Ask HN: Do you have any evidence that agentic coding works?

Is it just me, or does every post starting with "Great Idea!" or "Great point!" or "You're so right!" or similar just sound like an LLM is posting?

Or is this a new human linguistic tic that is being caused by prolonged LLM usage?

Or is it just me?

What is the actual value of using agentic LLMs (rather than just LLM-powered autocomplete in your IDE) if it requires this much supervision and handholding? When is it actually faster / more effective?

So review the code. Our rule is that if your name is on the PR, you own the code; someone else will review it and expect you to be able to justify its contents. And we don't accept AI commits.

What this means in workflow terms is that the bottleneck has moved, from writing the code to reviewing it. That's forward progress! But the disparity can be jarring when you have multiple thousands of lines of code generated every day and people are used to a review cycle based on tens or hundreds.

Some people try to make the argument that we can accept standards of code from AI that we wouldn't accept from a human, because it's the AI that's going to have to maintain it and make changes. I don't accept that: whether you're human or not it's always possible to produce write-only code, and even if the position is "if we get into difficulty we'll just have the agent rewrite it" that doesn't stop you getting into a tarpit in the first place. While we still have a need to understand how the systems we produce work, we need humans to be able to make changes and vouch for their behaviour, and that means producing code that follows our standards.

I used Ralph recently, in Claude Code. We had a complex SQL script that was crunched large amounts of data and was slow to run even on tables that are normalized, have indexes for the right columns etc. We, the humans spent significant amount of time tweaking it. We were able to get some performance gains, but eventually hit a wall. That is when I let Ralph take a stab at it. I told it to create a baseline benchmark and I gave it the expected output. I told to keep iterating on the script until there was at least 3x improvement in performance number while the output was identical. I set the iteration limit to 50. I let it loose and went to dinner. When I came back, it had found a way to get 3x performance and stopped on the 20th iteration.

Is there another human that could get me even better performance given the same parameters. Probably yes. In the same amount of time? Maybe, but unlikely. In any case, we don't have anybody on our team that can think of 20 different ways to improve a large and complex SQL script and try them all in a short amount of time.

These tools do require two things before you can expect good results:

1. An open mind. 2. Experience. Lots of it.

BTW, I never trust the code an AI agent spits out. I get other AI agents, different LLMs, to review all work, create deterministic tests that must be run and must pass before the PR is ever generated. I used to do a lot of this manually. But now I create Claude skills that automate a lot of this away.

I use agentic coding daily and rarely write any code by hand.

Here's what works for me:

Spend a lot of time working out plans. If you have a feature, get Claude Opus to build a plan, then ask it "How many github issues should this be", and get it to create those issues.

Then for each issue ask it to plan the implementation, then update the issue.

Then get it to look at all the issues for the feature and look for inconsistencies.

Once this is done, you can add you architectural constraints. If you think one issue looks like it could potentially reinvent something, edit that issue to point it at the existing implementation.

Once you are happy with the plan, assign to your agents and wait.

Optionally you can watch them - I find this quite helpful because you do see them go offtrack sometimes and can correct.

As they finish, run a separate review agent. Again, if you have constraints make sure the agent enforces them.

Finally, do an overall review of the feature. This should be initially AI assisted.

Don't get frustrated when it does the wrong thing - it will! Just tell it how to do the correct thing, and add that to your AGENTS.md so next time it will do it. Consider adding it to your issue template manually too.

In terms of code review, I manually review critical calculations line-by-line, and do a broad sweep review over the rest. That broad sweep review looks for duplicate functionality (which happens a lot) and for bad test case generation.

I've found this methodology speeds up the coding task around 5-10x what I could do before. Tasks that were 5-10 days of work are now doable in around 1 day.

(Overall my productivity increase is a lot higher because I don't procrastinate dealing with issues I want to avoid).