Microsoft CEO Satya Nadella recently spoke to Wired’s Steven Levy in an interview about the company’s heavy investment in generative AI and its recent deployment across most of its consumer and enterprise products. Nadella shared that it was only when the company was transitioning from GPT 2.5 to 3 that they started noticing impressive feats and realized that they could do and achieve more with this technology. The fact that no extra measures were put into place to train the model how to code, but it was still able to achieve impressive results was when he truly became a believer and realized that AI was transformative.
Nadella further recounted his eureka moment when the GPT-3 model helped him achieve one of his childhood dreams. While growing up in Hyderabad, India, he always wanted to read Persian poetry, and via the model, he was able to translate the work of Rumi into Urdu and then English.
Moreover, Nadella indicated that Microsoft has been investing in AI for a while now and even had its own entry dubbed Turing that was used across its offerings, including Bing and Azure. He disclosed that the company decided to join forces with OpenAI after making the discovery that they were both working towards the same goal. It was much easier for both companies to work together and combine resources to achieve this goal rather than trying to train multiple models simultaneously.
Levy inquired whether Microsoft considered buying OpenAI initially, but Nadella quickly responded and stated that it was a non-profit. Nadella further cited that OpenAI was able to come up with a complicated workaround where both companies were able to create a for-profit entry, allowing them to plow back profits from their ventures.
When asked about OpenAI’s Sam Altman belief on hitting AGI superintelligence benchmark, Nadella highlighted that his concern is more vested on how to make AI capabilities accessible to everyone.
The release of the new Bing chatbot caused a ripple effect, leading to key industry players jumping on the AI craze with their offerings headfirst, regardless of whether they were ready to ship. Several journalists cited instances where the chatbot had hallucinated or provided wrong responses outrightly. Microsoft maintained that the technology is getting better, and there are innovations that will help remedy instances where the chatbot hallucinates. Moreover, the tool sports several modes to help the user refine responses further.
When questioned about AI’s impact and whether it would destroy human life, Nadella highlighted that something that’s out of control is indeed a threat. However, elaborate measures have been put in place to establish control over the power grid and nuclear energy. Therefore, the same efforts and more can be used to manage AI.
The interview ended with Levy asking the CEO whether he thinks he’ll be remembered for his contributions to Microsoft, specifically the AI transformation.Nadella stated that it’s entirely up to people to decide what he’ll be remembered for and added that the company has been making strides in the past couple of years that continue to make it relevant in the technology space to date, with AI transformation at the forefront.
