OpenAI has launched a new search function within ChatGPT, allowing users to locate specific information from past conversations quickly. Available initially to Plus and Team plan subscribers, this tool enables users to search old chats by entering keywords. With this, users who frequently return to ChatGPT for project details can now find information more efficiently, reducing time spent manually scrolling through previous exchanges.
Search Tool and its Ease of Use
The search function, accessible via a magnifying glass icon atop the chat list, shows results in a separate panel, so users won’t have to re-enter keywords each time they revisit. This feature is rolling out in stages, starting with Plus and Team subscribers and expanding to Enterprise and Education accounts within the week.
Free users will gain access in November, following OpenAI’s approach of offering new features first to paid plans, then releasing them more widely. For regular ChatGPT users, this tool brings a streamlined way to retrieve information from lengthy chat histories.
We’re starting to roll out the ability to search through your chat history on ChatGPT web.
Now you can quickly & easily bring up a chat to reference, or pick up a chat where you left off. pic.twitter.com/YVAOUpFvzJ
— OpenAI (@OpenAI) October 29, 2024
Voice Mode Launches Across Europe
OpenAI recently started expanding ChatGPT’s Advanced Voice Mode to the European Union, including regions like Norway, Iceland, and Switzerland, following its initial release in the U.S. and U.K. Voice Mode provides a responsive, near-instantaneous interaction with ChatGPT, processing voice queries in under 232 milliseconds.
The feature offers five new voices, including Arbor, Vale, Sol, Maple, and Spruce, allowing users a choice in tone and accent that fits personal preferences. Voice Mode has a daily usage cap, alerting users when they near their limit, designed to balance platform demand.
OpenAI’s advances in ChatGPT’s voice capabilities present a challenge to other tech giants. Siri’s performance, which reportedly trails ChatGPT’s by around 25% in accuracy, has revealed gaps in Apple’s voice AI.
Google’s Gemini Live, released in September, has also seen limited success, with feedback highlighting issues around personalization and fluidity that remain unaddressed. The competitive gap underscores OpenAI’s advantage, as ChatGPT’s voice features offer natural conversation flow, customizable accents, and more responsive interactions.
Canvas and Enhanced Customization Features
In early October, OpenAI introduced the Canvas tool to ChatGPT’s Plus and Team users, adding precision to editing and debugging. Canvas allows users to isolate specific sections of text or code, making it easier to apply focused revisions, with shortcuts for adjusting length and readability.
For developers, this tool streamlines coding workflows, adding debugging comments with greater accuracy. Canvas, alongside the new search and voice tools, positions ChatGPT as an adaptable platform for professionals needing efficient, targeted AI support.
Pivot to TSMC and Broadcom Partnerships for Custom AI Hardware
On the hardware front, OpenAI has recently overhauled its strategy, moving from a $7 trillion global foundry plan to collaborations with TSMC and Broadcom for chip development.
TSMC’s A16 node technology, expected to begin large-scale production in 2026, will support OpenAI’s goal of making chips that process data faster with reduced power usage, helping ChatGPT handle high-demand tasks more efficiently. Inference chips, like those developed with Broadcom, are optimized for ChatGPT’s real-time responses by focusing on specific tasks, ensuring smooth performance during user interactions.
With over 11 million subscribers, ChatGPT generates over $225 million monthly, although OpenAI’s operational costs remain high, with a projected $5 billion loss this year. Compute costs, which cover server hardware, energy, and processing resources, represent one of OpenAI’s largest expenses.
By diversifying its hardware suppliers, including AMD’s MI300X chips integrated via Microsoft’s Azure platform, OpenAI aims to manage these costs while reducing reliance on Nvidia. Partnering with both TSMC and Broadcom to build custom chips enhances OpenAI’s flexibility, balancing the needs of its extensive user base with sustainable supply chains.
Last Updated on November 7, 2024 2:16 pm CET