OpenAI prepares to launch GPT-5 in August

Quick Note

AI is getting pretty good, very quickly — but tech companies are still trying to figure out how to make it easier to access and use. Today’s news dives into their ongoing efforts to do so.

OpenAI prepares to launch GPT-5 in August

OpenAI is reportedly planning to release its highly anticipated GPT-5 model by early August. The upcoming model is expected to be the most advanced AI in the world at launch.

Why it matters: GPT-5 will combine OpenAI’s core technologies into a single model for the first time. The company’s goal is to create a single AI system that excels at any task, unlike current models that specialize in different areas.

Diving deeper:

• ChatGPT currently hosts seven different models that each fall under one of two different families.

• OpenAI’s GPT models are general-purpose chatbots designed for tasks involving text, while its “o” series reasoning models are optimized for more complex tasks like coding.

• GPT-5 will combine the GPT series with the “o” series, ideally creating a single model that performs well across a wider range of tasks. 

Final note: Unified systems like GPT-5 are referred to as hybrid models. These types of systems are starting to emerge as the next step for most AI companies, with Anthropic debuting the first of its kind in February.

Why are hybrid models important?

Imagine you log into ChatGPT for the first time without having any experience with AI tools. You’re immediately greeted by a dropdown menu with seven different models to choose from, each accompanied by a one-line description that doesn’t fully explain what they do. You might pick one at random and walk away disappointed when it struggles with tasks that seem simple. 

OpenAI CEO Sam Altman doesn’t want first-time users to have that problem. He recently acknowledged that his company’s product offerings have gotten too complicated, and wants to create an experience where AI “just works” without users needing to navigate a model picker.

Microsoft Edge becomes AI browser with launch of Copilot Mode

Microsoft has launched a new feature in its Edge browser called Copilot Mode, allowing users to browse the web while being assisted by AI.

Key background: AI web browsers are designed to help users search the internet more efficiently. They can summarize web pages, extract important information, and even automate simple tasks like filling out forms or comparing products. It’s exactly like having an AI agent built directly into your browser.

The details:

• In Copilot Mode, Edge will watch what the user is doing, predict what they want next and then take action on their behalf.

• Microsoft says that Copilot can handle tasks like booking appointments, creating shopping lists and drafting content. 

• In the future, it will remember what users are working on and prompt them to pick up where they left off the next time they open Edge. 

Why it matters: None of these capabilities are new on their own, but integrating them directly into web browsers is a way to speed up interactions with AI. Similar to how OpenAI wants a single model to meet all of its users’ needs, Microsoft wants to deliver all of its AI capabilities in one place. 

More trending news

OpenAI launches Study Mode in ChatGPT

Spotify teases conversational AI interface for its music streaming platform

Microsoft renegotiates deal with OpenAI, aims to access its tech even after AGI milestone

Thanks for reading! Just a heads up — I might not hit your inbox tomorrow or Thursday. If that’s the case, we’ll catch up on everything we missed next week.

See you next time,

Chain of Thought