Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Gen AI will increase demand for software engineers

roarepally.com

Post reply on HN
Print view View original

51–60 of 62 posts

Re: Gen AI will increase demand for software engineers

#51

Post by Lio »

simpaticoder wrote: post #33

Yes, but it will strongly decrease demand for new ideas. AI is a "win more" technology for current software architecture incumbents. The more popular a tech is, the more content there is about it, the more capable an LLM is of giving a usable answer. A programmer who relies on LLMs for coding assistance will be far less likely to try something new, and it will have a chilling effect on the adoption of all new tools,…

I can kind of see that argument but it reminds me of when a new programming language comes out and there's no IDE support for it.

That definitely stops some developers from considering it, others are happy to use just a standard editor without intellisense, etc.

Re: Gen AI will increase demand for software engineers

#52

Post by ryandvm »

HarHarVeryFunny wrote: post #50
loudmax wrote: post #35

Earlier quoted context omitted.

What kind of time frame is "any time soon"? If you're thinking over the next 3-5 years, then AI is a novelty in most cases, a useful tool in some cases, and something truly disruptive in a fairly narrow set of cases. If you expand the time frame to 10-20 years, then AI is poised to become something that truly upsets most of the economy. Obviously, nothing can be said for certain, there could be black swan events, etc…

> What kind of time frame is "any time soon" I'm guessing that solving run-time learning may well require a different approach, and it's not clear that reasoning (in general form - ability to dynamically synthesize a problem-specific solution) can be just added to LLMs either (e.g. by adding tree search). There are also other missing components such as working memory that seem simpler to solve. Coming up with brand n…

But even now the LLMs absolutely have limited problem solving capability.

For example, yesterday I asked GPT-4o to write multiple alternate endings to the short story "The Last Equation". They weren't dramatically compelling, but they were logical and functional.

How is that not problem solving? And so help me, before anyone tells me it's just stringing together the next most likely tokens - I don't care. Clearly that is at least a primitive form of intelligence. Actually it's not even apparent to me that that isn't exactly what human intelligence is doing...

Re: Gen AI will increase demand for software engineers

#53

Post by simpaticoder »

Lio wrote: post #51
simpaticoder wrote: post #33

Yes, but it will strongly decrease demand for new ideas. AI is a "win more" technology for current software architecture incumbents. The more popular a tech is, the more content there is about it, the more capable an LLM is of giving a usable answer. A programmer who relies on LLMs for coding assistance will be far less likely to try something new, and it will have a chilling effect on the adoption of all new tools,…

I can kind of see that argument but it reminds me of when a new programming language comes out and there's no IDE support for it. That definitely stops some developers from considering it, others are happy to use just a standard editor without intellisense, etc.

Gen AI increases the demand for software engineers, but software engineers increase the demand for AI for code assistance. You are right, of course, but the number of programmers not using AI code assist will asymptotically approach some very low minimum, thanks for the productivity boost. I keep hearing "3x" and that feels about right to me. The same could not be said for intellisense, or at least is harder to quantify.

Re: Gen AI will increase demand for software engineers

#54

Post by HarHarVeryFunny »

lolinder wrote: post #45
loudmax wrote: post #35

Earlier quoted context omitted.

What kind of time frame is "any time soon"? If you're thinking over the next 3-5 years, then AI is a novelty in most cases, a useful tool in some cases, and something truly disruptive in a fairly narrow set of cases. If you expand the time frame to 10-20 years, then AI is poised to become something that truly upsets most of the economy. Obviously, nothing can be said for certain, there could be black swan events, etc…

> If you expand the time frame to 10-20 years, then AI is poised to become something that truly upsets most of the economy. Obviously, nothing can be said for certain, there could be black swan events, etc. But on the current trajectory [...] This is an interesting framing that I think exposes a key difference between people who view AI dominance as inevitable and those who don't. You view the progress of AI slowing…

I also think human level (and beyond) AGI is inevitable, but that in of itself says nothing about the time frame in which it will be achieved. People have been predicting AI/AGI ever since the advent of computers over 70 years ago, and estimates have always been way over-optimistic.

It's perhaps relevant to note that the reason estimates have always been off seems to be that people have always thought that the technology of the day was sufficient, and that it was just a matter of applying it. So far we've had procedural programming (SHRLDU), expert systems (CYC), general-purpose problem solvers (SOAR), PDP/connectionist approaches, modern neural nets, LLMs...

I do think we're getting closer, and that neural nets will get us there (maybe not with gradient descent as the learning method though). There have been lots of valuable lessons and insights learned from various neural net architecture, not least of which is the confirmation by LLMs that prediction seems to be the basis (or at least one of major pillars) of intelligence.

And yet, I can't help but feel that there is still a long way to go ...

Re: Gen AI will increase demand for software engineers

#55

Post by randomdata »

williamcotton wrote: post #26
nprateem wrote: post #9

Earlier quoted context omitted.

It's frustrating just how shit AI is at coding. The new gpt model seems even worse, splurging out a load of guff that makes it even more difficult to see the bullshit it's produced. AI a threat to humanity, my ass.

Weird, I find that I get very good results with GPT-4 and GPT-4o. It helps to be in a position where you can communicate very explicit instructions on what and how to build. Use it like you’re planning what to code each step but just letting this mechanism do the typing. General writing and technical communication skills are a must so I can understand why some developers struggle with LLMs.

> General writing and technical communication skills are a must

Perhaps that is what he is trying to get at? I am sure my communication skills are poor, but I found utility in using these tools a year or so ago. Whatever gibberish I was able to give it often produced a good result. These days I can't seem to get anything usable out of them. It does seem, like the parent suggests, that they have gotten worse.

It may very well be that said tools are no worse, if even better, where communication ability is stronger. However, if we have chosen to optimize these systems for those who are great at communicating, at the cost of those who are not, that doesn't help with the topic at hand. The non-software engineers – implying Average Joe – are almost certainly not going to be great technical writers.

Re: Gen AI will increase demand for software engineers

#56

Post by HarHarVeryFunny »

ryandvm wrote: post #52
HarHarVeryFunny wrote: post #50

Earlier quoted context omitted.

> What kind of time frame is "any time soon" I'm guessing that solving run-time learning may well require a different approach, and it's not clear that reasoning (in general form - ability to dynamically synthesize a problem-specific solution) can be just added to LLMs either (e.g. by adding tree search). There are also other missing components such as working memory that seem simpler to solve. Coming up with brand n…

But even now the LLMs absolutely have limited problem solving capability. For example, yesterday I asked GPT-4o to write multiple alternate endings to the short story "The Last Equation". They weren't dramatically compelling, but they were logical and functional. How is that not problem solving? And so help me, before anyone tells me it's just stringing together the next most likely tokens - I don't care. Clearly tha…

I would define intelligence as "degree of ability to use past experience to predict future outcomes" (which includes reasoning, aka problem solving ability, via repeated what-if prediction, then backtracking/learning on failure etc).

So, intelligence exists on a spectrum - some things are easier to predict given a set of learnt facts and methods than others. The easiest things to predict (the most basic form of intelligence) is "next time will be the same as last time", which is basically memorization and pattern matching, which is mostly what LLMs are able to do thanks to brute-force pattern/rule extraction via gradient descent.

Going beyond "next time will be the same as last time" is where reasoning comes in - where you have the tools (experience) to solve a problem, but it requires a problem-specific decomposition into sub-problems and trial-and-error planning/testing to apply learnt techniques to make progress on the problem...

Certainly a lot of human behavior (applied intelligence) is of the shallow "system 1" pattern matching variety, but I think this is over stated. Not only is "system 2" problem-solving needed for on-the-job training, but I think we're using it all the time when we're doing anything more than reacting to the current situation in mindless fashion.

So, sure, LLMs have limited intelligence, but it's only "system 1" shallow intelligence, gestalt pattern recognition, based on training-time gradient descent learning. What they are missing is run-time "system 2" problem-solving.

Re: Gen AI will increase demand for software engineers

#57

Post by PleasureBot »

kcplate wrote: post #36
lexicality wrote: post #21

Earlier quoted context omitted.

I'm curious as to what engineering solutions you think are available to solve this problem? All the solutions I've heard up until now is "we just need more data" which isn't particularly sustainable or something a SWE would be in charge of doing

It’s not my expertise. However, if it’s a barrier to adoption it’s a problem to solve and will be. AI is exponentially improving and there is no reason to believe that this won’t continue.

That's not how technology typically advances. Technology advances in waves. There's an initial breakthrough that leads to rapid progress in certain areas. But after some time there are diminishing returns as the the applications of that breakthrough are fully explored, and progress slows or stops until the next breakthrough. Predicting that the trajectory of progress for the next 20 years will match the progress that has been made in the first 2 years after the LLM breakthrough seems naive at best.

I would also disagree with the idea that LLM tech is "exponentially improving" now. There was the initial release of ChatGPT which was an enormous step forward, but since then its been small iterations on the same fundamental technology. We are already seeing significantly diminishing returns in our ability to improve these models. Most available training data has already has already been fed in, 10xing the compute results in only marginal improvements in performance, etc.

Re: Gen AI will increase demand for software engineers

#58

Post by HarHarVeryFunny »

bearjaws wrote: post #32
kachapopopow wrote: post #10

Eh, it's only a matter of time until new nvidia clusters are in production and we start seeing what hardware designed for training massive models (roughly 100x bigger than anything we know of at least) can do. With some additional architectural changes we might start seeing some sparks of real human stupidity.

Maybe 4x, 100x is many years away. The exponential graph people are posting is just to pump the stock price.

The exponential graph, to the extent it is measuring anything, is measuring progress on LLM benchmarks, which are based on tasks that LLMs can tackle and show progress on.

The set of benchmarks used to measure LLMs do not include things that require significant run-time learning or reasoning (e.g. Chollet's ARC test, although frankly that's a pretty simple test).

This exponential progress will flatten once LLMs can no longer improve along that LLM-benchmark axis (maybe they score 100% on all LLM benchmarks), but this "100% achieved on all tests" won't mean "mission accomplished - AGI achieved!", it'll just mean it's time for a new set of more challenging benchmarks, one step closer to what humans are capable of.

Re: Gen AI will increase demand for software engineers

#59

Post by WWWMMMWWW »

bob1029 wrote: post #23
noobermin wrote: post #7

This article feels like stating an opinion. I suppose the autopilot analogy is at the heart of their argument but the argument as of now feels a bit too flat. There are definitely some jobs that technology completely displaced, and there are some that merely made something more accessible. For example, grinding grain with the advent of modern milling equipment vs microwaves. Milling machines make hand driven milling…

> Milling machines make hand driven milling something of a niche whereas microwaves arguable have displaced no jobs in cooking and food preparation. I would argue that generative AI in the hands of an already skilled developer is equivalent to the mill scenario. One person's expertise leveraged by way of this tool can allow for output that is orders of magnitude greater. The mere inertia of the "machine" encourages m…

It's like having a team of hyperactive interns at your disposal.

They'll fetch you all the latest libs, best practices, code samples, etc.

But they have no idea how to apply it to the business case at hand, how to secure the code, what real world edge cases to protect against.

And half the time their code is suboptimal or doesn't quite compile ... and wait till they try complex async ...

It's a force-multiplier for good/experienced engineers, but it's going to devastate the market for white-collar sweatshops.

Re: Gen AI will increase demand for software engineers

#60

Post by nostrademons »

ch4s3 wrote: post #39
loudmax wrote: post #35

Earlier quoted context omitted.

What kind of time frame is "any time soon"? If you're thinking over the next 3-5 years, then AI is a novelty in most cases, a useful tool in some cases, and something truly disruptive in a fairly narrow set of cases. If you expand the time frame to 10-20 years, then AI is poised to become something that truly upsets most of the economy. Obviously, nothing can be said for certain, there could be black swan events, etc…

> If you expand the time frame to 10-20 years This feels like a lot of "20 years away" type predictions. There needs to be some sort of step change in the technology, and perhaps more than one or on more than one front to make this possible but its far enough away that 20 years feels like a safe bet. When people say this about LLM based AIs, I wonder how they imagine text prediction making those step changes. Is your…

I think the idea behind the "10-20 year" timeframe is that that's how long current technological changes will take to propagate into social & economic changes, like how HTTP was invented in 1989 but we didn't get e-commerce until 1995 or social networking until 2002.

LLMs dramatically reduce the price of bullshit. It follows that they'll likely reduce the value of bullshit too, but we don't know to what extent or how that'll affect the organization of the economy until entrepreneurs generate vast bullshit factories and counterparties start adjusting their behavior to adapt to it. A huge swath of the economy runs on bullshit, from media to PR to sales to venture capital to politics. How these industries are structured will likely change when writing vast quantities of truthy-enough-to-be-convincing text becomes insanely cheap.

Post reply on HN