/4 Anthropic reveals Claude uses “desperate” vectors to influence decisions Anthropic looks inside Claude Sonnet 4.5 and finds something unexpected: the model stores patterns that act like emotions. These are not feelings, but measurable activation patterns inside the network.
GENERATIVE AI
-
Stanford researchers evaluate visual reasoning in multimodal models
By
–
/2 Stanford shows Claude and GPT generate visual reasoning without image input Stanford researchers release MIRAGE, a method to test how multimodal models use images. They take GPT-5.1, Gemini-3-Pro, Claude Opus 4.5, and Gemini-2.5-Pro and remove every image from six benchmarks.
-
OpenClaw v2026.4.5: Video Generation, Memory Dreams, Cheaper APIs
By
–
explaining what OpenClaw just shipped:
— klöss (@kloss_xyz) 6 avril 2026
(most people don’t understand it yet)
→ 103 contributors on v2026.4.5
→ your agent can now generate videos mid-conversation… generate music. create content assets on command without leaving the chat… through providers like Runway,… https://t.co/q0jh2kRwNx pic.twitter.com/0NYLEw5ayTexplaining what OpenClaw just shipped: (most people don’t understand it yet) → 103 contributors on v2026.4.5 → your agent can now generate videos mid-conversation… generate music. create content assets on command without leaving the chat… through providers like Runway, xAI, Google Lyria, and MiniMax → your agent now dreams… literally… it promotes short-term memory into long-term recall in the background… light sleep, deep sleep, REM… it starts remembering what you talked about without you needing to remind it → your API bill likely just got cheaper… prompt caching got a serious overhaul… system prompts, tool ordering, image history all normalized so follow-up messages hit cache way more often… you didn’t need to change anything. it should cost less for you now → security got tightened across the board… plugin permissions locked down… browser exploits blocked earlier. if something crashes it fails safe instead of failing open… if your agent touches the internet this matters more than everything above → Telegram, WhatsApp, Discord, Slack… 20+ channels… voice notes, reconnect loops, threaded replies, disappearing DMs… the stuff that was quietly broken is now quietly fixed → control UI and docs now speak 12 new languages… Chinese, Japanese, Korean, Spanish, French, German, Portuguese, and more… OpenClaw isn’t an English only project anymore the people sleeping on OpenClaw tooling right now will regret it later get this installed today OpenClaw🦞 (@openclaw) OpenClaw 2026.4.5 🦞 🎬 Built-in video + music generation 🧠 /dreaming is now real 🔀 Structured task progress ⚡ Better prompt-cache reuse 🌍 Control UI + Docs now speak 12 more languages Anthropic cut us off. GPT-5.4 got better. We moved on. github.com/openclaw/openclaw… — https://nitter.net/openclaw/status/2040998570317197607#m
→ View original post on X — @ceobillionaire, 2026-04-06 04:34 UTC
-
Anthropic SDK Tool Limitations and MCP Alternatives
By
–
It doesn't support tools in the same flexible way the normal Anthropic SDK does, or Codex App Server now does with its dynamic tools, for instance. Your only options are the hobbled builtin ones, or MCP.
-
AI Models Under Development Show Future Promise
By
–
Their current models aren’t great, but I’m sure they are cookin’!
-
OpenClaw Agents Now Generate Videos with Google Veo 3.1
By
–
My OpenClaw Agents can now natively generate videos.
— Shubham Saboo (@Saboo_Shubham_) 6 avril 2026
Just hooked them with the new Google Veo 3.1 Lite (fast and cheap).
It is so much fun when my AI Agents can send videos on the fly. https://t.co/HpJtZt85hC pic.twitter.com/ujwXwZD3brMy OpenClaw Agents can now natively generate videos. Just hooked them with the new Google Veo 3.1 Lite (fast and cheap). It is so much fun when my AI Agents can send videos on the fly. Peter Steinberger 🦞 (@steipete) The next version of @OpenClaw comes with native video generation. To start, I added support for the following companies: – Alibaba – BytePlus – fal – Google – MiniMax – OpenAI – Qwen – Together – xAI docs.openclaw.ai/tools/video… — https://nitter.net/steipete/status/2040928953653744003#m
→ View original post on X — @saboo_shubham_, 2026-04-06 04:03 UTC
-
GPT and LLM Conference: Big Data, AI, and Machine Learning Topics
By
–
GPT, Generative AI, and LLM! @AverConferences #BigData #Analytics #DataScience #AI #MachineLearning #NLProc #LLM #IoT #IIoT #Python #RStats #TensorFlow #JavaScript #CloudComputing #Serverless #DataScientist #Linux #Programming #Coding #100DaysofCode geni.us/Aver
→ View original post on X — @gp_pulipaka, 2026-04-06 03:57 UTC
-
Hermes Agent: Early Creators Needed for Video Content Now
By
–
creators. if you're not creating videos on hermes agent, you're missing out. and if you are creating but not sharing in the community where real users exist, you're not serious. to anyone thinking about where to start video content hermes agent is the hottest thing right now. even if it's just a reaction video or a first-use walkthrough, share it with me. i'll amplify it to my 21k+ followers and the 4,100+ hermes agent community. this is your window. the early creators on any tool become the go-to names. don't wait until everyone else is already making videos. start now. Sudo su (@sudoingX) if you're creating video content on local AI or agent frameworks, hermes agent is the #3 fastest growing repo on github right now with 24,000+ stars and the X community just crossed 4,000 builders. i'm actively looking for creators to collab with on hermes agent video content. tutorials, walkthroughs, benchmarks, reactions, whatever format fits your style. i'll push your content to the community and help brainstorm angles that hit. drop your channel or your video below and i'll amplify it to 21K followers and 4,000 hermes builders. if you already have hermes agent content, share it here. let's make open source great again. — https://nitter.net/sudoingX/status/2040735907523784992#m
→ View original post on X — @scobleizer, 2026-04-06 03:55 UTC
-
OpenClaw 2026.4.5 Release: GPT-5.4 Integration and Dreaming Feature
By
–
How you and your @openclaw feels running gpt5.4 with personality upgrade and /dreaming…. with no random outages. https://t.co/bSWz8hNy2I pic.twitter.com/DfTTRiqPEr
— Vincent Koc (@vincent_koc) 6 avril 2026How you and your @openclaw feels running gpt5.4 with personality upgrade and /dreaming…. with no random outages. OpenClaw🦞 (@openclaw) OpenClaw 2026.4.5 🦞 🎬 Built-in video + music generation 🧠 /dreaming is now real 🔀 Structured task progress ⚡ Better prompt-cache reuse 🌍 Control UI + Docs now speak 12 more languages Anthropic cut us off. GPT-5.4 got better. We moved on. github.com/openclaw/openclaw… — https://nitter.net/openclaw/status/2040998570317197607#m
-
Claudeopedia: AI-Powered Knowledge Base System Goes Viral
By
–
Allie just took @slashlast30days and mashed it up with @karpathy's latest to make something special. Her post is going viral.. ! Allie K. Miller (@alliekmiller) I'm a knowledge base MONSTER in Claude Code right now. Introducing: Claudeopedia. 1) I took @karpathy's 'llm-wiki' idea doc (90% of this, so the biggest credit goes to @karpathy) and 2) Combined it with the /last30days skill (HT @mvanhorn) and 3) Added a /wiki skill with screenshot and download arguments to transfer raw inputs faster and 4) Built an interactive visualization to search my knowledge base (with date ranges to compare knowledge over time!) 5) Set up a "question your assumptions" cron job that runs my recent writing/client emails against the wikis All happening in Obsidian for now. All of this was done this weekend, including testing. Will keep adding more features. For now, the main topic I'm building out is (surprise, surprise) enterprise AI. I'm drooling. I need a Claude-branded bib. — https://nitter.net/alliekmiller/status/2040884878229565816#m
→ View original post on X — @scobleizer, 2026-04-06 03:43 UTC