Welcome to the newest edition of ID Tech’s AI update. Here’s the latest big news on the shifting landscape of AI and identity technology:
A platform for AI-driven voice agents has adapted Sesame’s AI model to create uncannily realistic synthetic voices that can be streamed in real-time. “We’ve spent the past couple of weeks rearchitecting CSM-1B from the ground-up to support realtime, low-latency inference, and we’ve finally cracked it,” the Vogent AI team explained in a post on Y Combinator.
American chip companies will escape China’s retaliatory tariffs on the U.S., at least for now. Amid the rapidly escalating trade war, the China Semiconductor Industry Association (CSIA) has clarified that it determines the country of origin for semiconductors based on where the fabrication plant is located—so, for example, AMD chips manufactured by Taiwan Semiconductor Manufacturing Co would be considered to be from Taiwan.
Speaking of which, TSMC’s quarterly sales jumped 42 percent for Q1 of 2025, beating market expectations somewhat as demand for AI-related chips has persisted. Of course, there are nevertheless growing concerns about market volatility and the trade war’s potential impacts on the sector.
Japan’s NTT Corporation is getting into the AI chip market, unveiling a large-scale integration (LSI) chip that is specialized for the task of real-time AI inference processing of ultra-high-definition video. This means AI tech that can analyze and understand video in real time and with a 20-watt power consumption, offering potential applications in drones and security feeds.
OpenAI has launched a countersuit against Elon Musk, the firm’s former co-founder who has been waging legal war on the company, ostensibly over its alleged betrayal of its founding mission to develop open source AI technology. “Elon’s nonstop actions against us are just bad-faith tactics to slow down OpenAI and seize control of the leading AI innovations for his personal benefit,” OpenAI posted on a Musk-owned social media platform.
OpenAI has dramatically reduced the amount of time and effort that goes into safety testing, according to a Financial Times report. Citing eight unnamed individuals who are familiar with its testing processes, the report says that for major new AI models, the safety testing period has dropped from months to weeks as the firm feels the heat from “competitive pressures”.
Andreessen Horowitz (a16z) is putting together a $20 billion “megafund” for AI investments, seeking to target growth-stage AI companies in the U.S. This would be the VC giant’s largest fund ever, and a significant portion of it would be allocated to supporting companies that are already in a16z’s portfolio, including Databricks, Mistral, Safe Superintelligence, and Elon Musk’s xAI.
The chatbot’s take: This week we tried to leverage GPT-4o’s knowledge cut-off for a bit of perspective on everything:
–
April 11, 2025 – by Alex Perala
