US EditionStudy: Humans Surpassed AI Models in Planning and Prediction Tasks
Executives remain confident in AI's long-term value as companies report strong revenue and productivity gains; 92% expect AI to deliver ROI within two years, IBM survey shows.
- At the Fortune Global Forum in Riyadh, global executives dismissed market skepticism over AI spending, with Jeremy Kahn reporting they focused on AI's long-term impact despite Wall Street's mixed reactions.
- At an IBM-sponsored dinner at FGF, Ana Paula Assis said firms weren’t held back by fear of an AI bubble and instead focused on adoption challenges, despite collective heavy capital spending,
- In recent arXiv research, Dreamer 4 was shown to support real-time interactions on a single GPU, said Hafner, demonstrating advanced world model capabilities, without attributing unverified learning methods.
- After quarterly reports, Wall Street investors showed impatience as Microsoft saw mixed responses despite 40% cloud growth and Meta’s shares plunged 9%. Jerome Powell, U.S. Federal Reserve Chair, said AI’s boom drives tangible growth unlike the dot‑com bubble.
- An IBM survey of 3,500 senior executives in 10 countries found two-thirds reporting productivity gains and 92% confident of ROI within two years, while companies plan to ramp AI spending through 2026.
16 Articles
16 Articles
TechXplore
Science Tech UpdatesDeepMind introduces AI agent that learns to complete various tasks in a scalable world model
Over the past decade, deep learning has transformed how artificial intelligence (AI) agents perceive and act in digital environments, allowing them to master board games, control simulated robots and reliably tackle various other tasks. Yet most of these systems still depend on enormous amounts of direct experience—millions of trial-and-error interactions—to achieve even modest competence.
Why world models are the next big thing in AI
This is an excerpt of Sources by Alex Heath, a newsletter about AI and the tech industry, syndicated just for The Verge subscribers once a week. Around the middle of last year, Pim de Witte started reaching out to a handful of prominent AI labs to see if they'd be interested in using data from Medal, his popular video game clipping platform, to train their agents. Within weeks, it became clear that Medal's data was more valuable to the labs than…
Coverage Details
Bias Distribution
- 50% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
