It was built by Levangie Labs, which is a whole new kind of AI (Claude is the engine, but isn't the car). The post you shared is human written, but the slide deck and other stuff was done by Grok, simulating a conversation, and sending that to Google's Notebook LM. Thanks!
LLMS
-

Microsoft Launches Copilot Advisors AI Debate Experiment
By
–

Microsoft released Copilot Advisors experiment in the US where multiple AI characters with live portraits will debate on a certain topic. Users will be able to select from various personalities and specify any topic they want. At the end, the debate will be summarized on a
-
Flex and Priority Inference Now Available for Tier 2 and Tier 3 Projects
By
–
This is available to Tier 2 and Tier 3 projects to start on most of our standard models! Hope this helps for folks trying to balance cost, reliability, and throughput! Read more: blog.google/innovation-and-a…
→ View original post on X — @officiallogank, 2026-04-02 20:03 UTC
-
Gemini API Service Tiers: Flex and Priority Options Launch
By
–
Today we are rolling out service tiers in the Gemini API! You can now (optionally) set "flex" or "priority". In the case of flex, this will save you ~50% on API costs (with lower reliability). In the case of priority, this will cost ~80% more but give you higher priority!
→ View original post on X — @officiallogank, 2026-04-02 20:03 UTC
-
X API Costs and AI Quality Comparison
By
–
X APi costs money and so do tokens. No other AI is as good
-
Gemma 4 Running Locally on Mac: Free, Safe, and Fast
By
–
this is Gemma 4 running locally on a 3 year old mac meaning:
— clem 🤗 (@ClementDelangue) 2 avril 2026
– free (=$0 no matter how much you use)
– safe (you're not leaking all your data via unsafe APIs)
– fast (as you can see) https://t.co/yX3gwJ3LF6this is Gemma 4 running locally on a 3 year old mac meaning: – free (=$0 no matter how much you use) – safe (you're not leaking all your data via unsafe APIs) – fast (as you can see) Georgi Gerganov (@ggerganov) Let me demonstrate the true power of llama.cpp: – Running on Mac Studio M2 Ultra (3 years old) – Gemma 4 26B A4B Q8_0 (full quality) – Built-in WebUI (ships with llama.cpp) – MCP support out of the box (web-search, HF, github, etc.) – Prompt speculative decoding The result: 300t/s (realtime video) — https://nitter.net/ggerganov/status/2039752638384709661#m
-
Claude and Zai.org adoption accelerates in AI landscape
By
–
Et encore, la plupart des gens qui utilisent #Claude n'ont pas encore essayé @Zai_org http://
Z.AI ça va trop vite! -

Three Key OpenClaw x Claude Code Breakthroughs
By
–
Day 2 of the OpenClaw x Claude Code project, and three game-changing features for autonomous AI agents have already been released. KAIROS enables 24/7 background operation, Governor slashes token costs by 85% for long tasks, and Docker-free Sandbox allows instant startup.
-
Hugging Face TRL and Async RL Training Landscape Guide
By
–
https://
github.com/huggingface/trl + https://
huggingface.co/blog/async-rl-
training-landscape
… https://
huggingface.co/blog/unsloth-j
obs
… You'll have to create an account on HF to share your experiments btw! -
Llama 31B Performance in llama-server: User Experience
By
–
Did you try the 31B in llama-server? Worked well for me