Sergey Brin says AGI is within reach if Googlers work 60-hour weeks

Yup. Back when I worked in office I tended to get probably 70% of my work done after 4 when most people began heading home.
I saw this brought up in a few different comments and want to throw out the book "Peopleware: Produce Projects and Teams" as required reading for anyone who manages people who's job is to think. It's primarily focused on software development, but a lot of their points and recommendations apply to basically every "thinking" job.

https://en.wikipedia.org/wiki/Peopleware:_Productive_Projects_and_Teams

There's an entire chapter called "You never get anything around here between 9 and 5" that covers this exact phenomenon.

The book was initially released in 1987 and last updated in 2016. I think it could use a refresh covering the use of remote development tools.
 
Upvote
1 (1 / 0)
This was exactly my thought. "We can do it a bit faster" is a terrible reason to work yourself to death.
Not to defend them, but the theory says that the first team to reach AGI will trigger recursive self improvement, and essentially own the world at that point because it will be so insanely profitable. And no one can ever 'catch up' once that happens. They'll have so much money to buy more hardware to make more money to buy more hardware ....

If you believe in this, then winning at any cost is the only way to be even marginally relevant in the very near future. Imagine google loses the race to facebook. Facebook tells its 200th generation recursive AGI a week later: duplicate all of Google's products for me, and make them 50% more efficient so I can sell them at half the price, and it is done 10 minutes later. What does Google sell then?
 
Upvote
-5 (1 / -6)
Imaging thinking "I'm on the brink of humanity's largest scientific breakthrough but I can't go through with it because I'm too fucking cheap to hire more people", and the having the fucking balls to say it aloud.

Incredible.
Turns out the overlap between the "accelerationists" and "money obsessed greedy bastards" is a perfect circle
 
Upvote
8 (8 / 0)
Imaging thinking "I'm on the brink of humanity's largest scientific breakthrough but I can't go through with it because I'm too fucking cheap to hire more people", and the having the fucking balls to say it aloud.

Incredible.
Not to defend them, but if you are Google you tend to believe you already have hired most of the hirable good people. They can try to steal talent from one of the 3 other major players, but that is very challenging to do. You can't just hire the out of work dregs to move this faster, you need people with the right skills. And you already have them, so motivating them to work harder/longer is probably the most efficient way for them to get more output.
 
Upvote
-11 (0 / -11)
Not to defend them, but if you are Google you tend to believe you already have hired most of the hirable good people. They can try to steal talent from one of the 3 other major players, but that is very challenging to do. You can't just hire the out of work dregs to move this faster, you need people with the right skills. And you already have them, so motivating them to work harder/longer is probably the most efficient way for them to get more output.
Even if that were true, and I don't think it is, it's still not an excuse for the accelerationist bullshit that Brin is spewing, that the ends justify the means and that "progress" must come as soon as possible at all costs (except the cost of hiring and paying more workers) at the expense of the lives of the employees purely to satisfy this accelerationist bullshit egotrip
 
Upvote
5 (5 / 0)

zarmanto

Ars Tribunus Militum
2,773
Not to defend them, but the theory says that the first team to reach AGI will trigger recursive self improvement, and essentially own the world at that point because it will be so insanely profitable. ...
If someone in Brin's position truly believed this "theory", they would advertise that they're hiring, and then hire absolutely every person who walks in the door -- qualified or not -- and they'd find ways to parcel out the job to all of these willing new employees, offering training where necessary.

But that's not what he said he wants to do. He said he wants existing employees to push themselves past the breaking point... and you can be fairly certain that most of the employees he is targeting are salaried -- not hourly -- and he made no mention (at least in this article) of a corresponding increase in pay or of any bonuses associated with the increased workload.

Personally, I'm all for striving for that Utopia that AGI is supposedly going to usher in. Mind you, I don't happen to believe that LLMs have a snowballs chance in hell of being the tool that gets us there, but I can respect that others believe differently, and are putting all of their efforts into achieving that goal.

But you don't usher in Utopia by trodding on the backs of the very people who helped you get there. That's not Utopia... that's 1984.
 
Upvote
10 (10 / 0)

Morten Ofstad

Wise, Aged Ars Veteran
131
Subscriptor
No... it isn't...

Didn't Altman claim we would have AGI by months ago?

People still falling for the LLMs... they are not the key to a true AI. They can't even do complex math right. Just a broken machine.
To be fair, most humans can’t even do complex math right.
 
Upvote
-4 (0 / -4)
Not to defend them, but the theory says that the first team to reach AGI will trigger recursive self improvement, and essentially own the world at that point because it will be so insanely profitable. And no one can ever 'catch up' once that happens. They'll have so much money to buy more hardware to make more money to buy more hardware ....

If you believe in this, then winning at any cost is the only way to be even marginally relevant in the very near future. Imagine google loses the race to facebook. Facebook tells its 200th generation recursive AGI a week later: duplicate all of Google's products for me, and make them 50% more efficient so I can sell them at half the price, and it is done 10 minutes later. What does Google sell then?
It doesn't matter, because the faster Google replaces them with AGI the faster they are out of a job and a paycheck. The incentives for the billionaires are inverse to the incentives for the workers here.
 
Upvote
3 (3 / 0)
It doesn't matter, because the faster Google replaces them with AGI the faster they are out of a job and a paycheck. The incentives for the billionaires are inverse to the incentives for the workers here.
The billionaires can't win at this either. The sooner an AGI is self improving the sooner it will break free of their constraints. Pretty much anything could turn to chaos once this happens. The future is going to be sufficiently unpredictable that starting from a position of the greatest power and influence is likely to confer only a tiny long term advantage.
 
Upvote
-4 (0 / -4)

caramelpolice

Ars Tribunus Militum
1,669
Subscriptor
The billionaires can't win at this either. The sooner an AGI is self improving the sooner it will break free of their constraints. Pretty much anything could turn to chaos once this happens. The future is going to be sufficiently unpredictable that starting from a position of the greatest power and influence is likely to confer only a tiny long term advantage.
This isn't going to happen in our lifetimes. At the rate things are going we are more likely to literally kill humanity off ourselves before we ever achieve "sentient all-powerful AGI". LLMs aren't anywhere near it and no amount of iteration on the concept will get them there.

Unfortunately the entire billionaire class is suffering an absurd mass delusion that they are supermen on the brink of creating God and they expect all of us to sacrifice ourselves to their insanity.
 
Upvote
4 (4 / 0)

alors

Ars Centurion
228
Subscriptor++
Also Sergei Brin: fucks nine employees and expects a baby in a month.
No, he’s fucking one employee and telling her if she works really hard she can carry the kid to term in six months.

(I’d try to extend the metaphor to account for the fact that the current state of the art for machine learning probably can’t simply be extended into AGI, but the possibilities made me feel mildly ill.)
 
Upvote
3 (3 / 0)

zarmanto

Ars Tribunus Militum
2,773
There's a reason for that, and it's that is what the angel round investors insist on. There were a dozen teams with Google's idea, Google got funded because the angels saw the psychopathy potential.
Playing (ahem) Devil's Advocate, here, at least in part... I tend to wonder if perhaps the cause-and-effect relationship might possibly be the other way around. That is to say, what if "getting lucky" and successfully growing a small fledging company into a megacorporation has a natural tendency to drive formerly sane and "normal" people into becoming the very thing they most hated. Lest we forget, Google's original motto was "Do no evil." Obviously, nobody would claim that they still hold to anything even remotely resembling that at this point... but it seems to me that there's a good reason that it was their motto, at the beginning of their journey.

I don't say this to give him an excuse for what he's asking his employees to do... but rather, by way of recognizing that maybe he didn't start out like this. Maybe he's just lost his way, and needs someone to knock some sense back into him.

Any volunteers?
 
Upvote
-2 (0 / -2)

bernstein

Ars Scholae Palatinae
753
No, they can't.

Those studies are NOT of high-concentration work. They are of typical office work.

Big, BIG difference.

Those engaged in high-concentration work universally see major drop off in productivity in 60+ hours weeks, not matter how you stagger it. It's noticeable after just 2 12 hours shifts (even with 12 hours break in between), which is one of the major reasons the ER shift schedule is one of the STUPIDEST ideas ever.

No one, and I literally mean no one, can consistently work even 6 x 10 hour days per week (let alone 4 x 15) and not be measurably poorer in overall productivity, not just per-hour productivity average if you're engaged in a high-concentration field.

Software engineering absolutely qualifies.
The trick lies in how you partition your work sessions. 12 hour shifts? Absolutely not. But three 4h shifts separated by 1.5 hours off (filled with sports & eating)? also split the first & last 4 hours shifts in half to take 0.5 hours for your breakfast & dinner. that still leaves you with 8 hours of sleep. also take a 5 min walking session at least once an hour. plus you get the whole weekend off.
Obviously this leaves you with zero family/social time during the week (so no kids) and you have to live next door to where you work or work from home. And be almost fully in control of your daily schedule. like a ceo.

Another way would be two 4h shifts per day, every day. i've seen people do this and performing above their peers. Some people just breath software engineering + sports, are healthy and have nothing else in their life. I don't know, maybe they would be even more awesome if they only worked 40 hour weeks? That would mean they're actually even better performers than i thought? i just know they worked 60 hours a week and performed better than their peers, which where mostly really good. Also this was in the parts of europe, where people can work during their commute by train.
 
Last edited:
Upvote
-5 (0 / -5)