OpenAI previews an exciting future during Reddit AMA

OpenAI has now completed an internet rite of passage: the Reddit Ask Me Anything (AMA).

CEO Sam Altman, Chief Product Officer Kevin Weil, SVP Of Research Mark Chen, VP Of Engineering Srinivas Narayanan, and Chief Scientist Jakub Pachoki answered questions about everything AI for an hour and a half earlier today.

Here are the highlights:

• Kevin says that although OpenAI will continue both the o series models and the regular GPT models, he expects the models to ultimately converge. This would check out with the report that OpenAI wants to eventually merge their models to create artificial general intelligence (AGI).

• They don’t have a release plan for their next text-to-image model, but it will be “worth the wait.” 

• Sam hyped up SearchGPT, OpenAI’s new search engine. 

• ChatGPT may be singing to you soon.

• Something really good is coming this year, but it won’t be called GPT-5. Perhaps…Orion?

• Sounds like autonomous AI agents are coming in 2025.

• And that it will be the next giant breakthrough.

• Longer context window coming soon.

• Sam believes they have all the hardware they need for AGI.

• The full o1 is a lot better than o1-preview.

  

• Sam wants to see OpenAI become more "open" again.

• He’s also very bullish on how AI will empower founders.

  

• Here’s what Srinivas would focus on learning if he was a new grad interested in AI.

• Last but definitely not least, we finally (sorta?) find out what Ilya saw.

So, there you have it. By the looks of it, OpenAI is very bullish on the future, especially regarding autonomous AI agents. 

If they can deliver anything close to what they are alluding to in the AMA, then 2025 promises to be a very exciting year.