1) I'm incredibly allergic to hype. To me, LLMs are very technologically impressive. I don't doubt that they're useful for many things – adversarial code review (including finding exploits), refactoring, search and math exploration are some shoo-ins in my view. However, these and other applications speak for themselves. They are impressive. They don't need people running around telling everyone how they should use more LLMs. How "the old ways" are obsolete, etc. Awesomeness does not need a fanclub.
2) Usefulness in some areas doesn't necessarily extrapolate as well as the fanclub seems to think.
3) The fanclub happens to be aligned with some pretty unsavory people, and some powers that have very little regard for our shared planet. This is, of course, not the fault of the fanclub, but many in the fanclub certainly could do a better job distancing themselves from certain people and acknowledging certain regulatory necessities.
4) I think this revolution has revealed a dichotomy in the set of people who enjoy programming: those for whom the end goal reigns supreme, and those for whom the journey is the point. You yourself seem to be in the former group. As a member of the latter, I have to say we feel a bit invisible. We're also often accused of wanting to halt progress so that we can keep doing what we want. I think that's an unfair characterization (I won't go into details of why here).
5) A lot of people in a geeky community like this are naturally skeptical of relying on things that we ourselves can't control. It's part of why the FOSS movement succeeded. This is all very much on a collision course with at least SOTA LLMs.
6) A lot of us do intellectual work. We therefore rely on a functioning system of intellectual property. It seems that a large fraction of the pro-AI crowd subscribe to the idea that passing IP through an LLM can strip it of its original ownership. For points 1-5, I believe we should have a nuanced discussion and try to understand each other. On point 6, though, I think these people have lost their minds. I completely fail to understand how they themselves don't think they'll soon encounter face-eating leopards if their worldview holds water. There seems to be very little acknowledgement of this, and it makes me angry.
I am not sure of that, may be in Europe, cannot see the U.S. banning. May be some states but highly skeptical there. Better chance of it flopping or doesn’t get to a good enough threshold to matter.
Are you saying you leave it up to the LLM to judge whether your idea is good or not? Are you even human anymore?
(I am not saying LLMs can't be a good tool in evaluating ideas. To me, it sounds like you're firing off ideas all over, letting the LLMs judge what's good and what's not. Insane.)
Not judge no. Implement and create a working MVP, yes of course.
And yes, I fire off ideas all over. Many require predicting the future to decide what to focus my individual effort on. This is a terrible way to do things because humans (and LLMs) are notoriously terrible at predicting the future. The gold standard is to try everything and eliminate what doesn't work. This is impossible using human labor. With LLM labor, it's simply a matter of relatively cheap money.
It's amazing. Technical problems are now no longer having to predict what the best implementation is. You can just try each one.
Again, no need to have an LLM judge, because the metrics that define 'better' are well-defined, and this is the interesting part of computer science, not the implementation.
I find it terrifying that people are willing to outsource thinking. Outsourcing thinking to an entity that is opinionated about what to think is beyond crazy.
What’s the difference between outsourcing thinking and using an LLM as a research tool?
An LLM with fetch/search is going to be a lot more effective than myself and Google. I would _never_ ask questions like this if the LLM wasn’t able to look up data
Sure, as long as all the code in all those unsafe blocks is sound! If not, the program is UB and all bets are off for every line of the codebase.
I think Rust strikes a perfect balance between a safe default and, as you say, "localized unsafety". Said localized unsafety is however only localized as long as you're "doing it right". I would absolutely not trust an LLM to do it right for hundreds of thousands of lines of translated code. This is insane.
Right, but the point is that you now have an explicit todo list of blocks to fix. You can trivially enumerate all sources of unsafety, and when a grep for unsafe blocks turns up empty you know your codebase is memory safe. When could you say the same about your Zig or C/C++ codebase?
I'm writing this as someone who doesn't even really like Rust; I'd probably prefer to write Zig! But those unsafe blocks definitely buy you something.
For normal use of unsafe in Rust, I completely agree, and I love the concept. But if you have a gazillion unsafe blocks written by someone (something) you don't trust to at least try to do the right thing, you're bound to have unsoundness in one of those blocks. And now your entire codebase is UB.
I don't see how this is any different from every line trailing with a comment of the form "FIXME: This line might be wrong".
And I say this as something of a Rust fanboy. I love the way unsafe blocks work, and the "locality of danger" they give you. But that all goes out the window if there's a gazillion haphazardly written such blocks.
They have a gazillion such blocks today. If they still have a gazillion such blocks in a year, sure. But presumably the plan is to replace them with safe Rust.
> > Evaluating whether it’s actually correct and maintainable is the hard part.
> But AI can also do that.
Citation needed.
> So, what’s the point?
The point is that there haven't been broad demonstrations of your claim.
> And if you think it can’t, wait one more year
You surely must understand that this isn't an argument? How many hundreds of billions have been burned through now? Yet we still have to suffer "soon" as an argument? I can't take any of this seriously anymore.
PS: Just to be absolutely sure you don't misunderstand me: I am NOT claiming that AI will never be able to do this stuff. Nor am I even claiming that it's too far off or too expensive. Just, for the love of god, you cannot build an industry on promises of how amazing it'll be in the future. Technology is evaluated based on how it performs. Not how you think it might perform in the future.
PPS: The last paragraph does also not mean that I think it's bad to invest in things that haven't yet paid off. On the contrary! What I am saying is you cannot claim success until there's success!
> I've yet to hear an argument that argues that software engineers can be replaced by AI that doesnt boil down to slop apologism, inability to detect slop or simple gaslighting.
That, or extreme extrapolation from events that form a vanishingly tiny part of the job of a software engineer. "Last week my AI solved this amazing software problem that I had struggled with" very quickly becomes "the AI is better at software than I am". Any pushback suggesting that the fact that something (or someone) did one tiny part of your job better than you one time does not mean you should be replaced, is quickly met with "yeah, but that's today, imagine how amazing the models will be in n years".
You can't win a debate with this much moving of goalposts.
1) I'm incredibly allergic to hype. To me, LLMs are very technologically impressive. I don't doubt that they're useful for many things – adversarial code review (including finding exploits), refactoring, search and math exploration are some shoo-ins in my view. However, these and other applications speak for themselves. They are impressive. They don't need people running around telling everyone how they should use more LLMs. How "the old ways" are obsolete, etc. Awesomeness does not need a fanclub.
2) Usefulness in some areas doesn't necessarily extrapolate as well as the fanclub seems to think.
3) The fanclub happens to be aligned with some pretty unsavory people, and some powers that have very little regard for our shared planet. This is, of course, not the fault of the fanclub, but many in the fanclub certainly could do a better job distancing themselves from certain people and acknowledging certain regulatory necessities.
4) I think this revolution has revealed a dichotomy in the set of people who enjoy programming: those for whom the end goal reigns supreme, and those for whom the journey is the point. You yourself seem to be in the former group. As a member of the latter, I have to say we feel a bit invisible. We're also often accused of wanting to halt progress so that we can keep doing what we want. I think that's an unfair characterization (I won't go into details of why here).
5) A lot of people in a geeky community like this are naturally skeptical of relying on things that we ourselves can't control. It's part of why the FOSS movement succeeded. This is all very much on a collision course with at least SOTA LLMs.
6) A lot of us do intellectual work. We therefore rely on a functioning system of intellectual property. It seems that a large fraction of the pro-AI crowd subscribe to the idea that passing IP through an LLM can strip it of its original ownership. For points 1-5, I believe we should have a nuanced discussion and try to understand each other. On point 6, though, I think these people have lost their minds. I completely fail to understand how they themselves don't think they'll soon encounter face-eating leopards if their worldview holds water. There seems to be very little acknowledgement of this, and it makes me angry.
reply