To my believe there was not a goal to write good code. The goal was maintainability and to keep it simple, so that people understand. People come and go, you constantly get to see foreign code and you have to do something with it.
Anyways, i see the maintainability hell coming onto us. I still wonder how i organize this with AI. I definitly do not want to touch it what is written by AI.
I think the industry-wide hope is that AI manages the AI-written code, but it’s unclear whether that’s actually going to work out in practice. Right now, my experience is that is dicey. I’ve had AI mess up a codebase to the point where I threw it away and restarted. Maybe I was doing it wrong, though, in that I was looking at the code and was increasingly horrified by the slop. I get the feeling that in this new world, we’re supposed to ignore how the sausage is made and just focus on the final outcome.
IME AI-native engineering requires a lot of infrastructure to make it viable. Teams who are just opening up cursor and putting it on "auto" and trying to one shot features may get stuff that works but is indeed slop.
Since the beginning of the year, I've been spearheading a low-stakes AI-native project (an internal tool). No one's written a single line of code. And we've learned so much from this experience. The first rule was our product manager, who is technical but isn't typically in the weeds, needs to be able to one-shot prompts with cursor auto. And so many rules stem from there, from e2e tests to ensure he doesn't break stuff, to custom linters to ensure that code lives in the right place, to architectural spec sheets so the LLM doesn't try to do raw DB queries from the client.
We're still not there, but we're getting closer and learning and improving every day.
I think the folks who are vibe coding a lot either aren't working in a team, or they are omitting the fact that they have spent a long time building harnesses to ensure the LLM doesn't run amok.
And I think the people who hate vibe coding are likely just asking Claude Code to do X without using Skills that have opinionated ways to do X.
All that said, I don't think we should ignore how the sausage is made at all. Part of what makes me able to move quickly in this project is knowing where stuff lives. I may not understand the line-by-line code, but if I know where to look to find out why I'm missing data that's in the DB, I can move a lot faster than if I have no idea what's going on in the codebase. Then when I find the problematic file or function, I can ask the LLM why it's like X and tell it it should be like Y.
Cool. Are you restricting the AI to be very focused on a function or an architectural blocks that is envisioned, or are you giving it more freedom? I seem to have less slop when I really constrain things, but that takes a lot of work (e.g., specs) and dialogue with the AI (“focus on X, now let’s design block Y,” etc.).
I give it freedom but with the predefined restrictions. I use a plug-in called Obra Superpowers. Whenever I want to start on a block of work, whether it's a ticket or if I just want to tackle tech debt, I start with the brainstorm command. I say something vague like "implement X" or "last time i tried to vibe code Y, Z happened. I don't want that to happen again. Let's improve the harness."
It'll ask follow questions, which I answer, then generate specs that I manually review. If it looks good, it'll generate a plan. If not then I'll give it feedback.
When the scope of work is well-defined (ie my boss says users should be able to do Y) then this process is fairly seamless.
When it's not well-defined then it does take a bit longer and more dialogue as you said. But because everything is documented and written down, we have a pretty good feedback loop (boss asks why it works like X, I can look at the generated spec/plan, or ask the AI to, to understand why).
Ok, so it’s constrained by specs, but you dialog with the AI and have it create the specs. I should try that. I’ve been creating my own specs and having it work from those and then iterating, but that’s not exactly quick and I find myself thinking, “At this rate I could do it faster myself.”
Yeah definitely agreed. I'm lucky I'm that my boss is willing to invest in this little experiment so the point isn't "can we do this faster manually" it's "how can we build our AI infrastructure such that it can actually be faster."
And also, I'm taking care of my infant daughter while working so my workflow is often "launch an AI agent from my computer while she's asleep, review plan on my phone while feeding or napping the little one, approve it and execute it" so it's often running when I'm not really in a mental space to be thinking deeply.
I use tailscale and mullvad vpn for a list of exit nodes i can choose from to work around restrictions, but also bad routing.
Like, when in asia and the route is to europe, sometimes it adds weird hops, while when i use an exit-node in Japan, i know, i have perfect routing to Japan and from there perfect routing to europe.
But the Mullvad VPN exit nodes often runs into problems like cloudflare blocking. So i am looking for alternative, not well known providers for exit-nodes.
Sometimes i even dream of sending my europe traffic via the internal aws network via regions, but hey...
> Sometimes i even dream of sending my europe traffic via the internal aws network via regions, but hey...
It's more work, but you can definitely do this. Inter-region traffic still carries egress charges though, so be aware of that in advance. This is a very common pattern in enterprise networking when building cloud-based SDWAN topologies: branch a,b,c connect to hub-1 in us-east-2; branch d,e,f connect to hub-2 in us-west-2; dc1 connects to hub-1 in us-east-2; dc2 connects to hub-2 in us-west-2; services in dc1 and dc2 can reach each other for DR and clients in branch f can reach services hosted in dc1.
Underlying all of these SDWAN technologies is essentially basic site-to-site VPN tunnels. Most still use IPSEC, although Wireguard is also used sometimes.
The only tricky part is the inter-region routing, and this can be managed largely within AWS using Transit Gateways (TGW), for a price, for more of a price AWS even makes it easier with Cloud WAN: https://aws.amazon.com/cloud-wan/
Basically if you just link your VPCs in each region with the appropriate routing policies, you can just connect to your preferred VPN server in each region and ultimately get routed correctly. This is what companies with cloud-based SDWAN do for providing SASE services to end-user clients.
I hope not. I am in general fine with Ads for content, even with profiling to what i listen and so on. But if you pay for subscription to get no Ads and still you get Ads from Spotify, then it is plain stupid.
I get the snark, but also - there is this "ideal Java code style" that most experienced devs tend towards. Unfortunately I don't see anything like that for Kotlin (yet?), and there are a bunch of patterns that I really dislike. I sometimes feel people just toy around, like "wow I can make this into an extension method, how cool" and leave that as the code.
In general, I really dislike extension methods, especially when paired with tiny objects with barely any functionality to begin with. Like people build a mental model of what a thing is based on how can it be used - but if you leave that empty and put every behavior at different files in the form of extension methods you make this understanding very hard to build up.
Add to it that it removes polymorphism and often actually hinders readability.. so my point is, having more ways to write code is not necessarily a positive.
Not in my city. Business is all dying, everyone avoids to go to the centre, everywhere the city fights cars, handy man charge extra just for comings, nah, it's basically gated communities now, well, they can have it, but life happens somewhere else then, where it can expand freely.
It is not convenient. It's freezing cold and icy, no walk, no bike, no scooter. Use mass-transit, sure, when you don't care about your life, when it's working, when it's coming regularly, when i don't have to exchange stations, but still, walking from home to a station and back, nah, it all sucks.
Imagining sitting in a cosy, warm pod, driving in a tunnel autonomously, point to point, and you have my vote.
That giant 5-level parking lot monstrocity could be a transport hub instead that has a warm metro stop, much better lighting and safety and perhaps even some light convenience retail.
> Imagining sitting in a cosy, warm pod, driving in a tunnel autonomously, point to point, and you have my vote.
Nope. The first thing i do in the warm car, is to turn the music on, making a stop at my favourite coffee shop and then i hit the road, humming my favourite songs. I barely drive anyways, its all automatic.
In mass-transit facilities all the people look at their screens, using headphones, waiting to be coughed on, scared to be not talked to, anxiety all around. Nope, not for me. Never looking into a dirty public toilet again, while the society yells: "but its free!".
Progress, folx, not regress. Come out of your bubbles, ignore the voices, that tell ya, hundreds of human bodies efficiently transported in an iron can is progress! Live! Expand! Use everything!
You should like a scary driver to share the road with. Your whole description of the “joy” of driving is all about how little attention you are spending on the actual driving part.
When you drive you are responsible for a massive complex device moving at high speeds. It must take your full focus
An unfortunate side effect of car dependence is people forgetting how to dress outside in the place they live, a skill humans had for thousands of years but apparently lost some time in the last ~100.
Anyways, i see the maintainability hell coming onto us. I still wonder how i organize this with AI. I definitly do not want to touch it what is written by AI.
reply