Peter van der Zee
@pvdz.ee
tafka @kuvos
eng @socket.dev
- 15yr js/ts
- rust
- ex vercel
- ex fb
- js1k-guy
eng @socket.dev
- 15yr js/ts
- rust
- ex vercel
- ex fb
- js1k-guy
Ah yes, the unknown category is legendary. Its effectiveness is un-de-fined.
🤦♂️
🤦♂️
November 11, 2025 at 5:25 PM
Ah yes, the unknown category is legendary. Its effectiveness is un-de-fined.
🤦♂️
🤦♂️
I guess screw you, Google play, and I'll never leave a review again. So not my problem.
And screw any random games demanding sign up to play without reasons.
And screw any random games demanding sign up to play without reasons.
November 1, 2025 at 7:45 PM
I guess screw you, Google play, and I'll never leave a review again. So not my problem.
And screw any random games demanding sign up to play without reasons.
And screw any random games demanding sign up to play without reasons.
Recognition for Sarah! So deserved! @sarahgooding.bsky.social
October 16, 2025 at 2:50 PM
Recognition for Sarah! So deserved! @sarahgooding.bsky.social
Good story bro, nay, Great story 😁 @voodootikigod.bsky.social
October 16, 2025 at 1:15 PM
Good story bro, nay, Great story 😁 @voodootikigod.bsky.social
Ahw. Tenko didnt make the cut, nor any of my other stuff. Im a failure.
October 14, 2025 at 2:24 PM
Ahw. Tenko didnt make the cut, nor any of my other stuff. Im a failure.
Love a small little idler that you can finish in an hour.
August 27, 2025 at 8:08 PM
Love a small little idler that you can finish in an hour.
Back in SF. Been a while.
August 18, 2025 at 12:31 AM
Back in SF. Been a while.
Self Driving chairs at Schiphol airport. Thats pretty cool.
Flying to sfo today. Want sure if I'd ever see sfo again but here we go.
Flying to sfo today. Want sure if I'd ever see sfo again but here we go.
August 17, 2025 at 10:26 AM
Self Driving chairs at Schiphol airport. Thats pretty cool.
Flying to sfo today. Want sure if I'd ever see sfo again but here we go.
Flying to sfo today. Want sure if I'd ever see sfo again but here we go.
Current status
Ok actually I should just go to bed :p
Ok actually I should just go to bed :p
August 6, 2025 at 12:13 AM
Current status
Ok actually I should just go to bed :p
Ok actually I should just go to bed :p
Had Cursor work batch of tasks
Also, I found a failure case of moves being parsed in correctly, stored it in a comment, no further description.
Mindlessly telling it to continue with next step.
"I finished tasks, but I see that you opened this file and let me fix the problem". Uh, what.
It did.
Also, I found a failure case of moves being parsed in correctly, stored it in a comment, no further description.
Mindlessly telling it to continue with next step.
"I finished tasks, but I see that you opened this file and let me fix the problem". Uh, what.
It did.
August 1, 2025 at 9:34 PM
Had Cursor work batch of tasks
Also, I found a failure case of moves being parsed in correctly, stored it in a comment, no further description.
Mindlessly telling it to continue with next step.
"I finished tasks, but I see that you opened this file and let me fix the problem". Uh, what.
It did.
Also, I found a failure case of moves being parsed in correctly, stored it in a comment, no further description.
Mindlessly telling it to continue with next step.
"I finished tasks, but I see that you opened this file and let me fix the problem". Uh, what.
It did.
Rest assured that I'm not exactly more confident in using libraries (over standards) when vibe coding.
It was so convinced this was correct.
It was so convinced this was correct.
August 1, 2025 at 1:47 PM
Rest assured that I'm not exactly more confident in using libraries (over standards) when vibe coding.
It was so convinced this was correct.
It was so convinced this was correct.
Here's the chess tool I was vibing.
It's a tool to find the best chess line based on a given position.
Built on Stockfish (Claude used the most time trying integrating this).
Nothing ground-breaking but I needed this tool for followup ideas :)
pvdz.github.io/bestchessline/
It's a tool to find the best chess line based on a given position.
Built on Stockfish (Claude used the most time trying integrating this).
Nothing ground-breaking but I needed this tool for followup ideas :)
pvdz.github.io/bestchessline/
July 29, 2025 at 7:42 AM
Here's the chess tool I was vibing.
It's a tool to find the best chess line based on a given position.
Built on Stockfish (Claude used the most time trying integrating this).
Nothing ground-breaking but I needed this tool for followup ideas :)
pvdz.github.io/bestchessline/
It's a tool to find the best chess line based on a given position.
Built on Stockfish (Claude used the most time trying integrating this).
Nothing ground-breaking but I needed this tool for followup ideas :)
pvdz.github.io/bestchessline/
Vibe coding experience so far:
- initial poc: 10 minutes
- implementing stockfish off-the-shelve: 2 hours plus tedious debugging and only solved after explicitly pointing it into an official example (despite telling it to read its docs)
- more feature work: 20 minutes
- polish: forever, of course.
- initial poc: 10 minutes
- implementing stockfish off-the-shelve: 2 hours plus tedious debugging and only solved after explicitly pointing it into an official example (despite telling it to read its docs)
- more feature work: 20 minutes
- polish: forever, of course.
July 27, 2025 at 6:49 PM
Vibe coding experience so far:
- initial poc: 10 minutes
- implementing stockfish off-the-shelve: 2 hours plus tedious debugging and only solved after explicitly pointing it into an official example (despite telling it to read its docs)
- more feature work: 20 minutes
- polish: forever, of course.
- initial poc: 10 minutes
- implementing stockfish off-the-shelve: 2 hours plus tedious debugging and only solved after explicitly pointing it into an official example (despite telling it to read its docs)
- more feature work: 20 minutes
- polish: forever, of course.
I thought, yay, I finally understand the chess problem that was thrown at me at my puzzle level.
But no. Turns out just made puzzles easy so now this kind of streak is completely devalued. I'm at puzzle ~2300 and getting puzzles 1000-1800. sigh.
But no. Turns out just made puzzles easy so now this kind of streak is completely devalued. I'm at puzzle ~2300 and getting puzzles 1000-1800. sigh.
July 10, 2025 at 5:26 PM
I thought, yay, I finally understand the chess problem that was thrown at me at my puzzle level.
But no. Turns out just made puzzles easy so now this kind of streak is completely devalued. I'm at puzzle ~2300 and getting puzzles 1000-1800. sigh.
But no. Turns out just made puzzles easy so now this kind of streak is completely devalued. I'm at puzzle ~2300 and getting puzzles 1000-1800. sigh.
Been implementing some array method conversion and it's highly satisfying. It's really just "simple" lambda-method-to-loop transforms, low hanging fruit, but it's just showing how much territory I have yet to explore
June 15, 2025 at 9:49 PM
Been implementing some array method conversion and it's highly satisfying. It's really just "simple" lambda-method-to-loop transforms, low hanging fruit, but it's just showing how much territory I have yet to explore
Even with an extensive agent .md prompt I can't seem to convey this to the LLM and it has to rediscover this fact every single time.
It's as if it doesn't expect any system to be as elaborate or thorough as preval is showing it to be.
I guess good for me. But makes test gen from LLM's much harder.
It's as if it doesn't expect any system to be as elaborate or thorough as preval is showing it to be.
I guess good for me. But makes test gen from LLM's much harder.
June 9, 2025 at 3:29 PM
Even with an extensive agent .md prompt I can't seem to convey this to the LLM and it has to rediscover this fact every single time.
It's as if it doesn't expect any system to be as elaborate or thorough as preval is showing it to be.
I guess good for me. But makes test gen from LLM's much harder.
It's as if it doesn't expect any system to be as elaborate or thorough as preval is showing it to be.
I guess good for me. But makes test gen from LLM's much harder.
I love it when a simple trick can do this. I sometimes get stuck on these hard problems so long that I forget to do this little ones.
June 2, 2025 at 9:21 PM
I love it when a simple trick can do this. I sometimes get stuck on these hard problems so long that I forget to do this little ones.
I may not be... Fit to print
May 24, 2025 at 7:13 PM
I may not be... Fit to print
Finally got nice detection for ternaries in Preval.
And by marking them, something like this becomes a lot easier. Before the binding was considered a let but now that I detect and mark them I can consider all reads for that binding as const. And so this sort of rule becomes a lot easier to do! :D
And by marking them, something like this becomes a lot easier. Before the binding was considered a let but now that I detect and mark them I can consider all reads for that binding as const. And so this sort of rule becomes a lot easier to do! :D
May 22, 2025 at 10:05 PM
Finally got nice detection for ternaries in Preval.
And by marking them, something like this becomes a lot easier. Before the binding was considered a let but now that I detect and mark them I can consider all reads for that binding as const. And so this sort of rule becomes a lot easier to do! :D
And by marking them, something like this becomes a lot easier. Before the binding was considered a let but now that I detect and mark them I can consider all reads for that binding as const. And so this sort of rule becomes a lot easier to do! :D
Let's crank up that brain of yours, Gemini. I didn't it see take three thinking passes before, heh. But it was still generating "is `void x` preserving x?" sort of hypothesis. All super basic stuff.
The new tests created are more in the desired direction (though Preval certainly supports these too).
The new tests created are more in the desired direction (though Preval certainly supports these too).
May 18, 2025 at 2:15 PM
Let's crank up that brain of yours, Gemini. I didn't it see take three thinking passes before, heh. But it was still generating "is `void x` preserving x?" sort of hypothesis. All super basic stuff.
The new tests created are more in the desired direction (though Preval certainly supports these too).
The new tests created are more in the desired direction (though Preval certainly supports these too).
Oof. Gemini. How does it go off the rails so hard here?
May 18, 2025 at 1:25 PM
Oof. Gemini. How does it go off the rails so hard here?
It's really like a junior programmer. AI was supposed to not have this. We've gone too far.
May 18, 2025 at 11:45 AM
It's really like a junior programmer. AI was supposed to not have this. We've gone too far.