6 ms·

Gen AI will increase demand for software engineers

• v3ss0n 2y ago
  Not for RAG and Automation applications But for Fixing Hallucinated code that written by Non Software Engineers.
  • joesimultaneous 2y ago
    [flagged]
  • drewcoo 2y ago
    > But for Fixing Hallucinated code that written by Non Software Engineers. And for testing to find those inconsistencies.
    • rpigab 2y ago
      And for reviewing failed tests that were hallucinated by LLM testing software. And for fixing the consequences of LMM testing giving false negatives.
    • candiddevmike 2y ago
      This sounds like hell, tbh.
      • naasking 2y ago
        Sounds like basically the same hell as reviewing code of junior SWEs.
  • kcplate 2y ago
    That is completely contingent on if present the day SWEs stop working on trying to eliminate the hallucinations.
    • lexicality 2y ago
      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
      • kcplate 2y ago
        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.
        • lelanthran 2y ago
          > AI is exponentially improving Not for the last two years at least. Over the last year it's been obvious that there are diminishing returns already... We increase the computational power by a factor of ten, and see a factor of 1.5 improvement.
        • randomdata 2y ago
          > AI is exponentially improving The theoretical research may have improved exponentially. Implementations available for use seem to be on the decline – at best, stagnant. I find less and less utility as time goes on. Things LLMs shined at a couple of years ago now produce garbage. And for creative work, the output is much too formulaic, which was all well and good initially while still novel, but one has to keep pushing new boundaries and the current crop of tools really seems to struggle with that.
        • PleasureBot 2y ago
          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.
  • nprateem 2y ago
    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.
    • williamcotton 2y ago
      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.
      • randomdata 2y ago
        > 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.
  • empath75 2y ago
    Human engineers of course never make mistakes, which is why we don't have PRs or unit or integration tests and canary deploys or anything that would catch mistakes before going to production because that would be a waste of time.
  • jayemar 2y ago
    Does RAG mean Retrieval-Augmented Generation? I've never heard of this term.
• andrewstuart 2y ago
  It’s weird that there’s people who seem to be “against” AI assisted programming. That’s like being against IDEs or git. It’s like writers who were brought up on typewriters being against word processors.
  • empath75 2y ago
    One way to think about ai programming is that it's another step on the continual ladder up from writing machine code to writing more in more 'human readable' programming languages. It's essentially a non-deterministic compiler from natural language to an intermediate representation like python. Even assuming it was able to near perfectly translate intent to code, that doesn't mean it will replace programmers, necessarily. The way to think about the impact on employment is comparative advantage. https://www.investopedia.com/terms/c/comparativeadvantage.asp https://www.investopedia.com/terms/c/comparativeadvantage.as... Even if AIs are better than humans at every programming task, there is still a limited amount of compute in the world and unlimited amount of potential work to do, so there will always be tasks available for humans to work on. If the cost of writing code becomes cheaper, there will be tasks to automate that aren't worth the effort to automate now, that will _become_ worth the effort to automate in the future.
  • xdennis 2y ago
    There are different reasons. Some people are against a technology which isn't mature enough. I don't know about word processors, but take digital photography for example. Some people stuck with film photography even after digital became cool because it was superior in some regards (such as dynamic range). Once digital improved in every way over film, those people moved too. Another reason is job security. Why would someone support something that will leave him homeless and destitute just because the software will be superior? Personally, I think that software engineering jobs will go up while AI assisted programming is perfected (10-20 years), and afterwards there will be a huge decline (10x shrinkage) once the tools won't need people anymore.
  • m0llusk 2y ago
    IDEs can be useful, but can also weaken developers. People get used to the IDE keeping track of bloated, overly complex interfaces. Autocompletion also leads to a style of development where every term must be defined before it can be used in any code. That may seem to make sense at first, but in practice means that it becomes impossible to freely sketch out code paths without first defining all terms. Depending on context this may all be appropriate and valuable, but it is worth keeping in mind that IDE saturated development has particular characteristics and may potentially amount to a crutch or even an impediment.
• noobermin 2y ago
  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 something of a niche whereas microwaves arguable have displaced no jobs in cooking and food preparation. We need more of an argument as to why GenAI will be more of a microwave technology vs a mill. Anyway, I actually am more towards this is a hype cycle that will pass, it's just the article itself didn't really drive home the argument in my opinion.
  • paulcole 2y ago
    > This article feels like stating an opinion You nailed it. This kind of thing should be saved for a personal blog and not printed in a scientific journal.
    • WhiteNoiz3 2y ago
      It literally says "my personal blog" at the top
      • fer 2y ago
        I could publish a peer-reviewed paper on how that was the joke.
        • paulcole 2y ago
          save it for your personal blog
          • BonitaPersona 2y ago
            wait, is Your Personal Blog a new peer-reviewed scientific journal?
  • 2y ago
• kachapopopow 2y ago
  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.
  • null_investor 2y ago
    That's incorrect, it isn't that better than h100s
    • kachapopopow 2y ago
      They chain a lot better and have much lower software complexity while also being able to share memory between two different nodes. Also the latency is much lower. (Also I'm sort of mocking LLM's here)
  • bearjaws 2y ago
    Maybe 4x, 100x is many years away. The exponential graph people are posting is just to pump the stock price.
    • HarHarVeryFunny 2y ago
      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.
• submeta 2y ago
  I for one used to hire people outside Germany for smaller projects, doing data analysis, writing interfaces, migrating data. With all the new tools I don't do this anymore. I can come up with solutions that are good enough, and I don't have the overhead with back and forth emailing, writing requirements, getting a result, giving feedback, etc. This is me, n=1.
  • randomdata 2y ago
    And now that your time and resources are freed up from working with data analysts, presumably you will shift that towards software engineers who can deliver new value to your customers. If not, your competition will.
• HarHarVeryFunny 2y ago
  AI isn't about to become capable of doing the job of a software developer any time soon, not least because two of it's major weaknesses - extremely limited reasoning ability and extremely limited learning ability (in-context only) are the two things that software developers most heavily lean on. It's not clear whether these are things that can really be addressed without major architectural change. There's still massive hype and group think about LLMs and what people project they will soon be able to do (or even what they can currently do - Dwarkesh's insider friends surprised it fails on Chollet's simple ARC tests!). A large part of this seems to be based on how new the tech is, and how capable it is, but people seem to be forgetting that the capability is based on brute force training time gradient descent. There is no runtime gradient descent or any alternative runtime learning method ... just limited in-context learning by example.
  • loudmax 2y ago
    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. But on the current trajectory of cheaper, more powerful compute and better understanding of the engineering around getting useful results out of AI, you'd be foolish to think any white collar profession won't be deeply affected by the technological shift that's coming.
    • ch4s3 2y ago
      > 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 thinking that text prediction is generalization enough that doing more of it better, faster, and on scaled hardware is sufficient to hit a tipping point?
• digger495 2y ago
  GenAI is a pipe dream.
• simpaticoder 2y ago
  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, architectures, and libraries.
  • passion__desire 2y ago
    If there is a way for LLMs to always suggest the best practices for a given task (e.g. upsampling on those tasks), then SE devs won't make trivial errors. So the knife can cut both ways. In fact it can become a source of uptodate information if augmented properly.
  • Lio 2y ago
    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.
    • simpaticoder 2y ago
      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.
• jillesvangurp 2y ago
  GenAI will indeed drop the cost of software creation and therefore increase demand for more software. Things that formerly were prohibitively expensive are now cheap enough to be doable. Which increases the market for software engineers that can get things done with the help of AI and other tools. These won't be the type of software engineers that specialize in things that should be automated but the type of engineers that can build a lot of stuff that formerly would have required them to delegate a lot of work to other engineers. This is how the software developer community has historically grown actually. It used to be that software engineers were faffing about with punch cards, assembly code, etc. Working months to produce a few kb worth of software. No libraries or anything. It all ran on bare metal. These days what engineers build is the tip of the iceberg because it all runs on frameworks, libraries, operating systems, etc. that do all of the heavy lifting. Which makes them way more productive than their colleagues from half a century ago.
• wiradikusuma 2y ago
  I'm writing a book about building mobile apps[1]. I've worked on it for over two years (I have a day job). When I started taking preorders, it coincided with the AI gold rush. I thought, "Fark, now people can just use AI instead of reading my book". But my sales are actually quite good (for a nobody). Thankfully, AI is still not there yet to make "professional" (as most people would define it) apps—unless maybe when people start feeding my book contents into GenAI, ha! (I guess it's a compliment?) Also, AI is like DB. Sure, you can use DBeaver or something to access it "raw", but 90% of us programmers are still hired to write a CRUD wrapper. Same thing with AI. It's more useful to have an app supercharged with AI wherever it makes sense. [1] https://opinionatedlaunch.com https://opinionatedlaunch.com