an agent harness built in rust with ratatui - checks out. i've built one myself. i don't maintain it, and continue to use opencode, but it was worth it to learn how agent harnesses work.
anyway, what's the real pitch on why i should move on from opencode to maki?
butterfly is interesting because it's faster than breaststroke (mentioned) but slower than freestyle. it also consumes far more energy than any other stroke.
to that end, i'm not sure why it exists, except that it's truly a unique style.
* i also still hold my high school's butterfly record, 20 years on.
> to that end, i'm not sure why it exists, except that it's truly a unique style.
Many people swim as a form of exercise. Fly is exercising different muscles and allows me to get my heart rate up higher than freestyle
Fly is useful to train for other strokes
Perhaps more importantly, I think that having a different stroke to do makes swimming more interesting. Whether doing sets as part of a swim team or on your own, it's more interesting when you can vary things. The more swimming is interesting, the easier it is to enjoy and keep doing it
The modern Olympics is at least as much for entertainment as it is for measuring human ability, and the butterfly is simply an awe-inspiring technique. The line of swimmers repeatedly shooting out of the water like flying fish is mesmerizing. Who cares if they're not going as fast as freestyle?
I think the objective is to show how strong you are. If you wanna go fast, use freestyle, if you want to conserve energy, use breaststroke or backstroke. I don't see a reason to use the butterfly outside of a butterfly competition.
In my case the reason is to exercise. I really like it as an exercise. Though I also practice the other strokes for more variety, and fun, when I'm short of time butterfly is my choice.
agree. it's strange reading the loud voices that are counter to my lived experience. llms just have seemingly infinite depth - or can at least debug and execute without fatigue.
You're not crazy. For a precious few things in my career in software, it's been really important to have an incredibly genius idea of how to design a system or how to actually implement it. Maybe 2% of the time. The rest of the time, you could enlist an engineer with 2 years of experience and enough sense not to do stupid things (like eval()ing user input, etc). They don't need to be a genius to trial-and-error their way through everyday coding problems -- you'd just have to give them enough time and feedback and they'd be able to build everything. To me that points to LLMs not needing to be AGI to replace 98% of our jobs, since they're so much faster and so cheap.
Note: I am mostly not comfortable with this world we find ourselves in, but I'm just saying what I've observed.
admittedly, i've not really cracked FE dev with LLMs at this point (and it's probably my big weakness). but, i'd heard somewhere that FE just isn't there yet - though i was suspicious of that claim.
i'm torn about sending screenshots to an LLM for debugging - seems imprecise. seems lossy, especially compared to inspecting the dom. however, it's always proved good enough (e.g. when messing with ratatui.rs and tui-pantry). similarly for web, maybe it's about decomposing into storybook. hmm. the next grand adventure i need to hack.
anyway, fascinating investigation of fable just automating that entire process and what it didn't automate, too.
Fable is really good at front end (Opus 4.8 is decent too) but it really needs a verification loop - it can't always infer the output from the code alone. Give it Playwright to check its work, and it'll generally do a good job. Also if you're using a framework, add to your CLAUDE.md to always rtfm before making changes!
the problem is scale. there's a tension between an individual developing technical skills (transfer cost is high, slow, expensive) and developing agent skills (transfer cost is low, instant, free).
so, just like a manager manages employees, or you consult a contractor, agents are a way of getting leverage over a system.
that said, if you want to learn to play saxophone, you're free to do so. just note your personal endeavors may begin to look more like hobbies than marketable skills.
When the cost for this leverage is more than an employee the math stops mathing.
Additionally, for tech work. There is a tension about doing work and not knowing that output is correct or not. I have seen ai spit out thousands of lines of opencv code for a simple color lut. The person doing this had no idea what was going on. If they continued, the token cost and time waiting for agents spinning only goes up.
Yes, agents get smarter and cheaper but the above example replays over and over again even on crud apps. You still need to dev the skills and transfer costs for it to be effective.
both points are fair, but operate at different levels. the former: willpower. the latter: constraints.
and, the latter is indeed dependent on the former. but, arguing that humans have no free will is an argument that should be tried independently of rebutting the former comment.
if you require a stable energy source, neither wind nor solar (nor both) provide a complete strategy. they can be a part of a composite strategy, though.
for both wind and solar, they're also quite taxing on the environment during manufacturing. the "true cost" is rarely reported.
anyway, what's the real pitch on why i should move on from opencode to maki?
reply