Weekly Summary:
ChatGPT gets a computer, enabling ChatGPT to book your flights, build business presentations and more
xAI’s releases Grok 4
Perplexity launches Comet browser for AI-first web
Feature Story #1
OpenAI releases ChatGPT agent, bridging research and action
Summary: OpenAI’s new “agent” version of ChatGPT represents a major leap, no longer just generating text, the model can now take real-world actions, such as using tools, browsing the web, and executing multi-step tasks.
Why It Matters: Previously, AI models worked in isolation, you provided context, asked a question, and received a response. Agents change that dynamic by allowing models to act autonomously across apps and systems. Multiple startups have explored this direction, but with OpenAI’s resources, talent, and infrastructure, this shift could redefine how AI integrates into our lives and work.
Our Take: As with all major tech shifts, missteps and hype are inevitable. But the potential here is real: models that don’t just answer but do. This could radically change workflows in knowledge work and personal productivity. We’re super excited to monitor use cases from the public.
Feature Story #2
Grok 4 passes Claude, DeepSeek in model rankings
Summary: Grok 4, the latest model from xAI, now tops various benchmarks for reasoning and comprehension. outperforming rivals like Claude and DeepSeek.
Why It Matters: The gen AI space is moving at breakneck speed. Each new model iteration brings leaps in intelligence and usability. Grok 4’s rapid improvement underscores the competitive intensity — and Elon Musk’s clear intent to make xAI a top-tier AI contender.
Our Take: We’re impressed by how quickly Elon and team have shipped a leading model. Smarter models open doors to real-world automation and embedded intelligence — and knowing Elon, Grok 4 will likely find its way into Tesla robotics, autonomous systems, and maybe even neural interfaces. We’re watching closely.
Feature Story #3
Perplexity launches Comet browser
Summary: Perplexity’s new Comet browser is an interesting move: a real-time AI-native browsing experience built from the ground up. This indicates their ambition to be the front door to the internet in an AI-first era.
Why It Matters: This isn’t just search with AI, it’s an entirely new way to use the internet, with the AI continuously contextualising and adapting to your needs. It’s a direct challenge to how Google has operated for decades, and a potential inflection point in how we interact with the web.
Our Take: If this is successful, this likely marks the beginning of all top tier LLM models producing their own browsers.