Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Agentic Coding Is a Trap

larsfaye.com

Post reply on HN
Print view View original

221–230 of 403 posts

Re: Agentic Coding Is a Trap

#221

Post by leptons »

ex-aws-dude wrote: post #9

The thing is the code quality is still ultimately up to you Nothing stopping you from iterating with the agent till the code is the exact same quality that you yourself would write

>Nothing stopping you from iterating with the agent till the code is the exact same quality that you yourself would write

Yeah, but in my experience, it takes the same amount of time or longer to cajole the AI to get it there. I'd rather write it myself and know how it works than insert an LLM as the middleman, especially when it isn't really proving to be any faster.

Re: Agentic Coding Is a Trap

#222

Post by ivolimmen »

Slightly off-topic.

The funny thing is that he mentions that Spec Driven Development is the future.

Technically we allready did this when we where doing Waterfall. I kind of miss it that we had good documentation. The last decade (maybe more?) I get jira tickets with a one liner. I often need to call people as they specify almost nothing.

I am still avoiding working with AI. I try to use some models locally for experiments. I refuse to pay for something that was build on ripping others off. And the local models are underwhelming thus so far.

Re: Agentic Coding Is a Trap

#223

Post by SpicyLemonZest »

dilyevsky wrote: post #218
SpicyLemonZest wrote: post #157

Earlier quoted context omitted.

> 1. models are now extremely good at totally automating tedious tasks such as updating dependancies, build/deploys scripts, unit tests, etc what used to take days now can takes minutes. Easily 50x speedup on this. This was non-trivial part of every engineer's day-to-day at an established company. "platform engineering" or whatever they call this now is dead. I confess that I don't understand why this isn't true, bec…

I think it’s because most PEs are in general conservative (the whole “mr no” meme) and also have limited experience writing software so they will be slow to adapt to this paradigm shift

It's not my experience that our production engineers are quick to say no or slow to adopt LLM agents.

Re: Agentic Coding Is a Trap

#224

Post by throwaway041207 »

bulbar wrote: post #217
throwaway041207 wrote: post #208

Earlier quoted context omitted.

> Just as "use code for contracts" failed for crypto currencies, "use AI output as prod" will fail for AI. Both is based on "just don't make catastrophic mistakes anymore". What I think will happen is AI will write code and it will do the best it can to mitigate mistakes prior to rollout, but once rollout time occurs, rollout will be incremental and it will self monitor by defining success conditions at rollout time.…

This assumes failures can be detected and fixed more easily than generating the corresponding change. I am not convinced that's the case. Counter points to my own arguments: 1. We don't know yet in detail what AI is good at. 2. AI doesn't need to be perfect, just "good enough", whatever that means for a specific project. More failures while saving hundreds of thousands dollars each year might be acceptable, for examp…

> 2. AI doesn't need to be perfect, just "good enough", whatever that means for a specific project. More failures while saving hundreds of thousands dollars each year might be acceptable, for example.

This I think is the unexplored aspect of what's happening right now. Guardrails around "good enough" systems is where the future value lies. In the future code will never be as good as when the artisans were writing it, but if you have an automated process to validate/verify mediocre code (and kick it back to AI for refinement when it fails) before it's fully productionized, then you have a pathway to scaling agentic coding.

Re: Agentic Coding Is a Trap

#225

Post by childintime »

ryandrake wrote: post #86
keyle wrote: post #79

As a senior developer, 25+ years, I have been thrown recently into a meeting "hey can you join in for 5 mins". I really don't like these meetings where you're dragged in in the middle of them without any clue. The questions came flying in fast, without any introduction, and this was about an external integration out of a dozen. They have their own lingo, different from ours, to make the situation worse. I had a _very…

What kind of place do you work where you get dragged into a meeting halfway through and then are peppered with technical questions without context, that you're expected to answer on the spot? Please let us know because I'm sure a lot of us want to avoid such a place. "I'll need to study the docs and code to answer these questions properly" is a perfectly fine (and very diplomatic) response to treatment like that.

> "I'll need to study the docs and code to answer these questions properly" is a perfectly fine (and very diplomatic) response to treatment like that.

By the time you said that, the AI could have given the 80% answer. So, no, this is no longer an adequate response. The right response would have been to take your tools and give an informed opinion on the AI answer, right there.

Re: Agentic Coding Is a Trap

#226

Post by thedevilslawyer »

WorldMaker wrote: post #184
dilyevsky wrote: post #132

Earlier quoted context omitted.

Respectfully, unless someone is really really bad at articulating what the quality standards are or works with a very niche stack that is definitely not the case anymore with SOTA models

Respectfully, the current models are all trained on everyone else's legacy code as of roughly six months ago and largely always will be. If I'm doing my job right an LLM cannot meet my personal quality bar on its own because I will always need innovation and excellence they will never see and thus cannot deliver. I also think that training these tools on my personal quality bar is more work than just writing it mysel…

LLMs are plenty innovative and generate good quality code by default, and great quality code when directed well.

If you're not seeing this, at best you're probably unable to direct them or use them well.

FWIW, if you don't believe the above, I challenge you to put up a quick git repo, where you are unable to get the deserved quality out, and we can quickly show you how the same quality is available via SOTA agents, within a fraction of hand-coded time.

Re: Agentic Coding Is a Trap

#227

Post by Terr_ »

AdieuToLogic wrote: post #130
ryandrake wrote: post #100

Using AI to go faster is optimizing the wrong thing. At every place I've worked, the "code writing" part takes the least amount of time, compared to all the other things you need to do in order to implement a feature. Let's examine a feature that takes a day to code: First, you've got to plan everything, using whatever Agile or Waterfall planning ritual your company uses, get the task breakdown, file the JIRA tickets…

> Using AI to go faster is optimizing the wrong thing. At every place I've worked, the "code writing" part takes the least amount of time, compared to all the other things you need to do in order to implement a feature. This reminds me of one of my software engineering axioms: When making software, remember that it is a snapshot of your understanding of the problem. It states to all, including your future-self, your…

> a snapshot of your understanding of the problem

Relevant: Programming as Theory Building (1985) by Peter Naur. The actual text is rather stuffy, but basically the code+docs cannot replace the richer in-human-heads ideas for what the real-world problem is and how computers should (or shouldn't) be used to face the problem.

Re: Agentic Coding Is a Trap

#228

Post by luodaint »

The trap isn't agentic coding, it's using it as fancy autocomplete. When you delegate whole tasks with real acceptance criteria, not just single functions, the economics change completely.

Re: Agentic Coding Is a Trap

#229

Post by redact207 »

I sometimes wonder if I'm in a different universe to other devs. Anytime AI coding is brought up, comments are overwhelmingly negative and often point out correctness, quality, slop, etc.

There's also the 'more stuff is being delivered, but it's not right, full of holes and papercuts'.

I'm 22 years into development and couldn't think of going back to non AI programming now. Not only has it sped up velocity by an order of magnitude, it's also helped me unlock side projects that I would never even begin in the past as I knew I didn't have that time.

It's just like any tool though, and I've found enormous differences in outcome depending on how you drive it. Launching into 'build this' and expecting it to output code that you would manually write would not get you there; and I feel this is where most developers stall out.

Getting the right outcomes takes a lot of harness set up - the same as if you wanted to hire new devs and get them productive without peering with them. You would set up linting, good test coverage and approaches, thorough documentation about what your project is, the domain, the architecture etc. This at least gets good code consistency for the most part.

For how to build, https://github.com/bmad-code-org/BMAD-METHOD is really good and I've onboarded a few Saas projects into it now. Tech speccing and multiple cycles of elicitation are what deal with all the edge cases that you normally only encounter during coding. It does front-load all of the planning brainwork; but condensing that into a couple days of solid speccing is far more productive than spreading it out over months.

It's taken a while to get to this point, and most agents aren't good for substantial work out of the box. Most of the time what the agent does will be a product of its environment.

Re: Agentic Coding Is a Trap

#230

Post by xg15 »

keyle wrote: post #79

As a senior developer, 25+ years, I have been thrown recently into a meeting "hey can you join in for 5 mins". I really don't like these meetings where you're dragged in in the middle of them without any clue. The questions came flying in fast, without any introduction, and this was about an external integration out of a dozen. They have their own lingo, different from ours, to make the situation worse. I had a _very…

> To be continued... But from now on, the work isn't done if I don't get a little 5 mins flash-card type markdown list of "what is this" and "what is that", type glossary.

Easy, just have the agent write them for you at the end, then never read them...

(/joking)

Post reply on HN