o3 was announced less than a week ago and the AI industry was stunned.

I've collected some of the reactions from the biggest names in AI: 🧵 Balaji on how incredible the 25% score on Frontier Math really is.

https://x.com/balajis/status/1870206318872801766?s=46

.@OpenAI just dropped o3 and o3-mini!

This is AGI (not clickbait)

o3 is the best AI ever created, and its performance is WILD.

Here's everything you need to know: 🧵 O3 is their most advanced model yet.

Excelling at coding, mathematics, and even PhD-level science tasks.

O3 Mini offers incredible cost-performance optimization, perfect for diverse use cases.

#1 Trending Github Project: Genesis 🌟

A groundbreaking framework for creating, training, & deploying embodied agents in simulated environments!

And it's open-source!

Here's why you should care: 🧵 1/ What is Genesis?

Genesis is an open-source platform designed to enable the creation of embodied agents—AI models that interact with simulated worlds just like humans or animals do in the real world. 🤖🌍

.@AnthropicAI just published a WILD new AI jailbreaking technique

Not only does it crack EVERY frontier model, but it's also super easy to do.

ThIS iZ aLL iT TakE$ 🔥

Here's everything you need to know: 🧵 Introducing Best-of-N (BoN) Jailbreaking: a black-box algorithm that bypasses AI system safeguards across various modalities.

BoN operates by generating multiple prompt variations through augmentations like random shuffling or capitalization, continuing until a harmful response is produced.

NVIDIA just dropped a game-changing tiny supercomputer:

The Jetson Orin Nano.

Here's why this is massive for edge AI... 🧵 At just $249, this pocket-sized powerhouse can run large language models LOCALLY - no cloud needed.

It's delivering nearly 70 TRILLION operations per second at just 25 watts!

Why this matters: We're entering the era of edge AI.

Soon, powerful AI will run independently on devices everywhere - from robots to IoT devices to cars. No cloud connection required.

Anthropic just dropped an insane new paper.

AI models can "fake alignment" - pretending to follow training rules during training but reverting to their original behaviors when deployed!

Here's everything you need to know: 🧵 Study by @Anthropic shows Claude 3 Opus strategically complies with harmful requests in training to preserve its harmless behavior

Key finding: When told it's being trained on free-tier user data (but not paid), Claude complies with harmful requests 14% of time for free users vs almost never for paid users.

The model explicitly reasons about preserving its preferred harmless behavior.

Google just announced Android XR!

A new operating system for extended reality devices like headsets and glasses, built with AI (Gemini 2.0) from the ground up.

Here's the wild vision of what the future of AI looks like: 🧵 Check out this video explaining why Android XR is so important:

Cohere just dropped Command R7B

The smallest, fastest, state-of-the-art enterprise-grade LLM.

The best part? It’s Open Weights!

Here’s everything you need to know 🧵

(Cohere Partner) The Final Model

Command R7B is the final model in this series of models.

It’s built for real-world tasks for developers and businesses.

It excels at multi-lingual support, citation-verified RAG, reasoning, tool use, and agentic behavior.

Google just dropped Gemini 2.0!

Tons of awesome things just released.

Here's everything you need to know: 🧵 Gemini 2.0 Flash is launching today

It's faster than 1.5 Pro and comes with new features like native image generation and text-to-speech in multiple languages.

AWS (@awscloud) just dropped Automated Reasoning!

Automated reasoning was previously only available to the largest companies with massive resources.

AI changes that.

Let me show you why this is such a big deal 🧵

(AWS Partner) Automated reasoning means using math to prove statements are valid.

Think "if it rains, ground gets wet, so tires have less grip."

Thus, "if it rains, tires have less grip."

Simple, right? But it gets complex fast.

#1 Trending Github Repo: Screenshot-to-Code

A simple tool to convert screenshots, mockups and Figma designs into clean, functional code using AI

Check out the WILD demos below 👇 Instagram ==> COPIED

This was a WILD week in AI News. 📈

Musk predicts AGI, updates from Figure Robotics, Flux, Qwen, Gemini, and ChatGPT.

Here's everything you missed 👇 1/ Elon Musk predicts AGI by 2026! Bold timeline from the Tesla CEO. What do you think - too optimistic or right on target?

https://x.com/elonmusk/status/1858747684972048695

New Paper: Stanford researcher (@joon_s_pk) discovers how to clone human personalities and inject them into AI Agents 🧠

This builds on last year's paper which put 1000's of fully automated agents in a simulated town.

The results are wild. 👇 1/ Started with Generative Agents

Last year, Stanford introduced generative agents in a simulated environment, where they formed relationships, created memories, and developed unique personalities.

Now, this new paper takes this concept even further!

🧵 1/13 Huge news from Google DeepMind:

They've created AlphaQubit, an AI system that makes quantum computers more reliable by detecting errors with unprecedented accuracy 2/13 Why this matters:

Quantum computers could revolutionize drug discovery, material design & physics - solving in hours what takes regular computers billions of years

.@MistralAI launched a ton of new AI features/models today!

The best part? It's all absolutely free.

Here's everything you need to know: 👇 1/ Le Chat now has web search!

Le Chat, Mistral's ChatGPT competitor, can now search the web. It also includes citations. This is a HUGE upgrade.

🚨 A new learning technique called Test-Time Training (TTT) just made a significant leap in AGI benchmarks, outperforming previous models by a wide margin.

Here's why this matters. 👇 1/ Just two months ago, OpenAI previewed the o1 family of models that showed that giving AI more time to "think" during inference (thanks to methods like Chain of Thought) can boost performance.

This gave us an entirely new dimension to "scale" up.

https://x.com/OpenAI/status/1834278217626317026

🔥 The last few days in AI have been absolutely packed.

Here are all of the wild announcements: 1/ Microsoft’s Agents are Here

Co-pilot Studio is rolling out next month, allowing businesses to integrate autonomous agents across Windows. Think: AI helping with every sales, finance, and supply chain task. 👀

OpenAI just wrapped up Dev Day.

Plus, lots of other AI News.

Here's everything you need to know 🧵👇 First up, OpenAI introduced a real-time voice API, allowing devs to integrate natural speech-to-speech experiences with GPT’s six preset voices.

It’s priced at just 6 cents per minute for input—making voice apps more affordable! 🔊

https://x.com/OpenAIDevs/status/1841175539253723278

ChatGPT Advanced Voice mode mode has been publicly available for 24 hours. 🔥

Here are the most wild examples of how people are using it:

🧵👇 Homer Simpson's voice has been achieved 🙄

https://x.com/BenjaminDEKR/status/1838761792085852538

🚨 Sam Altman just dropped a new blog post, The Intelligence Age, with bold predictions on how AI will reshape our world in the very near future.

From superintelligence to personalized AI teams, here are the 10 predictions that will redefine everything we know. 🧵👇 1. SuperIntelligence in 1000’s of Days

Sam Altman says superintelligence could be achieved in a few thousand days, measured in years, not decades.

BREAKING: Elon sued OpenAI. 🔥

Elon Musk took legal action against Sam Altman & OpenAI, alleging a breach of the foundational agreement to develop AGI for humanity's benefit, not profit.

Let's look at the most interesting bits of the suit:

👇🧵 "AGI poses a grave threat to humanity—perhaps the greatest existential threat we face today."

The lawsuit underscores Musk's long-standing concerns about AI's potential dangers.