Friday, March 7, 2025

Welcome to TC PM! Today we learn about a new way for X users to prompt xAI's Grok; we dig into a new AI bill in California — written by the author of the controversial SB 1047 bill; and we learn that Russia may be meddling with AI model answers. Have a great weekend!

TechCrunch Top 3

Image Credits: Jaap Arriens/NurPhoto / Getty Images

💅 More AI: X users are reporting that they can now ask xAI's chatbot Grok questions in their X replies and get an explanation about a post. X users already had access to Grok through a designated tab on X and through a button on every post, but we needed more avenues I guess. Read more

🦙 More models: Meta's next "open" AI model will reportedly have a voice focus. The model is expected to arrive "within weeks" and will be an "omni" model, which means it is capable of natively interpreting and outputting speech as well as text and other data. Read more

📅 Even more AI: Google is testing a new AI feature in Google Calendar that allows users to prompt Gemini AI to check their schedule, create a new event, or look up an event's details. It's meant to prevent people from having to search or manually add things to their calendar. Read more

A Message from Google Cloud

What businesses need to know about AI in 2025

Hear from 23 industry leaders, including Amin Vahdat (Google Cloud), Chamath Palihapitiya (Social Capital), Harrison Chase (LangChain), Crystal Huang (GV), and more. They share their perspectives on the most critical cloud and AI trends, opportunities, investment areas, and challenges that business leaders need to be aware of in 2025 in Google Cloud's Future of AI: Perspectives for Startups report.

Download Here

Afternoon must-reads

Image Credits: JuSun / Getty Images

🎙️ New AI bill: On today's episode of Equity, the hosts talk about a new AI bill in California that was written by the author of the controversial SB 1047 bill last year. They also dive into CoreWeave's IPO, Ramp's recent growth surge, and more. Listen here

😬 Intentional misinformation: A new report from NewsGuard found that a Moscow-based network called "Pravada" has been actively publishing false claims to affect the responses of AI models. Pravada published 3.6 million misleading articles last year. Read more

⚠️ A hack! NTT Communications, a Japanese telecom company, confirmed a cyberattack that affected the data of 18,000 corporate customers. It's unclear how many individuals were affected by the breach, which was discovered on February 5. Read more

👯 Less AI: Health experts are starting to warn about the overreliance on AI chatbots and companions for social connection. Recent research found that AI companion mobile apps saw 652% year-over-year revenue growth in 2024. Read more

🪆 Embedding models: Google added a new experimental model to its Gemini Developer API called Gemini Embedding that translates text inputs like words or phrases into numerical representations known as embeddings. Read more

Around the web

🙈 New online scam just dropped: There are numerous online scams through platforms like X and Telegram that target Donald Trump's supporters by promising to help them make money from their Trump-related items — for a price. Read more

💼 Over automation: Wall Street has adopted AI to help automate some of the "grunt work" tasks that make up a substantial portion of a junior banker's career. But junior bankers are worried that this automation means they aren't learning the skills they need to get to a higher role in the future. Read more