At my FAANG, there was a team of experienced engineers that proved they could deliver faster and more performant code than a complete org that was responsible for it earlier.
So now a lot of different parts of the company are trying to replicate their workflow. The process is showing what works, you need to have AI first documentation (readme with one line for each file to help manage context), develop skills and steering docs for your codebase, code style, etc,. And it mostly works!
For me personally, it has drastically increased productivity. I can pick up something from our infinitely huge backlog, provide some context and let the agent go ham on fixing it while i do whatever other stuff is assigned to me.
As a counter point, at the bigTech i work at, since Trump's H1B visa fee announcement all H1B hiring requires approval from pretty high up in the management chain.
> There are also false prophets—we all want quick solutions, and sometimes we get (too) excited about people who promise them. For example, in Germany, some claim they have “made Microsoft’s cloud European” through SAP. However, this turns out to be not quite the case. If the U.S. refuses to cooperate, they (by their own admission) would be down within a few months. Amazon also claims they are going to do something similar, but they’ve been saying that for years, and they conveniently avoid addressing what happens in case of a conflict with the U.S.
An EU sovereign cloud by Amazon though would be built in the EU and staffed by EU employees though.
The effect of the US trying to insist it won't co-operate would be similar to what happened to most Russian based American franchise locations: troublesome but it's not like the facilities, people or inventory just vanished.
Make all tech knowledge available, spin it into another completely separate company with 0 ties to the parent and then we can actually start talking. Rest is fluff and PR.
> An EU sovereign cloud by Amazon though would be built in the EU and staffed by EU employees though.
At that point, why is Amazon needed? Don't we really have the expertise and capital to do this in Europe without Amazon (or Google, MSFT)? I somehow doubt that very much.
It won't work that way. If you look at the internal structure of amazon, you realize that majority of the good engineers have been there for 5+ years. Amazon always churned through people, causing good engineers who couldn't deal with the culture to leave within 2-3 years, and mediocre engineers getting pipped over 3-4 years.
It's mostly money. If you are good at your job, amazon pays much better than most companies. I tried looking for a new job last year, and the only ones increasing my current comp were HFT and pre-ipo startups. Google wouldn't even match my current comp.
In terms of the rest, only Netflix, meta, snowflake and roblox (why?) might have offered better, but the wlb in the first two is similar to amzn, and i didn't like the outlook of the latter two.
Every time I talk to a roblox recruiter it's something about how they have 70mm monthly recurring users and they're "building the platform to build games on" or something, but they're a total ghost in the mainstream media. I don't see the value proposition. Maybe they're the next "it" social media company as the users turn 16-21. Whatever they're doing, they pay full price for talent, allegedly.
Apparently kids LOVE the shit out of them... them and Fortnite. An assload of kids paying a couple (or couple dozen (and some Twitch/whatever streamers paying several hundred to a few thousand)) bucks a month adds up.
Based on what little I've seen of it, it all seems like budget Garry's Mod to me... which is something that I bet that kids these days have never heard of.
Microsoft is stuck because anything they'd do would reduce the "freedom" of end users.
I work in big tech, and unfortunately we frequently need to have conversations about the smallest features because we have evidence about us giving users an inch and they taking a mile.
It depends on if that aligns with your management's views. In aws, ive seen teams spend some effort on keeping their backlog low (but no zero) since if it balloons out of control you know you're gonna get called out on it.
i recall a team that had all its feature releases struck down for the coming year so they would work through their backlog.
What this exec did was very illegal, but anyone in the startup scene would know that things like this are very common, but probably on a smaller scale.
Maybe there are no wads of cash changing hands, but I've personally seen contracts being earned less due to the feature set, but more because the startup got introduced to some C level and them applying downward pressure to choose that startup during the vetting process.
It's kinda noticeable when you're on a call with a big company's tech team and they sound defeated when talking about the success criteria and stuff
I think this blog fails to take into account that 2021 is not 2002. Computer Science is a much more formal/mainstream field of study now, and people don't operate "in a vacuum with no mentorship, guidance, or reference points.”
Some comments on the previous blog post raised important questions regarding minimum understanding/knowledge of technology one utilizes as part of their day job. And I would agree that indexing is a fundamental aspect while using relational databases.
But unfortunately this isn't very uncommon these days, way too often have I heard people in $bigco say, let's use X, everyone uses X without completely understanding the implications/drawbacks/benefits of that choice.
Just looking at the raw numbers they all seem outrageous. But if you take a look at some of these companies it makes a lot more sense.
The FAANG I work for makes a PROFIT of $2 mill/year for every employee. So it's easy for them to think, lets pay this lot $300-600k and use em to keep our advantage/keep them away from competition. For all I care they could be paying all of us $1 mill and still be making a good sum of money.
So now a lot of different parts of the company are trying to replicate their workflow. The process is showing what works, you need to have AI first documentation (readme with one line for each file to help manage context), develop skills and steering docs for your codebase, code style, etc,. And it mostly works!
For me personally, it has drastically increased productivity. I can pick up something from our infinitely huge backlog, provide some context and let the agent go ham on fixing it while i do whatever other stuff is assigned to me.