Learn from this investor’s $100m mistake

In 2010, a Grammy-winning artist passed on investing $200K in an emerging real estate disruptor. That stake could be worth $100+ million today.

One year later, another real estate disruptor, Zillow, went public. This time, everyday investors had regrets, missing pre-IPO gains.

Now, a new real estate innovator, Pacaso – founded by a former Zillow exec – is disrupting a $1.3T market. And unlike the others, you can invest in Pacaso as a private company.

Pacaso’s co-ownership model has generated $1B+ in luxury home sales and service fees, earned $110M+ in gross profits to date, and received backing from the same VCs behind Uber, Venmo, and eBay. They even reserved the Nasdaq ticker PCSO.

Invest for $2.90/Share

_{Paid advertisement for Pacaso’s Regulation A offering. Read the offering circular at invest.pacaso.com. Reserving a ticker symbol is not a guarantee that the company will go public. Listing on the NASDAQ is subject to approvals.}

Today:

• OpenAI Releases Open-Weight Models

• DeepMind Launches Genie 3 Simulator

• Anthropic Launches Claude Opus 4.1

• ElevenLabs Debuts Commercial AI Music

• Google Launches Gemini Storybook Tool

OpenAI Just Broke The Industry

OpenAI released open-weight models, gpt-oss-120B and 20B, under Apache 2.0. They score near o3 and o4-mini on coding and reasoning, run on an 80GB GPU (120B) or 16GB devices (20B), and support tool use. 

Training used trial-and-error learning plus an automated checker. Open weights boost access but are hard to recall and can be modified. Benchmarks are strong but debated. The drop decentralizes capability widely ahead of a GPT-5 launch.

WATCH THE VIDEO ON YOUTUBE

DeepMind Launches Genie 3 Simulator

Google DeepMind’s Genie 3 is a “world model”—an AI that simulates environments—that turns text prompts into 720p worlds running at 24 frames per second. Scenes stay consistent and remember locations from a minute ago. You can steer weather and objects with “what-if” prompts. It aims to train embodied agents, like robots, through longer tasks. Limits remain: short sessions, few direct actions, weak behavior with independent characters, and imperfect real-place accuracy.

Why this matters

• Trains AI by “learning through doing” in safe, cheap simulations—useful for robots and autonomy.
  

• Real-time, stable worlds enable longer tasks and better testing, pushing toward more general-purpose AI.
  

• Unlocks creative and training uses (education, games, film) while surfacing new safety and control questions.

Anthropic Launches Claude Opus 4.1

Anthropic released Claude Opus 4.1, an upgrade to its top model, improving coding, reasoning, and “agentic” tasks (multi-step, tool-using work). It scores 74.5% on SWE-bench Verified, a standard test of fixing real GitHub bugs. Early users report better multi-file refactors and pinpoint bug fixes. The model is available now to paid users and via API at the same price, and cloud partners. Bigger updates are promised in the coming weeks.

Why it matters

• Faster, safer software: higher bug-fix accuracy speeds maintenance and reduces failures.
  

• More capable agents: better multi-step planning and tool use moves AI toward reliable assistants.
  

• Easier adoption: same pricing and broad availability (app, API, clouds) lowers friction while signaling rapid near-term progress.

ElevenLabs Debuts Commercial AI Music

ElevenLabs launched an AI music generator it says is cleared for commercial use—meaning businesses can legally use the output. The company, known for text-to-speech, expands into songs. It announced licenses with Merlin and Kobalt to train on music from artists who opt in, with revenue sharing and safeguards. The move follows lawsuits against Suno and Udio over training on copyrighted works. Samples show style imitation that raises serious ethical questions.

Why it matters

• Creates a legal path for training data: artists opt in, get paid, and misuse is limited.
  

• Pressures clearer rules on copying style and voice, impacting artists and startups.
  

• Pushes AI from speech to full songs, enabling new tools for creators and businesses.

🧠RESEARCH

Qwen-Image Technical Report

Qwen-Image is a new AI model for generating and editing images. It uses step-by-step training, starting simple and moving to complex tasks, to improve text accuracy in pictures. A dual-encoder design helps keep meaning and image quality. It works well in English and Chinese, outperforming other models in tests.

SitEmb-v1.5: Improved Context-Aware Dense Retrieval for Semantic Association and Long Story Comprehension

SitEmb-v1.5 is an AI model that improves finding relevant text in long documents. It represents short text sections while considering their larger context, boosting accuracy without huge models. Trained with a new method, it beats bigger models in retrieving story details, works in multiple languages, and aids various applications.

CellForge: Agentic Design of Virtual Cell Models

CellForge is an AI system that builds virtual cell models from raw biological data. Using multiple specialized AI agents, it analyzes tasks, designs modeling strategies, and generates training code. Tested on six datasets, it predicts cell responses more accurately than current methods, aiding research on drugs, gene edits, and treatments.

🛠️TOP TOOLS

PicFinder AI - AI-powered image generation tool that transforms textual descriptions into unique visual content.

Unboring AI - AI-powered online platform that offers innovative tools for face swapping, photo animation, and video restyling, allowing users to create engaging and entertaining content with just a few clicks.

Crayo - Transform simple text inputs into viral-ready videos.

Dubverse AI - AI-powered platform that revolutionizes video dubbing and subtitling, offering creators and businesses the ability to make their content multilingual instantly across 30+ languages with human-like AI voices.

Bing Create - AI-powered tool that transforms text descriptions into visual art, offering users the ability to generate unique images from their ideas. 

📲SOCIAL MEDIA

🗞️MORE NEWS

• DeepMind’s Demis Hassabis predicts human-level AI within 5–10 years, calling it vastly faster and bigger than the Industrial Revolution. He envisions medical breakthroughs and abundance, urges fair distribution, and warns of major societal challenges ahead.

• Google’s new Gemini “Storybook” tool makes 10-page illustrated tales from user prompts, reads them aloud, and supports custom art styles. While creative, it sometimes produces odd mistakes, like misplaced objects or strange anatomy in images.

• Google and Kaggle’s new “Game Arena” pits top AI models against each other in strategy games like chess, aiming to replace outdated benchmarks with open, dynamic tests of planning, adaptation, and real problem-solving skills.

• A researcher found nearly 100,000 public ChatGPT conversations indexed by Google, revealing sensitive data like contract details, NDAs, and personal relationship discussions—highlighting privacy risks when users share chats without realizing their visibility.