Live data from Hacker News

Eight more months of agents

crawshaw.io

241–250 of 250 posts

Re: Eight more months of agents

#241

> Pay through the nose for Opus or GPT-7.9-xhigh-with-cheese. Don't worry, it's only for a few years. > You have to turn off the sandbox, which means you have to provide your own sandbox. I have tried just about everything and I highly recommend: use a fresh VM. > I am extremely out of touch with anti-LLM arguments 'Just pay out the arse and run models without a sandbox or in some annoying VM just to see them fail. W…

I'm doing web development in a VM (on exe.dev) and it works quite nicely.

Re: Eight more months of agents

#242

Earlier quoted context omitted.

Why is that staggering? That feels like a pretty dramatic expression. Is it a foregone conclusion that one must use agents?

one does not have to use anything at all… but if someone is “senior” and is incapable of using llms for some parts of her/his job then senior part is just age related and not tied to skill level

You originally said being "anti-llm", but you now refer to being "incapable". Surely you can see that those are different things?

Re: Eight more months of agents

#243
post #159
post #158

Earlier quoted context omitted.

Everybody keeps saying the models are getting better, the tooling is getting better, people are discovering better practices... So why not just wait out this insane initial phase, and if anything is left standing afterwards and proves itself, just learn that.

Because there's nothing to learn. "learning" to use claude code is less effort than learning how to use the basics of git. They provide value today so I'm using them today.

This. If you use a modern frontier model like Opus 4.5, there's nothing to learn. No special prompting techniques. You give it a task, and most of the time it's capable of solving a big chunk quickly. You still need to babysit it, review its plan/code and make adjustments. But that's already faster than achieving the same results manually. Especially when you're at low energy levels and can't bring yourself to look into a task and figure it out from zero.

Re: Eight more months of agents

#244

"In 2026, I don't use an IDE any more." Just a question? What IDE feature is obsolete now? Ability to navigate the code? Integration with database, Docker, JIRA, Github (like having PR comments available, listed, etc), Git? Working with remote files? Building the project? Yes, I can ask copilot to build my project and verify tests results, but it will eat a lot of tokens and added value is almost none.

> I can ask copilot to build my project and verify tests results, but [..] added value is almost none.

The added value is that it can iterate autonomously and finish tasks that it can't one-shot in its first code edit. Which is basically all tasks that I assign to Copilot.

The added value is that I get to review fully-baked PRs that meet some bar of quality. Just like I don't review human PRs if they don't pass CI.

Fully agree on IDEs, though. I absolutely still need an IDE to iterate on PRs, review them, and tweak them manually. I find VSCode+Copilot to be very good for this workflow. I'm not into vibe coding.

Re: Eight more months of agents

#245
post #46

Earlier quoted context omitted.

> We all write insecure code in our first pass through no, we don't

Yes we do, you don't just start a brand new web project and spit out CORS rules, authentication schemes, roles, etc in one sitting do you? Are you an AI?

I don't. I pull in or bake in the security and ACL on day 1.

Features come after I have tested authn and authz.

The first feature runs with full security.

Re: Eight more months of agents

#246
> I am extremely out of touch with anti-LLM arguments

Wow I know that feel.

I'm here using LLM for daily work and even hobbies in very conservative manners and didn't think much of it.

Now when I have casual discussions with other folks, especially non-tech people, the visceral hatred I get for even mentioning AI and the fact that I use it is insane. There's like an entire sub group of people who are so out of touch with these tools they think they're the devil like the anti-GMO crazies and the PETA psychos.

Re: Eight more months of agents

#247

> I deeply appreciate hand-tool carpentry and mastery of the art, but people need houses and framing teams should obviously have skillsaws. Where are all the new houses? I admit I am not a bleeding edge seeker when it comes to software consumption, but surely a 10x increase in the industry output would be noticeable to anyone?

It's ironic to me because I'm the Luddite who refuses to adopt agentic AI and still using only the Chat interface with Codex and Claude inside the VS Code extensions to help me with both work projects and personal projects. And I've had amazing results with only this. "Look at this codebase and tell me the best ways to integrate some new feature", "look at this source code file and tell me what's wrong with it", "show me how to implement this thing I want". Then I copy and adapt the results as needed and integrate it with the rest of my work. This has worked great and I've shipped a ton of projects much faster and easier. Clearly the AI could have written a lot of it itself but I'm not sure I'm really lacking in any benefits with this method. So this makes the whole agentic push especially seem like some kinda over hyped gimmick.

Re: Eight more months of agents

#248

Earlier quoted context omitted.

Org processes have not changed. Lots of the devs I know are enjoying the speedup on mundane work, consuming it as a temporary lifestyle surplus until everything else catches up. You can't saw faster than the wood arrives. Also the layout of the whole job site is now wrong and the council approvals were the actual bottleneck to how many houses could be built in the first place... :/

Basically this. My last several tickets were HITL coding with AI for several hours and then waiting 1-2 days while the code worked its way through PR and CI/CD process. Coding speed was never really a bottleneck anywhere I have worked - it’s all the processes around it that take the most time and AI doesn’t help that much there.

True story; I wanted to make a tiny update to our CI / CD to upload copies of some artifacts to S3. It took 1min for the LLM to remind me of the correct syntax in aws cli to do the upload and the syntax to plug it into our GitHub Actions. It then took me the next 3 hours to figure out which IAMs needed to be updated in order to allow the upload before it was revealed that Actually uploading to the S3 requires the company IT to adjust bucket policies and this requires filing a ticket with IT and waiting 1-5 business days for a response then potentially having a call with them to discuss the change and justify why we need it. So now it's four days later and I still can't push to S3.

AI reduced this from a 5-day process to a 4.9-day process

Re: Eight more months of agents

#249

Earlier quoted context omitted.

I bought Steve Yegge's "Vibe Coding" book. I think I'm about 1/4th of the way through it or so. One thing that surprised me is there's this naivete on display that workers are going to be the ones to reap the benefits of this. Like, Steve was using an example of being able to direct the agent while doing leisure activities (never mind that Steve is more of an executive/thought leader in this company, and, prior to LL…

You missed option c. C) keep all 8 engineers so the team can pump out features faster, all still working 8 hour days. The ceo will probably be forced to do it to keep up with their competition.

I didn't miss it, I just think it's going to be a rare outcome. I'm sure companies will become a little more bold about new products or new features, but I think there's an upper limit to the amount of change customers will tolerate. Once a product is stable enough to make money, for the most part, users don't want changes or new features and often rebel against them. Most people they think: I'm paying for this thing because it already does the stuff I need, please don't change it, I don't want to relearn this. I'm not saying that to be against evolving software of course, I just think competing on "we move a million miles per hour" will result in burnt out developers and overwhelmed customers. I mean, we're already seeing some burnout from people using AI tools, and I think part of that just has to be the pace of things.

Re: Eight more months of agents

#250
There is a lot of "my" floating around in this article. I always love getting peeks into experiences with this sort of thing, but I think the "mys" highlight something I've seen every day. These agents are really great at bespoke personal flows that build up a TON of almost personal tribal knowledge about how things get done if there is any consistency to those flows at all. Doing this in larger theaters is much more difficult because tribal knowledge is death for larger teams. It drives up the cost of everything which is why individuals or extremely new small teams feel so much more productive. Everything is new here and consistency doesn't matter yet.
Post reply on HN