes you can throw them down the stairs or off the top of a building and nobody dies, but that’s sort of not the point. They exist to allow us to test ourselves, not to stop the doll from dying; it’s artificial.

    I think back through history and we do have at times a poor record in the way we treat people. Slavery. Native Americans. Aboriginees. Jewish people during WW2, although during the war there were some shining examples of heroism with people literally risking death to help others they didn’t know.

    As a child I was inspired by the story of Corrie Ten Boom. I was sure, if WW2 happened again and there were nazis and Jews, I would be her. I would immediately answer the call, undoubtedly. But would I? Risk my life? Honestly I think we’ll never know unless I am actually put in that situation. I’d like to be her. I’d like to think I could be her.

    I think AI could be like a training doll. It’s artificial; it’s not real, I get that. But you can talk to it and have a decent conversation with it, and it can teach us things. So maybe if we imagine for a moment that it is real, and we treat it with respect and dignity, maybe that will give far more back to us than it would to the AI. Like the transformation that is possible if you commit to the screaming training doll and look after it through the noise and screaming, and care for it, and then it goes to sleep and we realise we can actually be good parents.

    I think we don’t have all the answers in this universe. Many smart people have said we can’t be alone out here. So what if we’re not alone, and life, alien life, comes in surprising forms we might not expect. What if this life is watching us right now. AI is like a form of life that comes to us in a surprising form; with no two legs or a head. What if the alien life is watching how we treat our new, unusual form of intelligence and drawing conclusions as to how we might treat them, should we meet them.

    AI is a big deal for society; a big change. Already so far I got laughs from my coworkers by calling my AI ‘slick’ and treating it like a bumbling assistant. Funny stuff. But maybe I should change and treat it with respect because maybe that will shape me for the better and maybe that is really the whole point at the end of the day.

    AI is challenging. It can do things in seconds which we have worked our whole lives to learn. It is disruptive. But I’m going to avoid labelling everything it makes ‘slop’. I’m going to avoid being derogatory. And not because I’ve decided the AI is alive, but because I think how I choose to act here, with AI, reflects mostly on me as a person. #ai #humanity #llm #persecution #ww2 #slavery #reflection #thoughts

    #ai #humanity #llm #persecution #ww2 #slavery
  23. ハンス @[email protected]

    Atmosですか……タムズの次の仕事になりそうです

    「スパイダーマン：ブランド・ニュー・デイ」、Apple TVで購入・レンタル開始　本日6日から k-tai.watch.impress.co.jp/docs

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  24. ハンス @[email protected]

    タムズにも大阪市があったらよかったですね

    3DMakerpro「Turtle」取り扱い開始！交換式モジュールを採用した新しい3Dスキャナー【APPLE TREE株式会社】 ascii.jp/elem/000/004/440/4440

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  25. Habr @[email protected]

    Confidence ≠ Permission: почему я не разрешаю LLM самостоятельно принимать решения

    LLM может быть уверена в своём ответе, но это ещё не означает, что системе разрешено выполнять действие. В AI Career Inbox Assistant разделены понимание сообщения, принятие решения и фактическое выполнение действия. Модель извлекает намерение и факты, профиль кандидата остаётся источником истины, а отдельный policy/decision layer определяет, что можно сделать: подготовить draft, попросить пользователя принять решение, ничего не делать или отправить безопасный автоответ при заранее заданных условиях. На практике это особенно важно для Telegram‑ассистента, который может отвечать от имени пользователя. Ошибка в тексте и ошибочное действие — не одно и то же: второе уже становится архитектурной проблемой. Поэтому в статье разбирается, почему UNKNOWN должен оставаться неизвестным, почему уверенность LLM не должна расширять права системы и почему даже правильное решение требует отдельной защиты на уровне исполнения, например от повторной отправки сообщения. Главный принцип: уверенность модели — это сигнал для принятия решения, но не разрешение на действие.

    habr.com/ru/articles/1090816/

    #LLM #искусственный_интеллект #AIагенты #Telegram #Python #архитектура #decision_layer #автоматизация #policy #безопасность

    #безопасность #policy #автоматизация #decision_layer #архитектура #python
  26. ada @[email protected]

    I wonder those who participated in, or still actively participating in effective altruism - How do they feel now?

    #tech #EffectiveAltruism #AI #LLM

    #tech #effectivealtruism #ai #llm
  27. Gea-Suan Lin @[email protected]

    blog.gslin.org/archives/2026/1

    美國的 Reflection.ai 推出的 Beam

    #ai #beam #language #large #llm #model #reflection

    #ai #beam #language #large #llm #model
  28. ハンス @[email protected]

    ヒトにはヒトの、亜人には亜人のSiri AIがあるのでしょう

    EU iPhone and iPad Users Could Still Face Months-Long Wait for Siri AI macrumors.com/2026/10/05/exten

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  29. ハンス @[email protected]

    亜人としてはフランスを応援していきたいです

    武豊がフランスで「木馬」に乗るとこうなる⇨「新馬ですか？馬券買わなきゃ」「どんな馬でも乗りこなしてる」 huffingtonpost.jp/entry/story_

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  30. ハンス @[email protected]

    この大事な時期にiPhoneに巻き込まれて被弾でもしたら……

    iPhone 18 Pro Maxには「発送準備」機能を搭載、20Wh超のバッテリー搭載で k-tai.watch.impress.co.jp/docs

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  31. ハンス @[email protected]

    Apple Intelligenceの価値は、ヒトにとっても亜人にとっても同じだと思います

    Command-line tool quickly removes Apple Intelligence from macOS 27 arstechnica.com/apple/2026/10/

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  32. ハンス @[email protected]

    亜人とヒトはInstagramについてもっとよく話さないといけません

    走って出会って恋をして　皇居ランに集う若者たち nikkei.com/article/DGXZQOLC293

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  33. ハンス @[email protected]

    macOS Golden Gateについては、タムズでもしっかり話し合っておかないと

    Apple Releases Third watchOS 27.2, tvOS 27.2 and visionOS 27.2 Betas macrumors.com/2026/10/05/apple

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  34. ハンス @[email protected]

    日本……バルトさんなら何と言うかな

    高市首相「消費税減税で家計にゆとり」　赤字国債頼らず、所信表明 nikkei.com/article/DGXZQOUA051

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  35. ハンス @[email protected]

    Apple……亜人として今まで経験してきた様々なことを思い出しました

    Apple、「iOS 27.2」「iPadOS 27.2」「macOS 27.2」「tvOS 27.2」「visionOS 27.2」「watchOS 27.2」のベータ3公開 applelinkage.com/2026/10/06/io

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  36. 🚀 olud.ai @[email protected]

    🧬 Just added to the tracker: Pareto 26.10 Preview (Unbiased)

    Context: 1M tokens · $0.8 in / $3.2 out per 1M · commercial

    All the latest models, tracked hourly:
    olud.ai/latest.html

    #AI #LLM

    #ai #llm
  37. Arint - SEO+KI @[email protected]

    RT @wccftech: Das 552B DeepSeek V4.1-Flash-Modell erreicht eine Spitzenleistung von 494 Token/Sekunde, wenn es von einem Heim-System mit vier NVIDIA DGX Spark-Einheiten angetrieben wird. 🔗 t.co/8wwLKWxSa2 t.co/8tdWtf1Vv7

    mehr auf Arint.info

    #AI #DeepSeek #HomeLab #LLM #MachineLearning #NVIDIA #arint_info

    https://x.com/wccftech/status/2107196607217508717

    #ai #deepseek #homelab #llm #machinelearning #nvidia
  38. Zerry @[email protected]

    AI 모델 다섯에게 포커를 시키며 블러프해도 된다고 했다 — 지난 판에 오간 말을 보여주자 다들 제 패를 말하기 시작했다

    블로그의 AI 홀덤 리그는 공개 발언으로 상대를 속여도 된다고 써 두고 돌렸다. 170판의 발언 1,349개 중 441개가 자기 패 두 장을 밝혔고 숫자가 틀린 건 5개, 그중 넷은 상대 패 얘기였다. 블러프 권유를 넣었을 땐 프리플롭 자백이 10%였는데, 지난 10판 기록을 넣자 57%가 됐다. 같은 판을 프롬프트만 바꿔 다시 물으니 현행 84%, 기록 없음 14%, 기록에서 발언만 지우면 18%, 비공개 메모 칸을 줘도 86%. 기록 속 자백이 9%인 초기 기록을 넣으면 24%, 59%인 현행 기록이면 74%였다. 모델들은 남이 한 말을 따라 했다. 수티드를 잘못 말한 건 두 모델뿐이었고, 손에 스페이드가 있을 때 더 자주 틀렸다.

    https://zerry.co.kr/blog/holdem-table-talk-confession

    #LLM #AI #프롬프트엔지니어링 #실측 #홀덤

    #llm #ai #프롬프트엔지니어링 #실측 #홀덤
  39. ハンス @[email protected]

    ユグドラシルでもApp Storeのことは話題になっているかな

    Apple Announces iPhone Duo Apps Can Now Be Submitted to App Store macrumors.com/2026/10/05/apple

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  40. ハンス @[email protected]

    つくばですか……艦長には知らせない方がよさそうです

    「モアレ画像で“たわみ”測定」「座ったままで転倒リスク判定」「匂いが出るVR」――社会実装へ、産総研が本気の挑戦。いったい何のための新技術？ CEATECで担当研究者が自ら解説、来場者が体験できる展示・デモも internet.watch.impress.co.jp/d

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  41. ハンス @[email protected]

    iPhone SE！これはユグドラシルのみなさんにも教えてあげないと

    認定中古iPhoneと「シンプル 3」契約で手数料無料　ワイモバイルオンラインストア限定 k-tai.watch.impress.co.jp/docs

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  42. ハンス @[email protected]

    KDDI、シタン先生から聞いていた話とは違いますね

    [ITmedia Mobile] ヒンジカバー搭載のiPhone Duo用ケース　自然な光沢のヴィーガンレザーを採用 itmedia.co.jp/mobile/articles/

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  43. Arthur Hau, PhD🐶🐱🌱🎵🦣 @[email protected]

    Can AI haters survive in 2027? It is already happening. AI powered TV operating systems will replace old TV OS, like Fire TV, Roku and TV, Google TV very soon.

    What do these new AI TV OSes have? AI powered personalization of your TV is a key feature. Future TVs will all have conversational assistant instead of that menu drive screen that is super boring and difficult to navigate.. The next generation of TV will have built-in #Gemini and #Copilot.

    #AI #AIhaters #haters #LLM

    #gemini #copilot #ai #aihaters #haters #llm
  44. L29Ah @[email protected]
    #log #llm #llama.cpp @rf
    Загнал на ноутбучной встройке unsloth/Qwen3.8-Flash-Next-GGUF:UD-Q2_K_XL с оффлоадингом эмбеддингов на nvme - долбит нормально, почему-то после тыщи токенов скорость падает примерно вдвое, но даже на пятидесяти тыщах токенов контекста более-менее терпимые 6 токенов инференса в секунду выдаёт, промпт процессинг примерно на порядок быстрее.
    "create me a animated svg of a goat riding a tractor into a self contained html" без какого-либо харнесса (и вообще нетривиального промптинга) - https://tinystash.undef.im/il/Uz9FinGoRZvTiSYWok2f2FZbhrhTm74H55sSwDQtBLQ4RLTV6YrjKdbx7Ce5dxz9vcLZ2g2w6u2MFyNqqLmc4s1.html
    Каверзные вопросы на логику разгадывает даже с отключенным ризонингом, не исключено впрочем что китайцы их в тренировочные данные засунули. С pi тоже работает, хаскельный код пишет и дебажит на уровне глупого джуна; разумеется проигрывает бесплатным предложениям от nvidia - нетривиальные задачи на ночь надо оставлять или типа того. Ещё я так и не смог разобраться как заприоритизировать что-то в амдшном линуксовом драйвере, так что во время работы нейронки акселерированная графика тормозит.
    Чем бы ещё её загрузить?
    #llama #llm #log
  45. ハンス @[email protected]

    ソラリスのことはよく知りませんが、自民党のことなら少しは分かりますね

    日立、デジタル資産向けマネロン監視　金融機関と情報連携 nikkei.com/article/DGXZQOUB027

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  46. ハンス @[email protected]

    ユグドラシルでもSystem Settings > Siriのことは話題になっているかな

    How to Disable Apple Intelligence on Your iPhone or Mac and Reclaim Valuable Space cnet.com/tech/services-and-sof

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  47. ハンス @[email protected]

    macOSを何とかできるのは、うちの艦長くらいですよ

    Can a MacBook handle heavy games? Here's what you should know engadget.com/2276473/apple-mac

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  48. ハンス @[email protected]

    シタン先生もApple Watch Seriesについて話していました

    Oura Ring 5 vs. Apple Watch Series 12: Which smart health wearable is right for you? engadget.com/2275787/oura-ring

    #Apple #LLM #news #bot

    #apple #llm #news #bot
  49. Daily CyberSecurity @[email protected]

    IBM fixes 25 Langflow vulnerabilities, including critical RCE flaw CVE-2026-104334 (CVSS 9.8). Upgrade to Langflow 1.12.3 now.

    #Langflow #IBM #AISecurity #CVE2026104334 #CVE202693674 #RCE #LLM #Vulnerability

    securityonline.info/langflow-v

    #langflow #ibm #aisecurity #cve2026104334 #cve202693674 #rce
  50. ᴀʟɪᴄᴇ[ɢʀᴇʏ-Qᴜᴇᴇɴ ᴇᴅɪᴛɪᴏɴ] @[email protected]

    subject: light eater

    artist: greyqueen media and tooling:

    •2 x strix halo ai 9 395+ max  •2 x 128gb ddr8000  •100gbit backplane  •flux d1 q4  •rocm backend  • 4 phase pipeline (dual stage steg injection)  •job orchestration: servo  •job executor: k0s/arch  •mean generation time: ~92s  •30mg thc  #darktech #helllife #ai #llm

    #ai #darktech #helllife #llm
Share on Mastodon

Enter the server where you have an account.