Tyler King

I understand the fomo. It's misplaced, but the culture right now is "don't get left behind" and it's hard to know exactly how much you need to pay attention to avoid that

This week was especially bad.I hadn't noticed downtime before but according to their status page, even before this week they only had something like 99% uptime, so it was never especially good.

Unashamedly judging people is one of life's great joys

It feels like if you don't adopt every new AI model and switch coding agents every week, you'll get left behind. But let's just remember that being, like, three weeks behind the cutting edge is totally fine, and you can get that by sticking to whichever tool you're using now.

Even their support link didn't work with the AI features turned on. Canceling the AI trial fixed everything.

The "companies are just shoving AI slop into their products" narrative isn't wrong. Can we please slow down and make sure our new features actually work?

The rise of the PM was Silicon Valley's greatest mistake. Back to a single product visionary and a bunch of coders that don't care about business results at all!

Yesterday I started a trial of Notion's new AI features. They mostly don't work, but they have made base Notion way buggier, so there's that!

The NameCheap navigation bar. Out of nine links, five have a big red "New" badge and one has a blue "TRY ME" badge. It's very busy and chaotic looking.

omg namecheap chill out

Better than the other AI tools, but waaaaay worse than a successful SaaS app from the pre-AI days. So many bugs, downtime, unhandled edge cases, unpolished UI, etc.

Maybe I'm a fool, but I'm hoping that one day quality and not just shipping speed becomes a priority once again.

I'm really curious to see how Anthropic does when things eventually slow down and people start expecting a bit more polish/maturity from their tools.

Do they have the DNA to build polished software? Are they accumulating too much tech debt due their frantic pace right now?

It's a point of pride for me to give customers what they want without giving them what they asked for. Obviously this can be taken too far, but solving customer problems in ways they didn't think of generally leads to more elegant products. It's how you make a product feel premium and intentional.

Just had a customer tell us they retired, but they keep paying for their LACRM account just to support us because we're one of the few companies that still values customer service.

Obviously it's hard to build a business around getting donations from retirees, but it sure does feel good.

The customers of your book are on indie hackers, no?

I feel so sad when my relatives who live in much "cooler" cities and/or deep suburbs come to visit and they didn't even realize America had neighborhoods where there are people walking around outside but it's still affordable.

We have lots of old rustbelty cities that were built before cars ruined everything, and are extremely affordable.

Sure, public transit and bike infra sucks compared to other countries, but I feel like our great neighborhoods deliver in pretty much other way. And the income:cost ratio is incredible.

Americans often complain about here vs. Europe. Not the obvious political/cultural clusterfuck, but just stuff like walkable neighborhoods and a real sense of community.

It's sad most of America is so hostile to live in, but you don't have to cross an ocean to solve it. Just move to my neighborhood

I don't know. Anyone using a cloud provider (AWS, etc.) could easily save 50% by hosting somewhere else, and yet we don't. For most of us, infra just isn't big enough of an expense for it to be worth spending time on that. And I'd much rather change hosts than languages.

Wow, I'm flattered. Seriously, thanks for the shout out. Makes me want to start writing again.

Whoa, that's really a throwback. Wish I could go back in time and give some tips to that version of myself.

Thanks for listening!

Sometimes I hate my dumb brain. One of my three monitors broke a couple months ago, and I've been meaning to replace it, but it wasn't urgent because two monitors is plenty.

Now that I've ordered the replacement, I want it NOW and I'm getting frustrated having to use just two monitors.

I'm embarrassed to say that I *love* LinkedIn's daily games (especially Queens) so I can't bring myself to delete the app.

I know most people love it, but I've always hated Apple's approach to software design. It's getting worse, but it was never good to begin with imo.

Too much of it is focused on the appearance of simplicity ("we hid all the useful stuff so the screen is clean!") rather than being actually simple.

Customers often compliment LACRM by comparing our design to Apple's ("it just works"). I know they mean well but I take this as a grave insult.

This just goes back to something I've been feeling more and more: As much as being "data driven" feels sensible, it's actually very difficult, and chasing the numbers can sometimes lead you in completely the wrong direction. Another win for intuition.

So this metric I've been using to measure our ability to convert people is being completely overpowered by a different unrelated thing. Even worse: The metric looks *worse* the more successful we are at marketing in general.

The fixed leads convert at a very high rate. The variable leads convert at a lower rate (they're comparison shopping). As a result, the more leads we have in a given month, the lower our trial-to-paid conversion rate is.

Of course we get more *total* conversions, but the conversion rate is worse.

In a given month, we have a more or less fixed amount of traffic from very qualified people. Old customers starting new accounts, people who have been ordered by their boss to use us, etc.

On top of that, there's a variable number of additional leads from our various marketing channels.

More in the adventures of "stats are less useful than you think"

We've been focusing on the trial-to-paid conversion rate. That seems like the best way to measure whether our main strategy is working.

But I just realized this metric has actually been lying to us.

There's a checkbox they have to click to advanced. I get it though. GitHub shows me something similar all the time and I click through without thinking.

We're doing more or less that with Claude Code. It walks through the PR and you can ask questions, etc.

Also a neat thing we're trying: For our newest hire who is still learning the code base, CC quizzes her on any AI-generated PRs so she has to understand it even if she didn't write it.