The cutting-edge next-generation GPT model has hit critical development bottlenecks stemming from unresolved AI safety and alignment challenges, pushing OpenAI into a nearly identical predicament that rival tech giant Google faced back in 2022, when the latter’s large language model rollout was stalled by unaddressed safety risks to avoid harmful public deployment. This setback has sparked widespread industry discussions on balancing advanced generative AI innovation with robust responsible…
Anthropic said its Qwen4-Max model scored near the top of coding benchmarks, outperforming Anthropic's Claude Fable 5.1 as well as homegrown rival Moonshot's Kimi K3.1 Pro
Indian startup Bharat Semi has raised Rs 215 crore to develop next-generation AI-powered image sensors for global markets. The company aims to build chips that will help cameras, drones and other devices process and understand visual information without relying on expensive additional hardware.
(Corrects November 13 story in bullet two to say that Shi Bingzhe is a Chinese national and not Singaporean) SINGAPORE, Nov 13 (Reuters) - Singapore has charged a fifth person in a case involving the alleged false representation of the end users of AI servers purchased from Dell,
OpenAI’s Codex engineering team slashes model capacity errors after months of complaints using six targeted backend tweaks, including smarter WebSocket retries. - Read in AI News on WinCentral