Bun's experimental Rust rewrite hits 99.8% test compatibility on Linux x64 glibc (twitter.com)

Article summary

Bun's experimental Rust rewrite has achieved 99.8% test compatibility on Linux x64 glibc, with the rewrite taking 6 days to complete. The rewrite was done using a large language model, with some commenters estimating the cost to be around $0.5 million in compute time. The achievement has sparked discussion about the potential of AI-assisted programming and its implications for the industry. The original article is not available, but the comments provide insight into the project's progress and its potential impact.

Main themes

• AI-assisted programming
• Rust rewrite
• Compute costs
• Industry implications
• Language models

What commenters say

• The use of large language models can significantly speed up the process of rewriting code in a different language, making it a potentially game-changing tool for the industry.
• The cost of using AI-assisted programming tools may be prohibitively expensive for some companies, potentially creating a divide between those who can afford it and those who cannot.
• The claim that the rewrite would have taken hundreds of engineers more than a year to complete is an exaggeration, and the actual time and cost savings may be less significant than claimed.
• The ability to automate coding tasks using AI could lead to a future where only tests and specs are committed to git repos, with the actual code being generated automatically.
• The use of AI-assisted programming tools does not necessarily guarantee better results, and human programmers may still be necessary to review and refine the generated code.
• The potential for AI-assisted programming to disrupt the industry is significant, but it is still unclear what the actual impact will be and how it will change the way companies approach software development.
• The cost of using AI-assisted programming tools may be worth it for some companies, even if it is at a loss in terms of revenue, due to the potential benefits and competitive advantage it could provide.
• The effectiveness of AI-assisted programming tools depends on the quality of the input and the specific task at hand, with some tasks being more suitable for automation than others.