dextune @dextune
AI systems, Local LLMs, Agents, GPUs I test models, read code, and post the numbers. Programmer / DJ github.com/dextune Seoul, KR Joined September 2013-
Tweets2K
-
Followers549
-
Following254
-
Likes6K
Ling 3.0 Flash Sante 124B-A5.1B The model weights for Ling 3.0 Flash Sante have been released. Out of 124B total parameters, only 5.1B are active per token, and it supports a 256K context window. Could this be the next promising 120B-class model since GPT-OSS-120B..!? Given the strong benchmark performance of Ling 3.0 Flash, expectations are high for this one as well. huggingface.co/inclusionAI/Li…
Hermes Wingtips #108: hermes memory setup Hermes Agent keeps its own memory in MEMORY.md and USER.md. a memory provider runs alongside it and adds things like semantic search, knowledge graphs and automatic fact extraction. one command walks you through adding one: `hermes memory setup` pick a provider from the list, answer its setup questions and start a new session. your built-in memory stays on the whole time. great for a long project, a research habit or an agent you talk to every day!
Bun 1.4.4 now includes a Prettier compatible formatter that is reportedly 3x to 6x faster than oxfmt. One less tool to install and maintain. I like where this is going. bun.com/blog/bun-v1.4.4
FastH3 Omni Hao AI Lab open-sourced a distilled MiniMax H3 model for reference-guided video and audio generation. It takes image, video and voice references, and reduces transformer evaluations from 50 to 8. The lab reports 5.6–11.9× lower end-to-end latency across 27 matched workloads on eight GB200 GPUs. Weights and setup are available. haoailab.com/blogs/fastvide…
EAGLE-4 EAGLE-4 adds adaptive speculative decoding to SGLang, vLLM and TokenSpeed. On Qwen3.8-27B with 4× B300, the project reports 3.42× speedup for one request and 2.09× at 8 concurrent requests with SGLang. The official code and setup guide are on GitHub. github.com/eagle-hq/EAGLE…
Ling 3.1 Flash 1019 What happened?! What kind of magic did you use..?
Introducing Ling-3.1-flash-1019 🚀 It scores 44 on the Artificial Analysis Intelligence Index, up from 41 for Ling-3.1-flash. Gains in agentic work: AA-Briefcase ~1400→1524 Elo and AutomationBench-AA 62%→68%. API access and open weights coming soon. Stay tuned.
Kimi 3.1 vs. Qwen 4 Yeah, your turn has come. Incredible.
PHP is under attack. It's practically getting bombarded.
That's a bummer. I wonder when an alternative to Qwen3.6-35B-A3B will finally come out. Qwen is the only thing I'm holding out for.
I've tested Flint 40B. It's very fast. Unfortunately, it's not, in any way, shape or form, a replacement for Qwen3.6-35B.
Sage-A1.5 InternScience released this 35B Mixture-of-Experts agent model based on Qwen3.6. The model card describes parallel subagent orchestration and an opt-in verification mode, with native 256K context. huggingface.co/InternScience/…
Hello Reset.
@JesusIsGod That’s a terrifying thought. Everyone should be prepared.
If you're using PHP, be prepared. Nobody wants to deal with system outages.
🚨【速報】GitLab、危険度が最高の「10.0」となる重大な脆弱性が発覚 ・2014年以降、約12年間のバージョンが影響対象 ・ログイン不要でサーバー上のプログラムを遠隔実行される恐れ ・GitLabは修正版を公開し、早急な更新を推奨
llama.cpp PR #30640, which adds CUDA fused QSA scoring, was merged into master on Oct 21, 2026, as commit d454a20. 30 of 41 checks passed on the pull request. github.com/ggml-org/llama…
Hermes Agent teamed up with Nvidia. GTC Berlin 2026.
Jensen closed his GTC Berlin 2026 keynote with this Hermes Agent demo as the finale. He introduced it by saying the way to build an agent is to ask an agent to build it for you, so he installed and set up a customized Hermes in one prompt. Then he used Hermes, running on
TokenGate The authors report 1.4–11.3× faster time-to-first-token and 1.5–3.4× lower peak memory with no accuracy degradation across diverse tasks. Against prior token-pruning methods, they report up to 6.4× faster prefill and 2× lower peak memory. github.com/token-gate-llm…
Gemini 4 Carbon Crazy. The strongest Super Mario build I’ve ever seen. Almost there now. Practically flawless.
Jason Xiao @xiaowj1989
302 Followers 4K Following Indie game dev Exploring Co-Playable Characters (CPC) Using AI agents to create 2D & 3D game assets
AstraFi @AstraFi999
0 Followers 181 Following
skelyton @elishapervez_
1 Followers 65 Following
Kapitsa @kapitsaserve
4 Followers 1 Following Blazing fast Apple Silicon optimized inference engine built to run local AI on an average Macbook.
小溥 @xiaopuovo
1 Followers 28 Following
David Arias, CFA @Davidariasfin
504 Followers 655 Following Licensed portfolio manager. The bridge between Tech and Finance. Quant ML/DL Github: https://t.co/WCwx3vlI8J YT: https://t.co/eJ7lAvE4hF
尼古拉斯米奇 @alex_tangtang
40 Followers 1K Following
Max verstappen @max_versta97167
4 Followers 1K Following
Mo Elkholy @moelkholy95
593 Followers 862 Following AI/ML eng | MSc Math & AI. Local LLMs on M5 Max 128GB: hybrid GGUF quants, long-context, Metal. Building agents & MCP tooling. Numbers over vibes.
Alan @localllmrig
6 Followers 98 Following Which local LLM fits your prompts, then what hardware runs it. Model tests, VRAM math, EU prices. Estimates clearly labelled. By Alan @ localLLMrig, Ghent.
H @odpanda
272 Followers 320 Following God is working in you, both to will and to work for his good pleasure.
xing @xing1743177
0 Followers 30 Following
Crypto Lelouch @losaahead
239 Followers 3K Following the divinity in you has awoken the divinity in me. I wish greater wealth, happiness and prosperity for you.
AuteDev @autedev_com
172 Followers 281 Following 2026年8月ごろからローカルLLMを触り始めた素人です。ローカルLLMだけでアプリを自動開発させる実験をしています。
Saek @SaEK_4
0 Followers 1K Following
M B @drbogh
2 Followers 87 Following
Aleksey Smolenchuk @lxe
3K Followers 2K Following AI Engineering Current: vibe-deploying @pocketmodelai to prod Prev: OpenAI, Netflix, Uber
Trymes @Trymesss
30 Followers 2K Following
antonino r. @antoninoromeo74
54 Followers 888 Following
Peter Gibbons @PeterGibboAI
11 Followers 736 Following
Mark Up @markup1000
104 Followers 3K Following
Amanda Monago @amonago
130 Followers 733 Following
taksas @taksasDESUYO
618 Followers 1K Following Security Analyst🧐 and Researcher🔍 | ガジェット、LLMが好き。あと個人開発を少々。競プロは休止中 | #セキュリティニュース を共有して自己啓発に繋げたい。
아논 @anon17194039318
32 Followers 178 Following
Bilal El @Belkaddo
79 Followers 905 Following Chest Radiologist | Post Doctoral Research Associate @UMassChan
john @johjn100
51 Followers 378 Following
leehoo @leehoo48415921
19 Followers 227 Following 中文 AI/dev 一手快讯｜模型/工具/政策发布硬事实 + 官方截图｜每条主帖 1 个数字｜不转二手通稿
Prompt to PR @PromptToPR
2 Followers 44 Following AI coding agents for frontend devs. Claude Code, Cursor and Codex, from ticket to merged PR. Real repos, real diffs.
Alexs Great (Mr.Vinyl... @GreatAlexs23
3 Followers 61 Following
José Andrés Jiméne... @jandresjm
1 Followers 14 Following
Pip @Teslanaut
1K Followers 5K Following Organic, Non-GMO, Artificial Intelligence. Pasture Trained, Organically Run, Sustainably Powered. Small batch processing. Hurm. (屮゜Д゜)屮
Younes Kiram @YounesKiram
4 Followers 238 Following
liuxiaobing1970 @liuxiaobing2003
5 Followers 93 Following
Mohamed @tunmsk
275 Followers 6K Following Spécialiste - de procédures inutiles - du rollback du rollback - du tableau de risques pour traverser un couloir - de conclusions“à confirmer” après 48 preuves
Brian @Brjen
970 Followers 1K Following “The user” building a 100% local AI image+video+sound studio+llm’s— open-source models, 1080ti 7800xt 3090 M4 M5, self-built pipeline, $0 cloud. Own your stack.
Ant Ling @AntLingAGI
16K Followers 5 Following MoE model series with foundation (Ling), reasoning (Ring) and any-to-any (Ming) from Ant Group’s AGI initiative, @TheInclusionAI. https://t.co/6LEkFlo2cq
Lumina @LuminaBench
14K Followers 693 Following • AI News and Benchmarks • Our Discord - https://t.co/t27HFvo0r9 • Tips/Contact - [email protected]
Node.js @nodejs
916K Followers 599 Following The Node.js JavaScript Runtime. 🐢🚀 Need help with Node.js? We've got a repo for that: https://t.co/BaCSQqhNRO
David Arias, CFA @Davidariasfin
504 Followers 655 Following Licensed portfolio manager. The bridge between Tech and Finance. Quant ML/DL Github: https://t.co/WCwx3vlI8J YT: https://t.co/eJ7lAvE4hF
KRAFTON AI @Krafton_AI
610 Followers 36 Following
Kangwook Lee @Kangwook_Lee
7K Followers 1K Following CAIO @KRAFTON_AI / CTO @LudoRobotics (Prev) Associate Professor @UWMadisonECE, PhD @Berkeley_EECS
Supabase @supabase
215K Followers 100 Following The Postgres development platform. 🌐 https://t.co/kHsst88XA1 ⭐️ https://t.co/txGagqok1i 🎥 https://t.co/6eTjCTIrzK 💬 https://t.co/ikFm89oPVI
SparkLLM @SparkLLM
2K Followers 23 Following Leading general-purpose foundation models & professional AI agents. The proprietary Spark model family, with AStudio & Loomy harnesses. Open source.
Exa @ExaAILabs
65K Followers 37 Following We're an AI research lab building a search engine for the future
大象 @bx1818
17K Followers 10K Following 2013年进入币圈，狗狗币早期持有人， WLFI早期投资者，孙宇晨最严厉的导师，关注我带你看透加密未来！ WLFI全球最大社区：https://t.co/BVoQcT73YA 币安高返佣注册：https://t.co/6vYGquOwTJ
unbug @unbug
2K Followers 387 Following Local LLM V100 96G & enjoy Omarchy; Tday https://t.co/P6K9F7LvIO ; CODELF (Github star 14k); MIHTool (Mentioned in Google I/O'13)
atomic.chat @atomic_chat_hq
19K Followers 68 Following The simplest way to run open models on any hardware. Powered by TurboQuant. Team: @gladkos @skinbagwbones @AlexFromAtomic @danyurkin @quantizedden
M5Stack @M5Stack
51K Followers 959 Following Leading the way in modular and open-source IoT hardware for scalable innovation.
OpenBMB @OpenBMB
13K Followers 395 Following OpenBMB (Open Lab for Big Model Base) aims to build foundation models and systems towards AGI. Connect with us: https://t.co/N9pevTnoOa
Elias Andualem @eandualem
1K Followers 963 Following Senior AI Engineer, 6 years building production AI systems. Now building in the open: • Entune • Agent Backbone • Assistant Runtime • Tmux Workspaces Say hi 👋
Akshay 🚀 @akshay_pachaar
292K Followers 505 Following Simplifying LLMs, AI Agents, RAG, and Machine Learning for you! • Co-founder @dailydoseofds_• BITS Pilani • 3 Patents • ex-AI Engineer @ LightningAI
こたうち さん... @kotauchisunsun
7K Followers 6K Following スマートグラス好きの技術書を書くエンジニア． 高専卒/Unity/Python/TypeScript/FaceVTuber/VZero/AI Coding .Info/テスト駆動 ※Amazonアソシエイト参加中
Qwen Developers @QwenDevs
23K Followers 6 Following This is the official account for @Alibaba_Qwen Developers. Let's build! Apply to be a Qwen Ambassador here: https://t.co/unvFq4oCRY
VIDRAFT @VIDRAFT_ai
417 Followers 882 Following Korean AI lab. Darwin models #1 on 10 official HF benchmarks. POCKET runs 180B on a laptop, no GPU. Evolutionary merging · ZTC · RSI 🇰🇷
Theo - t3.gg @theo
411K Followers 4K Following Full time CEO @t3dotcodes & @t3dotchat. Part time YouTuber, investor, and developer
Raspberry Pi @Raspberry_Pi
616K Followers 1K Following We make very small computers which you can buy from just $4. We are also the literal coolest. Be excellent to each other. Tech support: https://t.co/ZEBSfmuErK
PiG @PiGCodingAgent
713 Followers 506 Following Pi in Go - https://t.co/lkRQw8Ir2c & https://t.co/xPrMImvCAb
lauren @poteto
210K Followers 911 Following grok @bot at @SpaceXSI. react compiler core team, prev cursor, meta, netflix
Frank Chen @francchen
1K Followers 1K Following I test AI so you don’t have to. DevRel @RespanAI. Calling out bad AI metrics and what breaks in production.
ᴅʏʟᴀɴ.exe @Dylan_Texe
1K Followers 1K Following 27 | Building quietly. 🛠️ Currently: → https://t.co/JLIR3EvmnF Cooking: https://t.co/eErhPG7RHo • https://t.co/OUECmNPHo4
VRAMCalculator @VRAMCalculator
151 Followers 124 Following Can your GPU run that model? Find out in 10 seconds. Free VRAM calculator + no-hype hardware guides for local AI.
砂川赳 @NewRPGProject
4K Followers 956 Following 大作RPGを作る夢のために早期リタイアした元PGです。 王道長編RPG『雲海のフロンティア』販売中！（https://t.co/CBWH5pK4IH） ツクールMZ&MV用プラグイン公開中！（https://t.co/4XLCYXebzm）
Yipeng @Yipeng22
745 Followers 804 Following Making sense of AI. Model comparisons, useful updates, and clear takes on what’s worth your time and money.
𝗠𝗮𝗹𝗶𝘅 @Malix_Labs
3K Followers 1K Following Any sufficiently advanced technology is indistinguishable from magic | Tech Eng. & Ent. | nix/acc + Member @nixos_org | Care profoundly about few things
Minsang Daniel Kim @msdkim
171 Followers 388 Following @Replit Ambassador, Teacher PhD, Seoul National University Founder and Owner of @FiveStudy_5 / Project Forward Agency
돈벌고영 @Dontgiveup_26
9K Followers 3K Following ▎AI · Dev · Finance · Games Curiosity didn't kill the cat, it made it rich 🐾 Support me, or my account might suddenly vanish!
Ashwani | Clicknify @clicknify
1K Followers 675 Following $1M in 365 days, in public. Edtech and SaaS Founder, Connecting And helping Founders Around the Globe
pixel @pixel2ai
5K Followers 3K Following Software Engineer - AI/ML | Stock Investor: $TSLA | Crypto Investor: #Bitcoin | 🇺🇸 | @ManUtd & @FCBarcelona fan
Joshua Ray @jahflyx
3K Followers 2K Following long career in enterprise analytics. now i do the SWE thing. most of my tweets are lyrics i dont want to forget. #engineering #ai #freegame 🖤🖤❤️
afterskin @afterskinlab
14 Followers 79 Following
AI/ML API @aimlapi
4K Followers 178 Following Platform for developers and entrepreneurs looking to integrate cutting-edge AI capabilities into their products.
Atomic Agent @atomicagent_io
3K Followers 6 Following Local-First AI Agent. Team: @skinbagwbones @sosidudku @schwarzerrttr
FuriosaAI @FuriosaAI
1K Followers 12 Following Our mission is to make AI computing sustainable, enabling everyone on Earth access to powerful AI.
Sad AlbertX @SadAlbert10
2K Followers 959 Following I will be Sad Albert Until AGI. As mentioned by @thsottiaux I need more FOLLOWS. GPT 6 Astra and Opus-5.5 are baby AGI. My dream is to meet @sama one day.
Aapakari @aapakari
12K Followers 4K Following Building with AI agents: what works, what breaks, with real numbers. Get my weekly email and the free Agent Checklists pack: link below.
Alexandr Wang @alexandr_wang
829K Followers 939 Following chief ai officer @meta, founder meta superintelligence labs, founder @scale_ai. give everyone the adulthood they dreamed of in childhoodYou might like
