A new KPMG and University of Texas study of 523 professionals finds that workers who actively challenge and guide AI consistently outperform those who blindly accept its output, raising urgent concerns about AI overreliance deskilling the modern workforce.
A new AI fluency chart by Rowan Cheung is redefining workplace performance standards, outlining clear expectations for six professional roles — from developers to CEOs — on how effectively they must integrate AI tools to rank from unacceptable to exceptional.
Rumors swirl that Anthropic is acquiring robotics startup Physical Intelligence, even as its CEO denies the reports, while past acquisition talks and OpenAI's existing investment stake in the company add fuel to speculation and could complicate any potential deal.
A new study testing 1,008 AI-generated SVGs across 48 animal-vehicle combinations finds no evidence that AI labs are secretly 'pelicanmaxxing' — gaming Simon Willison's famous pelican-on-a-bicycle SVG benchmark — with pelicans ranking 6th out of 8 animals in drawing quality, suggesting broad SVG optimization rather than targeted benchmark manipulation.
Viktor, an AI employee built for Slack and Microsoft Teams, slashes agent thread costs by 81.8% on Claude Opus 4.8 through a prompt caching architecture that drops a 40-step thread from $11.35 to just $2.07 by keeping prefixes byte-stable, treating threads as append-only logs, and timing cache compaction strategically around …
Despite AI handling 75% of coding tasks, Replit's head of AI Michele Catasta says engineers remain irreplaceable for verifying code, making architectural decisions, and solving complex problems, as AI reshapes—not eliminates—the engineering role.
Replit raises $400M at a $9B valuation — tripling in just six months — while launching Agent 4, a 10x faster AI coding tool, as the platform targets $1B in run-rate revenue by 2026 with 50 million users and Fortune 500 clients worldwide.
Cisco launches Antares, a family of open-weight AI security models that detect vulnerable code locally without sending sensitive data to the cloud, outperforming larger models at a fraction of the cost.
ACP v2 draft releases with major breaking changes, overhauling session states to allow agents to send updates freely, introducing uniform message patching, redesigned file diff structures, and more flexible permissions — but implementers are warned to hold off on production deployment until the protocol is finalized.
A new Rust-based tokenizer called Gigatoken is shattering speed records, achieving 24.53 GB/s and running over 1,000x faster than HuggingFace's tokenizers by leveraging SIMD instructions and aggressive caching, with the ability to tokenize all of Common Crawl in just 6.5 hours.