Those are raw numbers. I would look instead at the job changes over total employment numbers. I don't have the numbers but I would wager we have many more people working in tech today (overall) than we did in 2008.
Also, that spike in 21/22 really did a number on people's expectations. The one constant in this industry is its cyclical nature.
Maybe I'm reading the graph wrong, but the decrease comes after years on continuous growth, so total employment numbers in tech should still be absolutely massive, compared to 18 years ago?
If it continues, then yes it could be bad, but so far it seems like a correction for over-hiring in 2021 - 2023. Seems a little weird to be focusing on a decline in 2024 - 2026, without addressing the large increase right in the years before.
There's a lot of dynamics where it's the short-term numbers that matter. If you're a developer who needs a new job after your spouse got transferred to LA or something, it does you no good that the absolute numbers are massive, nor that a different person looking for a job 3 years ago would have found it uncommonly easy.
I had no idea I was in such an exclusive group back in 2000. Everyone I knew was a software engineer or in tech one way or another so I suppose I got a warped sense that I belonged to a larger group.
I'm not sure the nation wide raw statistics are that reliable in the field of software engineering without interpretation.
In the 90s tons of people who were de facto software engineers were listed as "Information Technology Workers". I suspect a lot of that still hasn't been shaken out of the system.
According to the BLS in the year 2000 there were 3.4 million information technology workers.
BLS had some classification changes over the years. I think it's interesting in the "this is how people thought about the role over the decades."
Today there are computer programmers (15-1251), and software developers (15-1252), and web developers (15-1254).
In 2018, there was a reclassification - https://www.dol.gov/sites/dolgov/files/ETA/oflc/Presentation... where 15-1132, Software Developers, Applications and 15-1133, Software Developers, Systems
Software where reclassified into the software developers (15-1252) group.
The other thing that confuses this is that a lot of positions were classified as Computer systems analysts because that's a position that a TN visa can be hired for (there is no software engineer in there... and it wasn't until relatively recently that one could be a "software engineer" in Canada without being an Engineer.
Computer programmers 1010 15-1131
Software developers, applications and systems software 1020 15-1132, 15-1133
Where the "Computer programmer" was the more junior classification and Software developers working on a word processor were classified differently than a software developer working on the operating system... and they were the more senior positions.
Software Developers
Research, design, and develop computer and network software or specialized utility programs. Analyze user needs and develop software solutions, applying principles and techniques of computer science, engineering, and mathematical analysis. Update software or enhance existing software capabilities. May work with computer hardware engineers to integrate hardware and software systems, and develop specifications and performance requirements. May maintain databases within an application area, working individually or coordinating database development as part of a team.
Computer Programmer
Create, modify, and test the code and scripts that allow computer applications to run. Work from specifications drawn up by software and web developers or other individuals. May develop and write computer programs to store, locate, and retrieve specific documents, data, and information.
I think the idea that if you don't support white supremacy you should get off the site owned and run by a clear white supremacist applies regardless of how elections go.
I recommend _Culture in Nazi Germany_ by Michael H Kater. [0] It is very dry but goes into detail of the culture of the era from late 1920s to end of WWII.
One aspect he highlights at the end is that Fascism was not rejected by the current and former citizens, those that migrated, of Germany. In their mind it was incorrectly implemented. A number of Zionist that migrated from Germany to Palestine were supporters of Fascism. It was not until mid to late 1960s when people start realize and admitted Fascism was bad.
I personally will never fund Elon Musk. Anyone that says empathy is bad is a bad person at heart. Empathy is intelligence and those that lack it lack strong intelligence. There is no way to put yourself in the position others have gone through without empathy.
I recommend _Culture in Nazi Germany_ by Michel H Kater [0] for the fact it examines the complexity of reality during this time. He actually went and talked to the musicians, actors, and writers of the era to have a better understanding of the culture and view points that the people still held after the war. I will take his expertise over those of the modern era that have not engaged with these people that lived through it.
People scoff at the idea _Jews for Hitler_. Reality is that a number of Jewish Germany actually supported Hitler and Fascism in the 1920-1930s and latter. It latches onto the idea that that modern day people would be able to pick out Fascism and reject it ... which has shown to be the opposite with populism.
By the way, I consumed _Mein Kampf_ by Adolf Hilter. It was not to align with his ideology but to understand it and have an independent reference to _like Hilter_ that politicians, the media, and pop culture use, and as a base understand of Fascism. I in fact reject Fascism, Nazism, and Adolf Hitler.
Example is that Hilter's take over of Europe was under the disguise that resources are ours for the taking and are needed to support the country. He states this in his book. This is the same statement that Donald Trump and his executive branch uses against Greenland. Yet history proved Hilter wrong. Technology is what drives the economy not direct resource access. Japan, USA, South Korea proved this after WWII.
He didn't win a majority of the vote, just a plurality. And less than 2 of 3 eligible voters actually voted. So he got about 30% of the eligible population to vote for "yay grievance hate politics!" Which is way more than it should be, but a relatively small minority compared to the voter response after all ambiguity about the hate disappeared. This is why there's been a 20+ point swing in special election outcomes since Trump started implementing all the incompetent corrupt racist asshatery.
A 2025 study... Asking people if they "would have" voted for the winner of the election, a corrupt vindictive racist asshat already in power? Well, I guess that's one way to conduct a study. Fortunately the shift in sentiment is clear, growing, and reflected in special elections.
Your theory is that people who didn't care enough to vote are concerned that Donald Trump is going to come after them if they don't say they would have voted for him, when surveyed anonymously?
And then NPR was duped into credulously reporting on this polling?
I'm saying it doesn't take much for someone to say, "yeah, I would have voted for the guy already in power". I'm surprised it wasn't much higher than that.
So no, you definitely misrepresented my theory. It doesn't take a specific threat of violence for someone to say "sure, I would have cast a vote with the winner." And yet it was only ~1.5% higher than before the election. Are you saying you don't even recognize the bias of saying "yeah, I'm good with the winner"? Or the bias of a honeymoon period? I mean, June 2025 was before 90% of his craziest shit. But you go on.
Oh sorry, you made it sound like "corrupt" and "vindictive" were somehow relevant to the polling results.
The media seemed pretty surprised by the results, which indicates that your hypothesis is perhaps not accurate. But hey, keep doubling down, moving the goalposts, etc. I'll leave you to it.
Nah, just an observation. Or my hypothesis is accurate and they were just taking it at face value like you apparently did (assuming you are posting in good faith). The click bait appeal couldn't have hurt (although I agree with your expectation that they don't usually go for that). But dippy did pull their funding after all.
My goalposts never moved. Sorry you misinterpreted a few accurate adjectives.
I didn't get it either until I trained the algorithm to feed me what I want by just clicking the three dots and selecting Not Interested on anything I never wanted to see again... it listens, whats left is really unmatched anywhere, I've really looked, and occasionally still do out of curiosity.
Lots of info is shared there first. It shows up in news articles and podcasts 12-24 hours later. Not everything shared there is true, of course, so one has to do diligence. But it definitely surfaces content that wouldn't show up if I just read the top 2-3 news websites.
Publishing is more than just authoring. You have research, drafts, edits, source verification, voice, formatting, multiple edits for different platforms and mediums. Each one of those steps could be done by AI. It's not a single-shot process.
Not quite systems programming but this might give you some insight. Swift is memory efficient, and runs stable backend services. I've seen benchmarks showing that it's slightly more performant than typescript but twice as memory efficient (but not as efficient when it comes to memory management compared to Rust, C, and C++).
The other point I've seen is that its string library is slow and very accurate.
Besides that, the C-interop means you have quite a bit of flexibility in leveraging existing libraries.
>The other point I've seen is that its string library is slow and very accurate.
Swift strings default to operating on grapheme clusters, which is relatively slow. But you can always choose to work with the underlying UTF-8 representation or with unicode scalars, which is fast.
The only situation where UTF-8 incurs overhead is if the String comes from some old Objective-C API that gives you a UTF-16 encoded String.
Unicode scalars are not so fast. But yes working directly with UInt8/Data bytes is efficient.
That’s how I took over maintenance of SwiftSoup and made it over 10x faster according to my benchmarks. Besides various other optimizations such as reducing copying, adding indexes, etc.
Being only slightly more performant than an interpreted GC-only language is hard to believe (even though typescript is just a spec and you probably meant slightly more performant than v8).
I asked Claude to summarize the article and it was blocked haha. Fortunately, I have the Claude plugin in chrome installed and it used the plugin to read the contents of the page.
Codex works by repeatedly sending a growing prompt to the model, executing any tool calls it requests, appending the results, and repeating until the model returns a text response
Engineers are happy to pay for tools (hello, Claude Code). Libraries are quite different and it's a little uncomfortable to build a business on a closed-source, proprietary library.
Also, that spike in 21/22 really did a number on people's expectations. The one constant in this industry is its cyclical nature.
reply