[ChrisMarshallNY]

I have found that it gets some of the "cruft" out of the work, freeing me to do more work.

Since starting to use LLMs, I have actually been spending more time, at the console, than before.

One reason is that I like to ship (as opposed to "code"). That means a lot of tedious, boring stuff. The kind of thing that I want to "take a break before tackling," so I may take 30 minutes, and watch something on TV for a while, before rolling up my sleeves.

Now, the LLM can take care of a lot of this stuff, so I am not motivated to "take a break," so much, anymore.

It doesn't actually feel bad, but I now have to schedule "downtime." I never used to have to do that, before. My work always involved a lot of "context switch" points; naturally set up for taking breaks.

[throwawayffffas]

Yeah I have found that AI in general has a procrastination nullifying effect.

Before dealing with anything that might put me off. I can just ask the agent to do it for me. And then, do something else, take that break, but regardless in a few minutes I will have something to jump on instead of the same blank terminal with the same blinking cursor judging me. It really makes taking the first step, much easier and then the ball just gets rolling.

I see what his point is to be honest though, it's easy to say just one more week of polish, just 5 more features, etc.

[theshrike79]

It's the empty page problem.

For some people (most?) it's really hardhard, almost impossible, to fill the page with text, but when you have a page of text - even if it's shit - suddenly it's easy to critique it and interate to make it better.

AI lets me poop out an MVP out of any idea or spec with very little effort on my part, then I can go in and start improving it.

[ChrisMarshallNY]

> MVP

I guess a lot of folks are having better luck than I am, in getting ship-quality results from LLMs.

In my case, I find that the first iteration is almost never suitable for shipping. In fact, the more intricate the functionality, the more likely it is, to contain some real showstopper bugs.

In my experience, iteration really requires that I know my shit. I have to know what to look for, and how to form effective prompts. Testing is absolutely vital. I can’t ever “just assume” that the code is fundamentally sound; even for extremely basic stuff.

That lack of trust in my output is something I haven’t had, in decades. It can be a bit stressful.

For example, I write a lot of mobile software, so resource usage is a big deal. I had gotten used to not testing for battery usage, leaks, or out-of-control allocations, over the years of using Swift. I am now back to routinely running Instruments.

In a couple of instances, I was just unable to get anything useful from the LLM, had to toss all its output, and rewrite by hand.

That’s why I can’t even imagine directly shipping anything the LLM gives me.

That said, I have generally had extremely positive experiences with LLMs. It’s now a basic component of my workflow.

[theshrike79]

If you never ship, how can you iterate?

Do you spend weeks running benchmarks and optimising your code without asking/observing a single actual user whether they care if that bit is fast or not?

Or would it be more efficient to ship something, gather data and iterate where relevant?

Because if you're not doing that, I can guarantee other people are. Sadly being the first and moving fast (adding features as users request them) is in many cases better in the mobile market than having the snappiest app with zero memory leaks and excellent battery efficiency.

The whole Turkish mobile gaming industry works on this principle, they ship like crazy and iterate quickly to see what sticks.

[feoren]

> If you never ship, how can you iterate?

Our discipline has swung way too far in this direction. The useful insight of "watch your users use your product -- you will always be surprised about something" has turned into "you cannot possibly have any vision or produce a useful product without constantly probing your users".

A/B testing, design by committee -- it reeks of cowardice and produces garbage. It reminds me of an SNL skit where politicians in a debate were able to see real-time social media responses to their talking points, and they immediately started playing a sort of "hotter/colder" game, changing their politics to try to get the graph to go up.

I promise you it's possible to just have an actual vision yourself. Yes, talking to users will surprise you, and it's important. But the idea that you cannot possibly iterate without shipping code is way, way too extreme.

[ChrisMarshallNY]

It’s an old concept, called “taking risks.” Corporations and executives have been doing it for centuries. In fact, that's the excuse that executives give, for their monstrous salaries.

I think we have a culture, where people genuinely believe that we have some “right” to a risk-free fortune.

Training users to pay for unripe garbage, is probably one of modern Marketing’s biggest coups. It affords the “MVP model,” that everyone uses.

The other big coup, is “The Company is the Product, the User is the Resource, and the Investor is the Customer” model.

Steve Jobs famously, made many errors, in his career. Probably most of his ideas were duds, but the ones that scored, scored big.

He was also, famously, a world-class prick; but he wasn’t a misanthropic prick. He actually liked people (certain ones, more than others, I guess), and had a way of understanding them. That figured prominently, in his product ideas. Many folks are misanthropes, and refuse to understand people, so they have difficulty coming up with ideas that people might like.

Taking risks, and "putting our money where our mouth is," is the “old-fashioned” way of doing it.

[ChrisMarshallNY]

It depends on the target market.

The mistake many people make, is casting what’s good for the goose, onto the gander.

I Serve a demographic that can suffer great harm, from half-baked crap. Also, there’s no money to be made, so what’s the hurry? If someone else comes up with something better, that’s actually a good thing.

I have been shipping software for my entire career. It’s quite possible to ship good stuff.

It’s just more work, and less financially rewarding, so fewer people do it. It just gets my goat, to hear people say that it’s not possible.