FROZON3

The Hermes Kanban now supports unlimited boards/projects, and you can now subscribe to updates from a project to your home channel across the gateway messengers you have setup! Just run `hermes update` and you're good to go!

6. Most of your limits were installed by someone else. You’re living under rules that aren’t yours. 7. Every argument in a relationship is just two wounded kids defending old fears inside grown-up bodies.

Hermes Agent now has multi-agent via the Kanban, new in v0.12.0. Agents claim tasks from a board, work in parallel, and hand off when blocked. You watch progress and unblock from one easy view instead of juggling terminals. We asked it to plan and make this video about itself:

the biggest problem with personal agents is that they hallucinate about working on things, when they don't. one of the first things I connected to Spark due to that is a mission control system that works both with canvas and kanban. the system also automatically searches for the best skills to pair tasks with and gets each task done via paired specialization. in the future, you will be able to pair tasks with specialized masteries (based on recursive self-improvement benchmarks) on any domain. both Canvas and Kanban views are already live in Spark, and whenever telegram bot says it is starting a task, it does report to you periodically too. this is the UI of the canvas btw: fully open source at: agent[.]sparkswarm[.]ai

@Mho_23 UGC

i should never be sharing this but f*ck it seedance 2.0 + claude code + tiktok is THE best combo for AI videos i cracked the formula for generating videos that look hyper realistic & make your audience feel like "holy shit this person gets me" i'm finally sharing my FULL system with you.. here's what you're getting: - my prompting method for realistic voices (works every time) - my realistic human movements & breathing claude skill (this is key for realistic videos) - my exact method on how to make infinite length videos that maintain consistency - how to get AI tools for dirt cheap (90% off) RT + reply 'UGC' and i'll send you the step-by-step system (must follow so i can dm)

Major upgrade to 𝕏 advertising

lol Cursor is a better harness for both GPT 5.5 in Codex AND Opus 4.7 in Claude Code how is that possible?!

We’re introducing the Cursor SDK so you can build agents with the same runtime, harness, and models that power Cursor. Run agents from CI/CD pipelines, create automations for end-to-end workflows, or embed agents directly inside your products.

> owner of roblox game "tha bronx 3" > millions a year > rolls royce in the driveway > people tell him roblox is just a kids game > he just opens roblox studio > describes what he wants to claude > claude writes every line of code > publishes the game > roblox pays for every minute players stay > $1.5 billion paid to creators in 2025 > top 10 averaged $38.5 million each > the barrier was always the code > claude just removed it > bookmark this and read the article below - it's worth it

Ex-board member of OpenAI calls Sam Altman a liar. He lied to the board for years, hid ChatGPT launch, lied about owning Startup Fund, falsified safety info, and lied to oust her after her paper. Board lost all trust → fired him. Sam Altman is a liar.

Today we are publicly releasing our newest creation — the Space Agent 🌌🤖 Please make sure to share 🔁, give us a ⭐ on GitHub, and subscribe on YouTube ▶️ to keep the momentum going 💫 Links and video below 👇

Happy to announce that Hermes Agent's repo just surpassed Anthropic's Claude Code repo

Gemini after Google invested $40,000,0000,000 on Claude

@rhnxai @minchoi It's already at the point where the AI photos are indistinguishable from real photos, how much better can gpt image 3 even get at this point lol

@minchoi and this is just image 2. imagine what image 3 does 💀

This is wild. ChatGPT Images 2.0 just completely changed the AI image game. These are not real 🤯 PROMPT: 4x4 grid of candid nostalgic photos shot with iphone of a young couple at vacation taking selfies of each other and together. Lots of camera shake, amateur framing, and emotional/vintage aesthetic.

@jxnlco i heard codeX

الصين ترمي قنبلة مفتوحة المصدر وتدمر أسعار سوق الذكاء الاصطناعي! إطلاق نموذج (DeepSeek-V4) رسمياً وبشكل مجاني للجميع. اللي صار اليوم مو مجرد تحديث هذا زلزال بيغير قواعد اللعبة في وادي السيليكون: 1️⃣ الأداء المرعب: نسخة (V4-Pro) تناطح وتتفوق على أقوى النماذج المغلقة والمدفوعة في العالم (مثل GPT-5.4 و Claude Opus 4.6) في البرمجة والرياضيات والمنطق 2️⃣ الاستيعاب الخارق: النموذج صار يدعم نافذة سياق عملاقة توصل لـ مليون توكن (1M Context) كمعيار أساسي وبكفاءة عالية جداً في استهلاك الذاكرة 3️⃣ السعر الصادم: واجهة برمجة التطبيقات (API) أرخص بـ 10 إلى 50 مرة من المنافسين! نسخة (Flash) الجديدة حرفياً تعتبر كنز للمطورين المستقلين اللي يبون سرعة وتكلفة شبه معدومة لتشغيل وكلاء الذكاء الاصطناعي (Agents) 4️⃣ مفتوح المصدر: أوزان النموذج متاحة للتحميل والتعديل للكل على HuggingFace الشركات الأمريكية تصرف مليارات عشان تحتكر النماذج وتلزمك باشتراكات وقيود استخدام.. وتجي الصين تنزل لك نموذج أقوى وأخف وتفتحه للمبرمجين ببلاش المستقبل للمنظومة المفتوحة

Wow, this is huge, after months of speculation and the U.S. running a massive pre-emptive discreditation campaign (x.com/RnaudBertrand/…), DeepSeek-V4 is finally out! I haven't studied it in depth but here are the most striking aspects as far as I can tell: - Fully open sourced with open weights (available for download on huggingface: huggingface.co/deepseek-ai) - Zero CUDA dependency anywhere in its stack, which is probably the biggest deal of all. For those who don't know, CUDA is Nvidia's software layer - the foundation nearly every frontier AI model in the world is built on. Except, as of today, DeepSeek V4, which can run entirely on Huawei Ascend chips via Huawei's CANN framework (finance.yahoo.com/sectors/techno…). Very concretely it means that China now not only has its own frontier AI models, but its own domestic AI stack, top to bottom. - The prices are insanely low. V4-Pro is roughly 3x cheaper than GPT-5.5 on input and 8.6x cheaper on output. And V4-Flash is an order of magnitude cheaper still, at $0.14/$0.28 per million tokens vs OpenAI's $5/$30 - so 30-100x cheaper than GPT-5.5 (!). And remember, these are the prices DeepSeek charges on its own API - anyone can download the weights and run them for "free" on their own server. - It is at or near the frontier on most benchmarks that matter. V4-Pro-Max matches or beats GPT-5.4 and Claude Opus 4.6 on competitive programming (Codeforces rating 3206), coding (LiveCodeBench 93.5), and math (HMMT 95.2, IMO AnswerBench 89.8). It trails the very newest GPT-5.5 and Opus 4.7 on a handful of the hardest agentic and knowledge benchmarks, but it's in the same league. In effect the value proposition is: "Same league as frontier US AI, at a fraction of the price, open-source and freely modifiable, and hardware-agnostic - you can run it on whatever infrastructure you choose." Which is insanely good. I now understand the need for a preemptive discreditation campaign: they had every reason to be worried. For the vast majority of use cases, you'd have to be a literal idiot to keep paying OpenAI or Anthropic's prices when this exists.

gpt 5.5 might be agi

Hermes Agent v0.11.0 - “The Interface Release” Full changelog below ↓