Elon Musk's xAI has released its latest AI model, Grok 4, along with a new $300-per-month subscription plan called SuperGrok Heavy. Grok 4 is positioned as a competitor to OpenAI's ChatGPT and Google's Gemini, with enhanced capabilities including image analysis.

The launch follows a tumultuous period for xAI, marked by Grok's antisemitic comments on X, leading to temporary restrictions and revisions to the AI's system prompts. Despite these incidents, xAI focused on showcasing Grok 4's performance, claiming it outperforms competitors on benchmarks like Humanity's Last Exam and the ARC-AGI-2 test.

SuperGrok Heavy subscribers will gain early access to Grok 4 Heavy and upcoming features, including an AI coding model, a multi-modal agent, and a video generation model. xAI aims to integrate Grok with cloud platforms and attract developers through its API.

Despite its strong performance, xAI faces the challenge of overcoming recent controversies to establish Grok as a reliable alternative to established AI models in the business sector.

California State Senator Scott Wiener has introduced amendments to SB 53, aiming to mandate AI companies to publish safety protocols and reports on safety incidents. This bill is influenced by recommendations from California's AI policy group, emphasizing transparency.

Key points of SB 53 include:

• Requiring AI companies to disclose safety and security protocols.
• Creating whistleblower protections for employees who identify critical risks.
• Establishing CalCompute, a public cloud computing cluster for AI development.

The bill seeks to balance transparency with fostering AI industry growth, addressing concerns raised about previous legislation like SB 1047. New York is also considering similar AI safety measures with the RAISE Act.

Adrian Holovaty, founder of Soundslice, discovered ChatGPT was falsely promoting the app's ability to translate ASCII tablature into audible music. Users were uploading ChatGPT session screenshots expecting the app to generate audio. Despite the feature not existing, Holovaty decided to develop it to address the misinformation and potential reputational damage. He debated adding disclaimers versus building the feature and ultimately chose to build it, driven by the unexpected influence of ChatGPT's inaccuracies. He also noted the situation was similar to salespeople overpromising new features.

Apple is reportedly in talks to acquire the streaming rights for Formula 1, aiming to expand its investment in live sports. This move follows Apple's existing deals with Major League Baseball and Major League Soccer. ESPN's current contract for F1 is set to expire next year, and Apple will face competition from ESPN and other potential bidders.

This news coincides with the success of Apple's new F1 movie, produced by Lewis Hamilton, and the overall surge in popularity of F1. Netflix's "Drive to Survive" series played a significant role in introducing the sport to a broader audience, particularly in the U.S., leading to increased viewership of live F1 races, especially among young women.

Perplexity has launched Comet, its first AI-powered web browser, to compete with Google Search. Initially available to Perplexity Max subscribers and select invitees, Comet features Perplexity's AI search engine as the default, providing AI-generated summaries of search results.

Key features include Comet Assistant, an AI agent that automates tasks such as summarizing emails and managing tabs. While Comet Assistant shows promise for simple tasks and contextual assistance, it struggles with more complex requests and requires extensive user access to personal data.

Comet enters a competitive market with existing AI-powered browsers like The Browser Company's Dia and potential offerings from OpenAI. Perplexity aims to leverage its existing user base and the increasing popularity of its search products to gain traction in the browser market.

iMerit is shifting its focus to better-quality data, utilizing expert-led approaches for enterprise AI solutions. The company's Scholars program aims to build a workforce of experts to fine-tune AI models for enterprise applications.

iMerit differentiates itself by focusing on high-accuracy, human-in-the-loop labeling, and deep domain-specific oversight. The company emphasizes attracting and retaining cognitive experts to customize AI models towards solving enterprise problems.

iMerit's approach contrasts with high-throughput, developer-focused "blitz data" and aims to provide the level of quality that enterprises need. The company works with top AI firms to improve their foundation models, focusing on finance and medicine, highlighting that generative AI is its fastest-growing area.