Yeah -- 93 commits in there. Not to say AI can't make an improvement in a domain like this, but it's definitely not a YOLO situation right now. I use AI code heavily at work and it's quite prolific but also often creates bugs of various severities.
Having done a lot of optimizer work, there's definitely a category of things that I've avoided producing preliminary numbers for because I don't see how to get robust production code out of it. A lot of people are going to have to learn (and develop social scripts to explain) that not every 50% speedup is a good idea to pursue.
This article has a weird tone. It directly criticizes the results of this AI-driven coding effort (which the author admits is of bad quality), while at the same time it tries to reassure the reader that they're not here to criticize the author of the PR.
not if they are different authors; AI wrote the code and the human the PR. These used to implicitly be the same person but this is something important we have lost. You've now got a smart, former developer in a position of power vouching for code they essentially didn't write.
Instead of writing a blog post he could have written agentic instructions to fix the mr and prove it works and make another mr and send it back to the ceo
Yeah -- 93 commits in there. Not to say AI can't make an improvement in a domain like this, but it's definitely not a YOLO situation right now. I use AI code heavily at work and it's quite prolific but also often creates bugs of various severities.
Having done a lot of optimizer work, there's definitely a category of things that I've avoided producing preliminary numbers for because I don't see how to get robust production code out of it. A lot of people are going to have to learn (and develop social scripts to explain) that not every 50% speedup is a good idea to pursue.
This article has a weird tone. It directly criticizes the results of this AI-driven coding effort (which the author admits is of bad quality), while at the same time it tries to reassure the reader that they're not here to criticize the author of the PR.
Only of these can be true at the same time.
not if they are different authors; AI wrote the code and the human the PR. These used to implicitly be the same person but this is something important we have lost. You've now got a smart, former developer in a position of power vouching for code they essentially didn't write.
Instead of writing a blog post he could have written agentic instructions to fix the mr and prove it works and make another mr and send it back to the ceo
Why? Tobi doesn't need help wrangling his own bullshit.