Superstar coders are raking it in. Others, not so much(economist.com) |
Superstar coders are raking it in. Others, not so much(economist.com) |
Firstly, there's the massive confounding factor of Covid, the stock market went crazy, companies went on crazy hiring sprees etc. and the tail of that bullwhip effect is clearly still putting downward pressure on hiring from organisations that overhired.
But secondly, are we seriously saying that in the last 2 years, relatively slow moving companies adopted AI LLMs to help coders, integrated them into our work flows, and saw the results of those productivity gains in business outcomes?
I think it's unlikely. I think it's much more likely that CEOs love to watch where the crowd is going and then run to the front and shout "follow me". You don't actually need to have productivity gains for shareholders to reward you for saying how this is going to boost your margins and cut your costs. And this is even more true for companies like Salesforce for whom "AI" is a product they're selling. Marc Benioff isn't actually doing "AI is great I'll fire all my engineers", he's saying "AI is great, come buy Salesforce's AI products!". As for Microsoft, they employee almost a quarter of a million people, laying off 6,000 is a drop in the ocean, that scale of layoffs happen frequently at companies the size of Microsoft.
It's very much more just vibes than real data driving this. "<CEO who sells product> says product will cure cancer".
The underlying truth is that the competitive environment hasn't changed, if you can hire fewer engineers to do the same job, great, but your competitors are going to hire more engineers and out-compete you.
The more accurate frame is that employees produce more value than their costs, so each employee is actually a profit producer. If you can increase their efficiency, then they produce more profit. Firing profit producers decreases your profit, not increases it.
"if you can hire fewer engineers to do the same job, great, but your competitors are going to hire more engineers and out-compete you." - this is correct in aggregate across the market. Of course any individual company may have other constraints that makes hiring additional people unviable. I think this is the common mistake, its easy to look at an individual company and believe that the constraint applies to all companies simultaneously.
I've never worked for a company with thousands of devs but I imagine at a certain point the cost of simply collaborating and managing those devs becomes pretty high too.
Open source is hard to steal. It's intended to be taken for free with some license limitations. Second, most giants are already dead. Their work was used for free and that was/is considered normal. Third, many 'AI-whizzes' contributed to open source for long time.
In any case AI will be used on open source. For example to find bugs and backdoors at scale.
Moreover, many "AI" open source code bases are a mess and probably wouldn't even contribute in a positive manner to training.
the top of the pay scale were elite ai labs such as Openai, as well as hedge funds such as Jane Street that are also betting heavily on machine learning. In this tier, median pay exceeded $400,000 a year. Below that were tech giants including Alphabet, Microsoft and, until recently, Meta, where median pay was closer to $300,000. Experienced developers at most other companies earn much less. Their median was around $180,000 (see chart 2).
—-
A median of 180K is mostly definitely raking it in compared to the median of all US employees.
I’m well over double the household median income in my metro area and while I don’t feel like I’m raking it in, I guess I am when compared to others.
This just seems like a silly article.
$180k/year gives you $118,970 after tax
$300k/year gives you $186,880 after tax
I.e. the gross jump from $180k -> $300k is 67%, but the net jump is 57%. You have to increase the gross more to attract people, because take-home pay is what matters to them when switching job.
[0] https://www.mypaycalculator.net/us-paycheck-calculator/calif...
It's certainly true by comparing "senior" at G/Amazon/Apple to Microsoft, but is there level skew that compensates for this?
We’re just writing dumb gRPC services that use Postgres. I work probably 30 hours a week and still get awards, bonuses etc.
So it would be the exceptional case that firing employees because they were suddenly more productive makes sense.
What's more likely is they were already overhired and the new found "efficiency" is the excuse to resize.
Edit: btw this comes up IRL for me all the time in my work. Companies default play is to cut expenses to hit desired margin. It's much more beneficial in most cases to keep the team but increase their efficiency and handle more revenue. There are diminishing returns to profit by cutting back.