I think there’s a clear split amongst GenAI developers.

One group is consistently trying to play whack-a-mole with different models/tools and prompt engineering and has shown a sine-wave of success.

The other group, seemingly made up of architects and Domain-Driven Design adherents has had a straight-line of high productivity and generating clean code, regardless of model and tooling.

I have consistently advised all GenAI developers to align with that second group, but it’s clear many developers insist on the whack-a-mole mentality.

I have even wrapped my advice in https://devarch.ai/ which has codified how I extract a high level of quality code and an ability to manage a complex application.

Anthropic has done some goofy things recently, but they cleaned it up because we all reported issues immediately. I think it’s in their best interests to keep developers happy.

My two cents.

I kind of wonder if people with ADHD tend to fall into the latter group, as we are used to setting guardrails to keep us aligned to a goal.

ADHD may be a part of it but mentoring and some experience with TDD or strong unit testing is also a part.

FYI that prominent link to your sharpee repo on GitHub 404s

OMG (fixed) - I updated devarch to check all html links for accuracy. This came from a recent update to the website to simplify it and of course Claude completely hallucinated that URL.

You can NEVER stop being vigilant. This is why I still have no faith in things like OpenClaw. Letting an AI just run off unsupervised makes me sweat.

Dead on. Any company not thinking about this like the 2nd group is setting themselves up for a bad time (and sadly, anecdotally, that seems to be an emerging majority).

Sadly, I agree that the majority are undisciplined.

IME it seems that output quality is directly proportional to the amount of engineering effort you put in. If a bug happens and you just tell the model to fix it over and over with no critical thinking, you end up with an 800 line shell script meant to change the IP address on an interface (real example). If you stop and engage your brain to reason about bugs and explain the problem, the model can fix it in an acceptable manner.

If you want to get good results, you still have to be an engineer about it. The model multiplies the effort you put in. If your effort and input is near zero, you get near zero quality out. If you do the real work and relegate the model to coloring inside the lines, you get excellent results.

Even my guardrails can’t replace experience. You have to pay attention. This is exactly how some devs land in whack-a-mole loops.

OP here. I've been developing project-oriented guardrails for Claude (Code) since Anthropic released the first web version of Claude. It has since evolved with skills, directives, hooks, and guardrails based on Domain-Driven Design principles and my forty years of software engineering and architecture experience.

very cool guardrails :)

I used my moms Koss headphones as a microphone on her old stereo and used it to broadcast cast my teenage stupidity to the whole house.

This is the lie. Our economy generates 27 trillion gdp.

We can afford everything. We choose not to.

It's more likely both are true. We can afford to do more for the people, but at the same time we are over-spending. Streamlining some of these functions would be nice. One area we are vastly over-spending is highway and roadway construction, for example. Even if we can afford it, we shouldn't pay for it. There are other more politically hot topics here and both general sides of the debate have merit, but we should try to not be dogmatic about it and instead think in systems terms and long-term outcomes. When I see a city or state spending $400,000/each on units for housing homeless people, well, that's obviously a misuse of funds. That's not sustainable. We shouldn't do it even if we can afford it. When we spend $50 billion in a week of the Iran war (which I support but just as an example), well, that $50 billion could have paid off a lot of mortgages - so maybe we should or could do that instead.

Just to note. McKenzie Scott has given half of her divorce settlement away and still has more than when she started.

No one needs to be a billionaire. It’s inhumane.

I don't like or valorize billionaires, I guess (I mostly don't care about them), but I don't understand what's "inhumane" here. There aren't very many billionaires. Billion dollar companies are far more salient to ordinary people than billionaires are. And, obviously, you can't fund universal health care by liquidating the billionaires!

I've never really understood why people are so het up about billionaires. The distinction between them and decimillionaires seems mostly like comic book lifestyle stuff; like, OK, they fly their pets private for visitation with their ex-spouses or whatever, I guess that's offensive aesthetically?

Far, far more damaging to ordinary people is the Faustian bargain struck between the upper middle class and the (much smaller) upper class, which redistributes vast sums of many away from working class people into the bank accounts of suburban homeowners.

(Because fundamental attribution error guarantees threads like this will devolve into abstract left vs. right valence arguments, a policy stake in the ground: I broadly favor significantly higher and more progressive taxes, starting with a reconsideration of the degree to which we favor cap gains.)

I really applaud the work McKenzie Scott is doing. A lot of billionaires play the "aw shucks if only someone would tax me" - nothing is stopping them from just donating to the government if they really thought that. We have a housing problem, why not play Sim City in real life and build houses for people or something? Personally I think it would be a blast.

Similarly though, there's nothing stopping you personally from taking $50, $100, whatever and walking over to a shelter or food bank and donating. You don't need to wait for the government to stand up a program. Lead by example like McKenzie Scott is. We donate money to local organizations - again, no barriers here.

I don't care if someone is a billionaire, though of course we should tax them "appropriately". But if you're really mad about billionaires and you want these programs, you should be giving away your own money too and there's nothing stopping you. Waiting until you get just the right program or tax the right person is a bad strategy if you really care about some of these issues.

Yeah, it's always funny to see how MMT is a perfectly acceptable way to create tax cuts and enable corporate welfare but if you suddenly want universal medicare or childcare suddenly we care about budgets or MMT is suddenly impractical.

This should actually allow for a balanced budget and still affording everything. The problem is, the USA has the best government money can buy and it wasn’t bought by the people.

Not if you believe the (obviously wrong) laffer curve zealots

Eh, the way the US does a lot of things have significant cost problems.

Public spending on healthcare is around 8-9% of GDP once you add things up.

So you have already paid for a public healthcare system in many ways.

The New Deal effective tax rates on individuals and corporations plus subsidies for public college created the greatest middle class in history. It also barred banks from loaning Wall Street money for speculative and risky adventures.

In 1980 Reagan began unraveling all of the pillars of that middle class success.

46 years later you can see the damage. Housing is unaffordable even for professional couples. Public colleges are gated to the upper classes.

There is no middle class anymore.

> There is no middle class anymore.

There is a middle class, but it is the worst place to be. The best two places to be are rich or poor. The middle class are all suckers, they are making the rich richer and are supporting the poor.

(I am poor.)

The traditional economist/polisci view is that laws favor the middle class (the largest voting group) at the expense of the poor (who don’t benefit from most services) and rich (who pay most taxes).

See “Director’s Law”

The non-college educated working class was greatly reduced (in size and wealth), but mostly by trade and automation. In its place we have a smaller white-collar (college educated) middle class.

This is very silly. College attendance is at record highs, far above 1980.

At what expense to college loan sufferers?

If this isn’t an indictment of MS management (pun intended), I don’t know what is.

I don't understand. You're proposing a future where bugs are replaced with entirely new sets of bugs instead of fixed?

This sounds like one of the worst possible uses of generative AI that doesn't directly result in the loss of human life.

Well you make the assumption the new code will be worse.

I don’t think that will be true.

Note: if you haven’t used Claude since December then you’ve not seen the massive difference in quality code generation. All the studies prior to December are irrelevant.

No

I was a career Microsoft stack developer until Azure. Comparing it to AWS immediately forced me to make a decision to move away from their stack and towards AWS.

Just the networking and security infrastructure was complete trash compared to how those things worked in AWS.

Not one regret in my decision.

I run MalwareBytes on all my browsers and as my computer protection system.

LinkedIn is getting nothing.

Lol, you forgot the /s

I switched to Codex out of frustration with Claude Code and it has been surprisingly similar for my web and mobile coding needs

Except for censoring. It didn't allow me to use a screencapture library, thinking I was going to hack the world.

I sleep good at night, knowing that chatgpt saves me from the world wide web hackers

Gemini is a terrible product, I spent $15K on it. Anthropic and OpenAI make better models, it used to be that Gemini cooked but I don't feel that way anymore