Cooking a Steak Is Easy. Making a Good One With AI Is Not.
Almost no skill required to cook a steak
AI-generated software is like steak from a cheap restaurant: technically edible, rarely great. We churn out agents, prompts, and workflows, hoping for a perfect medium-rare, but often get charcoal. The fix? Learn to cook—understand the fundamentals, judge the output, and stop relying on luck. AI can speed you up, but it can't replace your judgment.
It can’t define quality for you, can’t decide which tradeoffs are acceptable, can’t always catch the moment when something is technically correct but wrong in every way that matters.
- MostlyStable
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.
- jifarooq
Like many other comments here, I take issue with the title. I thought I would learn how to cook a good steak. But instead it’s yet another AI article about how taste (read: judgement) matters.
- xtajv
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.
- asdfman123
God dammit I thought this was going to teach me how to cook a steak, not talk about AI.
- igiveup
This stuff should have some warning tag: rando's musings about LLMs. Lately, whenever I think I found an interesting article it turns out to be this.
- Havoc
>And what do we want?
>We want the perfect steak.
This premise seems incorrect. SWEs may aspire to produce perfect steaks as a matter of pride, but demand asks for mass produced steaks that trade off against other factor like time and cost.
There is a reason github isn't full of nasa lunar lander quality this can't crash code...
- fnoef
Either my brain is cooked or this is AI generated/assisted text.
The problem is I can no longer tell the difference between AI and non-AI writing. Maybe my brain got bad, maybe everyone now writes like AI writes.
I no longer know.
- dom96
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.
I hope to build software this way in the future.
- Night_Thastus
It would certainly help if it could taste what it produced. Or see, with its eyes, how the result is changing as its process changes. Or learn, from first principles, what is happening at each step and make deductions about how they will affect the result.
All it can do is randomly shuffle around the steps and ask if the customer likes it. If they don't, shuffle them around again and hope it's better this time.
- zippyman55
We would cook a roast yearly and I’d insert my thermometer and collect data points and then use R to produce. Regression curve and extrapolate a time dinner would be ready. It was fun as heck and the meal always came out right.