The current AI is not even close to discovering *real* Physics breakthroughs. I've tried using it on some problems, and it doesn't get very far except on stuff that's already well trod.
LLMS
-

Opus AI Model Performance Issues and Solution Quality
By
–
Did you use Opus for this? The fix isn't good.
-
Meta Open-Sources TRIBE v2 Brain Response Foundation Model
By
–
🧠 Meta just open-sourced TRIBE v2 – a foundation model that simulates how the human brain responds to sight, sound & language.
— Futurepedia – Learn to Leverage AI (@futurepedia_io) 28 mars 2026
Months of lab work. Now done in seconds.
Here's what makes it different:
🔹 Predicts fMRI-style brain responses across 70,000 neural points
🔹 Often… pic.twitter.com/lul9VgwjhmMeta just open-sourced TRIBE v2 – a foundation model that simulates how the human brain responds to sight, sound & language. Months of lab work. Now done in seconds. Here's what makes it different: Predicts fMRI-style brain responses across 70,000 neural points Often
-
Claude Code Team Feedback Discussion on AI Improvements
By
–
Yeah the hate is coming pretty one-sided. Heck I even sat down with the claude code folks for an hour and told them all the things I feel are important to improve.
-
Anthropic’s Operon Agent for Biology Research
By
–
BREAKING 🚨: Anthropic is working on a new Operon agent for Claude Desktop, built for scientific research in biology!
— 🚨 AI News | TestingCatalog (@testingcatalog) 28 mars 2026
Operon will have a "private environment" to work alongside you. Users will be able to create different sessions within Operon projects, manage generated… pic.twitter.com/BE5aBYDZIBBREAKING : Anthropic is working on a new Operon agent for Claude Desktop, built for scientific research in biology! Operon will have a "private environment" to work alongside you. Users will be able to create different sessions within Operon projects, manage generated
-

AI Agent Discovers Novel Adversarial Attacks Against LLMs
By
–
"Claudini: Autoresearch Discovers SoTA Adversarial Attack Algorithms for LLMs" This paper shows that an AI coding agent can autonomously invent jailbreak and prompt-injection attacks, and can even beat 30+ human designed methods. So on top of human red-teamers, the new LLM
-
Democratizing AI: 50K Models and Local Inference Freedom
By
–
Next steps:
– enable the 50,000 models available in inference providers
– enable the 3,000,000 models available on HF
– local free fast inference with llama.cpp
– train and bring your own model! We don't want a world where you're forced to choose between two or three lookalike -

Elon Musk Endorses Top 26 Papers for Mastering LLMs and Transformers
By
–
BREAKING Elon Musk endorsed my Top 26 Essential Papers for Mastering LLMs and Transformers Implement those and you’ve captured ~90% of the alpha behind modern LLMs. Everything else is garnish. This list bridges the Transformer foundations with the reasoning, MoE, and
-
Google Turbo Quant LLM Running Locally on MacBook Air
By
–
Google Turbo Quant running Locally in Atomic Chat
— Chubby♨️ (@kimmonismus) 27 mars 2026
MacBook Air M4 16 GB
Model: QWEN3.5-9B
Context window: 100000
Summarising 50000 words in just seconds..
You can do 3x larger context window, processing 3x faster than before!
They are first that have integrated Google… https://t.co/3hrWncqJtgGoogle Turbo Quant running Locally in Atomic Chat MacBook Air M4 16 GB
Model: QWEN3.5-9B
Context window: 100000 Summarising 50000 words in just seconds..
You can do 3x larger context window, processing 3x faster than before! They are first that have integrated Google -

Major AI Tools Coming Soon from OpenAI Google Anthropic
By
–
Google is also working on a new internal ai tool called „agent smith“. It seems very popular within the company. So we are waiting for – OpenAI „spud“ (few weeks)
– Claude Mythos (Q3/2026)
– Googles Agent Smith (later this year)