jona777than

I have to admit the merging of worlds here is wild. I never expected to read Mike Tomlin, long time coach of the Pittsburgh Steelers on HackerNews yet here we are. It was not on my Bingo card. Refreshing, indeed.

This comment gave me an uptick in hunger, instantaneously.

This article reminds me of a time when I was pushing my kid on a swing in the backyard after work. It was all I was doing. No background processes running mentally. With all the obligations and people waiting on me at work and deliverables due yesterday, I had managed to turn it off. It occurred to me that simply pushing my kid on a swing and thinking of nothing else was probably the most important thing I could ever do in that single moment. That was a freeing change of perspective for me.

I have flip-flopped more than ever in the last 365 days about prioritizing good code vs good product, in the AI age. This helps clarifies why.

I have come to the conclusion that we just do not know yet. There is a part of me that believes there is a point somewhere on the grand scale where the code quality genuinely does not matter if the outcome is reliably and deterministically achieved. (As an image, I like to think of Wall—E literally compressing garbage into a cube shape.)

This would ignore maintenance costs (time and effort inclusive.) Those matter to an established user base (people do not love change in my experience, even if it solves the problem better.)

On the other hand, maybe software is meant to be highly personal and not widely general. For instance, I have had more fun in the past two years than the entire 15 years of coding before it, simply building small custom-fitted tools for exactly what I need. I aimed to please an audience of one. I have also done this for others. Code quality has not mattered all that much, if at all. It will be interesting to see where things go.

Is it just me or could this apply to commentary as well? Sometimes, I set out to comment with all my thoughts and their intricacies related to the subject, but sometimes the simplest one contributes far more to the conversation. In my experience, simplicity enables others to more freely participate and contribute.

I highly relate to this. Code works or it doesn’t. My writing feels a lot more like self expression. I agree that’s harder to “let go” to an agent.

> If the original function isn’t working as expected, I suspect that the agent-created test will test the functionality as it exists, not as was intended, regardless of what it calls the test case.

I have experienced this on many occasions. It ultimately adds up to a sneakily false sense of code stability.

Love the simplicity. These sort of “less is more” solutions should become more prevalent as average code volume continues to rapidly increase

> It was at the same time depressing (I’m dumb), liberating (I don’t have to pretend I’m not dumb anymore) and exciting (I have a chance to be not dumb anymore).

> Learning is hard work, and if you don’t respect the process, it won’t happen.

These two ideas resonate well with me. My experience in pursuit of steady and sustainable growth in any area of interest has had these in common. You have articulated them well enough for me to realize that. I appreciate that.

I am also at a similar point in life that sits at the intersection of building consistent habits that support goals and balancing priorities like family life. “It’s a marathon, not a sprint” has never been more applicable.

I initially found Gemini Pro 2.5 to work well for coding. Over time, I found Claude to be more consistently productive. Gemini Pro 2.5 became my go-to for use cases benefitting from larger context windows. Claude seemed to be the safer daily driver (if I needed to get something done.)

All that being said, Gemini has been consistently dependable when I had asks that involved large amounts of code and data. Claude and the OpenAI models struggled with some tasks that Gemini responsively satisfied seemingly without "breaking a sweat."

Lately, it's been GPT-5 for brainstorming/planning, Claude for hammering out some code, Gemini when there is huge data/code requirements. I'm curious if the widened Sonnet 4 context window will change things.

This made me chuckle. Excellent comparison. I share your sentiment; it’s more exciting than distracting.

You have seemed to pinpoint where I believe a lot of opportunity lies during this era (however long it lasts.) Custom integration of these models into specific workflows of existing companies can make a significant difference in what’s possible for said companies, the smaller more local ones especially. If people can leverage even a small percentage of what these models are capable of, that may be all they need for their use case. In that case, they wouldn’t even need to learn to use these tools, but (much like electricity) they will just plug in or flip on the switch and be in business (no pun intended.)

One of my favorite use cases includes simple tasks like generating effective mock/masked data from real data. Then passing the mock data worry-free to the big three (or wherever.)

There’s also a huge opportunity space for serving clients with very sensitive data. Health, legal, and government come to mind immediately. These local models are only going to get more capable of handling their use cases. They already are, really.

I agree. It’s not improbable for there to be _more_ needs to meet in the future, in my opinion.

As a sidebar, I’m still not sure if GPT-5 will be transformative due to its capabilities as much as its accessibility. All it really needs to do to be highly impactful is lower the barrier of entry for the more powerful models. I could see that contributing to it being worth the hype. Surely it will be better, but if more people are capable of leveraging it, that’s just as revolutionary, if not more.

After 16 years of coding professionally, I can say Claude Code has made me considerably better at the things that I had to bang my head against the wall to learn. For things I need to learn that are novel to me, for productivity sake, it’s been “easy come; easy go” like any other learning experience.

My two cents are:

If your goal is learning fully, I would prioritize the slow & patient route (no matter how fast “things” are moving.)

If your goal is to learn quickly, Claude Code and other AI tooling can be helpful in that regard. I have found using “ask” modes more than “agent” modes (where available) can go a long way with that. I like to generate analogies, scenarios, and mnemonic devices to help grasp new concepts.

If you’re just interested in getting stuff done, get good at writing specs and letting the agents run with it, ensuring to add many tests along the way, of course.

I perceive there’s at least some value in all approaches, as long as we are building stuff.

It's nice to get a glimpse behind the scenes of JEfit.

I used this app when I got serious about my fitness journey around 8 years ago. I fell off from using it 4 or 5 years ago (no fault of the app.) I can honestly say, it made it really easy to stay consistent with my workouts and show up to the gym confident in my programming.

Perhaps what I love best about this story, and similar startup stories, is the purity of building something to solve a problem personally. Then when the success of that thing happens as a side effect, it seems more appropriate. Stories like this take me back to the simple joy of creating something useful.

> but then hit a brick wall

This is my intuition as well. I had a teammate use a pretty good analogy today. He likened vibe coding to vacuuming up a string in four tries when it only takes one try to reach down and pick it up. I thought that aligned well with my experience with LLM assisted coding. We have to vacuum the floor while exercising the "difficult skill [of] continually avoiding temptation to vibe"

> That may be good enough.

I would argue it is.

I have had discussions with peers recently around doing the big flash-y <insert revolutionary product>. An interesting analogy surfaced. The nuts in the studs of the infrastructure of the many structurally sound homes in existence are just as important (meaningful) as the doors, windows, and more flash-y features. They may be _more_ important in some cases. They all make up the home.

It made me realize it might not be all about maximizing ambitious pursuits. Maybe it is more about experiencing the joy of solving the next problem and the fulfillment that comes from simply being needed pretty regularly.

This is a good point. I thought about it, too.

I could see a frequent traveler using an AVP as a "full setup" on the go. In my experience, I can get away with most with a MacBook. Some projects really benefit from the extra screen real estate (and a mechanical keyboard.)

> work every day, ~8h/day

This is fascinating. What are your most used features?

> extended monitor

Do you also use a real monitor in the field of view?

I recently wrote my own invoicing application. I got wrapped up in the joy of adding the features I wanted. Many of these features come at a premium monthly fee for a big named product.

I needed to get an invoice out in a timely fashion. As much as I wanted to use my app, I found certain kinks I needed to work out (with styling, adding addresses, etc.) -- This was where I realized what you have articulated.

At some point, it becomes better to prioritize the "fun" in working on my bike, and the "usefulness" of the daily driver bike.

In doing so, perhaps the fun/usefulness of each converges over time.

I personally have found it effective to oscillate between having a goal and defining constraints in some _direction_. There are points on the journey where articulating a specific goal is helpful. At other times, I get more results with a few well-defined constraints. It can depend upon the season, but both have been solid tools for progress.

I would like to think of this as Operational Gravity, or something like that. Might help it stick. Good quote. Thanks for sharing.

I find that orienting around results can help unlock whether positive or negative space (a goal or constraints set) is the better focus. In my experience, there are times when goals do not serve me, but rather hinder me. This is purely from regularly observing results. In those cases, pivoting to a focus on some well defined constraints has yielded better results. As long as the direction is the same between the two, that might still be considered proactive.

I had this same thought immediately and had a good chuckle.

Ironically, my colleagues and I used “but can it center a div?” just today. It has been our running joke to measure usefulness of LLMs as they advance.

On a more humorous note, this ought to make for an interesting store checkout experience. “Would you like paper or… paper?”

This has been my experience. It has been something I’ve had to settle into. After some reps, it is becoming more difficult to imagine going back to regular old non-assisted coding sessions that aren’t purely for hobby.

Your model rankings are spot on. I’m hesitant to make the jump to top tier premium models as daily drivers, so I hang out with sonnet 4 and/or Gemini 2.5 pro for most of the day (max mode in Cursor). I don’t want to get used to premium quality coming that easy, for some reason. I completely align with the concise, thoughtful code being worth it though. I’m having to do that myself using tier 2 models. I still use o3 periodically for getting clarity of thought or troubleshooting gnarly bugs that Claude gets caught looping on.

How would you compare Cursor to Claude Code? I’m yet to try the latter.

This was the piece that stuck out to me as well. It articulates well the distinction between the good managers I have had and the bad ones as I recollect my work history. This is applicable in close personal relationships as well.

There will likely be more jobs because of AI. With more “knowledge”, comes more responsibility. Spam folders only exist because of automated emails. That classification process is more work. We may find there are more needs to meet as AI advances, not less.

The fallacy is in the statement “AI will replace jobs.” This shirks responsibility, which immediately diminishes credibility. If jobs are replaced or removed, that’s a choice we as humans have made, for better or worse.