ipsod

It might be that I haven't put enough time in, but, I haven't found it as ergonomic as SolidPython (OpenSCAD wrapper).

Have you found that it becomes more natural?

It's easier to install ublock origin and sponsorblock than to subscribe, and it's free.

Lol, that's awesome.

Have you met users?

under_construction.gif

Same, for sure. I'm not actually sure I've come out ahead. I've probably wasted at least a year of work in each direction.

In spite of my above-stated preference, I used Django for my most-recent project, and don't regret that decision a bit.

Same as people often say about AI writing bespoke applications, these days... Libraries (like applications) often have 98% stuff you don't need, and 2% stuff you do. You can end up better off with your own thing.

I've always liked writing my own libraries and minimal frameworks for PHP, which seems to be a very unpopular opinion, but it almost entirely removes churn from your stack, which is nice for tools that may stick around for years or decades. I also never switched off jquery, preferring simple techs. I'm almost definitely operating at a smaller scale than most web developers here, though.

Terminator, War Games, The Matrix...

Metropolis (1927) had robots that infiltrated a revolution to sabotage it for their masters.

Frankenstein (1818)?

...

> How do you measure that?

Vibes.

Reviews take way more time than new features, and often utilize 3 subagents at a time - it's not uncommon to have reviews take 3 hours when the feature only took 15 minutes. The numbers here are just estimates, though - I don't keep track. "10x to 20x tokens on review" is probably underestimating review cost, if anything.

I don't review all code myself, but I do spend down-time reading code. When I find obviously bad code/patterns/whatever, I don't just fix it directly - I work to create automated tests (like static analysis, almost) or adjust or create prompts and tools for agents (either for review stage or initial implementation) to solve that class of issue, instead of just that instance of it.

I never review the first iteration. I review it after the agents have already poured a ton of time into it, if ever.

I do sometimes watch agents work, and redirect them if I see them doing something dumb. But, usually, I have 5+ agents running, not counting sub-agents, and I'm just messing with the apps myself to find the next feature to create or polish.

> they don't value correctness

I spend ~10-20 tokens on correctness and codebase cleanliness for every 1 I spend on new features.

YOE is ~30, from my first PHP app to present. I've spent somewhere between 5 and 80 hours a week coding for the past 30 years, mostly self-employed, writing tools and creating custom CNC machines/robots for my own business, but more recently doing same thing for a company as a full time employee.

I still get to do some coding, where there are safety-critical production systems, but the work is mostly tooling that gets us to that point or supports the business otherwise.

It's an awesome job. Couldn't be funner, pretty much, except on data collection days. I get high level tasks, or come up with my own, and find or build solutions - no more running a whole business, doing production, keeping books, shipping orders, doing customer support - I only do the stuff I like.

Well, I used to pay more attention to the code, and then I found that if I spent more tokens having the AI review its work than actually creating new features, I was able to get more, better work done.

The AI review stage is where I've found prompting matters a lot more than that initial prototyping phase. I've iterated on prompts and tooling that help the AI keep the codebase clean-ish.

If you can find exercise that's "play" for you, that's the trick.

I like hacky sack and kettlebells. I can play hacky sack for 15 minutes, wherever I am, and easily keep my heart rate between 160 and 190 bpm. I can do kettlebells in my bedroom in 15 minutes, and stay pretty strong and capable. They're both genuinely fun for me, and leave me feeling better (lower stress, more alert, stronger, more agile, more attractive, less like a lazy slob).

I never could spend 10 minutes to and from the gym, and another 45 there, or anything like that. I feel lucky to have found these things I do like.

> It's not even close.

That's always been my experience. Not even close. One gives you legos that lets you make a plastic toy, and the other is an industrial machine shop for code.

I thought maybe I was holding VSCode wrong, though.

I'm paid to produce software that has features and works.

Nobody cares how I get there or how ugly the code is. They care how fast I can do it and how reliable the software is.

I look at code sometimes, in down time, to see what's been done terribly by AI, and to fix it and hopefully fix the process so that things don't get done terribly in that particular way in the future.

Code intelligence (find usages, go to definition, refactor/rename/move, etc.) and debugging are on another level - more powerful, more reliable, with more edge-cases covered.

At least, that was so in dynamic languages, like Python, and PHP, last I checked.

Has anyone had a good experience with Junie, to share?

I've tried it, and it seemed fine, but not compelling enough to even spend the free usage I get with my All Products Pack.

I tried 3.1 Pro over the weekend and you're right, it's great for chat. Easily my favorite. Thanks for the suggestion.

Wow. Thanks for sharing. I'll definitely try it. I've noticed the trouble you mention, and used 3.8 Flash in spite of it, because I like the model, but it's bit me more than once.

Also, 3.8 Flash for coding... I've done several experiments with the same prompt and same codebase against other models, and it's never turned out well for Flash against GPT Sol or Astra. No reason to use it for coding, either, except for quick scripts and prototypes, where it's nice because it's so fast.

You use 3.1 Pro?

I should try it, I guess - I always default to Flash 3.8.

Yeah, I'm not saying this is an example of great work. It's sufficient for the job - more time spent refining my instructions isn't necessary.

What I'm saying is, with just the effort it takes to notice and somewhat-sloppily mention things, they get fixed, and usually fixed well. Further testing will tell us whether it needs more work or not.

They may have never had these troubles at all, if I'd coded by hand originally, but then, this particular app would've never been written, because it wouldn't justify that amount of work.

More like google (when google just served content off of the website you were searching for), but also fed every form of media, turned up to 11, and given agency far beyond retrieval.

I'd love to know if you still felt that way after seeing what I do in a day. Company IP won't allow it. Maybe I'll stream some personal dev work or something.

I'm not saying you're wrong, but, I'd be shocked if you were still so adamant (assuming you're rational and not in your own psychosis).

Here's something I'm about to drop on AI. Like 90% of things like this, when I ask for a fix (with a few skills), get fixed with fewer edge cases left dangling than I'd have left if I'd spent all week on this stuff. And, it'll be done before tomorrow, with no further input from me. This is a list I put together in about 1 hour.

    Things to fix with OCR text labeling
        - vertical line in center breaks box-drag
            - nope, not line in middle... IDK why, but the drag pretty frequently snaps back to original position in middle of drag... lets me keep dragging, but have to start back from there
            - happens with "draw stamp box" too
                - forces me to re-click the button to start over
        - add zoomed view of text in bottom right corner of video
        - add ability to restrict OCR
            - current job is 0 1 2 3 4 X 6 7 8 9
                - 5 is an X, for some reason
            - is detecting full alphanumeric
        - add ctrl+z undo for modifications to drawn box
        - deleting character in middle of "your label" text box jumps cursor to end

3 years ago, these things became useful. Every year, they're drastically better. Most juniors are already dead weight. Most of the rest of us are next.

IDK how it could be otherwise. It's moving too fast and gone too far already.

I've always been either entirely self-directed, choosing my own goals, or almost entirely self-directed, being given high-level goals, and getting to choose my own means to accomplish them. So my opinions are probably different from somebody who was a smaller part in a chain.

You say that you liked learning how things worked inside out, whether it was Postgres or whatever. You can still do that. You can learn it at a high level, and then it becomes like a magic word in your arsenal. You don't have to know or remember the details of implementation, but only the nature of its costs, benefits, and synergies with other systems. You just have to remember that high-level abstraction, maybe just a single vocabulary word, of a low-level reality - the exact syntax is really no longer a concern.

The better the agents get, the less there is playing manager. Really your job now is to have and execute vision. Execution is largely a function of having a map of the high level vision and low-level realities in your head, or just having in-the-moment intuition about these things... then just speaking it.

Also, there is developing personal libraries. You may find that in your business you repeatedly tackle similar problems. You develop and refine a library that your future agents can utilize. You develop prompts that make your work more automatic and more robust. And when I say you, I mean, you direct your agents to do it.

It's not so different from what it always was, except that years get compressed into months. What was slow and tedious is now... less so.

I don't know how much advantage I have in this new world, having programmed (sometimes more, sometimes less) for 30 years. My brother has never written a line of code in his life. He's been heavy into AI for a while, and he has developed some really incredible stuff. I think he may have already outpaced me. His $1500/mo+ in subscriptions definitely help.

Uncle Ted warned us about this.

From one perspective, the 5% estimation is near-infinite orders of magnitude off, since they've trained on something approaching the sum-total of human knowledge.

Interestingly, the Amish shape their lives around this sort of thing. With any technology, they ask themselves, "Does this make us dependent?" If the answer is yes, they refuse it.

It can replace all work, IMO. I'm doing intellectual work that replaces manual labor with machinery and computer vision systems. It's massively sped up by AI.

Can you not seek mastery with the new tools, though?

It does feel a bit like building on quicksand, but coding has always been that way for me. Well, less so with embedded than webdev, but still.