91download.com supports a wide range of platforms, including YouTube, Facebook, Twitter, TikTok, Instagram, Dailymotion, Reddit, Bilibili, Douyin, Xiaohongshu and Zhihu, etc. Click the download button below to parse and download the current video
The world of artificial intelligence took a monumental leap forward with the advent of the Transformer model, a breakthrough credited to Jakob Uszkoreit and his team at Google. In this interview, we delve into the origins of this game-changing concept and explore its implications on technology, business, and our future.
Jakob Uszkoreit, the mind behind the Transformer, shares that the idea was born out of necessity. His team at Google faced a challenge with Long Short-Term Memory networks (LSTMs), which, despite their sophistication, were training at a snail's pace. This led to the birth of the attention mechanism, a concept that allowed for self-attention and faster training times, especially compatible with GPUs.
"It's you're always standing on shoulders of giants. But really what happened around that time is that my group was faced with an interesting conundrum..."
The Transformer model, published in 2017, was not just a step in the evolution of AI but a leap that opened up new avenues for applications ranging from machine translation to protein structure prediction.
Transformers have been silently working behind the scenes of our favorite Google products, from Search to Translate to Assistant. Jakob explains that these models were already in production, though their deployment was less flashy than ChatGPT's debut.
"ChatGPT hasn't actually been around as a consumer product for that long. And so there are actually applications, for example, at Google and other places where transformers have been deployed in production long before..."
This leads to the question of whether Google let a golden opportunity slip through its fingers by not being as quick to market with Transformer-based products. Jakob offers a nuanced perspective, highlighting the complexity and caution required when dealing with such transformative technology.
As one of the pioneers of the Transformer, Jakob feels a profound sense of responsibility towards the advancement of generative AI. He emphasizes the importance of data curation and quality control to mitigate bias and ensure that AI serves humanity's best interests.
"Deploying these responsibly in a way that actually maximizes benefit to humanity..."
Looking ahead, Jakob shares his vision for Inceptive, the biotech and AI startup he founded after leaving Google. He envisions a future where generative AI designs medicines that are in harmony with living systems, potentially saving countless lives.
In terms of profitability, Jakob believes that while media might not be the最容易 area to deploy AI profitably in the short term, there are numerous high-reward use cases, such as in the medical field, where generative AI could generate significant value.
"There's no question that over the next handful of years, certainly before the end of this decade, we'll see medicines hopefully saving many people's lives..."
Generative AI is poised to disrupt various industries, including education. Jakob suggests that it could make education more cost-effective and personalized, revolutionizing the way content is created and delivered.
"All sorts of processes requiring complex communication can be dramatically accelerated and or scaled using generative AI."
For students eyeing a career in AI, Jakob recommends a interdisciplinary approach, combining natural sciences, mathematics, and computer science.
In conclusion, Jakob Uszkoreit's journey from Google to Inceptive is a testament to the transformative power of AI. As we harness this technology, we must do so responsibly, ensuring it serves to enhance our lives and address some of our most pressing challenges.
Share on Twitter Share on Facebook