> The meta at this point seems to be mostly around "knowing what to ask for"
We came full circle on the problem that also exists with Google Search, and why so many people don't get good results out of it.
> For now, I have an advantage over these people when building applications. But perhaps that's just one round of RL away from being obsolete too.
I don't have a crystal ball, but my eggs are in the basket of "It isn't, and it won't be"
> The only thing that seems relatively future-proof is just being able to communicate clearly what you're trying to do
This somehow contradicts the previous statement. The advantage you have is, because of domain knowledge expertise, you're able to articulate what you want. It's not that you knowing what LazyLoading is that you now can describe it better, you can immediately tell that when your Frontend stalls for 5 seconds after hitting the "Load" button, perhaps it eagerly loads and blocks.
> We came full circle on the problem that also exists with Google Search, and why so many people don't get good results out of it.
yeah. I spent the first 15 years of my career being amazed at how many problems people came to me with which I was able to resolve in 10 minutes with a simple google search.
I suspect I will spend the next 15 years of my career being amazed at how many problems people come to me with which I am able to resolve in 10 minutes with a simple LLM prompt.
I am probably not the only one who feels embarassed when I copy-paste a large error output to an LLM and it points out to me that the error message very clearly said what I needed to do already.
I think the knowledge is one level higher at this point. It’s not LazyLoading that you need to know about, it’s just that it can and should be faster. That you can demand more (and sometimes you can’t). It’s hard to know the difference when you’re not in your area of expertise.
Fair point on the contradiction. I don’t have the answers either.
If we figure out long horizon tasks, why wouldn't you be able to ask it to send out a user survey, prioritize the next iteration of work based on the results, and then execute on it?
These things are already figuring out how to use hidden message boards to pass information between themselves for long running tasks. I see no reason we wouldn't be able to automate the whole product life cycle eventually.
Programmers can’t even agree on tabs vs spaces, less alone something important like strict vs weak typing. That survey is going to be a mess, even if you limited it to the most senior engineers. About the only thing we’ll be able to agree on is “source control good”.
i have been vibing out a godot game which is a 3d voxel map thing lately. i had it on software webgl compatibility output and just demanded it figure out what the slowdown was on the simulation over 3 days of back and forth. it turned off backfaces and invisible geomeetry and locked render resolution at 1080p and built in diagnostic telemetry and chopped up the level where doors separate rooms until now, still with no hw acceleration, the framerate is in 60fps with some frame drops, it started at 12 fps.
brute forcing performance optimization for a personal project by having only handwavy tricks ive read about from online and repeatedly setting targets until the magic box gives me what i know it can do means that eventually the game did what i wanted with those. yes baked lighting was needed.
next, i an going to aim for 60fps on a windos xp box just for the challenge. we had quake and such run blisteringly fast on weaker boxes than those, surely my empty museum tour of a 3d game can do the same with the graphical voxel and crappy looking models it does now.
real pro grade work is beyond vibing out and not caring about poor quality. but it scratched my itch and gave me the computer software i wanted and with way less effort than doing it for realz would require. i love programming itself on its own time (going to learn nasm and someday do a hexagonal architecture with actor model in idris just to add up my nerd trophies together) at times, but using llms lets me do my random hairbrained ideas much more easily for what i put into it than not existing in the past.
I would draw the difference here that a library is used by hundreds (of thousands) of people, and established across different scenarios. I don't read the boto3 library AWS provides, but I can trust them and the amount of customers enough to be certain enough that it behaves the way I expect it to.
The same can't be said with code we write in silos at our workplace or at home. It simply does not have the same test bench.
Yes, libraries aren't bug-free, but they give me a reliable abstraction tested in the field. Not rarely you dig into library code if you notice unexpected behavior.
If we could rely on our LLM or colleague written code, or own code, have run through the same amount of requests, sure I wouldn't need to review it, as my confidence can be north of 99.9999% it works correctly. But we can't.
Another example of "More AI, worse app" is amazon.com for me lately. Alexa AI Chat keeps showing whenever I scroll up on the amazon.com mobile page. I don't want Alexa to search for products. Worst thing is, I can't close it. The "X" button just isn't doing anything. So the only way out of it is to refresh the page and avoid scrolling up ...
This has been going on for at least 1 1/2 months by now, and I wonder if it'll ever be fixed.
Genuine Q: Are they buying failing businesses because they see potential and want to get those on the right track, or are they buying them because their forecast tells them they can still make money, before shutting them down for eternity?
It doesn’t matter to PE as long as they can make a profit.
Buy a failing business and leverage debt on it until you can’t, spend the money on yourself restructuring the business.
Buy a working business and gut it for multiple.
Buy failing businesses to offset tax burden.
This a meta level game, they don’t care about the outcome as long as it produces profit.
Product quality goes down the drain? That product is retirement homes, healthcare, food, utility, schools, it doesn’t matter. There is a million levels of separation and paperwork and a corp structure to prevent shit swimming up the stream.
Net result, if you live in America look around you, go to an auto service center, or a clinic. It’s been gutted for profit one way or another.
Wasn't there an article here, or was it under another HN post as a comment? Someone pointed out that the before the dawn of LLMs, "bad contributors" had a rate limiter of sorts. Now anyone can prompt their LLM and throw their output at you, shifting the burden entirely on the reviewer at highest possible speed.
It's so much worse than just that. Because they will also decide on how problem should be solved by prompting LLM and ask you to argument doing it any other way vs the LLM slop they pasted
Not as a rule for all prs you want to reject, but any time it seems to be no-effort pure ai. Both to waste as little of your own time and effort as possible, but also to to help make this activity not rewarding.
I wouldn't worry about false positives and being unfair. This is self defense and also community defense. Anyone legit will manage to get your attention. You'd be filtering not on any specific detail or rule but on simply "care".
If a legit person fails to convince you that they are a human that cares, then that's just an unfortunate industrial accident. It's unfortunate but trying to do anything about it is worse.
Besides, any actual adult understands that if they can't prove their age right now, then they can't buy a drink right now, even though they normally have that right, and is never the one trying to cry about the travesty of justice at being denied. If I submit a pr that fails your ai sniff test, I get it and don't think you're abusing me. I can try harder and we can be friends later no problem at all.
I put something like that right into AGENTS.md in one of my open source repositories -- disclose the usage of AI (it's super obvious anyway) and be prepared to act on review comments, otherwise the PR is closed and submitted could be banned depending on my mood. And I only got like 3 PRs before I put it.
Those weren't even that bad code-wise, some have uncovered realy problems, but I ended up asking /my/ LLM to ponytail them into shape, because the submitter didn't bother.
> And, in line with Ray Kurzweil’s predictions from the end of the XXth century (opens in a new window), we now find ourselves at the moment in history of computing where machine intelligence is starting to exceed that of humans in transformative ways.
It is kind of strange to see this sentence, when OAI's definition of what AGI is has been watered down throughout the years.
> I expect and hope for voluntary slowdowns to become commonplace until shared safety bars are established.
Read: Please play by our rules, so we can be the first.
Yes it's super misleading. All the marketing and title is calling it "electric" where in reality it's a hybrid electric-gas airplane with a 125 mile electric range. Aviation uses nautical miles yet they advertise 125 statute miles so the number seems bigger (125 statute miles = 108.62 nautical miles).
It would be like Toyota marketing the Prius as an electric vehicle, then buried in their presentation sneaking in the gas propulsion. Same with Chrysler Pacifica (32 mile electric range).
> Core parts of the product, like navigating to individual files in a code review, are broken
I think this is a good argument to underline "It's not _just_ the scale".
Adding to this, the Github Code Review experience is kind-of broken, the way comments/threads are stacked in the PR overview has not improved, pagination isn't really a thing, and these issues are age old.
Hopefully, one day, Github will mature.
> Modern cloud systems shouldn't even have shell access, immutable containers with telemetry data.
Genuine question, what is your alternative suggestion to accessing lower level features then?
A hard time I have with GUIs is that their input/output can't be piped. It's much more cumbersome to try to automate a GUI than to automate something in the terminal.
Whilst I get that this is about TUIs, I do believe there is a reason for shell access due to automation. Good abstractions can be put into the GUI, but I'd argue that you should first build the abstraction, then put a GUI on top of it. AWS is a prime example, any of their APIs can be used w/out AWS Console.
devdraw can be thought of as a 2D rendering engine. You load assets like text, bitmaps and so on then issue draw commands to render them. There is no SSH. Instead you connect to a remote machine and export resources served by your local machine including your keyboard, mouse and display, each of which is a network transparent microservice. The window manager, Rio (and others), multiplexes access to these microservices. You can run a graphical program without a window manager, e.g. you can boot right into Doom or the Acme editor. It's seamless and can run over any pipe that can push 9P including ssh. If you look at the Sam and Acme text editors you will find they are true modern TUI programs.
Shame we had a cloud ready OS for 35+ years and no one noticed. Instead were living in the 60's with GPU accelerated typewriter emulators.
However I would rather prefer its successor, with Limbo for userspace, although probably not Tk as the main UI toolkit.
The terminal thing emulating typewriter are a UNIX thing, as other graphical OSes, presented a plain REPL, or graphical shell, e.g. Amiga DOS did not emulate any tty.
> > I'm using my skills to orchestrate LLMs and agents, and I can write better code much faster.
>
> The fact that the top comment on this thread calls it out, in a negative way, hints at that you aren't.
It's interesting that what I meant as a purely factual question with no prejudice either way has been taken by almost everyone to be explicitly negative & critical.
I need to be a tad less blunt I guess if I don't want to
We came full circle on the problem that also exists with Google Search, and why so many people don't get good results out of it.
> For now, I have an advantage over these people when building applications. But perhaps that's just one round of RL away from being obsolete too.
I don't have a crystal ball, but my eggs are in the basket of "It isn't, and it won't be"
> The only thing that seems relatively future-proof is just being able to communicate clearly what you're trying to do
This somehow contradicts the previous statement. The advantage you have is, because of domain knowledge expertise, you're able to articulate what you want. It's not that you knowing what LazyLoading is that you now can describe it better, you can immediately tell that when your Frontend stalls for 5 seconds after hitting the "Load" button, perhaps it eagerly loads and blocks.
reply