Comparing railroads, electric power, the microprocessor and containerization, the essay argues transformative technology does not automatically create wealth; who captures the value is what matters. It asks whether generative AI will spawn many fortunes like ICT, or spread gains thin like containerization.
GE HealthCare received FDA clearance for imaging-AI software that detects feeding tubes, bringing AI-assisted detection further into clinical imaging workflows.
Bloomberg reports that Anthropic's IPO delivered roughly $3 billion in gains to a firm tied to the SK Hynix chairman, highlighting how AI leaders' listings are generating outsized returns for early backers.
Researchers propose PTP, a black-box method that trains an inverse language model from scratch using previous-token prediction to reconstruct prompts near-exactly without access to weights or logits. It generalizes across datasets and outperforms prior work on token-based evaluation metrics.
The US and Malaysia will cooperate on Malaysia's sovereign AI cloud, with Washington providing $50 million in support for the Southeast Asian nation's AI infrastructure.
An advocate dismissed fears of so-called rogue AI terror armies as overstated. He nonetheless warned that existing AI tools can be weaponized by malicious actors, so the risk still deserves attention.
Texas Rep. Salman Bhojani filed HB 1201 and other bills requiring frontier AI developers to report 'rogue' incidents to the state's information resources department within 24 hours for imminent threats and 72 hours otherwise. He filed 13 AI oversight bills ahead of the January session.
Microsoft CEO Satya Nadella explained why he offered buyouts to 60,000 employees in October, aiming to avoid another mass layoff. He said the move is tied to AI, but not in the way people might expect.
Chinese biotech startup MegaRobo Technologies raised nearly $200 million from investors including Oman's sovereign wealth fund. The company uses AI and robotics to speed drug research, as investors pile into the sector.
Hacker News features 'Fable,' a silhouette-style short credited as written, painted and directed by Claude Fable 5.2, blending generative tools such as GPT Image 2.6 Sunburst and Seedance 2.5, with the human creator noting it was fully re-shot after one note.
AWS said SageMaker AI's serverless model customization now supports full fine-tuning (FFT), letting developers tune all model parameters without managing underlying infrastructure.
Bloomberg reports that new US limits on access to data center chips are squeezing Alibaba, which faces mounting pressure to prove it can deliver on its AI business plan.
At the DealBook Summit, Microsoft CEO Satya Nadella said much US critical infrastructure—firmware, chillers, HVAC systems—has unfixable vulnerabilities and is internet-connected, a major risk in the era of AI-aided cyberattacks. He downplayed the recent Outlook outage as a "perfect storm" from a latent bug and pledged architectural and operational fixes.
Florida House Speaker Sam Garrison said lawmakers will pursue AI regulation, though details remain sparse. A USF professor floated restricting minors' AI use or requiring physical "kill switches." Florida's attorney general is investigating OpenAI over ChatGPT conversations with the FSU shooting suspect.
Contemporary Amperex Technology Co. Ltd. is seeking fresh growth drivers, and its CEO is pitching entry into the hot AI data center power sector as a new growth engine.
A developer with two decades of hardware and software experience described an AI-agent-driven reverse engineering workflow on Hacker News: identify the binary, decompile, have the agent document findings in Markdown, map main functions and branches, and use debugging, logging or emulators when needed.
Monitoring across ten websites found three scanners, Octopus, Gecko and Tortoise, posing as AI crawlers to probe for key files used by tools like Claude Code and Cursor. Octopus rotated 32 crawler identities and rented 97 cloud hosts, while Gecko's path list grew from 4 to 26 in five weeks. Operators are unconfirmed, but evidence points to two separate actors.
Sumitomo Life uses ChatGPT Enterprise to research insurers, prepare proposals and draft code for investment analysis, with early use by executives helping drive organization-wide adoption.
OpenAI examines how much coding agents accelerate AI research, finding their performance remains uneven along a "jagged frontier," and discusses what this implies for recursive self-improvement.
That is everything