Mc91

Tired of recruiters asking me if I had a BSCS, I went back to college after years working in IT.

Had a conversation with the other seniors in a CS 300 class one day, I would say a few of them were in the top half of the class. Was talking about software version control. "What's that?" they all said. "You know, like git", I said. "Git?" they said. None of them knew what software version control or git was.

I would suggest that if you have a class where you learn say C++, toward the end of the class, or right after the class, write your own C++ program - doing whatever you want it to do. If you take a graphics class where you learn OpenGL, write your own program using OpenGL - doing whatever you want it to do. You tend to learn these things more doing it.

A lot of people are flooding open source projects with AI right now, but that aside, also if you just learned C++ or Java or whatever, look at actively maintained open source projects on Github or Ubuntu or Debian or freedesktop.org or whatever. Look through bug reports. Try to reproduce them. If you reproduce, you can reply to the report with what versions you are using (latest is best usually) and that you reproduced it. You can try to fix the bugs too. One thing with this is - it's similar to what an intern or associate/junior SWE would be doing at a company. If you submit a pull request and it gets accepted - it is like what you are doing at a big company.

Years ago I went to a Debian bug squashing party. I didn't know if I'd know anyone there, but getting there I did know one guy who did IT work in my city. Anyhow, a lot of the people there were pretty good. I remember a fellow named Daniel talking and he seemed to know a lot. Plus others. Then someone else walked in and everyone started saying "Pablo is here". It was a fellow named Pablo who was doing a doctorate in computer science at Columbia, They were all looking toward him and curious what he was working on. It seemed like they thought he was one of the most knowledgeable person there, among a group of knowledgeable people.

The point of this is - it was a group of people who were pursuing technical knowledge, and who respected technical knowledge. This guy Pablo from what they said was very devoted to this, and they respected him for it. This kind of social reinforcement is important too - it's talked about all the time, how being in San Francisco or YC or whatever with like-minded people pushes people ahead. I have seen it in many circumstances. Being part of one or more social groups like this focused on getting technically good and eventually trying to get to be a SSWE in a FAANG or whatever is something that helps. It can be easier to find in some places than others, but there are hackers everywhere.

My work status hasn't changed since 2022 when I got all those recruit attempts. I log in at least once a week, thumbs up people's accomplishments and sometimes reply "Congratulations", same as I always did.

Other people I talk to said there has been a dropoff as well.

I am a programmer in the US. Linkedin mails to me from recruiters -

2026 - one (recruiter did not name the company, hybrid six month contract in another city)

2025 - none

2024 - none

2023 - four (none after March)

2022 - twenty-seven (including Amazon, Google)

I remember Gnucleus and followed the release of Gnutella and used to read the old Gnutella Developer Forum on Yahoo where the major Gnutella clones hashed out the next steps for the network.

I agree - competition from BitTorrent lowered Gnutella's popularity.

Another thing about BitTorrent - I go to the Pirate Bay right now and look at top 100 for music, and "Pink Floyd - Discography 1967-2014" is one of the top torrents - 2.86 gigs, 131 seen seeds, 10 seen partials (seeds means they have 100% of the blob). You can download all their albums if you want, or one album, or one song on one album. Also most bittorrent clients have anti-leech constraints, so someone might start out accepting and sharing any part of that blob, making the one song they want a high priority request for them - they get the song they want quicker but they're also contributing to the torrent while they're online. So this sort of thing has all kinds of benefits for the network.

Whereas Gnutella tends to be look for and get one song at a time.

You are looking at this from your perspective, which is either pass or fail.

From their perspective, they might be interviewing, say, six people. As you say, they've already weeded out people from their resumes before they even get to the interview. From my experience, and I have heard people note this before, interviews tend to be a Gaussian curve with a normal distribution. People are weeded out by resumes and such. However, if I interview six people, usually someone slips through the cracks who knows nothing or next to nothing. That leaves the remaining five.

Of the five, four are usually interchangeable. They're like you - they get the technical questions right, or right enough. It's obvious they've been writing features for code for a company like yours. But of the five, often one person seems to not just know the easiest questions, or normal questions, but has a very good understanding of the subject matter. You keep probing how much they know, and they have in-depth knowledge about a lot of things. They know how registers on a processor work, they know about cache, they know the big O space of various algorithms, they can explain different approaches to concurrency in depth, or testing, or a lot of things. So you got the answers right, they just did better.

It could be something else - you might be just as good as someone else, but they were recommended by someone already on the team, or on an adjacent team, and they get brought in.

I guess personalities are on a bell curve as well. Maybe one out of six people fail on this. Maybe they're disorganized, or immature, or arrogant. Sometimes they miss basic social cues, or don't follow instructions, or even seem like they have a screw loose. Then four out of six people seem nice enough - professional but friendly. Then maybe one out of six just seems very sharp and smart, or avuncular, or what have you. A lot of it ties together - someone who has done the work to learn a programming language more than the other candidates, you assume is going to be hard-working on features as well, and they also seem sharp because they know so much (about IT, but other things as well).

Some things are contrasting. The hard-working person who knows the programming language in and out, and who gets a lot of feature work done is probably willing to sacrifice a little comity within the group to get a feature out. On the other hand, some people are so stubborn and argumentative, their presence would be a negative, even if they have technical skills. But some personality traits can contrast - I've working with friendly, supportive leads with great technical skills, but if they are a little bit hard charging this type of thing might be expected to come with the package of being very good technically.

I program at a non-tech Fortune 100 company. Our team is on a pilot program to try out AI-assisted programming at the company, and Cursor with OpenAI models are mostly what we are using. I have it integrated into my standard IDE workflow and try to write unit tests and the like with it.

I got a gig from who's hiring a few years ago. I also interviewed at some other places which were not a fit. Some places never got back to me. It has been a mix.

I went to a state school (cheap) and took one class a semester - usually at night, sometimes on a weekend. So I'd work until 5, then take a class from say 7 to 830 every Tuesday and Thursday.

In 2009 the economy sank and I was laid off. I was able to go back full time for a while. I was already enrolled and everything, I just started taking 4-5 classes a semester instead of 1, I was also able to take them during the day for a while.

$80k over four years is a certain amount, $80k over more years is less. I think I paid less per than $20,000 for a year's worth of classes. Some people got Pell grants and financial aid, although I just paid for it cash.

One thing I do is go to Leetcode, see the optimal big O time and space solutions, then give the LLM the Leetcode medium/hard problem, and limit it to the optimal big O time/space solution and suggest the method (bidirectional BFS). I ask for the solution in some fairly mainstream modern language (although not Javascript, Java or Python). I also say to do it as compact as possible. Sometimes I reiterate that.

It's just a function usually, but it does not always compile. I'd set this as a low bar for programming. We haven't even gotten into classes, architecture, badly-defined specifications and so on.

LLMs are useful for programming, but I'd want them to clear this low hurdle first.

I have been working in IT for over 30 years. What is happening is not new. Late 1999 was a very go-go time, early 2022 was a very go-go time. Alternatively, things were dead in 1991, in 2001, in 2009. Things were briefly dead in some ways for some people in spring/summer 2020 when Covid hit. 2022 went from go-go in the spring and summer to massive FAANG layoffs in November. Massive FAANG layoffs continue into early 2023, and things have kind of been stagnant since them. Things seem to have gotten worse at the beginning of this year, although it varies, some people with certain AI-related skills are doing well.

In 2000-2001, dot-com startups were hit harder than the rest of the economy. I worked for Internet startups and dot-coms from 1996 until the end of the summer 2000, where I started consulting for a large investment bank. I figured Internet-related startups were not going to make a comeback in the short term and I was right. The Fortune 500 was kind of starved for technical talent at the time, especially outside the Bay Area, so you could shift. There were some difficulties getting hired - I knew a lot about Red Hat Linux and Apache web servers and Java application servers, and I moved into a world of Solaris e4500 servers and NFS mounts and RAID 10 arrays and middleware. There were later shifts - for backend, things began shifting from monoliths and SOAs to microservices. Ruby on Rails was big from 2007 until 2013 until Javascript web front end began picking up more. Then native/hybrid mobile began cutting into the dominance of web front end. Now AI is coming in. So there are economic ups and downs, but what skills they are hiring for shift as well.

Unless society is entering some major transformative period like the 1930s, these shifts of the business cycle will keep happening. While the general tech market has been stagnant since November 2022, Nvidia stock has gone up 800%, as it has gone from the 15th most valuable company in the world (by market cap) to the 2nd, behind Apple. I have a strong feeling it will surpass Apple in the coming months and years as the most valuable company in the world. Programmers programming CUDA for them and whatnot, programmers programming Pytorch for FAANG and AI startups, and these kinds of jobs are open now, and in a few years companies might be offering $200k TC to people coming out of college who can program that. Or maybe LLMs will hit a wall in the short term and that won't happen. But something will happen - I've seen IT hit a slump a bunch of times and it always comes back (unless, as I said, we get into a situation like the 1930s).

> Sometimes that means making dumb business decisions like sacrificing an evening to a company that doesn't care, but IMO that sort of thing is worth it now and then.

I sacrifice an evening - but not to my company, but to studying Leetcode to move on to the next company. I also have side hustles that I devote time to.

> when layoffs come your next job often comes from contacts that you built up from the current job, or jobs before. If people know you are a standout contributor then you will be hired quickly into desirable roles. If people think you are a hired gun who only does the bare minimum that next role will be harder to find.

I am helpful to most people when they need help, and they remember this. My code is clean and well architected and well tested, and they can see this too. They also know that I know the language and platform we're using, and general programming (and business) knowledge. Few care whether I'm a "standout contributor" in terms of getting many stories done. Actually if I have a good lead or manager I might go above and beyond for them in terms of doing more.

> a company will never love you back. But your co-workers will.

Well, this is correct. I help my co-workers.

Things are situational. If I got a job helping set up LLM's or something, I might dive in and work a lot of hours just because I feel it is benefiting me too. On the other hand I can be somewhere where it doesn't make sense to work more than forty hours (if that) a week.

We say grind out Leetcode, because if someone can't do a Leetcode medium (or hard) like finding a cycle in a linked list in a few minutes, it diminishes their chances of getting into FAANG. It might be the bare minimum nowadays, but if they can't do even that, it is where they should start.

We're not telling people to grind Leetcode because we think Leetcode is great, we're telling people to grind Leetcode because this is usually one of the steps you need to do to get in.

> it’s much more realistic to start a business that is going to convince enough people to pay you enough to support yourself.

I only have to convince one place - Google Ads. Plus bring in the "eyeballs" with my free app, but I have accomplished that more than once.

> Also you have to convince companies to do business with you instead of a well known company.

Just one company in my case (actually several, but 90+% of the money comes from Google)

> Oh and to be competitive you need to have some type of funding. I have to be competitive enough to make a few thousand a month, and with my programming (and database design, and UX, and SRE etc.) skills, I have achieved that.

> And you need to make enough to pay for health care. In the US you do.

Corollary: This role can also be filled by someone of Asian heritage who was raised in the US, and who speaks English very well, as well as some Asian language proficiently.

I live in an apartment building and was given a fob that opens my apartment garage door and doors in the building. The fob takes specific lithium coin batteries and lately has been acting worse, I practically have to replace the batteries every day to use it. I had my Flipper Zero read the signals (the fob does one for garage door, one for the other doors) and I now use that to get into my building. Instead of having to buy two new lithium coin batteries every day, I just have to charge my Flipper Zero every few weeks. I heard the Flipper Zero had other uses, and I did scan some other things with it, but I got it in the hopes that it could handle my garage door, and it does.