Their profile generally comes up here on HN very often with Dunning-Kruger effect like comments so it makes me believe it is no AI. AI would do a better analysis, for the better or worse.
Small prompts leading to large programs has absolutely nothing to do with programming languages and everything to do with the design of the word generators used to produce the programs — which ingest millions of programs in training and can spit out almost entire examples based on them.
People learn by example. They want to start with something concrete and specific and then move to the abstraction. There's nothing worse than a teacher who starts in the middle of the abstraction. Whereas if a teacher describes some good concrete examples the student will start to invent the abstraction themselves.
Based on what I observe as an occasional tutor, it looks like compiler warnings & errors are scary for newcomers. Maybe it's because it shares the same thing that made math unpopular for most people: a cold, non-negotiable set of logical rules. Which in turn some people treat warnings & errors like a "you just made a dumb mistake, you're stupid" sign rather than a helpful guide.
Weirdly enough, runtime errors don't seem to trigger the same response in newcomers.
Interesting angle: Compiler errors brings back math teacher trauma. I noticed Rust tries to be a bit more helpful, explaining the error and even trying to suggest improvements. Perhaps "empathic errors" is the next milestone each language needs to incorporate.
I suddenly understand part of why experienced programmers seem to find Rust so much more difficult than those who are just beginning to learn. Years of C++ trauma taught them to ignore the content of the error messages. It doesn't matter how well they're written if the programmer refuses to read.
Interesting. I think over the long term many people come to realise it's better to know at compile time (when they mistype something and end up with a program that runs but is incorrect it's worse than not running and just telling you your mistake). But perhaps for beginners it can be too intimidating having the compiler shout at you all the time!
Perhaps nicer messages explaining what to do to fix things would help?
That's surprising because runtime debugging depends on the state of the call stack, all the variables, etc. Syntax errors happen independent of any of that state.
I think languages with strong support for IDE type hints as well as tooling that takes advantage of it are a fairly recent phenomenon, except for maybe Java and C# which I think are regarded by the wider hacker community as uncool.
C++/C IDE support is famously horrible owning to macros/templates. I think the expectation that you could fire up VS Code and get reliable typescript type hints has been a thing only for a decade or so - for most of modern history, a lot of people had to make do without.
It’s a remarkably powerful computer for a decent price. I like this new direction from Apple and also think chromebooks were a great idea even if the execution on them hasn’t always been great. I don’t see how they are disposable, perhaps for you their price makes them so? For many it just means they are actually attainable unlike other Mac products.
If you’re looking to a corporation to save you from corporate lock-in I’m not sure what to say.
Probably. One common feature of LLM output is grammatical features that indicate information density, like nominalizations, longer words, participial clauses, and so on. Perhaps training tasks that involve asking the LLMs for concise explanations or summaries encourage the use of these features to give denser answers.
They can work well when sparingly used and well thought-out, unfortunately LLM use is more on a par with:
‘It’s not mashed potato. Its potatoes lovingly mixed to perfection with butter and milk which quietly dominate the carrots beside them.’
The words are in the right order, th grammar is ok, but the subject is so banal as to undermine the melodramatic style chosen and they often insert several per paragraph.
At this point it’s pretty easy to detect unaltered LLM output because it is such bad writing. That will change over time with training I would hope. At some point I imagine it will be hard to tell.
I honestly don’t know what sites like this will do when that happens and the only way of detecting LLMs is that they are subtly wrong or post too much, we’d be overrun with them.
Not sure if we should be hopefully or fearful that they will improve to be undetectable but I suspect they will.
> That will change over time with training I would hope.
There's precious little training material left that isn't generated by LLMs themselves.
Consider this to be model collapse (i.e. we might be at the best SOTA possible with the approach we use today - any further training is going to degrade it).
> There's precious little training material left that isn't generated by LLMs themselves.
Percentage-wise this is quite exaggerated.
> Consider this to be model collapse (i.e. we might be at the best SOTA possible with the approach we use today - any further training is going to degrade it).
You consider this above factor to lead to model collapse? You’ve only mentioned one factor here; this isn’t enough. I’m aware of the GIGO factor, yes. Still there are at least ~5 other key factors needed to make a halfway decent scaling prediction.
It is worth mentioning one outside view here: any one human technology tends to advance as long as there are incentives and/or enthusiasts that push it. I don’t usually bet against motivated humans eventually getting somewhere, provided they aren’t trying to exceed the actual laws of physics. There are bets I find interesting: future scenarios, rates of change, technological interactions, and new discoveries.
Here are two predictions I have high uncertainty about. First, the transformer as an architectural construct will NOT be tossed out within the next five years because something better at the same level is found. Second, SoTA AI performance advances probably due to better fine-tuning training methods, hybrid architectures, and agent workflows.
> There's precious little training material left that isn't generated by LLMs themselves.
> Percentage-wise this is quite exaggerated.
How exaggerated?
a) The percentage is not static, but continuously increasing.
b) Even if it were static, you only need a few generations for even a small percentage to matter.
> You consider this above factor to lead to model collapse? You’ve only mentioned one factor here; this isn’t enough. I’m aware of the GIGO factor, yes. Still there are at least ~5 other key factors needed to make a halfway decent scaling prediction.
What are those other factors, and why isn't GIGO sufficient for model collapse?
I wouldn't say it's "bad writing", but rather that the sheer volume of it allows the attentive reader to quickly identify the tropes and get bored of them.
Similar to how you can watch one fantastic western/vampire/zombie/disaster/superhero movie and love it, but once Hollywood has decided that this specific style is what brings in the money, they flood the zone with westerns, or superhero movies or whatever, and then the tropes become obvious and you can't stand watching another one.
If (insert your favorite blogger) had secret access to ChatGPT and was the only person in the world with access to it, you would just assume that it's their writing style now, and be ok with it as long as you liked the content.
Attention grabbing tricks like binary opposites overused constantly
Overuse of adjectives and adverbs in particularly inappropriate places.
Lack of coherence if you’re generating large bits of text
General dull tone and lack of actual content in spite of the tricks above
Re your assertion at the end - sure if I didn’t know I’d think it was a particularly stupid, melodramatic human who didn’t ever get to the point and probably avoid their writing at all costs.
> At this point it’s pretty easy to detect unaltered LLM output because it is such bad writing.
And yet people seem to still be terrible at that. Someone uses an em-dash and there's always a moron calling it out as AI.
> I honestly don’t know what sites like this will do when that happens and the only way of detecting LLMs is that they are subtly wrong or post too much, we’d be overrun with them.
My personal take is that it doesn't really matter. Most posts are already knee-jerk reactions with little value. Speaking just to be talking. If LLMs make stupid posts, it'll be basically the same as now: scroll a bit more. And if they chance upon saying something interesting then that's a net gain.
“The underlying model that brings Grok to life is a voice-to-voice model which understands the expressive range of human speech... The model is able to do this because of how it internally, within a single model, processes speech (including paralinguistic cues) and generates expressive speech output.” https://blog.livekit.io/xai-livekit-partnership-grok-voice-agent-api/
reply