We got AI because we weren't working from home
Innovation happens in person
My plan going forward it so post twice per week, Tuesday morning for everyone and Thursday morning for premium subscribers. Today’s is a little late due to my ambitious run in the Greek sun… This week premium subscribers will get two posts on Expedia’s new loyalty program. You should subscribe!
The FT has a fantastic profile (ungated) of the eight Google researchers who wrote the modern foundational AI paper “Attention is all you need”. That paper became the core insight and building block that has led to the current generative AI revolution. The article is worth reading in whole to better understand the breakthrough that came from the paper.
For today’s purpose I want to dig into the description of the serendipity that led to the discovery. While the article does not say it directly, it is clear from the profile that if the eight researchers had been working from home it is unlikely the paper would have ever been written. Consider the following quotes from the piece (bold emphasis mine)
In early 2017, two Google research scientists, Ashish Vaswani and Jakob Uszkoreit, were in a hallway of the search giant’s Mountain View campus, discussing a new idea for how to improve machine translation, the AI technology behind Google Translate…
…The three Google scientists surmised this [scanning and analyzing an entire sentence at once rather than word-by-word] would be much faster and more accurate than existing methods. They started playing around with some early prototypes on English-German translations, and found it worked. During their chat in the hallway, Uszkoreit and Vaswani were overheard by Noam Shazeer, a Google veteran who had joined the company back in 2000 when Google had roughly 200 employees… The chance conversation formalised a months-long collaboration in 2017 that eventually produced an architecture for processing language, known simply as the “transformer”…
…As they all remember it, they were originally working as three separate groups on various aspects of self-attention, but then decided to combine forces... “The idea for the transformer formed organically as we worked and collaborated in the office,” says Jones. Google’s colourful open-plan working environment, complete with campus bicycles, proved fruitful. “I recall Jakob [Uszkoreit] cycling up to my desk and scribbling a picture of a model on a whiteboard behind me and gathering the thoughts of whoever was in earshot.”
One of the ongoing themes of Marketing BS is that while working from home seems to make some people (most people?) more productive, it does so at the cost of collaboration, learning, and innovation. When you are running an existing process working from home makes a lot of sense. But when you are trying to create something new or solve hard problems, nothing beats being in the office.
I have previously shared that the biggest single value-driver we created at A Place For Mom happened due to a hallway conversation. It seems like the greatest value-creating research paper of the last decade happened due a series of hallway conversations.
The question is, what type of company is Google?
On June 8th Google has told 80% of their employees that they need to spend at least 3 days per week in the office. This means that 20% are still fully working remotely, and everyone else is still working from home ~40% of the time.
That might be okay. Google is a $1.5 Trillion dollar company with almost all their revenue coming from a single product (advertising). It could be that Google’s workforce should be run as a slow moving business driving incremental productivity improvements and ensuring nothing breaks. Working from home could be the best way to make that model run most efficiently.
But it is not a way to be innovative.
And the researchers that wrote the foundational paper all realized that. All eight have now left Google. From the FT:
…they also found that Google was not structured in a way that allowed for risk-taking entrepreneurialism, or launching new products quickly. It would require building a “new kind of software . . . computers you can talk to,” Vaswani adds. “It seemed easier to bring that vision to light outside of Google.”…
“The reason why I left Google was that I actually didn’t see enough adoption in the products that I was using. They weren’t changing. They weren’t modernising. They weren’t adopting this tech. I just wasn’t seeing this large language model tech actually reach the places that it needed to reach,”
Where are the researchers now?
Illia Polosukhin: Left in 2017 to co-found Near (Blockchain)
Aiden Gomez: Left in 2019 to found Cohere (applied generative AI)
Jacob Uszkoreit: Left in 2021 to co-found Inceptive (biotech using transformers)
Noam Shazeer: Left in 2021 to co-found Character.ai (Consumer generative AI)
Lukasz Kaiser: Left in 2021 to join OpenAI
Ashish Vaswani and Niki Parmar: Both left in 2021 to co-found Essential.ai (AI business applications)
Llion Jones: Left in 2023 to co-found a [still stealth] start-up
I don’t know for a fact, but I would bet that the companies these people founded have employees who all work in the office.
Keep it simple,
Edward