GXO and Agility Robotics debut humanoid Digit robots in a Spanx warehouse, a first for commercial robots-as-a-service (RaaS) deployment.
Category: robotics/AI – Page 311
Transcend us, AI friends!
👉 Researchers from Harvard University, UC Santa Barbara, and Princeton University show in a new study that generative AI models can outperform their human trainers through “transcendence”
New research shows that generative AI models can surpass their human trainers. The researchers call this phenomenon “transcendence” and demonstrate it using the example of chess.
AI models are typically trained to imitate human behavior. However, is it possible for these models to outperform their human “trainers” in certain areas? Researchers from Harvard University, UC Santa Barbara, and Princeton University show in a new study that this is possible through what they call “transcendence.”
The researchers trained a transformer on chess games played by players with limited skill levels. The resulting model, called “ChessFormer,” was able to play better than all the players in the training dataset in some cases.
A team led by NCI researchers has developed an artificial intelligence (AI) tool that uses data from individual cells inside tumors to predict whether a person’s cancer will respond to a specific drug. Learn more about how these findings hold promise for optimally matching cancer drugs to patients:
Precision oncology, in which doctors choose cancer treatment options based on the underlying molecular or genetic signature of individual tumors, has come a long way. The Food and Drug Administration has approved a growing number of tests that look for specific genetic changes that drive cancer growth to match patients to targeted treatments. The NCI-MATCH trial, supported by the National Cancer Institute, in which participants with advanced or rare cancer had their tumors sequenced in search of genetic changes that matched them to a treatment, has also suggested benefits for guiding treatment through genetic sequencing. But there remains a need to better predict treatment responses for people with cancer.
A promising approach is to analyze a tumor’s RNA in addition to its DNA. The idea is to not only better understand underlying genetic changes, but also learn how those changes impact gene activity as measured by RNA sequencing data. A recent study introduces an artificial intelligence (AI)-driven tool, dubbed PERCEPTION (PERsonalized single-Cell Expression-based Planning for Treatments In ONcology), developed by an NIH-led team to do just this.1 This proof-of-concept study, published in Nature Cancer, shows that it’s possible to fine-tune predictions of a patient’s treatment responses from bulk RNA data by zeroing in on what’s happening inside single cells.
One of the challenges in relying on bulk data from tumor samples is they typically include mixtures of like cells known as clones. Because different clones may respond differently to specific drugs, averaging what’s happening in cells across a particular patient’s tumor may not provide a clear picture of how that cancer will respond to a drug. Being able to capture gene activity patterns all the way down to the single-cell level might be a better way to target clones with specific alterations and therefore see better drug responses, but so far, single-cell gene expression data haven’t been widely available.
Influencer makes AI clone of herself. But it turns out badly.
Caryn Marjorie is a social media influencer whose content has more than a billion views per month on Snapchat. She posts regularly, featuring everyday moments, travel memories, and selfies. Many of her followers are men, attracted by her girl-next-door aesthetic.
In 2023, Marjorie released a “digital version” of herself. Fans could chat with CarynAI for US$1 per minute – and in the first week alone they spent US$70,000 doing just that.
Less than eight months later, Marjorie shut the project down. Marjorie had anticipated that CarynAI would interact with her fans in much the same way she would herself, but things did not go to plan.
In the creepiest video of 2024, a scientist explains how his system, Cognify, could ‘fix’ prisoners by implanting AI-generated artificial memories in their brains. And it somehow gets weirder.
The Japanese government is planning to connect major cities with automated zero-emissions logistics links that can quietly and efficiently shift millions of tons of cargo, while getting tens of thousands of trucks off the road.
From human intelligence collection to information gathered in the open, the CIA is leveraging generative artificial intelligence for a wide swath of its intelligence-gathering mission set today, and plans to continue to expand upon that into the future, according to the agency’s AI lead.
The CIA has been using AI for things like content triage and “things in the human language technology space — translation, transcription — all the types of processing that need to happen in order to help our analysts go through that data very quickly” as far back as 2012, when the agency hired its first data scientists, Lakshmi Raman, the CIA’s director of AI, said during an on-stage keynote interview at the Amazon Web Services Summit on Wednesday in Washington, D.C.
On top of that, AI — particularly generative AI in recent years — has been an important tool for the CIA’s mission to triage open-source intelligence collection, Raman said.
“If you look at the trajectory of improvement, GPT-3 was maybe toddler level intelligence, systems like GPT-4 are smart high schooler intelligence and in the next couple of years we’re looking at PhD level intelligence for specific tasks,” she said during a talk at Dartmouth.
Some took this to suggest we’d be waiting two years for GPT-5 but looking at other OpenAI revelations, such as a graph showing ‘GPT-Next’ this year and ‘future models’ going forward and CEO Sam Altman refusing to mention GPT-5 in recent interviews — I’m not convinced.
The release of GPT-4o was a game changer for OpenAI, creating something entirely new from scratch that was built to understand not just text and images but native voice and vision. While it hasn’t yet unleashed those capabilities, I think the power of GPT-4o has led to big changes.
NVIDIA’s CEO Jensen Huang believes that the AI frenzy will automate a whopping $50 trillion worth of companies, stating that Blackwell will play a dominant role.
NVIDIA Isn’t Taking The Foot of The AI Accelerator Pedal Any Time Soon, Plans To Take Blackwell’s Adoption To a Whole New Level
NVIDIA has undoubtedly managed to pick up a market that will progress rapidly in the future. Not only is every big tech firm, whether Microsoft or Amazon, forced into the race of “AI automation,” but the demand for adequate computing power is rising massively.