Published in AI

OpenAI's Sam Altman declares AGI "achievable" with existing kit

by on05 November 2024


It might cost a bit

OpenAI CEO Sam Altman claims that current hardware can achieve Artificial General Intelligence (AGI). However, this optimistic vision requires $7 trillion price tag and many years to construct 36 semiconductor plants and additional data centres.

Altman did not hold back in his critique of the current model, GPT-4, describing it as "the dumbest model any of you will ever have to use again, by a lot. It's important to ship early and often, and we believe in iterative deployment."

Despite his candid remarks, Altman assured GPT-5 would surpass its predecessor, promising a more competent and efficient iteration.

Addressing the highly anticipated GPT-5 development timeline, Altman remarked, “We have some good releases coming later this year! Nothing that we are going to call GPT-5, though," thus dispelling rumours of an imminent launch.

Reports last month speculated about the release of a next-gen AI model, Orion, by December, but Altman swiftly labelled these claims as "fake news."

OpenAI seems to be moving away from the traditional naming of its models, suggesting that GPT-4's successor might be introduced under a new moniker. Altman hinted at a future where the AI functions like a "virtual brain," signalling a significant leap in the capabilities of their technology.

Chief Product Officer Kevin Weil responded enthusiastically to a suggestion for a hands-free way to end voice conversations with ChatGPT, stating, "I love this idea. Sharing with the team now!" This aligns with OpenAI's continuous efforts to enhance user experience and introduce advanced features.

Reflecting on the launch of ChatGPT, Altman shared, "The big one was how much people liked it. The early version of the product was a very, very larval prototype. Now it's pretty good, but it sure wasn't then. It wasn't very accurate, it didn't have many features, it went down all the time, etc. And yet people found enough value to stick with it."

Marking its second anniversary, ChatGPT is set to receive more GPUs to support its sophisticated functions.

Earlier this year, OpenAI introduced SearchGPT, a prototype search tool that aims to revolutionise the search landscape.

Altman said: "For many queries, I find it to be a way faster/easier way to get the information I'm looking for. I think we'll see this, especially for queries that require more complex research. I also look forward to a future where a search query can dynamically render a custom web page in response!"

Last modified on 05 November 2024
Rate this item
(0 votes)

Read more about: