I’ve said it before and I’ll say it again. Businesses operate on three timescales. Monthly, quarterly and annually. Anything beyond that they’re not even thinking about.
When a company talks about long term strategy you can ignore everything they say after. It means nothing.
So AI can make them more money in the short term. They don’t understand the problem even if it’s right in front of them.
From what I’ve noticed, with AI companies, it’s usually not “annual” but “annualized”, which is just an estimate. And it doesn’t seem like there are any rules to calculating that estimate, you can just grab the best 30-day period of operation, multiply it by 12 and claim millions in annualized revenue. And then they usually just sell to someone based on this estimation, since it’s almost never profitable to build something on top of existing infrastructure of the main players in this market and maintaining your own is just not feasible.
Anything beyond that they’re not even thinking about.
Every place I’ve had a white-collar office job, there’s been a 5-year plan. But the plan is typically a flippant and unrealistic goal, I suppose the purpose of which is to temporarily boost morale and to be a cult-like litmus test to see which employees can ignore reality.
Two of the more egregious examples: I worked in what was considered a medium sized company (maybe 30ish employees at the time). Management decided their 5-year goal was that we were going to grow large enough to own/build a skyscraper downtown. Another was at a relatively unknown regional retain chain that couldn’t even compete with Walmart whose 5 year goal was to become a household name around the world.
A lot of businesses operate at a 5-10 year time scale or more. Take all these AI datacenters being built. That takes a long time. You need to obtain planning permission from the local government. You need to buy the land, get all the utilities you’ll need to the site (they use a lot of water and electricity). You need to build the actual buildings. You need to put in an order for the computer hardware so it’s available when the building is done. You need to actually build the buildings.
Some of these can be done in parallel, but a lot can’t. You can’t lay out the electrical or cooling paths until the building is built. You can’t start installing racks until the electrical and cooling systems are in place. You can’t start installing servers until the racks are all ready. You can’t start connecting the servers and everything until all the electrical, cooling and networking setup is done. And, at every step there are going to be setbacks, especially when you’re building something new and innovative, and not just plopping down something you’ve built 100x before.
I’ve said it before and I’ll say it again. Businesses operate on three timescales. Monthly, quarterly and annually. Anything beyond that they’re not even thinking about.
When a company talks about long term strategy you can ignore everything they say after. It means nothing.
So AI can make them more money in the short term. They don’t understand the problem even if it’s right in front of them.
I feel this is always true for public companies, while there are some exceptions in private ones.
Part of it is that corporations have decentralized and siloed ethical responsibility. It’s a lot of greased cogs.
Yeah, I usually say “large organisations”. So it’s a good point.
From what I’ve noticed, with AI companies, it’s usually not “annual” but “annualized”, which is just an estimate. And it doesn’t seem like there are any rules to calculating that estimate, you can just grab the best 30-day period of operation, multiply it by 12 and claim millions in annualized revenue. And then they usually just sell to someone based on this estimation, since it’s almost never profitable to build something on top of existing infrastructure of the main players in this market and maintaining your own is just not feasible.
Every place I’ve had a white-collar office job, there’s been a 5-year plan. But the plan is typically a flippant and unrealistic goal, I suppose the purpose of which is to temporarily boost morale and to be a cult-like litmus test to see which employees can ignore reality.
Two of the more egregious examples: I worked in what was considered a medium sized company (maybe 30ish employees at the time). Management decided their 5-year goal was that we were going to grow large enough to own/build a skyscraper downtown. Another was at a relatively unknown regional retain chain that couldn’t even compete with Walmart whose 5 year goal was to become a household name around the world.
A lot of businesses operate at a 5-10 year time scale or more. Take all these AI datacenters being built. That takes a long time. You need to obtain planning permission from the local government. You need to buy the land, get all the utilities you’ll need to the site (they use a lot of water and electricity). You need to build the actual buildings. You need to put in an order for the computer hardware so it’s available when the building is done. You need to actually build the buildings.
Some of these can be done in parallel, but a lot can’t. You can’t lay out the electrical or cooling paths until the building is built. You can’t start installing racks until the electrical and cooling systems are in place. You can’t start installing servers until the racks are all ready. You can’t start connecting the servers and everything until all the electrical, cooling and networking setup is done. And, at every step there are going to be setbacks, especially when you’re building something new and innovative, and not just plopping down something you’ve built 100x before.
You are so correct. They will, in very short time now be more or less doing exactly what I suggested!