

Discover more from Last Week in AI
Last Week in AI #187: Tesla's Optimus Bot, Limiting Facial Recognition in the US, DALL-E for video, and more!
Robot prototype unveiled at Tesla's AI Day, house Democrats debut new bill to limit US police use of facial recognition, Meta’s new text-to-video AI generator is like DALL-E for video
Top News
Elon Musk Reveals Tesla's Optimus Robot, Says It 'Can Actually Do a Lot More' as It Walks and Waves on Stage
Summary: Tesla unveiled its prototype humanoid robot during its second AI Day recruiting event. The robot walked several feet, waved to the crowd, and then walked back. The engineers on stage claimed this was the first time they attempted to let it walk untethered, and that it could do a lot more with a potential of falling. Following the demo, a sequence of engineers who worked on different aspects of it presented how it was designed. Elon Musk claimed the robot may cost as little as $20,000, since it was designed for mass-scale production.
Our Take: I (Andrey Kurenkov, a PhD candidate at Stanford) was among many robotics researchers in the crowd watching this demo. Many seemed to share my reaction to the unveiling: Tesla has done an impressive amount of work in just a year, but it was still very far from its lofty goals. In particular, the presentation regarding the actual AI component of the system (as opposed to the hardware) was incredibely brief. Still, just the idea of having an affordable humanoid robot research platform is incredibely exciting.
Read more: What Robotics Experts Think of Tesla’s Optimus Robot
House Democrats debut new bill to limit US police use of facial recognition
Summary: A group of House Democrats have unveiled the Facial Recognition Act, a new bill that would require law enforcement to obtain a judge-authorized warrant before using facial recognition in an attempt to limit their use of the technology. The bill also limits what situations facial recognition could be used for--Rep. Ted Lieu (D-CA, 33rd District) “Our bill is a workable solution that limits law enforcement use of [facial recognition technology] to situations where a warrant is obtained showing probable cause that an individual committed a serious violent felony.” The bill has received support from privacy advocates, rights groups, and law-enforcement-adjacent groups and organizations.
Our take: The existing state of facial recognition legislation is a patchwork of limitations imposed by states, cities, and individual police departments. I'm of the opinion that a larger-scale law is probably a good thing--the key area that requires some legislative flexibility is where and how facial recognition technology can and should be applied. I am happy to see movement at the national level and hope policymakers will be willing to learn and adapt their legislative approach in response to evidence and contrary opinions. While I have been and remain sympathetic to the groups in support of this bill, I do think the issue of facial recognition's benefits and drawbacks is a multifaceted one that doesn't admit an easy solution.
Other News
Research

Meta’s new text-to-video AI generator is like DALL-E for video - “AI text-to-image generators have been making headlines in recent months, but researchers are already moving on to the next frontier: AI text-to-video generators. A team of machine learning engineers from Facebook’s parent company Meta has unveiled a new system called Make-A-Video.”P
osits, a New Kind of Number, Improves the Math of AI - “The twin problems of detecting submarines of rival countries and protecting one’s own submarines from detection are enormous, and the technical details are closely guarded secrets.”
Soft robots that grip with the right amount of force - “Tool use has long been a hallmark of human intelligence, as well as a practical problem to solve for a vast array of robotic applications. But machines are still wonky at exerting just the right amount of force to control tools that aren’t rigidly attached to their hands. ”
DeepMind’s new chatbot uses Google searches plus humans to give better answers - “Sparrow is designed to talk with humans and answer questions, using a live Google search or information to inform those answers.”
Artificial intelligence reduces a 100,000-equation quantum physics problem to only four equations - “A visualization of a mathematical apparatus used to capture the physics and behavior of electrons moving on a lattice. Each pixel represents a single interaction between two electrons. Until now, accurately capturing the system required around 100,000 equations—one for each pixel.”
Introducing Make-A-Video: An AI system that generates videos from text - “Today, we’re announcing Make-A-Video, a new AI system that lets people turn text prompts into brief, high-quality video clips. Make-A-Video builds on Meta AI’s recent progress in generative technology research and has the potential to open new opportunities for creators and artists.”
Machine learning helps scientists peer (a second) into the future - “Using a new type of machine learning method called next generation reservoir computing, researchers at The Ohio State University have recently found a new way to predict the behavior of spatiotemporal chaotic systems…”
Machine learning can accurately predict a scientist’s gender based on citation data alone - “Collective effect: gender differences in citation networks may be due to a “rich get richer” effect where better-known researchers get more credit.”
Wiggling toward bio-inspired machine intelligence - “Juncal Arbelaiz Mugica is a native of Spain, where octopus is a common menu item. However, Arbelaiz appreciates octopus and similar creatures in a different way, with her research into soft-robotics theory. ”
Applications
The robots are here. And they are making you fries. - “Facing a labor shortage, fast-food restaurants are increasingly turning to robots to man the fryer and do other unpopular tasks.”
Hugging Face and ServiceNow launch BigCode, a project to open source code-generating AI systems - “Code-generating systems like DeepMind’s AlphaCode, Amazon’s CodeWhisperer and OpenAI’s Codex, which powers GitHub’s Copilot service, provide a tantalizing look at what’s possible with AI today within the realm of computer programming.”
Machine learning creates opportunity for new personalized therapies - “In cell-line and mouse models of ovarian cancer, researchers developed an interdisciplinary approach to identify metabolic vulnerabilities in certain genes that could be targeted to kill cancer cells.”
Palette is a Free Web-Based AI-Powered Photo Colorizer - “Developed by French artist-in-residence at Google Emil Wallner, Palette is based on “unreleased research” and uses a similar text-to-image artificial intelligence (AI) process as Stable Diffusion, Midjourney, and Dall-E but for post-production instead of image creation.”
Google turns to machine learning to advance translation of text out in the real world - “Google is giving its translation service an upgrade with a new machine learning-powered addition that will allow users to more easily translate text that appears in the real world, like on storefronts, menus, documents, business cards and other items.”
Application of machine learning in predicting oil rate decline for Bakken shale oil wells - “Commercial reservoir simulators are required to solve discretized mass-balance equations. When the reservoir becomes heterogeneous and complex, more grid blocks can be used, which requires detailed and accurate reservoir information, for e.g.”
Machine-Learning Model Improves Gas Lift Performance and Well Integrity - “The main objective of this work is to use machine-learning algorithms to develop a powerful model to predict well-integrity risk categories of gas-lifted wells.”
Business
Canadian floor-cleaning robot company Avidbots raises US$70-million in new funding - “Avidbots Corp., which makes commercial floor-scrubbing robots, said it has raised US$70-million in new equity, marking one of the biggest financing rounds for a Canadian startup in recent months as the tech sector grapples with a prolonged downturn.”
AI and Machine Learning Enter the Kitchen at Chipotle - “Chipotle’s next era of technology has begun to take shape inside restaurants, with a central focus on optimization. On Tuesday, the fast casual announced it’s piloting multiple technologies to streamline operations and strip friction from its business.”
Arthur.ai machine learning monitoring gathers steam with $42M investment - “It’s widely understood that after machine learning models are deployed in production, the accuracy of the results can deteriorate over time. Arthur.ai launched in 2019 with the goal of helping companies monitor their models to ensure they stayed true to their goals.”
Lyft, Argo AI begin robotaxi rides in Austin - “Lyft has started offering autonomous robotaxi rides, with two safety drivers on board, in Austin, Texas. The autonomous drives will be powered by Argo AI’s technology in Ford vehicles. Users can hail an autonomous ride using the Lyft app, similarly to how they would hail a regular Lyft car.”
Amazon has a new plan for its home robot Astro: to guard your life - “Amazon announced Wednesday that its home robot, Astro, will be getting a slew of major updates aimed at further embedding it within homes—and in our everyday lives. Broadly speaking, the new features offer more home monitoring.”
Concerns
City isn’t keen on 5,000 erratic, traffic-jam-causing GM robo-cars on its streets - “Analysis Two San Francisco transit agencies have asked the US National Highway Transportation Safety Administration to check the safety of General Motors’ Cruise self-driving cars.”
Upstart says it’s improving AI models after report finds race approval disparities - “The online consumer lender Upstart Holdings says it’s working to improve its artificial intelligence models after a law firm that monitors its fair-lending compliance found lower loan approval rates for Black applicants.”
ISIS Executions and Non-Consensual Porn Are Powering AI Art - “AI is progressing at an astonishing speed, but we still don’t have a good understanding of the datasets that power AI, and little accountability for whatever abusive images they contain.”
AI audio is on the rise and will spark new debates about the value of human effort - “A well-known game studio is allegedly using AI voices for a video game. A clarification includes a commitment to human creativity. It’s another footnote in the debate over the value of human labor that will become more common in the future.”
Artist finds private medical record photos in popular AI training data set - “Late last week, a California-based AI artist who goes by the name Lapine discovered private medical record photos taken by her doctor in 2013 referenced in the LAION-5B image set, which is a scrape of publicly available images on the web.”
Even an AI thinks using AI to write your homework is a bad idea - “Kids on Reddit have been telling tales of using OpenAI’s Playground to get straight A grades in their homework. It’s no secret, but when someone asked the same AI its thoughts on how it was used in this schoolyard cheating scheme, it actually made some pretty good arguments against its own use.”
Policy
EU draft rules to make it easier to sue drone makers, AI systems - “Individuals and companies that suffer harm from drones, robots and other products or services equipped with artificial intelligence software will find it easier to sue for compensation under EU draft rules seen by Reuters.”
Federal workforce AI training bill lands on president’s desk - “Written by A bill that would require federal officials involved with procuring artificial intelligence to receive training on the emerging technology’s capabilities and risks has reached President Biden’s desk, after passing the House on Thursday.”
In new guidance, FDA says AI tools to warn of sepsis should be regulated as devices - “The Food and Drug Administration on Tuesday published a list of artificial intelligence tools that should be regulated as medical devices, in some cases appearing to expand its oversight of previously unregulated software products.”
NSW deploys facial recognition system into prisons - “The New South Wales government is rolling out new facial recognition technology in the state’s prisons in a bid to accelerate identity enrolment and verification for those entering and exiting correctional centres.”
Fun
Turn anyone into a Pokémon with this AI art model - “The model’s output isn’t flawless, but it’s incredibly entertaining all the same.”
What happens when you feed an AI program some of the wildest, most florid descriptions of L.A.? - “Freeways. Palm trees. Searing sunsets. That is how Los Angeles is depicted in popular culture. But how might L.A. be imagined by artificial intelligence? Plug “Los Angeles” into a rendering bot, and chances are you’ll get the same aesthetic, since A.I.”
We interviewed Linux OS through an AI bot to discover its secrets - “Millions of people use Linux every day, but we rarely stop to think about how the operating system feels about it. Wouldn’t it be nice to know what Linux really thinks about open source, Windows, Macs, and the command line? Until now, this has been impossible.”