Pedro Domingos has been tweeting about various topics lately. On November 25th, he discussed threats to Nvidia, including the development of more efficient learning algorithms, tech giants using their own chips, chip designs better suited for deep learning than GPUs, and the possibility of deep learning being replaced by algorithms unsuited to GPUs. He also mentioned non-silicon computing and the potential for an AI winter.
In a tweet on November 22nd, Pedro Domingos expressed his belief that a grad student somewhere is inventing the algorithms that will make OpenAI obsolete. On the same day, he responded to a tweet about the sex of OpenAI board members, stating that it is irrelevant and criticizing the fixation on this topic.
On November 17th, Pedro Domingos mentioned Sam Altman founding NextAI and speculated that OpenAI would fall behind as a result. He also expressed a preference for firing the board rather than Sam Altman.
In previous tweets from September 10th, July 24th, May 6th, April 3rd, December 23rd of previous years as well as August 2022 through January 2021 , Pedro Domingos touched on topics such as influential people in AI according to TIME's list versus his own compilation, authors leaving Google related to transformer paper publication and needing more engineers in government.
Based on these tweets alone without context from other sources or recent data analysis trends in AI are not clearly identifiable.
In terms of sentiment analysis regarding AI's progress overall based on these tweets alone without additional context or information it seems Pedro Domingos has mixed feelings. While he acknowledges threats and potential challenges in the field (e.g., efficient learning algorithms replacing current technologies), his tone remains realistic rather than overly negative or alarmist. Some tweets even express optimism (e.g., comparing AI to reading books and the advantages it brings). Overall, his sentiment towards AI appears to be neither strongly positive nor negative.
Books By Professor Pedro Domingos
The Master Algorithm