I’d love some thoughts from the team on what we should be building towards, given the progress of LLMs to date, and what will happen over the next 12-18 months. With our goals to democratize publishing and make WordPress and as much of its ecosystem of extensions available to people regardless of their preferred language, what gets us there the fastest? What should be automated, and when should humans come in? Feel free to bring in folks from the AI team for this brainstorm. Let’s also assume that costs of the frontier-level machine translation for tens of millions of strings will be either donated by companies like Google since we’re open source or in the low tens of thousands of dollars of inference so easily covered by sponsorships.
It’s a good moment to do a ground-up reset of our strategy and thinking here.