OpenAI is taking significant strides in expanding the capabilities of its desktop apps for PC and Mac, bringing ChatGPT closer to becoming a fully agentic assistant capable of piloting your computer. During Day 11 of its "12 Days of OpenAI" event, OpenAI unveiled new integrations with a host of applications, including Integrated Development Environments (IDEs), terminals, and text apps.

Key Highlights:

• New IDE Integrations: ChatGPT now supports BBEdit, MatLab, Nova, Script Editor, TextMate, and VSCodium. These join the JetBrains family of IDEs, such as IntelliJ IDEA, PyCharm, and Android Studio, as well as popular tools like VS Code, Xcode, and iTerm 2.

• Beyond Coding Apps: Additional integrations include productivity tools such as Apple Notes, Notion, and Quip, enabling broader utility across project contexts.

• Advanced Voice Mode Enhancements: ChatGPT's Advanced Voice Mode now works seamlessly with supported apps, adapting to the user's workflow and providing contextual assistance based on ongoing projects.

Why It Matters:
Kevin Weil, Chief Product Officer at OpenAI, highlighted that these desktop app integrations represent a pivotal step toward a more agentic user experience for ChatGPT. “As our models get increasingly powerful, ChatGPT will more and more become agentic. That means we’ll go beyond just questions and answers; ChatGPT will begin doing things for you,” Weil said.

By enabling deeper access to desktop applications and workflows, OpenAI aims to unlock new levels of productivity and automation for users. These features place ChatGPT in closer competition with Anthropic’s Claude Computer Use, which offers similar capabilities like screen navigation and text typing.

At the DealBook Summit, Google CEO Sundar Pichai addressed key challenges and opportunities facing the tech giant, including its position in the AI race, ongoing antitrust lawsuits, and the ethics of data usage in AI model training.

Key Highlights:

• Google leverages proprietary infrastructure, products like YouTube and Gmail, and Nobel Prize-winning research to lead in compute, data, and algorithms.

• Over 25% of new Google code is AI-generated but engineer-reviewed, potentially broadening programming accessibility.

• Pichai advocated for leveraging existing regulatory frameworks rather than introducing excessive new laws to govern AI applications.

• The CEO hinted at the possibility of some Alphabet subsidiaries spinning off in the future.

Why It Matters:
Google's stance on balancing innovation with responsibility and societal impact will likely set a benchmark for the tech sector as AI's role in reshaping industries deepens and regulatory scrutiny intensifies.

A word from our sponsor:

RAG in just a few lines of code? We’ve launched a predefined RAG tool on our developer platform, making it easy to bring your data into a Knowledge Graph and interact with it with AI. With a single API call, writer LLMs will intelligently call the RAG tool to chat with your data.

Integrated into , it eliminates the need for complex vendor RAG setups, making it quick to build scalable, highly accurate AI workflows just by passing a graph ID of your data as a parameter to your RAG tool.

AI search startup Perplexity has secured a massive $500M funding round, tripling its valuation to $9B as it continues to challenge traditional search engines.

Key Highlights:

• Perplexity's valuation has skyrocketed from $1B in April to $9B in this latest round.

• The startup has attracted over 15M active users since its 2022 launch.

• Recent feature additions include one-click shopping and financial analysis.

• Perplexity has inked revenue-sharing deals with major publishers to address content usage concerns.

Why It Matters:
Perplexity's rapid rise demonstrates the ongoing shakeup in the search landscape, with AI-powered alternatives gaining traction against established industry giants.

DeepMind has introduced a new benchmark for evaluating the factual accuracy of AI models, with Gemini 2.0 Flash emerging as the most accurate model available today.

Key Highlights:

• The new tool helps determine which models are best at catching their own errors.

• Gemini 2.0 Flash outperforms other models in factual accuracy.

Why It Matters:
This benchmark could become a crucial tool for assessing and improving the reliability of AI models, potentially leading to more trustworthy AI applications across various industries.