thehucklecat

I have no idea how many hours I work.

Just no idea how to measure it.

Hours at laptop? Hours thinking about it? A 1 min slack reply at 8pm, does that count as 1min?

I could argue that I work anywhere between 25 and 60 hr / week.

Do most people use something like stripe terraform? or are you all making products / prices in the Stripe UI? or custom code to create these objects?

I have always used the UI, but felt guilty about it.

"A key function of good enough parenting is to provide the essential background to allow for the growing child’s disillusionment with the parents and the world, without destroying their appetite for life and ability to accept (external and internal) reality."

is a banger quote.

In practice it is so hard to just go along with this disillusionment. I can't do it. Kids come home totally bored at school and I'm supposed to say "get used to it, your life will be full of drudgery!" Nah.

i got 6. surprised that it was that few, I figured it would take forever.

I appreciate that the MAU sliders are powers of 2. It really needs the enterprise pricing though. ld pushes you into those plans in surprising ways so it's easy to think you're going to get to stay on pro, but then surprise! it's enterprise time.

No standardized testing to get into an ivy league college was a ZIRP.

In my startup we track another number in order to normalize our compensation.

We do a quick "expected valuation" calculation on the share price and use that, instead of the 409a or the valuation. This is how I actually value it in my own head, so we just kind of canonicalize.

example: I think there's a 1% chance 1billion a 5% chance we make 100M. A 40% chance we make 20M. So that's a 23M and then I calculate the value of an option based on that.

Using that, I can then try to "match" a salary from a public company. So we set our comp as 80% of the google levels.fyi data. (Chose google bc it has the fullest levels.fyi data).

This gives us a full compensation benchmark for any roles / levels. ie a Sr engineers makes 263k. But we pay 140 in cash and 123k in equity.

Then I can explain to an engineer. We feel like we're paying you as well as you would be paid at google, but you need to believe that we have a 1% chance at a billion. 5% chance at 100M etc. They can easily tweak these expectations too so they can compare offers. If they think there is a zero percent chance of 1B they can adjust the offer themselves in their head.

I think those are all good choices and I don't judge you for k8s. It all depends if you know it already. We use GKE with auto-scaling kubernetes and it's great and took 2 day to get it terraformed sufficient to where it feels like what I had at my old job.

If/when you need to start "real company stuff": AngelList Mercury Ramp Rippling HubSpot

I've been pretty impressed with all of those.

I really enjoyed geoguesser.

Don't forget: https://fmworldcup.com/excel-esports/microsoft-excel-world-c...

Mistral is the one that made it really click for me that OpenAI may not be the winner at all.

Anything can work.

If the employees are mediocre it takes the chance of success from 11% to 5%. Or something like that. The point being that it's unlikely to "do well" regardless.

More important is to figure out what you're learning. If you're on the steep part of the learning slope that's all that matters. If it's early stage and it starts to work a bit, you'll get more responsibility and you can be the one to let these people go.

But if you aren't learning or are just spending time with bozos, then it is a waste of your time.

I was a non-believer for a while too. My first impression of node was definitely that it was a shell game.

But ease of use matters and I have to admit that a lot of node apps really scream and scale for developers of a lot of different skill levels.

I am really excited about java project loom virtual threads. This feels like the most have your cake and eat it too thing I've seen in a while.

I definitely feel this pain. Changing environment variables at works sucks. JIRA ticket and then back and forth with SRE about whether we really need it.

So it all goes in code and defeats the whole purpose of having the app be configurable.

I don't know if secret management should be part of the same system though.

I want a cookbook that only has recipes for my family ie: -gluten free -vegetarian -no mushrooms -no goat cheese

I hate having to skip past 60% of the recipes in a book. Feels like this is a fun GPT project waiting to happen.

NPC is a funny term here. That makes me think it's putting the character inside the game.

Is this just like a personalized walkthrough helper?

what kind of application changes are you thinking it would equire?

my policies are definitely too broad, but feels like I should be able to tighten them up without changing code. (just potentially breaking things if I get it wrong and go too tight).