In general, I agree. But they are related but slightly different things in some contexts. Like when I say "I am running an open-weight model", I could be using OpenRouter or ollama cloud models.
TOOLS
-
Presentation of auto-search for arXiv articles via autoarxiv
By
–
Introducing autoresearch for arXiv papers
— alphaXiv (@askalphaxiv) 18 juin 2026
Change 'arxiv' to 'autoarxiv' in any paper URL
An agent deploys to resolve setup issues on the codebase, run a minimal reproduction, and estimate full replication cost. Read more below pic.twitter.com/2UHVqbT7EuPresentation of auto-search for arXiv articles. Replace 'arxiv' with 'autoarxiv' in any article URL. An agent deploys to resolve configuration problems in the codebase, run a minimal reproduction, and estimate the replication cost.
-
OpenKB solves RAG’s amnesia problem based on Karpathy’s concept
By
–
TRADITIONAL RAG HAS A MASSIVE AMNESIA PROBLEM. It rediscovers knowledge from scratch on every single query, and nothing ever accumulates. OpenKB is a new open-source alternative that finally fixes this. Based on a brilliant concept outlined by @Karpathy
, OpenKB treats -
Desk Lamp Becomes Live Claude Code Indicator with Color Changes
By
–
THIS DEV TURNED A DESK LAMP INTO A LIVE CLAUDE CODE INDICATOR 🤯
— Charly Wargnier (@DataChaz) 18 juin 2026
It changes colors in real-time based on exactly what Claude is doing:
🧠 Thinking = one color
💻 Writing code = color shifts
✅ Finished = changes again
❌ Error = instant alert
You know exactly what your AI is… pic.twitter.com/xWnADIYEyhTHIS DEV TURNED A DESK LAMP INTO A LIVE CLAUDE CODE INDICATOR It changes colors in real-time based on exactly what Claude is doing: Thinking = one color Writing code = color shifts Finished = changes again Error = instant alert You know exactly what your AI is
-
PhysX-Omni generates simulation-ready 3D objects from descriptions
By
–
What if creating 3D assets for simulation was as simple as describing them? NTU and ACE Robotics present PhysX-Omni: a unified system that generates simulation-ready rigid, deformable, and articulated objects. It uses a new efficient geometry representation for vision-language
-
PhoneWorld: Automating Mock Apps for Phone Agent Training
By
–
What if you could scale phone-use agent training without hand-building every environment? Tencent Hunyuan and collaborators present PhoneWorld: a pipeline that automatically turns real screenshots and user interactions into controllable mock apps, executable tasks, and
-
Dreamina Seedance 2.0 Mini: faster and cheaper AI video
By
–
The biggest challenge with AI video isn't creativity anymore. It's how much time and money it takes to generate content at scale.
— AI Highlight (@AIHighlight) 18 juin 2026
That's why Dreamina Seedance 2.0 Mini caught my attention.
It's about 2x faster than Seedance 2.0 Fast, around 30% cheaper, and the quality remains… pic.twitter.com/ItKjFC48h0The biggest challenge with AI video is no longer creativity. It's the time and money needed to generate content at scale. That's why Dreamina Seedance 2.0 Mini caught my attention. It is about 2x faster than Seedance 2.0 Fast, about 30%
-
AI video speed: creators turn images into cinematic sequences instantly
By
–
What stood out to me most is not the quality. It’s the speed. AI video systems now let creators: → Turn static images into moving cinematic sequences
→ Add or remove objects with prompts
→ Instantly change lighting, style, and environments
→ Generate video + audio together -
Grok Imagine API revolutionizes video creation via text prompts
By
–
The biggest disruption in filmmaking is not better cameras.
— Ronald van Loon (@Ronald_vanLoon) 18 juin 2026
It’s this:
With tools like the Grok Imagine API from @xai, you can now create and edit cinematic video scenes with simple text prompts.
No production crew.
No reshoots.
No weeks of iteration.
Just ideas → video in… pic.twitter.com/lpjWTtKuuHThe biggest disruption in filmmaking is not better cameras. It’s this: With tools like the Grok Imagine API from @xai
, you can now create and edit cinematic video scenes with simple text prompts. No production crew.
No reshoots.
No weeks of iteration. Just ideas → video in -
Gemma 4 26B demo: one orchestrator, 10 parallel agents, all local
By
–
🚨 One orchestrator. 10 parallel agents. 100+ tokens a second.
— Charly Wargnier (@DataChaz) 18 juin 2026
All local.
The @googlegemma team just dropped a MASSIVE demo for Gemma 4 26B.
They built a concurrent workflow that lets the 26B model coordinate an entire team of sub-agents on your machine.
Out of the box, the… pic.twitter.com/KoljxPbJheOne orchestrator. 10 parallel agents. 100+ tokens a second. All local. The @googlegemma team just dropped a MASSIVE demo for Gemma 4 26B. They built a concurrent workflow that lets the 26B model coordinate an entire team of sub-agents on your machine. Out of the box, the