@bcardarella @CaesarDavid_ @dhh To be clear, I am not dismissing your overall question, which is "how well AI can do the job". That's a valid question but I am skeptical the benchmarks answer that either. Just compare the AGENTS.md files of Elixir and Go:
github.com/basecamp/once-campfire-elixir/blob/5a8…
github.com/basecamp/once-campfire-go/blob/9b0ef06…
Elixir is asked to use Rails as a reference, Go is asked to use Rust. So even if a given model knows how to improve Elixir, we are basically pruning that search space by saying "Preserve the pinned Rails reference behavior" which points to a checkout of the Rails app. And then the verifier is rejecting pull requests that try to deviate from it.
Anyway, I sent a PR to address that, we will see. No one is asking for special attention, just to be treated the same.
19 likes · 2 reposts · 679 views
How do you know that @DHH is a liar and a grifter? You just watch what he does.
He (well, his AI slop bot) closed the PR that @ZachSDaniel1 made to fix Elixir with "doesn't maintain compatibility with Rails" when his own Rust slop doesn't either
once-campfire-elixir#1 t.co
90 likes · 3 reposts · 4,735 views
@dhh When Rust optimisations are applied to Elixir, they seem to be almost equal when it comes to speed.
No surprise though! Benchmark has no writes. I believe both tuned versions have like 99.9% cache hit ratios, so it's not very realistic.
once-campfire-elixir#5 t.co
13 likes · 2 reposts · 590 views
@ZachSDaniel1 Here's a PR with the code: once-campfire-elixir#5
I also tested it against an older rust commit (3fe9f61) - though not the oldest. It looks like there has been a lot of performance improvements made to the rust repo.
P.S. The screenshots I posted are missing "req/s". t.co
19 likes · 1 repost · 1,416 views
@dhh Here you go once-campfire-elixir#2
got a 10x speed up on the Room page requests, and sped up all the other requests significantly as well. the biggest gains came from some no brainers: implement caching and pool the db connections with proper async t.co
3 likes · 255 views
Mentions
@bcardarella @CaesarDavid_ @dhh To be clear, I am not dismissing your overall question, which is "how well AI can do the job". That's a valid question but I am skeptical the benchmarks answer that either. Just compare the AGENTS.md files of Elixir and Go: github.com/basecamp/once-campfire-elixir/blob/5a8… github.com/basecamp/once-campfire-go/blob/9b0ef06… Elixir is asked to use Rails as a reference, Go is asked to use Rust. So even if a given model knows how to improve Elixir, we are basically pruning that search space by saying "Preserve the pinned Rails reference behavior" which points to a checkout of the Rails app. And then the verifier is rejecting pull requests that try to deviate from it. Anyway, I sent a PR to address that, we will see. No one is asking for special attention, just to be treated the same.
How do you know that @DHH is a liar and a grifter? You just watch what he does. He (well, his AI slop bot) closed the PR that @ZachSDaniel1 made to fix Elixir with "doesn't maintain compatibility with Rails" when his own Rust slop doesn't either once-campfire-elixir#1 t.co
@dhh When Rust optimisations are applied to Elixir, they seem to be almost equal when it comes to speed. No surprise though! Benchmark has no writes. I believe both tuned versions have like 99.9% cache hit ratios, so it's not very realistic. once-campfire-elixir#5 t.co
@ZachSDaniel1 Here's a PR with the code: once-campfire-elixir#5 I also tested it against an older rust commit (3fe9f61) - though not the oldest. It looks like there has been a lot of performance improvements made to the rust repo. P.S. The screenshots I posted are missing "req/s". t.co
@dhh Here you go once-campfire-elixir#2 got a 10x speed up on the Room page requests, and sped up all the other requests significantly as well. the biggest gains came from some no brainers: implement caching and pool the db connections with proper async t.co