Airbnb's Brian Chesky to Establish New AI Company
Airbnb CEO Brian Chesky is planning to establish a new AI company, according to reports. TechCrunch reports that Chesky is considering launching an "Airbnb AI Lab".
The move is rooted in Chesky's assessment of existing LLM products. Last year, he held off on partnerships with major LLM companies, arguing that "current LLM products are not yet sufficiently mature." Now, having judged that AI technology is reaching a practical level, he has decided to move forward with establishing his own AI lab.
It is noteworthy that the head of a company with a massive platform like Airbnb is launching its own AI company rather than simply forming partnerships. The development of AI agents specialized in the travel and accommodation domain is anticipated, and it could reshape the industry's structure.
Trump Administration Backing Accelerates the Rollout of "AI Doctors"
According to the Washington Post, efforts to introduce AI doctors into US healthcare settings are gaining full momentum with the support of the Trump administration.
Specifically, programs are being promoted to bring AI-powered diagnostic support, treatment plan recommendations, and patient monitoring into medical institutions. The aim is to deliver high-quality healthcare even in areas facing physician shortages by deploying AI on the front lines of medicine.
However, many issues remain unresolved, including where responsibility lies for AI medical decisions, how to handle misdiagnosis legally, and how to protect patient privacy. Healthcare professionals and ethicists have voiced caution, and the balance between technological progress and institutional readiness is now being called into question.
AI Water Consumption to Reach 1.3 Billion People's Worth by 2030 — A Warning on Environmental Impact
According to an analysis reported by the Spanish newspaper El País, AI water consumption is projected to reach a level equivalent to that of 1.3 billion people by 2030.
Training and running inference on AI models require vast amounts of cooling water, and water consumption is surging as data center demand expands. Training large language models in particular consumes enormous computational resources, with massive amounts of water consumed behind the scenes.
This issue is an aspect that has received insufficient attention in debates over AI's environmental footprint. While electricity consumption is drawing growing scrutiny, the impact on water resources is also becoming serious. For sustainable AI development, innovation in cooling technology and the adoption of water recycling systems are urgent priorities.
Dashlane Explains How Password Vaults Were Exfiltrated — AI Attacks Grow More Sophisticated
Dashlane, a password manager, has released a detailed explanation of how attackers managed to download encrypted password vaults.
According to Ars Technica, the attackers employed a tactic of targeting large numbers of users to raise the odds of successfully cracking some weak passwords. Although the vaults were encrypted, the exfiltration of the vaults themselves constitutes a serious security incident.
The incident suggests that AI-powered password attacks are becoming increasingly sophisticated. As attackers leverage AI to make password guessing more efficient, password manager vendors are likewise being forced to strengthen their defenses.
Sumitomo Rubber and Fujitsu Cut Tire FEM Analysis Time to One-Ninth with an AI Surrogate Model
Sumitomo Rubber Industries and Fujitsu have jointly developed a surrogate model that uses AI to predict tire performance.
According to ITmedia, in proof-of-concept experiments, the analysis time for predicting tire deformation behavior was reduced from roughly 45 minutes to about 5 minutes. High-precision analysis at a scale of approximately 600,000 elements was also achieved. FEM (Finite Element Method) analysis had been a bottleneck in product development due to its high computational cost, but the AI surrogate model has delivered a dramatic speedup.
In Japan's manufacturing sector, AI-driven simulation acceleration is one of the most promising application areas. Beyond tires, a wide range of applications is expected, including structural analysis and fluid analysis.
Higgs Audio v3 — An Open Speech Synthesis Model Supporting 100 Languages Debuts
Higgs Audio v3 TTS 4B has been announced in Reddit's LocalLLaMA community. It is a 4B-parameter text-to-speech model that supports 100 languages and also offers inline control.
Optimized for voice chat use cases, it enables control over emotional expression and speaking rate. Its release as open source is expected to further energize the development of multilingual voice AI applications.