What impact? None. I don't think many actual programmers take Musk seriously, despite his claims to possess expertise in software development, engineering, logistics, management, etc. A dilettante who takes credit for the work of other people remains a dilettante.
People with actual expertise and experience programming have talked and written about simplicity and complexity and managing it for decades. Brian Kernighan wrote "Controlling complexity is the essence of computer programming" back in 1976 (in Software Tools), and at the time he just expressed what programmers had discovered a long time before.
Listening to Elon Musk talk about the ideal manufacturing process as "A Genie turning your Raw Material into Rocket Engines"
...and then comparing that vs how much time and money Space X wastes hammer sheets of metal into shape
https://m.youtube.com/watch?v=t705r8ICkRw
I saw a tweet that a twitter home-page visit is 100 poorly-batched RPC calls, and I believe it
Not many bytes are necessary to deliver some text and images and a css layout
Technology markets are very interest rate sensitive, and there is a definite trend towards more complex, larger-team centric specialization during bull run periods as organizations pile on the people and resources and teams specialize as they scale. At the end of these periods is often a “clearing out” and “simplifying down” as teams downsize and need more efficiency and integration in their frameworks.
I would say that Elon gutting all the services in Twitter is not unprecedented. In fact it’s not that uncommon after and M&A for the previous management teams initiatives and technology platforms to be gutted.
Do not conflate Elon’s management style of needing to know the details on everything and throwing out oceans of services with the simplification trends that always come at the end of bull markets or the throwing out of the predecessors bull market bloated technology stacks.
Pretty standard stuff honestly — getting rid of old services and whatnot, it’s just being done very dramatically and in public.
I also suspect that many microservices shops can do without large parts of their technology stacks, as managing microservices is hard in large organizations and tends to have a lot of sprawl. You have to go in and clean house occasionally, and I imagine that’s also a factor.
Anyone who work(ed) at Twitter have perspective?
I’m doing my best to steer away from commentary on what’s going on because it’s not HN appropriate. I’ll just say my heart goes out to folks through all this.
The complexity is not recent and isn't accidental. And even a person who is as smart as Musk can not comprehend the true complexity of legacy systems that have resulted from many layers of acquisitions, technology cycles and just plain mismanagement that happens at any technology company that has been in existence for more than a few years.
Musk can not simplify Twitter's architecture, and he doesn't care to. All of this is theatre while he trims the fat. His companies are well-known as places with poor work cultures, especially for software people, who enjoy a wide variety of choices in where they work. He simply wants to be seen as doing things and garnering publicity while he searches for the next CEO for Tesla, Twitter, or both. One of his original reasons for buying Twitter, which is to reinstate Trump (and thus buy favor with the possible next President/Republican Party), has been accomplished now.
Jeff Bezos bought The Washington Post, Elon Musk is showing that he is infinitely cooler by buying Twitter.
what impact will that have on software development as a field
Ignoring the fact it's Elon and Twitter, because those are incredible devisive subjects, your question is really "What would happen if someone successfully implemented a massive change to simplify how a tech company works?" We're also going to have to assume that it does work out which will take anything up to a couple of years to play out. But hypotheticals are fun so I'll play along.
Basically nothing will change unless people can figure out how the change worked. You can't just cargo cult firing half your staff, driving another quarter out, and radically slash and burning your internal services. That would immediately end most companies. Twitter seems to have been doing a lot of "non-core" things that could be shut down without impacting the main product so it ould work there though.
I strongly suspect a business that did this will have a hard time growing afterwards. Recruiting people to a company that recently lost 75% of the head count would be difficult, so doing things that the current staff aren't capable of will be a challenge. That will probably lead to stagnation for a while if the business survives. In time, that will alienate investors and users.
tl;dr its a hell of a risky thing to copy. Some companies will be looking at Twitter and thinking they could do the same thing, and I think we'll see a few businesses try it and fail.
Side note: Something really important to remember in all of this is that most businesses don't employ people to babysit their products. They build their product to be resilient and able to run without many people. Babysitters are a waste of money if you can automate the process instead. Twitter appears to have done that pretty well. What that should tell you is that Twitter's staff were good at their jobs, and losing them hinders Twitter's ability to make more good, resilient products in the future. Twitter will find it harder to grow. To use a car analogy - it's a bit like saying Tesla could fire 75% of its staff because the Model S factory runs well. Sure, but how do they then make the next model?
7 comments
[ 3.1 ms ] story [ 33.0 ms ] threadPeople with actual expertise and experience programming have talked and written about simplicity and complexity and managing it for decades. Brian Kernighan wrote "Controlling complexity is the essence of computer programming" back in 1976 (in Software Tools), and at the time he just expressed what programmers had discovered a long time before.
I saw a tweet that a twitter home-page visit is 100 poorly-batched RPC calls, and I believe it Not many bytes are necessary to deliver some text and images and a css layout
To be fair though, if that's what you think Twitter is then you need to think about it some more.
Technology markets are very interest rate sensitive, and there is a definite trend towards more complex, larger-team centric specialization during bull run periods as organizations pile on the people and resources and teams specialize as they scale. At the end of these periods is often a “clearing out” and “simplifying down” as teams downsize and need more efficiency and integration in their frameworks.
I would say that Elon gutting all the services in Twitter is not unprecedented. In fact it’s not that uncommon after and M&A for the previous management teams initiatives and technology platforms to be gutted.
Do not conflate Elon’s management style of needing to know the details on everything and throwing out oceans of services with the simplification trends that always come at the end of bull markets or the throwing out of the predecessors bull market bloated technology stacks.
Pretty standard stuff honestly — getting rid of old services and whatnot, it’s just being done very dramatically and in public.
I also suspect that many microservices shops can do without large parts of their technology stacks, as managing microservices is hard in large organizations and tends to have a lot of sprawl. You have to go in and clean house occasionally, and I imagine that’s also a factor.
Anyone who work(ed) at Twitter have perspective?
I’m doing my best to steer away from commentary on what’s going on because it’s not HN appropriate. I’ll just say my heart goes out to folks through all this.
Musk can not simplify Twitter's architecture, and he doesn't care to. All of this is theatre while he trims the fat. His companies are well-known as places with poor work cultures, especially for software people, who enjoy a wide variety of choices in where they work. He simply wants to be seen as doing things and garnering publicity while he searches for the next CEO for Tesla, Twitter, or both. One of his original reasons for buying Twitter, which is to reinstate Trump (and thus buy favor with the possible next President/Republican Party), has been accomplished now.
Jeff Bezos bought The Washington Post, Elon Musk is showing that he is infinitely cooler by buying Twitter.
Ignoring the fact it's Elon and Twitter, because those are incredible devisive subjects, your question is really "What would happen if someone successfully implemented a massive change to simplify how a tech company works?" We're also going to have to assume that it does work out which will take anything up to a couple of years to play out. But hypotheticals are fun so I'll play along.
Basically nothing will change unless people can figure out how the change worked. You can't just cargo cult firing half your staff, driving another quarter out, and radically slash and burning your internal services. That would immediately end most companies. Twitter seems to have been doing a lot of "non-core" things that could be shut down without impacting the main product so it ould work there though.
I strongly suspect a business that did this will have a hard time growing afterwards. Recruiting people to a company that recently lost 75% of the head count would be difficult, so doing things that the current staff aren't capable of will be a challenge. That will probably lead to stagnation for a while if the business survives. In time, that will alienate investors and users.
tl;dr its a hell of a risky thing to copy. Some companies will be looking at Twitter and thinking they could do the same thing, and I think we'll see a few businesses try it and fail.
Side note: Something really important to remember in all of this is that most businesses don't employ people to babysit their products. They build their product to be resilient and able to run without many people. Babysitters are a waste of money if you can automate the process instead. Twitter appears to have done that pretty well. What that should tell you is that Twitter's staff were good at their jobs, and losing them hinders Twitter's ability to make more good, resilient products in the future. Twitter will find it harder to grow. To use a car analogy - it's a bit like saying Tesla could fire 75% of its staff because the Model S factory runs well. Sure, but how do they then make the next model?