I mean... this is very much not nice to the people on the ground, but let's face it, this is not healthcare or utilities or food production. Even if whole Meta burns itself to the ground I don't see too many people shedding tears for it 🤷♂️
Actually, WhatsApp is considered a utility in many parts of the world! The 2021 outage - when all Facebook services went down - it caused issues in the developing world: https://news.trust.org/item/20211005204816-qzjft/
So it does have impact: daily, about 50% of humanity uses one of Meta's services, so outages will be varying level of inconvenience and frustration.
There's a reason their core services rarely have SEV0s, and they historically invested so much in reliablity.
From a perspective I've always admired the balls they have going all in a direction. Have worked/know about a lot of companies whom have lost their moat because they never developed any risk appetite, instead they were clinging on what they had.
You cannot sustain your moat by clinging, or at least not to the same margin, which meta knows very well. However, if you invert the problem and ask how could you asap loose your moat, you could say that exactly by alianating the people getting you that mote. More specifically your employees and your customers.
Thank you. This is all bleak, from the min-maxing reviews to the wider industry psychosis.
> Bosworth, admitted to staff that the AI reorg was atrocious
From the Wired article, it looks like he only said they "did an atrocious job explaining the vision" and supporting employees through the transition, not an admission about the reorg strategy itself. It seems like the door is left open to double down and make deeper cuts:
> Bosworth also suggested that drafting people onto the AI team in the name of speed was the correct call ...
> ...he said, “We should heed the saying, ‘AI won’t take your job but someone who knows AI might.’” Employees’ performance will be based not on just using AI, but having an “impact” with it, Bosworth said.
Do you know more about the memo or reaction to it? Are people mollified by social events and micro kitchens when the core strategy is unchanged?
> The biggest problem: people stop caring about real work and focus on performative work.
Incisive. You've discussed performance reviews (and "impact") at Meta before, but not (as far as I know) so critically. Were the metrics truly serving the company well pre-AI and only recently have become "weaponized"? Or have they always had less than helpful effects?
One gets the sense that the appeal of AI workers isn't in employees that don't need sleep and sick days, as execs like to say, but in a fantasy of employees that work how Welch-brained managers have always imagined. You can evaluate them by metrics, totally surveil them, set quotas, motivate at will, brook no autonomy, allocate them fungibly by the man-month. It will come true, ignoring the massive waste and quality loss, and after all, why not, why shouldn't we treat humans the same way? That's the root of the psychosis.
You’re writing in the past tense, as if the result is settled. I feel like we see the downside of all the decisions completely here and now, but whatever is the desired upside, will take time to reap and reveal. That counterpoint is never taken into account in the article, surprisingly for your usually more well rounded analysis.
Zuck's track record is not great here with Metaverse and all.
Doing "obvious" things like switching to mobile and investing in ads worked well. Big bets didn't. Although now the heavy AI pivot feels "obviously right" to many people so yeah it might all work out in the end.
Actually, we don't know his track record based on this example. Metaverse and VR simply didn't catch on. If they did, we might have been saying "Oh Zuck was such a great visionary to point the whole company in the right direction and execute hard on it, and now they're dominating this new way of living"
> To me, it looks obvious that Zuckerberg doesn’t care how engineers feel about the massive
> changes, and that Bosworth likely ignored the chaos
But of course! And this didn't start two months ago. Zuck stopped listening to engineers some time around 2022. There was a lot of push back on overhiring first but he ignored it. Then came the Metaverse, out of nowhere. If he listened to engineers none of this would have happened. But he is the boss and he knows best. What happened recently is just the logical conclusion.
Bosworth became CTO but most engineers didn't like him. Essentially Zuck put his friend in the position and there was no real process to select a CTO as far as I am aware.
If they're cutting 3-5 people from each 10-people team, isn't this forcing the remaining organisation to move from "2 pizza teams" to "1 pizza team", which means they need to be more effective with AI, not just tokenmaxx?
I mean... this is very much not nice to the people on the ground, but let's face it, this is not healthcare or utilities or food production. Even if whole Meta burns itself to the ground I don't see too many people shedding tears for it 🤷♂️
Actually, WhatsApp is considered a utility in many parts of the world! The 2021 outage - when all Facebook services went down - it caused issues in the developing world: https://news.trust.org/item/20211005204816-qzjft/
So it does have impact: daily, about 50% of humanity uses one of Meta's services, so outages will be varying level of inconvenience and frustration.
There's a reason their core services rarely have SEV0s, and they historically invested so much in reliablity.
Good point! WhatsApp team was always famous for being tiny for their impact, wishing their boat the best of luck in the storm!
WhatsApp is the de facto OS of Latam. Here in Mexico telecom companies offer unlimited WhatsApp use for a low price.
From a perspective I've always admired the balls they have going all in a direction. Have worked/know about a lot of companies whom have lost their moat because they never developed any risk appetite, instead they were clinging on what they had.
You cannot sustain your moat by clinging, or at least not to the same margin, which meta knows very well. However, if you invert the problem and ask how could you asap loose your moat, you could say that exactly by alianating the people getting you that mote. More specifically your employees and your customers.
Thank you. This is all bleak, from the min-maxing reviews to the wider industry psychosis.
> Bosworth, admitted to staff that the AI reorg was atrocious
From the Wired article, it looks like he only said they "did an atrocious job explaining the vision" and supporting employees through the transition, not an admission about the reorg strategy itself. It seems like the door is left open to double down and make deeper cuts:
> Bosworth also suggested that drafting people onto the AI team in the name of speed was the correct call ...
> ...he said, “We should heed the saying, ‘AI won’t take your job but someone who knows AI might.’” Employees’ performance will be based not on just using AI, but having an “impact” with it, Bosworth said.
Do you know more about the memo or reaction to it? Are people mollified by social events and micro kitchens when the core strategy is unchanged?
> The biggest problem: people stop caring about real work and focus on performative work.
Incisive. You've discussed performance reviews (and "impact") at Meta before, but not (as far as I know) so critically. Were the metrics truly serving the company well pre-AI and only recently have become "weaponized"? Or have they always had less than helpful effects?
One gets the sense that the appeal of AI workers isn't in employees that don't need sleep and sick days, as execs like to say, but in a fantasy of employees that work how Welch-brained managers have always imagined. You can evaluate them by metrics, totally surveil them, set quotas, motivate at will, brook no autonomy, allocate them fungibly by the man-month. It will come true, ignoring the massive waste and quality loss, and after all, why not, why shouldn't we treat humans the same way? That's the root of the psychosis.
Good reporting about self-made disaster in progress.
There is an aspect of it that I haven't seen discussed at all.
Why Meta is so hellbent to have its own SOTA models?
What competitive advantage it has over other players in this space?
Or what advantage it would bring to its existing business that it can't realize by using 3rd party models?
You’re writing in the past tense, as if the result is settled. I feel like we see the downside of all the decisions completely here and now, but whatever is the desired upside, will take time to reap and reveal. That counterpoint is never taken into account in the article, surprisingly for your usually more well rounded analysis.
Zuck's track record is not great here with Metaverse and all.
Doing "obvious" things like switching to mobile and investing in ads worked well. Big bets didn't. Although now the heavy AI pivot feels "obviously right" to many people so yeah it might all work out in the end.
Actually, we don't know his track record based on this example. Metaverse and VR simply didn't catch on. If they did, we might have been saying "Oh Zuck was such a great visionary to point the whole company in the right direction and execute hard on it, and now they're dominating this new way of living"
We know it, it’s just short haha
> To me, it looks obvious that Zuckerberg doesn’t care how engineers feel about the massive
> changes, and that Bosworth likely ignored the chaos
But of course! And this didn't start two months ago. Zuck stopped listening to engineers some time around 2022. There was a lot of push back on overhiring first but he ignored it. Then came the Metaverse, out of nowhere. If he listened to engineers none of this would have happened. But he is the boss and he knows best. What happened recently is just the logical conclusion.
Bosworth became CTO but most engineers didn't like him. Essentially Zuck put his friend in the position and there was no real process to select a CTO as far as I am aware.
@Gergely I wonder with all these layoffs, why hasn't the software eng market been flooded with cheaper labor?
Why do you think it hasn't been? The market is pretty bad.
If they're cutting 3-5 people from each 10-people team, isn't this forcing the remaining organisation to move from "2 pizza teams" to "1 pizza team", which means they need to be more effective with AI, not just tokenmaxx?