Trending Articles
Trending Topics
Trending
Sergey Brin thinks Google staff should work harder to help the company win the race to Artificial General Intelligence – and ditch remote work for in-office.
Now, if you’re a Google staffer, you may be wondering why you should listen to Brin. He co-founded Google, but stepped back from running the company in 2019 – then got so excited by the AI boom and/or disappointed by Google’s progress that he returned.
Last week, Brin sent a memo around Google’s AI team that was seen by The New York Times: “I recommend being in the office at least every weekday,” he wrote, adding that “60 hours a week is the sweet spot of productivity.” He warned against working longer than that, as any more will apparently lead to burnout.
“Competition has accelerated immensely and the final race to AGI is afoot,” he reportedly wrote. “I think we have all the ingredients to win this race, but we are going to have to turbocharge our efforts.”
Google says it has no plans to change its hybrid working system in response to this memo.
Do we really live in a time where hard graft is all that’s holding up AGI? Just because Brin believes it doesn’t make it true.
Consider what actually held back deep learning in the first place. A slowdown in funding research in the subject, which Geoffrey Hinton et al ignored. A lack of data to train these massive systems. And a need for the right type of compute power, which Nvidia eventually supplied.
All hands have been on deck ever since but it hasn’t necessarily been enough to win the AI race.
That’s because it’s not actually about hard graft and long hours. Two years before Brin departed – in other words, on his watch – Google actually published the key paper that outlines a network architecture called transformers that enable large language models to leap forward to the level we’ve seen.
Despite this early advantage, Google failed to take advantage of ideas that should have given it a big lead. And that was well before working from home became a pariah policy to pin blame upon.
Human productivity is a funny thing, but much of the research suggests that there is a maximum amount of time most of us can be productive without burning out or productivity starting to drop off. And it’s well below 60 hours. Sorry Sergey.
One study suggested that work quality becomes worse the more hours are worked, in particular comparing those toiling for 40 hours versus 55; plus, output falls sharply after 55 hours, with someone working 70 hours producing no more useful output in that time.
Others argue that five or six hours is the most any of us can really focus productively each day; add in a few breaks and meetings, and you’ll come naturally to that 40-hour work week. One writer, who studied Nobel Prize-winning authors as well as musicians and athletes, suggested “highly motivated” people can concentrate at top levels for around four hours a day, and that should be punctuated with breaks.
Oh, and working long hours increases the risk of illness by 35% and early death by 20%; another study from the WHO suggests that working 55 hours or more a week increases your risk of stroke by a third and death by heart disease by 17%, versus a normal work week.
Of course, it differs by person. Brin himself might be super productive at 60 hours a week, but it’s foolish to assume others can follow that lead.
If a team is working to crack a problem, it might make sense to work longer hours for a period of time to keep focus and avoid that challenge holding up other parts of the process.
But if you want people to work 60 hour weeks all the time, update their contracts and their pay and stock options – after all, you’re asking them to damage their health for you, so it’s effectively danger pay. Google’s parent firm Alphabet posted operating income above $100 billion last year, so it can afford a few performance-based bonuses and beefier health benefits for the inevitable outcomes.
And that means AI is going to need to actually prove itself useful one day to make up for all this overtime.
