Daily issuepublished

AI learns to rearrange objects, like building LEGO

A surveey of 13,000 people in the workforce—from executive suite leaders to frontline employees

By 5 min read
AI learns to rearrange objects, like building LEGO

In today’s email:

  • 🧱 AI learns to rearrange objects, like building LEGO.

  • 🔥 A surveey of 13,000 people in the workforce about AI.

  • 📱 Tutorial: Creative QR Codes with ControlNet

  • 🛠 Various AI-related tools and platforms, including CopyCat, Tutor GPT, Duply, Receiptor AI, Adobe Generative Recolor, ReviewWriter, Octopulse AI, Composer, SuperDash, and more.

Story illustration

Sign Up | Advertise | Extension

Highlights💡 

LEGO-Net: Learning Regular Rearrangements of Objects in Rooms [Link],[GitHub]

Story illustration

Humans universally dislike the task of cleaning up a messy room. If machines were to help us with this task, they must understand human criteria for regular arrangements, such as several types of symmetry, co-linearity or co-circularity, spacing uniformity in linear or circular patterns, and further inter-object relationships that relate to style and functionality. Previous approaches for this task relied on human input to explicitly specify goal state, or synthesized scenes from scratch – but such methods do not address the rearrangement of existing messy scenes without providing a goal state. In this paper, we present LEGO-Net, a data-driven transformer-based iterative method for LEarning reGular rearrangement of Objects in messy rooms. LEGO-Net is partly inspired by diffusion models--it starts with an initial messy state and iteratively “denoises” the position and orientation of objects to a regular state while reducing distance traveled. Given randomly perturbed object positions and orientations in an existing dataset of professionally-arranged scenes, our method is trained to recover a regular re-arrangement. Results demonstrate that our method is able to reliably rearrange room scenes and outperform other methods. We additionally propose a metric for evaluating regularity in room arrangements using number-theoretic machinery.

Augment AI - Imagine ChatGPT, but it knows you and works across all your apps [Link]

  • Compose messages in your own tone & style, in the apps you already use (Slack, iMessage, GMail, GDocs, Notion... literally anywhere).

  • Automatically summarize Zoom/GMeet meetings.

  • Stop copy-pasting into ChatGPT. Just reference anything you’ve come across.

Augment AI is briefly re-opening its early access exclusively to Superpower readers. Sign up for Augment AI.

AI at Work: What People Are Saying [Link]

Story illustration

How do nearly 13,000 people in the workforce—from executive suite leaders to frontline employees—in 18 countries feel about AI?

  • Survey respondents are more optimistic than concerned about AI.

  • But there are vast differences in perceptions of AI between leaders and frontline employees, and companies need to address them.

  • Employees recognize the need for training and upskilling that this new era will require, but few have actually received it.

  • Almost everyone sees the need for responsible AI and regulation.

Tutorial: Creative QR Codes with ControlNet [Link]

Story illustration

Story illustration

Tools & Links 🛠️

Empower Your AI Journey: Key Resources, Software, and Innovations

Editor's Pick

CopyCat - Make your favorite celebrity say anything [Link]

Use Demucs to split your favorite song into tracks, then send one to MusicGen to get a completely new music piece! [Try here]

Story illustration

Tutor GPT - Dynamic few-shot metaprompting for theory-of-mind-powered tutoring. [Link]

Story illustration

Duply - Automate your Visual Creation [Link]

Editor's Pick

Receiptor AI - Extract receipts and invoices from emails with AI [Link]

Story illustration

Adobe Announces ‘Generative Recolor’ AI Feature For Adobe Illustrator [Link]

Story illustration

ReviewWriter - Respond to reviews with AI [Link]

Story illustration

Octopulse AI - A Platform for Growth teams to optimize Activation, Conversion and Retention. [Link]

Story illustrationEditor's Pick

Composer - Build trading algorithms with AI, backtest them, then execute—all in one platform. No coding skills required. [Link]

Story illustration

SuperDash - Develop and manage an AI-powered app using Ready-to-Use
Components

Story illustration

GPT-4-generated pitches are 3x more likely to secure funding than human ones [Link]

Story illustration

Paper: Large Language Models as Tax Attorneys: A Case Study in Legal Capabilities Emergence [Link]

Story illustration

Unclassified 🌀