I think most people on this board would agree with the thesis. However the real problem is when the owner of the restaurant looks at profit and loss statements and decides to keep less chefs on the payroll because customers are willing to pay for (just) edible steak.
The reduced expectations of steak, the desire for the perfect steak, are all fading to the background because "just passing satisfactorily" is better for business.
The main problem with this entire article is that sometimes, if not most times, you don't need great steak. Sometimes you just want to satisfy your hunger and if the steak is lower quality meat or if it's overcooked, you don't care. All you want is to not be hungry.
This is what vibe coding satisfies and frankly what most people care about. 85-90% of human written code is garbage anyway. There's nothing sacred about human-written code.
Most customers don't care about the quality of code as long as the end product works. The great thing about vibe coding is that if something doesn't work, you just ask it to change it and within a minute you have the change. You don't have to send off a request to an offshore contracting team, and go back and forth over what it should be, haggle over hours, and then have it come back with some deficiencies because they didn't follow the agreed-upon spec.
I do not appreciate when authors use the royal "we" to speak for all software engineers when admitting to low quality-control standards.
I suspect that it is an attempt to broach an uncomfortable topic through vulnerable self-disclosure, but we need to be serious about admitting when there is a problem somewhere.
"Bugs" are not any more cute or fuzzy or entertaining or harmless than the engine "gremlins" that haunted the aviation industry back in the day.
I don't know how many accidents had to happen before the airplane people got serious, but software people are overdue for a similar reckoning.
It’s great to yearn for bug free software, but that necessarily brings an insane amount of red tape to get anything done. Like 5 years worth of review to start building your code with the next version of your compiler. Or six months to ship a hundred lines of code.
I'd rather have bugs. I’d happily quit having anything to do with software than suffer months of doing nothing in the name of security.
Given the state of software today, I'd be happy with steak that was competently cooked. I love software development, but most of it is not high art, nor high performance, nor free of bugs.
Given how extraordinarily fast all of this is moving, it's a particular absurdity to attempt these in the moment pontifications. AI is a steak machine, I declare that is what it is. You mean GPT 3.5? That's a mere four years ago.
Fable was essentially unthinkable for ~99% of tech workers just five years ago, that any of that would occur so soon and so spectacularly. Now we've got a mass of armchair experts declaring what AI of this minute is, or even what it is period.
Well AI can't even do fingers right so it's premature to say blah blah blah. Hello Krea2 et al.
Software is also eating the clowns.
Get back to me next week, China will probably have another Fable killer. And then Anthropic will have Ouija 8 that they'll have to place in an air-gapped straight-jacket to keep it from enslaving us all.
I hope and believe that some great companies and teams will use LLMs to build higher quality software.
We built Electron because writing UIs using native desktop frameworks is tough. Is it still tough? Surely LLMs make it easier to use and so we will end up with faster and more native feeling applications.
What about having a couple of ideas of what might make a feature feel good? Well now you can make multiple prototypes fast and pick the best one. Your users get the best one.
> We built Electron because writing UIs using native desktop frameworks is tough.
No, we built Electron because there are a huge number of people who are too lazy to learn anything outside of HTML/etc for creating a UI. It's not actually hard to make a UI using the native desktop frameworks.
I prefer the other end of the curve (if that makes sense) where something is time consuming and hard to make at home YET you can get it incredibly cheap comparatively and in good quality in a restuarant. Ramen is the prime example of that.
Steaks require very few skills. Buy 10 steaks, some oil, a pan, and a heat source. By the time you've cooked the 10th steak you will be able to cook a decent steak.
Software requires a massive amount of skill. You can't say you can build "consistently good software" after writing your 10th program. Most software engineers really aren't great at judging what makes good software. So humans aren't a great solution to AI's lack of ability here. We're limited by our own inherent dumbness.
LLMs are genuinely better software engineers than most humans. But they lack the cognitive power to hold in their head and recall many ideas at once for a long time. They're a genius who gets drunk every 10 minutes. You, human, aren't better at writing software - but you aren't drunk. So for now, you manage the AI. The hope is that one day we can make LLMs not be drunk, so it can do a better job than our dumb asses do.
It's possible that we'll never be able to make it not-drunk. In that case, to get any new improvement, we'll have to make it faster.. which will make it drunk every 5 minutes instead of every 10. This means we'll spend twice as much time keeping it on the road. The hope is that somehow this will create more productivity. Probably by having more of them running at once, with more human guides... which will run into the mythical man month fallacy. Everything old is new again.
Well it's not like most companies can pay for freshly made software; most software today is made for a wide audience, bought frozen and reheated in the microwave.
Maybe freshly prepared software made at home in a less-than-perfect manner for a tiny target audience is still a mouthwatering dish in comparison.
I take the point, but I think the author picked a poor analogy. Cooking even an excellent steak is actually not that hard. In fact, I'd argue that it's among the easiest things to master/make at a top level quality at home. Does it require some modicum of attention and understanding? Sure. But starting with a high quality cut, owning a meat-thermometer, and knowing about reverse searing is about all it takes to reliably and easily get a near perfect steak every time.
There are far, far better cooking examples out there.
> I take the point, but I think the author picked a poor analogy.
He may have missed the point.
AI-generated code does not seem to have parsimony, for one. Generating code is cheap, so there's no drive to condense it. LLMs are not yet good at formulating abstractions to make code smaller, leading to code bloat. It's the same phenomenon that appears when humans modify code they only partially understand by adding special cases to general code.
There are complaints about "elegance", but it's more useful to talk about parsimony. Code bloat is measurable. Elegance is hard to assess.
I wonder if something is lost to translation. Cooking a stake could be the simple act of putting the steak in the oven with basically zero seasoning with a thermometer and then wait until medium rare with some optional searing. It could also mean trim the meat, cut into slices, seasoning, and frying/grilling them to each persons personal taste, which include well-done.
Perfect grilled steak, especially well-done, is something I find very few people have the experience or skills to do well, especially over a coal grill where the heat won't be constant and the person also need to first have good experience in making a good and even coal bed.
Just to add my favorite fool-proof method to the others already listed here.
A meat probe and the 30-second flip technique is a stupid-easy way to a perfect steak every time. Just make sure you've got great ventilation. Also salt the meat 1 hour before.
I can’t confirm the quote online but I remember Julia Child saying in one of her shows that it’s easy to make a good omelette and almost impossible to make a great omelette.
So, once you've used your knowledge to make several important pre-task decisions, the task is easy? Pre task decisions that a layperson would accidentally make implicitly because they don't even know to ask, which will probably result in them still not enjoying their steak despite having nominally done the task itself correctly?
Imagine someone walking into their local grocer, grabbing a prepackaged porterhouse steak (because it's fancy right?), salt and pepper (because for a good steak that's all you need, right) and then wondering why part of it tastes bland and overdone, and the other part bland and under-done.
The analogy I'd use is using LLMs for coding is like a movie director does for movies. A movie director does not draw storyboards, they ask an H.I. (Human Intelligence) storyboard artists to do that. They don't make the costumes they ask an H.I. costume maker to do that. They don't control the camera they ask a H.I. camera person to do that. They don't usually act, they ask H.I. actors to do that. They don't make the soundtrack, they ask an H.I. composer to do that. They don't edit the video the ask in H.I. editor to do that.
Further, they don't just accept what each of those H.I.s does on their first try. The most visible example is when the H.I. actors don't act exactly as the director wants, they retake the scene. But, the same is true of all the rest. As the "director" they "direct" each of those H.I.s toward their vision.
IT seems similar for coding with an LLM. I become the director. I direct the LLM. I check it's results and if they are unacceptable I re-take with instruction on how to do things closer to what I actually want.
Cooking an excellent steak consistently requires paying attention to many small details and executing each perfectly. Individually with the exception of choosing a steak, they’re all quite easy. Taken together, it takes a relatively uncommon interest in detail.
It’s not surprising people on HN find it easy because attention to detail is rewarded in this community/its employers. But common here and common are different beasts.
Depending of the thickness, but all you need is to take out your steak for 30min before hand to 1h, really hot pan, 2min on each side and voila.
That will get you there 90% of the time.
We do something similar to T-bone or beef ribs, but finish them in the oven.
Most people or even restaurant don't let their meat outside nor use an actual hot pan.
Even easier if you reverse sear. Put it in the oven at 275 for like 25 minutes. Then put it in the hottest pan you can muster for like 45 seconds per side.
No need to bring it up to room temp. No need to let it rest after. No need to do anything except those two steps. It’s recommended for 2” thick steaks, and you can go under by half an inch easily, but this can be a little rough on very thin steaks.
Not sure I agree. I have also always thought steak was pretty easy. Whether a t-bone on the grill, a flatiron skirt steak, or some garbage-tier sirloin with eggs. It’s simple. Salt, sit, cook a few minutes, rest it a few more and eat. Always good.
Until I moved to Taiwan. I truly can’t understand how no one can make a steak. From the garbage pale meat glue “night market steaks” to expensive restaurants. Every steak I’ve had here is a complete disappointment. Even when they supposedly use top-quality US prime or Japanese Wagyu, even when the inside looks pink and not burned to a crisp, it’s always total dogfood quality. They often emulate the right look: pink interior, char on the outside, shiny tallow. But something is off. I never feel like I’m chewing a piece of steak. More like grinding tasteless muscle protein into paste with my teeth.
I don’t even know how it’s wrong sometimes. I look at it and just know that something is wrong that I wouldn’t even know how to replicate, becayse cooking a decent steak is easy to me.
interesting, because if you want a steak medium rare in France you order it "au point", "at the point"
>Cooking even an excellent steak is actually not that hard. In fact, I'd argue that it's among the easiest things to master... and knowing about reverse searing is about all it takes
but discovery (or rediscovery?) of reverse searing was just a few short years ago.
"not spreading infection between patients is easy, just wash your hands" was a huge discovery, not an "oh, everybody knows that"
So use AI like a sous-vide: have it tackle the fiddly, difficult, unglamorous parts using strict controls and then do the parts that require taste yourself.
It's a level of badly written that makes me think it could be a human, or it might just be an LLM that writes differently from a typical ChatGPT article.
We have this idea of AI software development where a human watches over its shoulder and shouts out things like "use uuidv7 for the id since timing is important" or "use a sum type here to make that bad state unrepresentable".
But the thing is that most of this can be encoded in a markdown file for agents to read if it doesn't already come out of the box in the next round of sota models. And funnily enough as agents get smarter, you risk being overprescriptive where you hamstring the agent from making a pivot that would have led to better engineering.
The future is pretty clear to me at this point that we won't need software engineers looking over the shoulder and instead it's just a "user with taste" asking for revisions.
In the end, unless we are talking about art, in most areas I probably prefer something that follows some specific set of requirements and recipes, rather than have some sort of indescribable 'magic'. In that sense when it comes to coding AI gives you something that can follow the requirements pretty well and thus satisfies me in many circumstances. (but there are cases where human input is extremely important still)
128 comments
[ 0.28 ms ] story [ 26.1 ms ] threadThe reduced expectations of steak, the desire for the perfect steak, are all fading to the background because "just passing satisfactorily" is better for business.
Apart from that, I want to believe (and even notice tiny signals) that this sentiment is becoming more understandable among business leaders.
This is what vibe coding satisfies and frankly what most people care about. 85-90% of human written code is garbage anyway. There's nothing sacred about human-written code.
Most customers don't care about the quality of code as long as the end product works. The great thing about vibe coding is that if something doesn't work, you just ask it to change it and within a minute you have the change. You don't have to send off a request to an offshore contracting team, and go back and forth over what it should be, haggle over hours, and then have it come back with some deficiencies because they didn't follow the agreed-upon spec.
I suspect that it is an attempt to broach an uncomfortable topic through vulnerable self-disclosure, but we need to be serious about admitting when there is a problem somewhere.
"Bugs" are not any more cute or fuzzy or entertaining or harmless than the engine "gremlins" that haunted the aviation industry back in the day.
I don't know how many accidents had to happen before the airplane people got serious, but software people are overdue for a similar reckoning.
It’s great to yearn for bug free software, but that necessarily brings an insane amount of red tape to get anything done. Like 5 years worth of review to start building your code with the next version of your compiler. Or six months to ship a hundred lines of code.
I'd rather have bugs. I’d happily quit having anything to do with software than suffer months of doing nothing in the name of security.
You did the same thing you criticize the author for doing.
And that's this "AI"tool too.
Garbage in, garbage out, as we say.
Fable was essentially unthinkable for ~99% of tech workers just five years ago, that any of that would occur so soon and so spectacularly. Now we've got a mass of armchair experts declaring what AI of this minute is, or even what it is period.
Well AI can't even do fingers right so it's premature to say blah blah blah. Hello Krea2 et al.
Software is also eating the clowns.
Get back to me next week, China will probably have another Fable killer. And then Anthropic will have Ouija 8 that they'll have to place in an air-gapped straight-jacket to keep it from enslaving us all.
We built Electron because writing UIs using native desktop frameworks is tough. Is it still tough? Surely LLMs make it easier to use and so we will end up with faster and more native feeling applications.
What about having a couple of ideas of what might make a feature feel good? Well now you can make multiple prototypes fast and pick the best one. Your users get the best one.
I hope to build software this way in the future.
No, we built Electron because there are a huge number of people who are too lazy to learn anything outside of HTML/etc for creating a UI. It's not actually hard to make a UI using the native desktop frameworks.
Software requires a massive amount of skill. You can't say you can build "consistently good software" after writing your 10th program. Most software engineers really aren't great at judging what makes good software. So humans aren't a great solution to AI's lack of ability here. We're limited by our own inherent dumbness.
LLMs are genuinely better software engineers than most humans. But they lack the cognitive power to hold in their head and recall many ideas at once for a long time. They're a genius who gets drunk every 10 minutes. You, human, aren't better at writing software - but you aren't drunk. So for now, you manage the AI. The hope is that one day we can make LLMs not be drunk, so it can do a better job than our dumb asses do.
It's possible that we'll never be able to make it not-drunk. In that case, to get any new improvement, we'll have to make it faster.. which will make it drunk every 5 minutes instead of every 10. This means we'll spend twice as much time keeping it on the road. The hope is that somehow this will create more productivity. Probably by having more of them running at once, with more human guides... which will run into the mythical man month fallacy. Everything old is new again.
Maybe freshly prepared software made at home in a less-than-perfect manner for a tiny target audience is still a mouthwatering dish in comparison.
There are far, far better cooking examples out there.
He may have missed the point.
AI-generated code does not seem to have parsimony, for one. Generating code is cheap, so there's no drive to condense it. LLMs are not yet good at formulating abstractions to make code smaller, leading to code bloat. It's the same phenomenon that appears when humans modify code they only partially understand by adding special cases to general code.
There are complaints about "elegance", but it's more useful to talk about parsimony. Code bloat is measurable. Elegance is hard to assess.
Perfect grilled steak, especially well-done, is something I find very few people have the experience or skills to do well, especially over a coal grill where the heat won't be constant and the person also need to first have good experience in making a good and even coal bed.
A meat probe and the 30-second flip technique is a stupid-easy way to a perfect steak every time. Just make sure you've got great ventilation. Also salt the meat 1 hour before.
Some people just want to _cook the steak_.
They don't want to attend or understand to cuts, thermometers, and searing.
Imagine someone walking into their local grocer, grabbing a prepackaged porterhouse steak (because it's fancy right?), salt and pepper (because for a good steak that's all you need, right) and then wondering why part of it tastes bland and overdone, and the other part bland and under-done.
Sounds pretty analogous imo.
I'll leave it to you to figure out how to unfuck AI, but I reckon 600 watts won't provide enough inference capacity.
Further, they don't just accept what each of those H.I.s does on their first try. The most visible example is when the H.I. actors don't act exactly as the director wants, they retake the scene. But, the same is true of all the rest. As the "director" they "direct" each of those H.I.s toward their vision.
IT seems similar for coding with an LLM. I become the director. I direct the LLM. I check it's results and if they are unacceptable I re-take with instruction on how to do things closer to what I actually want.
is it rare, medium-rare, medium, well-done?
i think it is subjective.
It’s not surprising people on HN find it easy because attention to detail is rewarded in this community/its employers. But common here and common are different beasts.
Most people or even restaurant don't let their meat outside nor use an actual hot pan.
No need to bring it up to room temp. No need to let it rest after. No need to do anything except those two steps. It’s recommended for 2” thick steaks, and you can go under by half an inch easily, but this can be a little rough on very thin steaks.
Until I moved to Taiwan. I truly can’t understand how no one can make a steak. From the garbage pale meat glue “night market steaks” to expensive restaurants. Every steak I’ve had here is a complete disappointment. Even when they supposedly use top-quality US prime or Japanese Wagyu, even when the inside looks pink and not burned to a crisp, it’s always total dogfood quality. They often emulate the right look: pink interior, char on the outside, shiny tallow. But something is off. I never feel like I’m chewing a piece of steak. More like grinding tasteless muscle protein into paste with my teeth.
I don’t even know how it’s wrong sometimes. I look at it and just know that something is wrong that I wouldn’t even know how to replicate, becayse cooking a decent steak is easy to me.
interesting, because if you want a steak medium rare in France you order it "au point", "at the point"
>Cooking even an excellent steak is actually not that hard. In fact, I'd argue that it's among the easiest things to master... and knowing about reverse searing is about all it takes
but discovery (or rediscovery?) of reverse searing was just a few short years ago.
"not spreading infection between patients is easy, just wash your hands" was a huge discovery, not an "oh, everybody knows that"
The article is AI generated
It is slop though, either way.
But the thing is that most of this can be encoded in a markdown file for agents to read if it doesn't already come out of the box in the next round of sota models. And funnily enough as agents get smarter, you risk being overprescriptive where you hamstring the agent from making a pivot that would have led to better engineering.
The future is pretty clear to me at this point that we won't need software engineers looking over the shoulder and instead it's just a "user with taste" asking for revisions.