ChatGPT’s 700 million users will from today be using its new model, GPT-5, which its creators say is faster, more accurate, and less likely to produce fabricated or incorrect answers [1], [2], [3].
The new OpenAI model comes three years after the launch of the first ChatGPT and two years after its last major upgrade (GPT-4). Experts note that while it performs better in programming and in solving mathematical and other scientific problems, the leap from GPT-4 is not as “dramatic” as the leap from GPT-3 to GPT-4.
Sam Altman claimed the model responds like a PhD-level expert, with Elon Musk making similar claims about his own AI’s latest version, Grok, just days ago. The companies with the largest AI investments — Alphabet, Meta, Amazon, and Microsoft — are expected to spend $400 billion on AI this year.
The companies with the largest AI investments — Alphabet, Meta, Amazon, and Microsoft — are expected to spend $400 billion on AI this year.
Even so, GPT-5 still cannot replace humans, as it lacks the ability to learn independently. On his popular AI podcast, Dwarkesh Patel compared current AI training to teaching a child the saxophone by having them read notes from a previous student.
“A student gives it a try,” he said. “The moment they make a mistake, you send them home and write detailed notes on what went wrong. The next student reads your notes and tries to play Charlie Parker right away. When they fail, you refine the notes for the next student. That just wouldn’t work.”
