DrStone's recent activity
-
Comment on Control the ideas, not the code in ~comp
-
Comment on Control the ideas, not the code in ~comp
DrStone (edited )Link ParentI'm a software engineer with 15+ years in the industry. Over the years, I have occasionally tried AI through whatever the Intellij IDEA Ultimate IDE offered when I was really stuck on something...- Exemplary
I'm a software engineer with 15+ years in the industry. Over the years, I have occasionally tried AI through whatever the Intellij IDEA Ultimate IDE offered when I was really stuck on something fairly self contained (e.g. sql query optimization). Every time I was disappointed and wondered how anyone got anything worthwhile out of AI.
Earlier this year, my boss got us all a 1-month trial to Claude Code (Opus 4.6 I think), and told us genuinely no-pressure "give it a shot, see if it helps". I went in very skeptically, with a baseline zero trust for the AI output. The first few tasks were self-contained, dead simple, and easily verifiable - stuff like "translate this legacy Twirl template to React". I reviewed every line it produced and tested every result. It did pretty well, so I started using it a bit more.
It's been a few months now and I find that I use Claude Code (now Opus 4.8, thinking high) for the majority of my work. I've created some tight agent files for the projects and a short global one based on actual usage and results. I've heard my process called "architecture driven development", though I don't really participate in the AI industry? culture? fandom? Detailed design docs / plans vastly improve the AI implementation results and subsequent review, like @stu2b50 mentioned. Anyway, here's what's gotten me good results:
- Personally draft as detailed of a requirement set as I can, with high level design/architecture. Reference local code precedent where possible.
- Give that to Claude and prompt it to draft a plan. Anything beyond a known simple fix, always draft a plan. It'll ask me various questions during the planning, and state any assumptions or decisions it made at the end.
- In a clean session (so it's a cold-read), prompt Claude to critically review the plan it just wrote. I created a skill for this after a while based on how I would usually prompt it, refined with Claude itself. This digs in to everything from verifying all claims against live source / data, reviewing high level architecture and cost, preemptively looking for contradictions and ambiguities and edge cases, etc. It'll come back to me with a list of findings, organized by estimated severity. Giving it a read-only user for the database and several other key integrations has helped a lot here, and sometimes even surfaced bad assumptions on my part.
- Manually walk through each point in the plan review findings with Claude, prompting a resolution or further investigation, until all are resolved.
- Depending on the nature of what was found and how much the resolutions altered the plan, repeat step #3 as many times as reasonable.
- Manually skim the final plan. At this point, I'm pretty familiar with it from the drafting and reviews, so it's mostly a sanity check.
- In a new session, have Claude implement the plan and check against whatever automated verification the specific thing might have (compile, linting, the rare tests, etc.)
8 - 10. Same as step #3-5, except for the implementation. In a clean session audit the implementation. I've got another skill for this based on what I'd usually ask it to do. Faithfulness to plan, trace impacted code flows, etc. Walk through findings, rinse and repeat as necessary. - A skim of the actual code diff. Working with a large, mature codebase gives Claude a lot of precedent to follow, so it usually organizes and writes things consistent with our existing code. Sometimes I don't like exactly how it organized things, or the specific implementation of something, and I'll either directly change it or prompt Claude to redo that bit.
- Manual end to end testing. Usually some adjustments fall out of this.
Note that the various clean sessions are important, or having claude run the review skill with subagents in clean sessions. Aside from growing large contexts eventually degrading results, a cold reading of the plan and the implementation are key. They don't let old assumptions skip over investigating potential faults, dropped branches from investigation get brought back in, or anything like that.
If nothing else, it's a good way to force yourself to articulate and record all of your requirements, all of the desired flows, and account for "all" edge cases. Like extreme rubber-ducking.
EDIT:
Oh, and one other thing that I don't think people often consider: Maybe once every few months, ask the AI to review all of it's memories and agent files for consistency, redundancy, verbosity, and organization (different project memories, memory vs agent file, project agent file vs global agent file). Over time, especially if you've got multiple related projects, it'll accumulate a lot. Cleaning up can make a difference in result quality and context token usage. -
Comment on Why does WhatsApp drain so much phone battery? in ~tech
DrStone Link ParentiPhone 11 Pro here, with poor battery health at 72% capacity, living in a country where WhatsApp is the communication tool for people and businesses. Normal battery usage for WhatsApp at 2-5% over...iPhone 11 Pro here, with poor battery health at 72% capacity, living in a country where WhatsApp is the communication tool for people and businesses. Normal battery usage for WhatsApp at 2-5% over the past week. I also have a second device (macbook) logged in to whataspp, if that impacts the phone app at all. I am not in any large groups, just a few small family/friend groups.
-
Comment on Give me your culture clash stories in ~travel
DrStone Link ParentDon't forget that the Buddhist vegetarians also avoid alliums (garlic, onions, etc.), so you might want to specify which subset of no-meat vegetarian you are. Allowing alliums will open up more...Don't forget that the Buddhist vegetarians also avoid alliums (garlic, onions, etc.), so you might want to specify which subset of no-meat vegetarian you are. Allowing alliums will open up more options on the menu.
-
Comment on Modern, abstract art makes me angry in ~arts
DrStone Link ParentThis is the point where the human intentionality as a key factor breaks down for me personally. If their intended message is lost or completely different for a reasonable observer, and the piece...And it's also not to say that I'm sure to pull the same things out of art as were put into it.
This is the point where the human intentionality as a key factor breaks down for me personally. If their intended message is lost or completely different for a reasonable observer, and the piece is too abstract to have at least an agreed upon foundation to derive from, I don't see the human creator's intentions as being relevant.
I know art, like many things, has boundaries nearly impossible to pin down, but I'd like to push a bit further since you've clearly spent a lot of time reflecting on art. Do you get anything out of art made with intentional techniques, but unintentional results? Like arranging some basic unhardened clay outside before a heavy storm arrives to naturally alter and erode it. Or intentionally selecting the state of something already occurring to preserve identically (unstylized). Is it simply the intention to create art that matters, or the techniques, or must the intention flow through to the result (even it it doesn't make it to the observer)?
-
Comment on Modern, abstract art makes me angry in ~arts
DrStone Link ParentWhat do you think the difference is, if there is one, between looking at and pulling meaning out of (abstract) art and doing the same with nature (pareidolia). Finding everything from simple...What do you think the difference is, if there is one, between looking at and pulling meaning out of (abstract) art and doing the same with nature (pareidolia). Finding everything from simple shapes to elaborate scenes and figures in clouds is a common leisurely pastime. Seeing faces in things is especially common.
-
Comment on Modern, abstract art makes me angry in ~arts
DrStone Link ParentI think seeing the Mona Lisa has even less to do with the historical context, the artist, and the techniques for the vast majority of the masses at this point compared to simple FOMO and momentum...I think seeing the Mona Lisa has even less to do with the historical context, the artist, and the techniques for the vast majority of the masses at this point compared to simple FOMO and momentum because millions of people go to see it. No different than many of the most famous tourist attractions. Taken to a more depressing extreme, it's like how there are some celebrities that continue to be famous and popular primarily for being famous popular.
-
Comment on Give me your culture clash stories in ~travel
DrStone LinkPrevious discussion: What's a culture shock that you experienced? (January 13)Previous discussion: What's a culture shock that you experienced? (January 13)
-
Comment on Cambria, California banned fireworks. Then came the dogs. in ~society
DrStone Link ParentProbably for similar reasons that some people use white noise machines when sleeping. It's really easy to notice, and relatively jarring to hear, noises in an otherwise quiet environment. The...Probably for similar reasons that some people use white noise machines when sleeping. It's really easy to notice, and relatively jarring to hear, noises in an otherwise quiet environment. The white noise evens things out a bit so that loud noises don't stand out as much, and quieter noises are drowned out completely. Fireworks are exceptionally loud and sporadic, so regular white noise would have to be really loud. A bunch of tennis balls providing a consistent "thump boom thump" in a dryer maybe could smooth over the booms of fireworks better.
Either that, or maybe part of some lower volume similar-sounding conditioning before the actual big day. (I don't know if this would work)
-
Comment on How important is sexual chemistry/ability/quality to you when you date/marry/whatever? in ~life
DrStone Link ParentSometimes the drive is there, but other things stand in the way. For example, I’ve struggled with a sensitive gag reflex, to the point where I dry heave sometimes just taking out my retainer. A...Sometimes the drive is there, but other things stand in the way.
For example, I’ve struggled with a sensitive gag reflex, to the point where I dry heave sometimes just taking out my retainer. A pube in the teeth would be a guarantee. I’ve managed to overcome it, but it’s taken time, a lot of desensitization practice, and some cooperation from my partner.
And then there’s partners who have their own obstacles - mental or physical - around being eaten out. You could know and be able to do everything “right” and it still wouldn’t be a pleasure until you and your partner work through their hangups.
-
Comment on bubbles.town: Tildes but exclusively for blogs in ~tech
DrStone LinkI like the idea, but skimming through some of the top posts... the quality really isn't there. Hot takes, glorified tweets, "i had a one-off bad interaction", platitudes stretched out to posts,...I like the idea, but skimming through some of the top posts... the quality really isn't there. Hot takes, glorified tweets, "i had a one-off bad interaction", platitudes stretched out to posts, brief surface-level observations, and so on. I get that these are small personal blogs (I wrote my own many years ago) and quality spans the spectrum, but I don't get the impression that bubbles.town is really helping to surface good ones.
-
Comment on Not so empty nesters: record-high number of US adults under 35 live at home, new data says in ~life
DrStone (edited )Link ParentI would bet that it's more the definition of "home" being "where a person lives", so to someone who's only learned English as a second language, it it seems tautological to say "record-high number...I would bet that it's more the definition of "home" being "where a person lives", so to someone who's only learned English as a second language, it it seems tautological to say "record-high number [...] live at home". Of course they do. You'd probably need to be more familiar with how "home" is used in different contexts, how common independent nuclear family households are compared to multi-generational ones (leading to different "homes" for one person), and so on to really get what the headline means without additional context.
-
Comment on If you are asking for human attention, demonstrate human effort in ~comp
DrStone Link ParentYeah, there's a big difference between a reactive "it's the tool's fault" defense after getting called out on garbage results or blown deadlines, and a proactive "the tool we use at the company...Yeah, there's a big difference between a reactive "it's the tool's fault" defense after getting called out on garbage results or blown deadlines, and a proactive "the tool we use at the company for this process is [unusable/not producing good results] for [already investigated reason]. We can wait for it to be fixed, or we can instead..." or just figuring an alternative solution out yourself without the broken tool to meet the agreed upon deadlines if the tool usage is your personal choice.
-
Comment on Epic Games announces Lore open-source version control system in ~tech
DrStone Link ParentHaving used cvs and svn in a professional context, I absolutely love git. The cli took a little while to learn, but now all of my common workflow, troubleshooting, and branch/history hygiene...Having used cvs and svn in a professional context, I absolutely love git. The cli took a little while to learn, but now all of my common workflow, troubleshooting, and branch/history hygiene commands are muscle memory. Throw in a few custom aliases for convenience and I’m good to go. The only time I leave the official git cli is to use IntelliJ’s conflict resolution GUI for complicated conflicts (the “magic” resolve gets me 90% of the way there, and I’m already using IntelliJ as my primary IDE, so it’s right there); I still commit after resolving and do everything else with the cli.
-
Comment on Test your medical knowledge with daily clinical scenarios in ~health
DrStone Link ParentI see two cookies for the domain after playing doctordle_cookies_accepted: true doctordle_stats: stats that appear to be cumulative (gamesPlayed, lastPlayedDate, guessDistribution, and some...I see two cookies for the domain after playing
doctordle_cookies_accepted:truedoctordle_stats: stats that appear to be cumulative (gamesPlayed, lastPlayedDate, guessDistribution, and some others)
Seems weird to require it just for stats. I had thought maybe they were using cookies for the current state, but that appears to be in local storage, where I see:
doctordle_ga_consent:granteddoctordle_archive_335_completed:lost(I played one archived game)doctordle_336_Thu Jun 18 2026: game state
-
Comment on What are some seemingly silly things in your life that have practical purposes? in ~life
DrStone Link ParentThe NYT has an article on this topic Do You Really Need to Clean Your Fridge Coils? (2025)The NYT has an article on this topic Do You Really Need to Clean Your Fridge Coils? (2025)
The majority of us don’t need to clean them, ever.
[...]
In the past, said Amiel, dust and lint would collect in the fan and the condenser coils that lived under or in the back of your refrigerator, and that dust would block the air flow that a fridge needs to keep things cool. But for the past two decades, fridges have been designed to fully protect those elements so you don’t need to worry about cleaning them, said Amiel. -
Comment on How one plant murdered a continent in ~enviro
DrStone LinkOfftopic/meta: I have no idea where that "SG" is coming from. I selected the title on youtube, copied, and it pasted with an "SG". Copying just part of the title doesn't come with the "SG". If...Offtopic/meta:
I have no idea where that "SG" is coming from. I selected the title on youtube, copied, and it pasted with an "SG". Copying just part of the title doesn't come with the "SG". If someone could remove it here, that'd be great. -
Comment on How one plant murdered a continent in ~enviro
DrStone LinkThis is an interesting history of an invasive species nearly taking over a continent and the efforts to combat it - the prickly pear cactus in Australia. This is the first video I've seen from...This is an interesting history of an invasive species nearly taking over a continent and the efforts to combat it - the prickly pear cactus in Australia. This is the first video I've seen from this guy, but the presentation is fun and interesting beyond being a wild story.
-
How one plant murdered a continent
11 votes -
Comment on Caught the cycling bug. Anyone else? in ~hobbies
DrStone LinkI don't bike much due to the weather (consistently 28-32C, 70%+ humidity, UV 6-7), though I've considered picking up a helmet so I could at least do a bikeshare or something when my spouse has the...I don't bike much due to the weather (consistently 28-32C, 70%+ humidity, UV 6-7), though I've considered picking up a helmet so I could at least do a bikeshare or something when my spouse has the car. There is an excellent resource from Virginia Tech rating bicycle helmet safety. They've got other sport/activity helmet sections too.
Always comforting to hear people arrived at a similar usages and conclusions. Either we're doing something right, or we're at least in the same wrong boat together!
Yeah, memories are a bit goofy. They seem to be useful short term, helping to bridge the gap between workdays and task switches, but can come back to bite weeks down the line when they're no longer in sync with the current state of the world. I try to get it to write as much as possible and useful to the plan files, and much more rarely to the agent files after hitting the same problem or deriving the same information over and over again.
For skills and the agent files, I had Claude write the prose, but built them with nearly the same flow as the feature flow. Detailed objective scope and requirements, known bits of information or where to gather them from, desired behavior, etc. up front, iterative back and forth, cold reading reviews by both me and Claude, and tweaks over time based on their performance. I think it's important for people to use the AI tools and harness as vanilla as possible for a while to find what fits you personally rather than cargo-culting a bunch of skills and prompts and agent file "hacks".
Definitely agree about the design docs for large areas of the code base, as well as the boundaries and interactions between code bases if your project has multiple repos/applications. Before I did that, anything non-trivial would either get confused if I didn't tell it explicitly where things lived, or would spend time and tokens poking around every time. For the same reason, I keep a database schema dump on hand so it can do some quick grepping to find most things.
Oh, and I just remembered another tidbit, agents are great with CLI tools. I rarely need to reach for some feature-subset MCP when the thing has a mature cli tool I can install. Plus I don't have to keep authentication info in some plaintext MCP config.