Microsoft co-founder Bill Gates’s skepticism of Artificial Intelligence took on a new dimension as he predicted chatbots may have already seen their peak.
While tech entrepreneurs like Elon Musk have cautioned against the unbridled development of AI technologies, fellow billionaire Gates has routinely positioned himself in support of a mutually assured destruction mindset when it comes to the emerging technology. Now, almost a year after OpenAI introduced ChatGPT to the public in late Nov. 2022, he predicted development has likely “reached a plateau.”
The notion was suggested during an interview with German business newspaper Handelsblatt from Dakar, Senegal where the tech mogul spoke to the advancements specifically in Generative Pre-trained Transformers, the GPT of the well-known AI tool.
“Microsoft founder Bill Gates predicts rapid progress for artificial intelligence (AI),” wrote the newspaper. “In the ‘next two to five years,’ the accuracy of hits of AI software will increase and the costs will drop so much that new and reliable applications become possible.”
Though he noted “The potential is great” and he could be wrong, the outlet went on to describe how he “predicts stagnation in development. ‘We have reached a plateau,’ explained Gates with a view to the GPT AI model from OpenAI, which caused a sensation worldwide.”
“The next version will not be much better than the current GPT-4, you have reached a limit there,” they wrote.
Though he was negative on the overall differences that could be noticed looking ahead toward GPT-5, he did not describe the advancements from GPT-2 to GPT-4 as “incredible.”
Gates also spoke to the fact that with AI chips currently running as much as $30,000 from Nvidia, there was an expectation that production costs would go down and thus distribution of the technology would spread.
“Well, it’s pretty expensive to train a large language model. But the actual usage costs were once ten cents per query. Today it’s probably more like three cents. The costs for computing power or semiconductors remain enormous,” he told Handelsblatt.
Earlier in the year, while Musk had been signing his name to a letter of industry leaders encouraging a pause to AI development in order to let society catch up with technology, Gates had been interviewed by ABC News’ Rebecca Jarvis to whom he suggested the need to plow ahead.
“We’re all scared that a bad guy could grab it,” he explained. “If you just pause the good guys and you don’t pause everyone else you’re probably hurting yourself.”
“You definitely want the good guys to have strong AI,” continued the Microsoft co-founder.
“The field, more than any field I know, is actually putting in a lot of the smart people into, OK, what comes next and how do we make sure that’s beneficial,” Gates told Jarvis. “I see that AI, used properly, is providing a lot of benefits that I wouldn’t want to throw away.”
Some of the benefits he spoke of with Handelsblatt included in healthcare, toward developing pharmaceuticals and vaccines quicker, as well as toward the green agenda in producing better climate models.
On more advanced AI systems like the hypothetical Artificial General Intelligence (AGI), more in line with the sentient machines so often depicted in SciFi, Gates admitted there was no telling when that might actually come about.