Tech is a rare industry where you can easily pivot into it later in your career, start at an entry level position, work your way very high up the ladder and get paid very well along the way.
I know folks without a college degree that started as tech support reps who worked their way up to senior roles in big tech that make multiples more than most physicians I know or others that had to go down a rigid path to get to where they are.
The industry favors eager people who are hungry to continually learn and grow. While there can be short term bumps in the road, those that fit these characteristics have and will continue to do extremely well.
"The industry favors eager people who are hungry to continually learn and grow. While there can be short term bumps in the road, those that fit these characteristics have and will continue to do extremely well."
That wasn't my experience.
The people you describe making so much money are statistical outliers who got lucky.
I'd love to live in whatever bubble you do. All I'd have to do to make stupid level of money is be myself (enjoy learning and growing) and it would easily happen...
Relative to other fields. Compare it to wanting to switch careers into the military, plumbing, captaining a boat, or becoming a doctor or lawyer or civil engineer after the age of 40.
Technology has no age limits, educational requirements, no government enforced credentials or inspections required, no apprenticeship or residency requirements, no significant union presence, no commonality for time in title, no professional tests to pass, and a trivially small tool investment cost.
You can plausibly learn the craft to an employable level with just internet access, a couple hundred bucks in computer hardware, an insatiable amount of curiosity and desire to build things and the right personality match against the interview team. It's not a guarantee but the barriers to entry are laughably nonexistent compared to many careers.
Yeah, and it's still not easy, especially if you're talking about being in the higher income levels. It's also not enough to just enjoy learning and growing to get you there.
What we are discusing is the distorted view in the original comment. The original comment was comparing the top earners of one field to the median earners for another. It would be better to compare the medians. For example median dev salary is about $110k. Median physician salary is about $205k. So no, these are not comparably paying jobs. And one does not "easily" earn more than a physician simply by enjoying learning.
I have an employee right now with Bachelor of Art in Music whose career in music fizzled out making $190k today in software outside of SV. That's it. That's his software credentials. A physician lost far more money in life getting their degree and spends about that gap in pay in malpractice insurance and professional attire required to work in a hospital, and generally works more hours at a higher stress level.
Sorting by annual income, software developers are surrounded by more highly credentialed career options that are much harder to get into or are surrounded by similar paying roles that much more sparsely available per capita.
What sob story? Are those immediately adjacent roles physicians - no, they're making almost double nased on median hourly wage. There are many near that level that do not require extensive training, including ones that have sizable representation (look at all those manager positions, and FYI magistrates are near there and require no law degree, etc).
Had that person used nurses, PAs, etc, then they would have been correct. Those do have training requirements and make similar money.
FYI, that's the mean annual income. I'm talking about median (use the median hourly wage column).
It's more than that. It's the continuation of a distorted view of reality in which the world is a meritocracy and everyone can be rich if they just try hard enough. That's there's nothing stopping anyone from taking a job that doesn't require years of training and making 2x+ the median for that role and using just an anecdote to back it up. This type of stuff is detrimental to the choices and mental health of the next generation - expecting things that don't exist and then being crushed by reality.
It would be interesting to see what that looks like adjusted for location/COL. It's possible the comp due to cost of living with a high concentration of devs in one area might be enough to influence this. Until we understand that and compare it to the distribution for physicians, it would be hard to draw any conclusions from it.
Bay Area bias is almost certainly skewing this entire conversation thread. In my average non-coastal metro area, all the nice homes in the nice areas are owned by classes of people like lawyers and doctors, not software developers. Most developers I meet here drive average cars and live in average housing.
> Bay Area bias is almost certainly skewing this entire conversation thread
Definitely.
> In my average non-coastal metro area, all the nice homes in the nice areas are owned by classes of people like lawyers and doctors, not software developers.
Is the area benefiting from brain gain or suffering from brain drain?
I'm not sure how that can be true. I would expect the if there were demand for the higher quality devs then we would see them skewing the distribution locally (like on Glassdoor). There are certainly some higher paid devs in the distribution, but the overall distribution is lower than SV. If the demand is low, then I would also assume the pay would be lower.
> I would expect the if there were demand for the higher quality devs then we would see them skewing the distribution locally (like on Glassdoor).
Glassdoor or salary surveys aren't the most reliable source of information for outliers. There's a proportion of devs that are flying completely under the radar and who are completely invisible to most local companies, even if they are in the same "local market" [0]. Because they are extremely valuable, they don't interview a lot and tend to hop between companies where they know people (or get fast tracked internally). When hiring is red hot, they might completely disappear from the hiring pool by junior year.
Something else to factor in is stock compensation. That probably won't be reflected in the BLS data.
What I've also been told is that due to the legal complexity of hiring international folks remotely, a lot of the people working remotely for US companies are legally contractor in their home country (so they would be reported in the local stats as self-employed and not necessary in the same category as other software devs).
There are alternatives, such as the military programs that will pay those bills for about an 8 year commitment.
But yes, I agree that being a programmer is better from an opportunity cost perspective. What I'm disagreeing with is the claim that devs make the same as docs. When looking at the medians, it's closer to half. If they had compared to to nurses or PAs, then it would be more appropriate.
Keep in mind almost all “salary” data I’ve seen for tech is just that, “salary”, it doesn’t include RSUs which begin to make up 50% or more of your TC which brings it to parity (or more) in your examples.
Someone else posted a link to actual income which still showed the median was about half. "Wages", according to the BLS, include bonuses and other comp. Stuff that vests or is deferred might not be counted. Even if it isn't, that's a small percentage of devs even earning them and wouldn't change the median much.
Real estate has licensing requirements, a national association you kinda need to participate in (MLS) and tons of regulatory requirements. I can't just wake up tomorrow and sell real estate.
The real estate exam is way easier than any IT interview. The MLS is a minor fee for access to the listing service. Tech has tons of regulatory requirements, some segments more than others. Most of the requirements for real estate are in the standard template already. Anything more difficult, and an agent will refer you to see a real estate attorney (much higher barrier to entry).
An easy way to look at this is how many real estate agents could be devs vs how many devs could be agents...
But those are all hurdles you must jump even if the skill required to do so is low. My dog groomer can fiddle with writing Excel formulas and then decide that tomorrow she wants to apply for an entry level software role, write a convincing resume and bullshit the right mix of interviewers well enough and literally nothing would stop her from making it through the application process or beginning an entry level job and turning that into a career if she was successful in the role. There is no standard or required process or governing body or common association or union or group to say who has what it takes to be a dev.
Tech can also have tons of regulatory requirements for the product or business but almost none exist for the employees themselves in the majority of jobs. You can hire a whole team of convicted felons that have never written a line of code into your software department, hand out the associate software engineer title to them and still pass most compliance, regulatory and audit requirements. It would probably be a dumb decision but nothing will stop me from calling them software developers and paying them as such and these roles are common (even if they don't pay entry level FAANG wages, they certainly pay living wages).
That doesn't make writing software easy, and doing nothing to set yourself up for success is going to lead to failure, stalling and washing out most of the time, but it's effectively a walk-on playing field of meritocracy in most cases.
Meritocracies don't exist, only the illusion for those that benefit.
Yes, it seems you're using the lowest theoretical barriers. That's fine. I'm talking about the common practical barriers. Of course our conclusions will differs.
I think an important part they left out is- if you want to make a lot of money, you need to optimize for money.
That means everything you do should have the goal of landing a higher paying job. Instead of studying things you find interesting, focus on learning skills that are frequently mentioned in job postings. Learning how to make your own compiler, for example, might not be the most profitable investment of your time.
To a salary-maximizer, work is simply a way to improve their resume. Before committing to a project, consider whether talking about that project will help you in an interview.
Most importantly, to make more money you need to constantly apply to better jobs even if you already have a high paying job you are happy with. This is where the majority fail, as most people won't apply for other jobs unless they are unemployed or unhappy. Some people are so deluded that they think they can get a better salary through promotions, simply by working hard at their current job, rather than through job hopping.
I personally know multiple people that transitioned into IT well into their 30s, from a variety of professional backgrounds.
They might not be employed at a FAANG making 500k a year, but they make enough money to live comfortably. Certainly above average for the rest of the population.
Easy? I don't like to undervalue other people's achievements, so I wouldn't call it easy. They certainly put their fair share of elbow grease into it. But it was not an unreasonable, insurmountable difficulty.
Now this seems like a fair assessment I can agree with.
I do want to add that in my 30s I'm seeing my mental abilities diminish pretty rapidly, so I don't know how long I can keep doing this. I've also seen people switch to other careers as they get old because they can't handle it anymore. So it seems to go both ways.
FWIW - I'm 55 and still going strong after 40 years in tech...
You sure you're not just burned out ? I go thru phases every 5-10 years where I don't even want to look at a computer :-P
When people say "easily", they don't mean "everybody gets to it". They mean "The chances are better than elsewhere". Which is true. And also fyi if you actually enjoy learning and growing you wouldn't sound so bitter about not making stupid amounts of money
Are the chances better than elsewhere? It seems sales, real estate, etc would be the "easier" place to make money as there's even less involved in barriers to entry.
"And also fyi if you actually enjoy learning and growing you wouldn't sound so bitter about not making stupid amounts of money"
FYI these aren't mutually exclusive. You can enjoy learning and growing while being bitter about getting screwed over. And it's not bitter to point out to other people that initial comment about just learning and growing being enough to easily reach top 5% income is a lie.
The only people I’ve seen who have major troubles are people who don’t actually like to program, or who have very staunch/stagnant attitudes. Do you fit into either of those buckets by chance?
"who don’t actually like to program, or who have very staunch/stagnant attitudes."
Were those in the comment? Seems these would be more than just enjoying learning and growing.
I might be getting there now. That wasn't the case in the past. Office politics and poor management ruined me early in my career. I'm just a burnt out low performer now. I have seen many others who also fall into the category of enjoying learning and growing (and higher performing than I was) who are not making $150k. Remeber, the national median dev salary is about $110k. So most devs by definition are making substantially less than the numbers claimed. I guarantee many of them also enjoy learning and growing.
Take a sabbatical dawg. Or save up to take one. Being burnt out is no way to live, and is blinding you to your potential. You appear to have little to no professional confidence.
I might be in the lower end of the distribution. That's fine. But that doesn't change the inaccuracy of original comment from ignoring medians to make an inappropriate comparison.
Yeah I think the drawbridge is getting pulled up on that one. This won't be as common in the future and from the other side we're at the point where most new hires went to college to get a degree specifically to get a tech job. This creates another problem where the new crop isn't curious about tech for it's own sake - it's just a respectable career option :-/
> This creates another problem where the new crop isn't curious about tech for it's own sake
This is not a problem and has never been a problem. Some of the best developers that I've worked with have absolutely no interest in turning on a computer when they leave work.
I know someone who was working an amazon warehouse job (as a regular entry-level warehouse worker, not as a manager or anything like that), did the internal amazon "learn to be an swe" training program (forgot the actual name, but iirc it is a 6-9 months long program) right before covid, and then converted to a SWE through that program. Currently, she is still employed as a SWE and makes typical amazon SWE salary rates (significantly above $150k). No tech background whatsoever.
This is still definitely possible, with a caveat that it just might not be as doable right at this specific point in time (given layoff waves and recession and all).
I know a guy who started as Java developer in some entrylevel company. No education, on previous job he used to drive trucks. He already has more job-specific knowledge than I do, though lacks some fundamentals.
I work for a FAANG, and I don’t know of a single IC here that was hired in their 40s or older. Almost everyone we’ve hired is in their 20s, with a handful in their early 30s. As a 40+ IC, I’m praying I don’t get laid off because it looks rough for older workers.
I got hired by a FAANG when I was 38, but I was definitely an outlier (although someone older than me got hired a few months after me, both as IC's). Looking back, I was lucky to hop on one of the very last trains leaving the station.
If I get fired or laid off, I will probably try to open a restaurant instead of looking for another tech job.
If it helps any, I'm in my mid-fifties and got hired as a full-time consultant at a F500 like 2(3?) years ago. consistently have clients select me for open projects. Current client is looking at extending another 3 years...
Hang in there! I'm a firm believer there's more of a 'cost bias' then an 'age bias' and if your on the side of the shop that brings in revenue, age isn't an issue...
> I don’t know of a single IC here that was hired in their 40s or older
Quite likely these older, experienced devs are either moving into management, or own their own consulting/contracting business, rather than be hired as an IC.
The problem with the current way large FAANG hierarchical pay structure is that ic have a ceiling in compensation.
Since we go for anecdata - I work for a FAANG, and I am the youngest person on my team in my late 20s. Half the team is above 40, the other half is above 30. Some of the people from both age groups were hired after I got hired (in the past 2 years).
Isn't that more to do with the fact that FAANG often hires straight out of college so they can pay you $80k a year for a degree from a top university and you won't realize you are getting shafted because you are a dumb kid?
15years ago I was reading the same ancedotes in slashdot comments. Tech is familiar with ageism, and the occasionally crusty greybeards refusing to go into management.
I'm at a FAANG adjacent company and we do hire older ICs. Though I wouldn't want to be interviewing right now. Senior level interviews can be demanding, and junior positions hard to find.
It's crappy answer to age discrimination, but people pivoting can leave less relevant experience off of resumes. I've done that.
I dunno...I'm 55 and was laid off earlier in the year and finding a new job wasn't terribly hard. I don't have a FAANG background and don't necessarily work with the absolute latest tech, but I try to work in one or two areas that seem marketable.
Took me a about three weeks to get two offers, but ended up going with a third that paid a bit less, but would let me learn some new skills.
Tech is a rare industry where you can easily pivot into it later in your career, start at an entry level position, work your way very high up the ladder and get paid very well along the way.
That's not likely to be possible anymore. Microsoft's net income went 9x in the last two decades. Apple's 475x. I don't think this kind of growth is going to continue.
Just because the outsized equity awards are likely a thing of the past doesn't discount anything from the sentence you quoted.
I've seen it time and time again, what the tech industry really craves is very detail oriented people who can communicate and who are proactive (i.e., if they see that something has fallen on the floor, they pick it up, they don't just assume someone else will). That said, as easy as it is to conceptualize those qualities, they are quite rare in my experience. Can't say how, as a software engineer, it's been extremely disappointing to see so many product managers who don't seem to understand that having a detailed knowledge of their product at the lowest levels is a requirement for the job.
This isn't really true for most people. Sure if you're already at one of the big companies they provide means for shifting your career. But most people aren't fortunate enough to work at one of those places, so changing careers can be much more challenging. You often need someone to take a chance on you or you need to take a more entry level position and the pay cut that comes with it.
But seriously - I learned HTML on AOL when I was 13 and dropped out of community college (where they were teaching Java and assembly).
I’m glad I had the very basic post-high school experience but much of what I know I was self taught as the technology came out.
I’m no electrical engineer and while I have a basic understanding of how transformers work I’m not about to revolutionize AI.
That being said I’ve been successful enough without incurring the debt of a bachelors degree while still making a lower (for HN) 6 figure salary.
Cost of entry is lower for tech (compared to some of the other fields mentioned here) and I’ve been able to pivot/expand into different areas within tech by continuing to teach myself with real world experiences.
It’s hard to let some rando on the street revolutionize healthcare/finance/whatever “established” system because they don’t get rebuilt every ~5 years.
I'm amazed by the negativity in the replies to this post.
I transitioned into programming as an older worker and it was really easy. It's much easier to be a successful programmer than a successful scientist.
Also I don't think that ageism in tech is a thing. The reason that there are very few older workers is because the number of jobs is increasing exponentially (doubling every 5 years since the 60s). All of those new jobs have to be filled with inexperienced people who are mostly young.
No. Everyone I've met who learned how to program after e.g. age 40, they had a hard time.
There's a certain mental agility which comes with youth which makes for good coders. I haven't met anyone who learned to program late who had the same. Over time, it gets replaced with depth, maturity, and experience, so people who start in tech do okay older. However, people who come in late have neither of those. Everyone I've seen try to make the hop late-career felt clumsy at programming, and at the same time, inexperienced.
It wasn't a failure due to age discrimination, although goodness knows that exists too. It was a failure due to simply not being able to become any good.
This isn't to say it's not worth doing, especially for people with a different prior skill set. An experienced teacher who learns just a little bit of programming can become awesome at ed-tech. An experienced biologist who learns a little bit of biology can do product work or manage a team of programmers for a biotech. Virtually any discipline with a little bit of programming becomes an awesome combo.
You can work your way up in most businesses or start your own once you hit the glass ceiling of credentialism. A tech job may speed that up by paying more, and having lots of docs to choose from, but it is a similar mechanism.
It's insulting that people think most of HN are highly paid. I dont really fit in anywhere. I suppose I don't even fit in on here. $150k would be a very nice raise for me, one that I'm not capable of abtaining.
HN is a very Silicon Valley focused website. $150k in SV isn’t very much at all. You get more than that even at startups as a senior engineer. New grads at FAANG will get at least that or more.
Isn't $150k the standard number for the common top end grade? They make adjustments based on location for COL. So it could be more there. Even so, the majority of developers live outside of SV where even the $150k number is attractive (given the national dev median is around $110k).
We're talking about federal jobs. Compensation is infact laid out in regulation. Stuff like a GS15 step 10 (top) has a set base number, but is then adjusted for duty location.
I don't think this has anything to do with any specific money figure.
I'm not from US or another international tech hub, I work remotely as a freelancer from the Italian country side.
Still, I make 93k, which puts me 4 times above the italian average. I'm in the top 1% globally, top 2% in Italy/Europe.
Not that I care much that I'm richer, but to reiterate that I care to struggle less for money than the average worker which was the best financial part of being in tech.
Going back to the median is not an appealing option to be honest.
Those of us who don't live in San Francisco that have been job searching have found salaries average about $125k. Over here in the real world tech doesn't make that much money.
Also, it's crazy that anyone would complain about that salary. By any metric $125k is more than enough money for a decent living.
> Those of us who don't live in San Francisco that have been job searching have found salaries average about $125k. Over here in the real world tech doesn't make that much money.
Incorrect. Most people I know in this industry never lived in SF and make well over that amount. Putting in some time on the West Coast is not a bad idea though - doesn’t have to be SF.
Big news headlines seem to be positioned in a way to scare people away from tech jobs right now, but the hard reality is that every organization relies tech people, not just those few big tech companies in Silicon Valley that over-hired during the pandemic.
Plus, I would argue that any argument trying to scare people away from tech jobs based on the current economic circumstances is worthless if it does not suggest a more promising alternative.
Will an individual who has the appropriate qualifications and experience find that tech jobs are harder to find, less stable, and less well compensated than they were 5 years ago? Yes, absolutely!
But will that same individual find that tech jobs are harder to find, less stable, and less well compensated than jobs in another industry right now? If so, which industry?
good systems have disconnected incentives and connected outcomes.
the same way advertisement upholds political correctness and same-thinking and make criticism harder, based solely on self interest and "not being demonetized by youtube", a journalist having clicks as the only metric will benefit whoever can spend money not caring for clicks to push an agenda to select the click-driving crisis du jour.
If you were an average engineer destined to work in cookie-cutter roles in tech consultancies, would you still pursue a tech career, knowing that your job will sooner or later be taken over by AI?
Remember that careers last 30 years, and if you're the average engineer, you're more likely to fall into the 80% of the pareto half than the 20%. Can you be confident that AI won't displace you in 10, 15, or even 20 years?
This doesn't apply to Stanford grads destined for high-tech roles in FAANG tier companies. I'm taling about the Idaho State grads who will work for IBM and Accenture and likely won't make the $500k salaries to retire from their tech roles in 10 years.
As long as we have human using computers, we'll have more skill humans helping the humans.
We were told that inventing web search engines and giving kids ipads would completely eliminate IT. We see how that turned out LOL.
From a financial standpoint, as long as the top 1% of "IT skills" population can make the other 99% of the company more than 1% more profitable, they'll never get fired, well, completely anyway. At big companies its a pretty low bar to surpass, and small companies can contract with MSPs.
From a cultural / psychology standpoint, its pretty deeply ingrained that some folks are dumb/lazy/ignorant and thats OK, some folks have intense learned helplessness as a character trait, and some folks get really excited about forcing others to "do their work for them" WRT primate dominance. We as a culture will never avoid the appeal of BYOD, even with a miracle of 100% BYOD in IT, we as a culture would STILL hire cloud admins just so the job can be delegated if nothing else.
The pay in IT in most of the country is ... microscopic compared to software dev. New grads or people with nothing but an A+ are not getting new Stanford grad software dev salaries, its more like $22 to $25/hr in most of the country for a new help desk jockey. And it's pretty boring work for 99% of the employees 99% of the time. However, if chatgpt can fizzbuzz better than humans, doing "IT stuff" beats starvation for us computer people.
Its also worth pointing out that AI as a tool is just biased single answer web search. Web search more or less resulted in more employment. Seems likely this variation on the theme will result in more employment.
Dentists are already 3:1 dental hygienists these days and I would expect AI can make that ratio 5:1 or 6:1. Technicians will also be enabled to be more effective with AI (faster diagnoses, remote diagnosis, predictive ordering so you don't have to come back twice with a special part, etc...). Also AI will drive higher reliability of the underlying equipment so the number of technicians needed will be lower.
If AI becomes capable of replacing the bulk of software developers, many trades will be replaced by a guy earning a little over minimum wage wearing AR goggles with an AI showing+telling them what to do and correcting their mistakes.
No need for the AR goggles, just hire day labor to work for a skilled tradesman as a labor multiplier, or try the old fashioned "here's a mobile phone do what your boss says"
In the trades themselves, we've been doing this for centuries... have a master mason supervise the journeyman masons supervising the manual day labor. AI will not be able to compete with experienced personnel. In that field they've already dumbed down the entry level requirement to "pick up a brick and move it there" and it can't go much lower.
On the other hand, if I were manufacturing survey equipment I'd worry about AR goggles as the competition.
Medicine, law, accounting, professional engineering. Notice something in common with all of these?
With the exception of law, automation has already arrived in all of these fields (to accounting first, 1-2 generations ago) and there are now more people in each field rather than less.
Answering my own question: each has a comfortable regulatory moat and competent professional society to defend it.
Yeah, but only certified humans can legally practice law. And they're also the group of people with the best ability to prevent changes in that legislation
I'm not sure on that. the the pieces seem to be there as far as hands that can manipulate rubix cubes and stuff, catch baseballs, boston dynamics robots agility, other humanoid robots.
Do I think white collar jobs go first, yes. also, the dynamics who buys all the stuff that blue collars workers install and build? without a white collar workforce the demand for that goes way down, and the supply of workers goes up.
Not saying it's going to happen just think other jobs aren't that far behind.
Perhaps in the long run everything will be automated, but in the meantime there will be a progression of what AI can do, and software engineering might be "lower-hanging fruit"
Any career with strong regulation is going to take longer to automate. Hands-on work as well.
For example, there is strong regulation preventing computers from practicing driving/trucking/delivery, law, medicine, nursing, physiotherapy, and certified engineering
Also, build me a robot that can do the following. I'll wait: construction, woodworking, metalworking, glassblowing, painting (painting houses), electricking, plumbing, delivery, service industry, cooking.
If you’re entering college right now, you’ll be around 33 in 15 years.
We have no way of knowing what this tech will be like in 15 years.
The iPhone was released about 15 years ago. Even the most prescient of us couldn’t have predicted how fundamentally that would change life, culture, work and business.
Lol, I talked about very short snippets and you counter my argument with a video showing very short snippets.
If I wasn't scared before, I'm even less scared now.
But if all you do is write trivial code anyone can learn to write in 1 week, with extremely detailed instructions, perhaps your job is in danger indeed. That's not what a software engineer does anyway.
According to HN, technology never improves. The first version of something is always the way it will ever be. OpenAI hasn't improved at all from GPT-3 to GPT-4, and all the billions flowing into this field will cause zero improvements in the next 30 years.
This place needs to change its name to LudditeNews.
The main concern of the luddites was not "technology = bad" like most popular caricatures of them, but that their livelihood was obliterated by the first wave of factory owners taking all the profits that were previously shared with the workers for themselves. Really they were pushing in the same direction as those who today are calling for machine taxes or UBI.
Ehhh I'd argue they had more in common with socialists/communists, they were literally destroying the means of production in worker solidarity to improve their condition and bargaining ability. Marx even refers directly to the situation of the guilds in industrialization.
The belief is that new means of production should be embraced, owned, and operated by workers to benefit the many and not just the few exploitative capitalists (which are those who own means of production).
Instead of exploiting lay-offs and long shifts, and disregarding safety to save the capitalist money, the factory would instead be built and operated for the benefit of workers and society, prioritizing safety and sharing shorter shifts among a larger workforce to preserve jobs.
From a socialist/communist perspective machine taxes and UBI are like cheap bribes to avoid the issue and preserve the exploitative system in place. The luddites probably would have felt the same had they been offered a pity sum or machine tax from factory owners to "offset" their condition.
boggles my mind that we had a tool that was smart when it was first launched in November 2022, experienced a massive generational improvement in the next iteration (GPT-4) within 4 months...and people on HN are ready to assert that this is the best it will ever be.
It improved drastically in quality within 4 months, yet for some reason, the nex 10 years will see 0 progress in a field with literally tens of billions and intellectual capital flowing in.
Most revolutionary technical changes like this tend to hit a plateau in progress right after they hit the top of the hype cycle. It’s unlikely that AI/ML will be an exception.
Not bad news for anyone involved in this work, though. That period is when the real business and societal changes happen. It’s also when most of the money gets made.
People and businesses need time to test new technology and adapt it so that it works reliably for their specific use cases. GPT and other LLMs are no exception.
Again, the point is that careers are built over decades, and this proof of concept tech is already good enough to do the job of a coding intern.
That might not be a threat to you if you’re a 35 year old developer, but if you’re an 18 year old exiting high school, there’s a non zero chance that by the time you’re 35 - 17 years - the tech will be good enough to replace you.
The hype ratio between GPT2 and GPT3 is much higher than the hype ratio between GPT3 and GPT4. If we project this into the future then the hype ratio will be even lower between GPT4 and GPT5. The key innovation was RLHF and everything since then has been incremental improvements.
>knowing that your job will sooner or later be taken over by AI?
There's a few layers that make me not worried:
- automation will shift jobs
- an engineer is a generally smart employee (yes, even the average engineers), so the company will be a sinking ship if they have to eliminate jobs and get rid of people instead of moving smart people around
- even if you do end up jobless, as a generally smart person, you can learn a new skill. Plus, now AI can help you learn more quickly.
The catch is, of course, predicting which next job won't be replaced by AI.
My perception is that at least in Europe, mass layoffs have had almost no impact on the tech sector. I would say there is still a shortage of people for desired roles and therefore job offers and salaries are still high.
Not talking about FAANG positions (or salaries) of course, more run-of-the-mill roles that companies struggle to fulfill.
I‘m currently on the market and the situation is definitely very different from the same time last year. Companies are hesitant/waiting to see where the economy is heading, a lot if startups ran out of money. It‘s not that easy anymore, at least in my bubble.
Yeah last year my inbox was flooded with recruiters, but I had a rough round of technical interviews back to back and succumbed to imposter syndrome. I postponed interviewing to learn and prep and try to add more to my resume.
Now a year later, I'm trying for a new job again and I hardly hear from recruiters, and I actually have to apply for jobs. It's a huge contrast from this time last year and it makes me regret waiting.
Side Note:
Having seen other coworkers move on and knowing what I know now, I've realized I was more than ready to get a new job a year ago, I just lacked the confidence to keep trying. Hopefully this serves to motivate others paralyzed by imposter syndrome.
The pandemic hiring blitz was more like seasonal hiring in a sector that doesn’t normally do that, and nobody had the vocabulary to describe the phenomenon.
Combined with talent hoarding the whole thing became a luxury companies can’t afford without free money from either investors or the fed.
Even with all this cutting, headcount is still growing. But it’s harder than ever for the entry level guys. Painting a rosy picture here is wrong.
I'm not sure what would be a better bet? It's not like automation / the economy / capitalism aren't also coming for other workers. Maybe a good time to retrain as a bailiff?
ChatGPT was based on GPT3, which was released in 2020. The fundamental advance appears to have been InstructGPT which was 2022. GPT4 was 2023, three years after version 3. Id expect to see GPT5 sometime in 2026.
This thing is out in the public now and has massive mindshare. Funding and intellectual talent is flowing in at a much faster pace. The big incumbents are now under pressure from their shareholders to come up with competitors.
I'll be very, very surprised if the growth rate doesn't accelerate from here.
It depends on what improved GPT-4 relative to 3. If it was adding more data, then we'd expect progress to speed up in the short-term and then come to a screeching halt as we run out of text.
If it was algorithmic/model improvements then it could be much quicker to improve or much slower.
I think the big problem is that nobody (except Open-AI) knows what GPT-4 does and how it does it, so it's hard to predict.
I'm generally a fan of a naive model given the uncertainties around this, hence my prediction of three years.
SV is really failing their community lately. First with SVB and now with the job market. People outside the valley are largely still hiring developers. East coast is pretty hot still.
How would this all go if it is necessary to start rationing access to AWS, Azure and GCP, driving up the cost of hosting applications on them to the point that only well funded organizations can continue doing so? What percentage of "Tech jobs" are in companies that exist at the whim of a couple shareholders or poorly funded organizations that are unlikely to withstand 3000% cloud price increases? (Spoiler: MOST.) Was any of this discussed by NPR?
What could cause such a black swan event? Does NPR want us to even think about the possibility of the sudden "closure" of certain fabs in Taiwan and Shenzen (making the chip shortage a semi-permanent event)? https://nypost.com/2023/04/26/us-to-send-nuclear-submarine-t...
Nobody in Tech has time for NPR, there are a million YouTube videos to watch, websites, podcasts and Safari books to read, plus management of your personal life and ACTUAL Sphere of Influence. The best you can do is vote for someone who you (not NPR, not Communist-controlled social media platforms or the revolving door of bloviators on TV) think is able to effectively manage issues in a way that will bring about the best economic outcome for you and everyone you know. https://youtu.be/bPpcfH_HHH8
I am convinced that the mass layoffs are occurring because forward-thinking executives can see that AI agents are coming to replace employees faster than most people expect. By doing layoffs early they can cut staff before public opinion shifts to questioning AI as the motive, as opposed to "the economy".
From what I can see, GPT-4 is incredibly useful but you have to be very careful to lead it in the correct direction otherwise it produces bad outputs. I say this as somebody that probably uses it every 10 minutes. Do you think that it will stop making errors of judgment and be able to behave as an agent soon?
I've not encountered anything yet that gives me faith in that, other than hype about it increasing in ability exponentially which seems unlikely.
> I say this as somebody that probably uses it every 10 minutes
Do you feel like your ability to produce correct output with GPT improved with practice? If so, then eventually you might reach a point where you get 99.99% accuracy with it, and this can easily 10x your productivity.
No, that's not really what has happened. If you ask a question, it tends to mean you lack knowledge about something, and therefore you don't have enough context to produce a question that perfectly teases out the correct answer from GPT. Instead what happens is that you have to carefully engage in conversation and iterate towards improving GPTs context and your understanding of what it is saying well enough to ask it more incisive questions or to correct obvious inconsistencies in what it has communicated.
That’s a level of confidence, preparation and forward thinking that would be uncharacteristic. The reality is that everyone else is cutting staff so it’s become expected. And because it’s expected and happening throughout the industry it’s a lot easier to do a layoff or two now than it was a year ago
It’s not because it’s “expected” in the abstract. It’s because everyone is competing for money with the US Treasury, and the rates on those Treasury bills are looking better than they have in a long long while.
Did you have a growth business that’s going to be 50% bigger in 10 years? A few years ago that would sound not-amazing but reasonable. Today that’s nothing. You can get that kind of returns from bonds with like zero risk.
Firms therefore prioritize savings like they haven’t done in years.
The point is that if they are not letting go of employees for economic reasons i.e. they are no longer economically viable because of the environment, then they must be getting rid of them while the employees are actively valuable? otherwise the AI agent argument holds no water, because the employees aren't creating value anyway, so their dismissal is orthogonal to the AI agent question.
I know folks without a college degree that started as tech support reps who worked their way up to senior roles in big tech that make multiples more than most physicians I know or others that had to go down a rigid path to get to where they are.
The industry favors eager people who are hungry to continually learn and grow. While there can be short term bumps in the road, those that fit these characteristics have and will continue to do extremely well.