The question is so loaded… these are tools. I’m evaluating them like any other tool and using them where they’re helpful and not where I perceive them to be unnecessary or cause trouble.

It also seems to me that some people are leaning on them for things they should not be. It can help you research psychology; it cannot be your therapist.

In these strange times, apparently that makes me “resisting” or “a Luddite.”

This isn’t going to happen.

I can already build a ticket tracker in a weekend. I’ve been on many teams that used Jira, nobody loves Jira, none of us ever bothered to DIY something good enough.

Why?

Because it’s a massive distraction. It’s really fun to build all these side apps, but then you have to maintain them.

I’m guessing a lot of vibeware will be abandoned rather than maintained.

Teams skip DIY trackers because maintenance is expensive now if agents cut that cost a lot more teams will tolerate the distraction.

Who said you’re building it? You’re telling your AI to build it while you go play golf or something.

> Who said you’re building it? You’re telling your AI to build it while you go play golf or something.

If AI can build your software for you, you aren't going to be playing golf, you're going to be standing in line in the soup kitchen.

The hard part has always been shipping, buttoning things up, doing the design. Not the idea per say. And then if any of it is successful and starts making money guess who you're gonna call to maintain it?

These are local systems. Think of it like vibe coding your personal GUI or CLI. Each programmer uses their own custom build. There's no maintenance except only for themselves.

You typically use an off the shelf project management software because it's too time consuming to build one catered to your own preferences. But with AI, it just does it for you. I'm talking about custom one off personal solutions readily done because of AI executing on it for you.

That's not how anything works.

Seriously. For starters I don’t even play golf!

And it’s all downhill from there…

That's how vibe coding works.

But they don’t. Seriously, do you read?

Books encode skill.

I’m not a hater. LLMs on search is the best research tool I’ve ever used because it’s read everything and can find minutia buried in places it would take me a long time to find.

But there’s a huge difference between using it to assist focus, or as a study aide, and offloading the whole act of thinking itself.

This isn’t true. Firefox users get really salty about this. They really will drop your product.

Absolutely. I will refuse to do business with anyone who is hostile to the open web.

Wait, most of the major services I use operate on inquiry forms, email, and texting.

Where are robots and unreturned voicemails the only two options?

This is highly regional within the US. You’ll see a lot more of it amongst NYC professionals than anywhere in the Midwest.

Yeah and I think it reflects the sorts of jobs people have. A lot of my buddies have these jobs they themselves think are pretty stupid but hey, they were hiring. They aren't going to identify as some salesperson of payroll software though. Probably literally no one in that industry does. It isn't rocket surgery. It just pays the bills.

So I would like to know how it found the proof. Because it’s much more likely to have been plucked from an obscure record where the author didn’t realize this was special than to have been estimated on the fly.

This makes LLMs incredibly powerful research tools, which can create the illusion of emergent capabilities.

Here's the PDF: https://www-cs-faculty.stanford.edu/~knuth/papers/claude-cyc...

It wasn't Knuth who used Claude, but his friend. Nevertheless, Knuth was quite impressed.

If you read https://www-cs-faculty.stanford.edu/~knuth/papers/claude-cyc... it was more of a guided effort to write a program to find examples that helped with moving the proof along

This answer really isn’t good enough. The providers can’t both aim to replace search and claim PhD level intelligence that will do all the jobs, but hide behind “it makes mistakes” in small print.

I think it's the fluency. Other tools fail visibly. A bad search result looks like a bad search result. A hallucinated quote reads exactly like a real one. There's no signal in the output itself that something is wrong. You have to go back to the source to check, and the whole point of using the tool was to not have to do that.

I'm not making excuses for the providers either. But seeing through the inflated claims of commercial service providers is not a new skill.

True, and fair. But there also seems to be something about these that really seems to throw off otherwise skeptical people.

I suspect it was sneaky.

The old Nokia in school wasn't a problem. You get in trouble for playing snake. The iphone 1 wasn't really a problem. There weren't that many, and it served as a calendar.

But year after year, release after release, the industry deliberately loaded more and more addictive machinery, pushed more and more boundaries, until it's beyond unacceptable.

As an aside, it's amazing how hard it is to turn the modern phone into a no-nonsense tool, and I'm an adult with self-control, a deep understanding of dark patterns, and a fully-functioning brain after 3 cups of coffee.

Completely. I'm a software engineer that has a better shot at this than just about anybody, and I have no idea how to give a child a phone that's not just digital crack. If you think ScreenTime etc will do the job you probably have no idea what's actually happening on your child's phone.

You can buy a dumbphone. For example a Nokia 3210 4G.

Interestingly dumb phones are making a comeback.

They disappeared for a few years, but now you can buy a dumb phone, for example running KaiOS, that charges with USB-C and supports modern cell networks. You can even get a Nokia!

There is absolutely no need to buy a smartphone to any kid younger than 15. Now for high school students it's a bit different, they should be old enough to have self control and respect rules to keep their phones in their bags during class.

Oh, entirely. But the hype cycle is such that if you find a legitimate criticism or run into the hard limits of human cognition (there are real limits to multitasking), a lot of people blame themselves.

My pet theory is we haven't figured out what the best way to use these tools are, or even seen all the options yet. But that's a bigger topic for another day.

With the trend going towards devs coordinating multiple agents at once, I am very curious to see how cognitive load increases due to the multitasking. We know multitasking reduces productivity and increases the likelihood of mistakes. Cal Newport talked about how important is to engage in "deep work." We're going in the opposite direction.