My experience is that AI is easier to work with than inexperienced junior coders. I designed database structure and wrote function subs, and some comments, then asked AI to 1) fill the functions, 2) write unit tests while I work with interesting parts or read stuff.
I think the result is much less errors compared to junior coder. You have to review the code just like with human. Time it takes is just little bit less than working with other person, but it's orders of magnitude cheaper. I'm certain that many will cut corners and ship code not reviewed by humans, but that's on them.
Interesting question is, can inexperienced coder get more or less benefit from AI compared to experienced programmer, or is it skill multiplier.
_aavaa_ · 8h ago
Replacing junior positions with AI is eating the seed corn.
3vidence · 5h ago
I think the whole "AI will replace developers" thing makes sense to people in silos.
I want to do X I can ask AI to help me with X.
But the thing is that as businesses grow and scale one person cannot even keep in their head all the things that have to get done, let alone write it in such a specific way to get an AI to accomplish it.
Even Jr developers (if used correctly) are not just code assistants, they can actually take your mental load away as they grow and then provide projects / ideas that you didn't even originate.
People have mental limits to how much they context they can keep in their head at the same time, AI does not magically solve this fact.
Hire and develop talent, share the load.
zbentley · 5h ago
Nothing you said is wrong, but:
> makes sense to people in silos
Investors, especially PE investors, are in a motherfucker of a silo.
The sign on their silo says "short term growth at all costs. ZIRP/dotcom boom days are never coming back. The last train is leaving the station, and even if it isn't the real deal like eCommerce/app stores/social media, you had better be on it, because there won't be another one."
That last train is AI.
NoPicklez · 8h ago
It's what many are saying about AI in general for the workforce, it frees up the "easier" tasks to allow people to do more high level work.
I do think that as a junior using AI could speed up your learning in programming significantly.
You essentially have a tutor in your back pocket who can help teach and also do some of the doing if you need it to.
jvanderbot · 9h ago
The comments here don't seem to address TFA.
There's a lot to unpack on this subject and tons of anecdata but the article manages to unpack an infinitesimal amount of real data.
The result shown is that the total number of software jobs has increased the last 3 years (of the data), even increasing relative to total USA population.
It shows a 23% increase in employment the last 3 years overall (1.6/1.3 million = 1.23).
Data ends before 2024.
It mentions BLS analysis projects a 17% growth until 2033. (Useful if LLMs are not a paradigm shift, or if BLS took that into account?)
This comment is longer than the article if you remove the preamble discussion of LLMs.
b3ing · 7h ago
The main stream media mentioned 30% of job postings are fake and another 30% are going overseas, not sure this addresses that or the 1 million plus layoffs.
jvanderbot · 7h ago
It's not built on job postings or openings, just employment. So it would account for layoffs.
BandButcher · 7h ago
Yea no mention of the layoffs the past couple of years although this blog post was published today. Satire? Fluff piece? Or maybe an advertisement for cursor and copilot??
Article supposedly published by a CS professor but if I had published something like this in uni I feel I may have gotten laughed at.
As per the topic,
I could see software jobs increasing in total number due to the emergence of machine learning jobs but to generalize and say the demand is as strong as ever is a little disingenuous especially when ever other CEO claims ai will and has already reduced the workforce
jvanderbot · 7h ago
The post has a few stats and that's it. I repeated them here, because I thought it might be interesting to cut through sentiment a bit.
Yes, ending before 2024 might mask a lot of change from all kinds of factors.
And, not going back prior to 2021 might mask a downturn prior to that from which we haven't recovered.
Edit: 2019 levels were 1.4m, making the % increase from 2019 to 2023 17%.
It shows 1.65m, so basically level with 2023. That makes growth from 2019 to 2024 17% (1.64/1.4) as well.
leeroihe · 10h ago
If you ignore data regarding H1B, OPT and O1 your analysis is meaningless. I support immigrants coming to this country to work, my girlfriend is one of them! But what has happened to tech in terms of Americans being displaced is candidly astounding.
As a black man, I honestly think more of the apparent "discrimination" against American minorities is a direct result of offshoring and h1b.
nyarlathotep_ · 8h ago
> But what has happened to tech in terms of Americans being displaced is candidly astounding.
My tenure at a BigCo came to an end recently-ish and in the 2 years before I left I saw:
- multiples of new openings in LatAm/India vs Domestic
- increasing use off offshore staffing firms even on big name stuff
- presumed increase in offshore teams based on my random heuristic of clicking usernames making recent posts in various Slack channels "team: $FOO, Location: Bangalore" and etc. The volume here was much higher than last year, and far higher than it was in the years prior.
- lots of open internal talk about "leveraging partners"
Majority of the work I did was working with private sector F500s and the extent to which those companies outsourced IT/Software was staggering. I encountered several major companies that had basically no "actual" IT/Software in the States: outside of a handful of meeting-takers the actual implementation work was almost 100% outsourced/staff-aug via H1B shops.
The talk about "AI" taking jobs is a total distraction from the actual source of the jobs disappearing.
nunez · 6h ago
I don't think we have enough evidence to conclude whether the field is collapsing or not _due to AI in the workplace._ This is a recent phenomenon.
jvanderbot · 6h ago
For it to be collapsing for any reason it must first collapse. It seems not to be, is all the article was saying.
Belopolye · 9h ago
The sentiment in my circles seems to be that senior-level positions are still going strong.
My experience is that AI is easier to work with than inexperienced junior coders. I designed database structure and wrote function subs, and some comments, then asked AI to 1) fill the functions, 2) write unit tests while I work with interesting parts or read stuff.
I think the result is much less errors compared to junior coder. You have to review the code just like with human. Time it takes is just little bit less than working with other person, but it's orders of magnitude cheaper. I'm certain that many will cut corners and ship code not reviewed by humans, but that's on them.
Interesting question is, can inexperienced coder get more or less benefit from AI compared to experienced programmer, or is it skill multiplier.
I want to do X I can ask AI to help me with X.
But the thing is that as businesses grow and scale one person cannot even keep in their head all the things that have to get done, let alone write it in such a specific way to get an AI to accomplish it.
Even Jr developers (if used correctly) are not just code assistants, they can actually take your mental load away as they grow and then provide projects / ideas that you didn't even originate.
People have mental limits to how much they context they can keep in their head at the same time, AI does not magically solve this fact.
Hire and develop talent, share the load.
> makes sense to people in silos
Investors, especially PE investors, are in a motherfucker of a silo.
The sign on their silo says "short term growth at all costs. ZIRP/dotcom boom days are never coming back. The last train is leaving the station, and even if it isn't the real deal like eCommerce/app stores/social media, you had better be on it, because there won't be another one."
That last train is AI.
I do think that as a junior using AI could speed up your learning in programming significantly.
You essentially have a tutor in your back pocket who can help teach and also do some of the doing if you need it to.
There's a lot to unpack on this subject and tons of anecdata but the article manages to unpack an infinitesimal amount of real data.
The result shown is that the total number of software jobs has increased the last 3 years (of the data), even increasing relative to total USA population.
It shows a 23% increase in employment the last 3 years overall (1.6/1.3 million = 1.23).
Data ends before 2024.
It mentions BLS analysis projects a 17% growth until 2033. (Useful if LLMs are not a paradigm shift, or if BLS took that into account?)
This comment is longer than the article if you remove the preamble discussion of LLMs.
Article supposedly published by a CS professor but if I had published something like this in uni I feel I may have gotten laughed at.
As per the topic, I could see software jobs increasing in total number due to the emergence of machine learning jobs but to generalize and say the demand is as strong as ever is a little disingenuous especially when ever other CEO claims ai will and has already reduced the workforce
Yes, ending before 2024 might mask a lot of change from all kinds of factors.
And, not going back prior to 2021 might mask a downturn prior to that from which we haven't recovered.
Edit: 2019 levels were 1.4m, making the % increase from 2019 to 2023 17%.
https://www.bls.gov/oes/2019/may/oes151256.htm
And 2024 is available, but slightly more difficult to query. https://data.bls.gov/oesprofile/
It shows 1.65m, so basically level with 2023. That makes growth from 2019 to 2024 17% (1.64/1.4) as well.
As a black man, I honestly think more of the apparent "discrimination" against American minorities is a direct result of offshoring and h1b.
My tenure at a BigCo came to an end recently-ish and in the 2 years before I left I saw:
- multiples of new openings in LatAm/India vs Domestic
- increasing use off offshore staffing firms even on big name stuff
- presumed increase in offshore teams based on my random heuristic of clicking usernames making recent posts in various Slack channels "team: $FOO, Location: Bangalore" and etc. The volume here was much higher than last year, and far higher than it was in the years prior.
- lots of open internal talk about "leveraging partners"
Majority of the work I did was working with private sector F500s and the extent to which those companies outsourced IT/Software was staggering. I encountered several major companies that had basically no "actual" IT/Software in the States: outside of a handful of meeting-takers the actual implementation work was almost 100% outsourced/staff-aug via H1B shops.
The talk about "AI" taking jobs is a total distraction from the actual source of the jobs disappearing.