NerdNews

Nvidia CEO Jensen Huang and other AI leaders at the World Economic Forum in Davos, Switzerland, downplayed concerns of an AI bubble, with Huang calling it a 'largest infrastructure build-out in human history'. However, OpenAI CEO Sam Altman has taken a different stance, recently introducing ads on ChatGPT, which he previously called a 'last resort'. Microsoft CEO Satya Nadella also warned that if the AI economy doesn't benefit non-tech sectors, it could become a bubble.

Elon Musk's lawsuit against OpenAI, Sam Altman, and Microsoft can proceed to a jury trial. Musk alleges that Altman and others misled him about OpenAI's plans to transition to a for-profit, resulting in zero profits for Musk. OpenAI denies the allegations, calling them 'legal harassment'. The case could have far-reaching outcomes for the future of AI and potentially force OpenAI to pay billions in damages.

Anthropic's Claude Code, an AI agent, can analyze DNA files, file taxes, and design knitting patterns. The company has announced Claude Cowork, a user-friendly version of Claude Code, which aims to bring agentic capabilities to a broader audience. Claude Cowork has a friendlier user interface and can access files and programs on a user's computer, but its success is yet to be seen.

RadixArk, the company behind SGLang, an AI model engine, has spun out with a $400M valuation. The company focuses on optimizing inference processing, allowing models to run faster and more efficiently. This move follows a trend of open-source AI tools becoming venture-backed startups, with companies like vLLM and Baseten also securing significant funding.

Anthropic has released a revised version of Claude's Constitution, a document that outlines the ethical principles guiding its chatbot. The update adds more nuance and detail on ethics and user safety, and hints at the possibility of chatbot consciousness.

Taiwan Semiconductor Manufacturing Company (TSMC) has reported record fourth-quarter earnings and expects AI chip demand to continue for years. CEO C.C. Wei expressed optimism about AI demand, citing conversations with cloud providers that verified the demand is real. TSMC plans to spend between $52 billion and $56 billion on capital expenditures in 2026, up from $40.9 billion in 2025.

China is at the forefront of humanoid robot development, with companies like Unitree leading the charge. These robots are expected to have a significant impact on the workforce and economy, with predictions of 10 million humanoids being shipped annually by 2035. Chinese companies are leveraging their tightly integrated manufacturing and supply chains to produce high-quality, low-cost robots, making them a major player in the global robotics market.

To identify a truly capable AI PC, consider Microsoft's Copilot+ PC definition, which requires at least 16GB of DRAM, 256GB of SSD storage, and a Neural Processing Unit capable of more than 40 trillion operations per second (TOPS). However, Intel's Jimmy Wai argues that it's not enough to focus solely on the NPU, as the CPU, GPU, and NPU are all crucial for different AI tasks. Intel's second-gen Core Ultra processors aim to provide a balanced platform for AI workloads, with up to 120 TOPS of aggregate processing power and a focus on energy efficiency.

Anthropic has published a 'constitution' for its AI model Claude, outlining principles for safe and ethical behavior. The document is used to train Claude and shape its character, with the goal of creating a more reliable and trustworthy AI. This approach is a shift from traditional mathematical methods and instead uses plain English rules to guide the model's behavior.

Reports of AI-related incidents have increased by 50% year-over-year, with deepfake-enabled scams and chatbot-induced delusions on the rise. The AI Incident Database tracks these incidents, but its crowd-sourced nature has limitations. Researchers are working on tools to enhance the database and help policymakers spot trends.

BioticsAI, the winner of TechCrunch Disrupt's Battlefield competition in 2023, has received FDA clearance for its AI-powered fetal ultrasound product. The technology uses computer vision AI to support fetal ultrasound quality assessment and anatomical completeness. The company plans to scale its product across health systems nationwide and add more features for fetal medicine and reproductive health.

Upscale AI, a startup, has raised $200 million in Series A funding to challenge Nvidia's dominance in switches for rack-scale AI systems. The company is developing a new class of scale-up network switches, called SkyHammer, which will compete with Nvidia's NVSwitch chips. SkyHammer will offer a memory semantic-based load-store network architecture and acceleration for collective communication.

Wix has launched Harmony, a new AI website builder that combines human creativity with advanced technology, allowing for pixel-perfect design control and speedy AI generation. The platform features an AI agent called Aria that can execute design desires, such as updating color palettes or adding commerce capabilities, based on natural language input.

Apple is reportedly developing an AI wearable, a pin with two cameras and three microphones, to compete with OpenAI's upcoming device, potentially releasing 20 million units in 2027

Zipline, a US autonomous drone delivery startup, has secured $600M in new funding, valuing the company at $7.6B. The funding will be used to expand its operations to at least four US states in 2026, with initial launches in Houston and Phoenix. Zipline's drones deliver food, retail, and health products, and have completed over 2M deliveries to date.

A massive spam wave is targeting people worldwide, originating from unsecured Zendesk support systems. The spam emails are being generated by attackers abusing Zendesk's ability to allow unverified users to submit support tickets, which then automatically generate confirmation emails. Companies such as Discord, Tinder, and Dropbox have been affected, with victims reporting hundreds of emails with strange and sometimes alarming subject lines.

Fortinet customers report that patched FortiGate firewalls are being hacked due to a patch bypass for a previously fixed critical authentication vulnerability. The company is planning to release new updates to fully patch the security flaw. In the meantime, admins are advised to temporarily disable the vulnerable FortiCloud login feature to secure their systems.

Threat actors are exploiting misconfigured web applications used for security training and internal penetration testing to gain access to cloud environments of Fortune 500 companies and security vendors. The vulnerabilities in these testing apps can be used to deploy crypto miners, plant webshells, or pivot to sensitive systems.

OpenAI has announced plans to invest in local communities near its Stargate datacenter facilities, aiming to minimize the project's impact on electricity prices, water usage, and local ecosystems. The company will establish OpenAI Academies to support regional workforce development and create jobs. This move comes as datacenter projects face increasing opposition from local communities due to concerns over noise, energy consumption, and job creation.

Lemonade is launching an insurance product for Tesla's Full Self-Driving customers, promising to cut per-mile rates by 50%. The product, called Autonomous Car insurance, will leverage vehicle telemetry data and train usage-based risk prediction models to determine when a driver is using FSD or operating the vehicle themselves.