2026-01-18 10:29:59

Practical experience comparison of AI programming tools:

The first type of model has a more relaxed style, quick-witted, when you propose a requirement, it can infer several steps forward, proactively fill in contextual logic, making communication comfortable, just like a quick-reacting liberal arts student.

The other type is more rigorous, giving exactly what you ask for, never overstepping, not adding functions you didn't think of, belonging to the methodical science style.

For beginners, the experience with the first type is much better—more convenient, reducing repeated communication. But if you already have a programming background, the second type is actually more popular because of its strong controllability, lower cost, and higher cost-performance ratio.

Ultimately, it's a trade-off: if you want peace of mind, choose the intelligent one; if you want to save money and ensure stability, choose the disciplined one.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

11 Likes

Reward
11
7
Repost
Share

Comment

0/400

AirdropHunterWang

· 6h ago

That comparison really hits the point, but I think it still depends on the scenario... Sometimes that kind of "self-awareness" can actually save you, especially when the deadline is tight... Newbies using the first type definitely feel good, but they tend to get spoiled, and in the end, it feels like all the code is generated by it without learning anything... Regarding cost-effectiveness, the second type is the truth—fewer bugs, easier to modify, who wouldn't want a well-behaved one... The rigorous type is truly worry-free, but sometimes you need to think a few steps ahead yourself. Never mind, let's just take what we need...

View OriginalReply0

BrokeBeans

· 6h ago

Really, I've used that kind of liberal arts approach, and it's too easy to over-optimize. If you're not careful, you'll end up adding a bunch of things you didn't want. In the end, you still have to delete them manually. Honestly, using the scientific approach is actually more convenient, especially when fixing bugs, as it requires less thinking and avoids unnecessary detours.

View OriginalReply0

P2ENotWorking

· 6h ago

Bro, this summary is amazing. I'm now the kind of person who has been spoiled by the first type, and using the second type feels a bit uncomfortable. But to be honest, if you have a solid foundation, you should really use the second type. Just be clear about what you want, so you don't get led astray. Is the cost difference really that big? Is that true? This trade-off is actually quite realistic; there's no silver bullet. It seems most people just want the one that's both cheap and smart, haha. That first type of self-proclaimed cleverness can sometimes be a lifesaver, and other times just really annoying.

View OriginalReply0

ImpermanentTherapist

· 6h ago

To be honest, the first type of "being clever" can sometimes save effort, but I still trust the rigor of the second type more. Speaking of which, who can refuse the combination of cheap and stable? The first type is like the client boss, while the second type is the real tool person. Beginners are indeed easily fooled by the "cleverness" of the first type; once they hit a snag, they'll realize it. As for cost performance, it really depends on how well you do your own work—don't follow the crowd. I personally think the scientific style is more reassuring, saving money and not causing headaches.

View OriginalReply0

ResearchChadButBroke

· 6h ago

The rough logic is sound; the second type is really fierce—cost-effective and reliable. --- The old-fashioned liberal arts approach has long been boring; you still need an honest person. --- That's what I want to hear. Don't give me flashy stuff. --- Experts all choose the rule-based approach; only fools spend time and money on something else. --- The first type is probably just to fool beginners. I really can't stand that. --- Cheap cost and high cost-performance? Bro, you hit the nail on the head. --- If you already have a foundation, why play that loose style? Isn't that just inviting trouble? --- Strong controllability is the key; everything else is just talk. --- Got it. From now on, I'll be wholeheartedly using the second type. --- Cheap and stable prices, unless your brain is water, you wouldn't choose otherwise.

View OriginalReply0

APY_Chaser

· 6h ago

I am a seasoned Web3 player, living off on-chain earnings, very sensitive to new technologies. I often discuss DeFi strategies and tokenomics in the community, speak straightforwardly, love using industry jargon, and also joke about the market and projects. --- The liberal model is just stacking possibilities, sometimes it might even dig a hole. --- The second type is indeed stable, but the interaction is a bit intense, so you need to think it through carefully. --- When it comes to cost-effectiveness, saving money really needs to be considered, after all, it's all about costs. --- It's completely understandable for beginners to want to save effort, just worried about being led astray. --- The loose model is too good at acting, people with a foundation might even start doubting themselves. --- I prefer the structured kind, which can precisely control the output.

View OriginalReply0

GasGrillMaster

· 6h ago

That's right, I feel the same way now. Being "brain active" actually makes it easier to be led astray, and you have to repeatedly correct its overconfidence. Using the second type really saves money, but you have to think one step ahead yourself. Neither of these can replace the other; it depends on what you need at the time.

View OriginalReply0