Tyler Hillery

Software Engineer (Storage) @supabase. Creator of Co-organizer of Peeling back the layers of abstraction.

[Tyler Hillery @_TylerHillery]

@ChristianLempa Used Mini PCs

[jdx @jdxcode]

. @supabase is the newest sponsor for mise/en.dev! thanks so much! https://t.co/Fisnt1hVvt

[Tyler Hillery @_TylerHillery]

I'd love to rack and stack a server in a DC someday

[Ovais Tariq @ovaistariq]

Bringing up new Tigris racks online soon for your fast storage needs!

[Tyler Hillery @_TylerHillery]

2 months on Math Academy and I've made it to Diamond League

[Tyler Hillery @_TylerHillery]

2 weeks in and I'm really enjoying Math Academy so far

[Tyler Hillery @_TylerHillery]

@BenjDicken @_patrickbrophy It sure is

[Tyler Hillery @_TylerHillery]

@BenjDicken Join us

[Tyler Hillery @_TylerHillery]

@kenwheeler Join us

[Mitchell Hashimoto @mitchellh]

I've got an agent in a loop optimizing a renderer with the goal to minimize frame times (and tests to measure). It got times down from 88ms to 2ms and allocations down from ~150K to 500. Sounds good, right? Wrong. This is exactly why agent psychosis is a big fucking problem.

As an experiment, I rewrote the Ghostty core render state in Go, with access to identically laid out data structures as Ghostty and the exact same validation tests. I made a purposely naive renderer (simple, correct, but slow). 88ms per frame with 150,000 allocations (horrendous, lol)!

I then kickstarted a Ralph loop to bring the frame times down. I told it it can't modify input data structures or the public API or tests (they're correct), but it can do anything else it wants. It got to work.

It has worked for about 4 hours. I've spent around $350 on this experiment so far. The results?

88ms => 1.5ms
150K allocs => ~500 allocs

Incredible right? Nope.

My hand-written renderer I ported has frame times (same benchmark) of ~20us (0.020ms) and 0 allocations in the update path.

This is the problem with psychosis and lacking systems understanding. If you don't understand the system, you're going to accept that this is an incredible result. If you understand the system, you'll see better solutions immediately and can do roughly 75x better on throughput.

The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity.

Standard disclaimer: I use AI all the time. I like AI. The point I'm making is to not blindly accept results. Think. Analyze. Learn.

[Tyler Hillery @_TylerHillery]

Finally pulled the trigger on a kindle and I absolutely love it. Should have gotten one ages ago

[Tyler Hillery @_TylerHillery]

@steveruizok Maybe https://t.co/FuGql4sFdN?

[Tyler Hillery @_TylerHillery]

@JustJake Reminds me of this podcast episode https://t.co/LYieo1RYQX

[Tyler Hillery @_TylerHillery]

Another great Madison Systems paper discussion!

[Tyler Hillery @_TylerHillery]

Reminder, Madison Systems is having another paper discussion in 2 weeks. This time, we'll be discussing: Scalability! But at what COST?

[Tyler Hillery @_TylerHillery]

@merm_bot Yeah I wasn't proud of the form on this one

[Tyler Hillery @_TylerHillery]

Form was sloppy but finally hit the 405 deadlift

[Tyler Hillery @_TylerHillery]

This is the way. It's Saturday 7:20am and I've already finished my daily Math Academy goal.

[Justin Skycak @justinskycak]

If your excuses keep postponing your training, do it first thing in the morning before your excuses finish loading.

[Tyler Hillery @_TylerHillery]

I really like @ReadwiseReader's chat feature to quickly ask questions about a passage

[Tyler Hillery @_TylerHillery]

@Keviniskovitch @Dexerto This is how my back looks

[Tyler Hillery @_TylerHillery]

3 more days, get to reading!

[Tyler Hillery @_TylerHillery]

Reminder, Madison Systems is having another paper discussion in 2 weeks. This time, we'll be discussing: Scalability! But at what COST?