24% of increased productivity (yes, this is assuming of course that the “proxy” of merged PRs reflects productivity) is actually a pretty big deal. Given the salary of developers, this translates to tens of thousands of dollars per year, per developer.
My guess is they used # of PRs as a measure as it’s easy to obtain, while other measures are hard, may be due to other factors, etc.
FWIW I saw a similar number for myself, around 30% more PRs in the last 6 months, compared to the 6 months before that (I picked up agentic coding around at the start of the year). And a similar increase for closed issues.
In my case this clearly doesn’t translate to as much value for the organization, or rather, it’s hard to say, as many of those PRs were things I wouldn’t even have done without AI support. This means they were low priority. However, many were of the cleanup/refactor type, so they might result in speedups later.
IMHO Anki (the software) has very little value now with coding agents. The spaced repetition algorithm is quite simple and just 100 lines of python (or even less). Every coding agent can write that for you in minutes. The value of Anki comes from (1) its flexibility, and (2) the pre-made decks, but building an app that’s tailored to you is better, and you can find a lot of word lists online if that’s what you need.
If you have a Claude (or similar) subscription, you can build yourself a language learning app that’s better for you than most products out there, in a small number of days. TTS, dictionary lookup, you name it, the AI will build it. Azure TTS has a generous free tier, so all you need is the AI subscription (one month is plenty) and a place to host your app. It’s a simple prompt to ask the model to modify cards so it plays the audio first before showing the text. Off the top of my head I don’t have a good idea for your specific issue of having audio for the same word in multiple contexts. Maybe for this you could use LLM-generated sentences.
About your footnotes, personally I think vocabulary becomes more important with progress, because you just won’t be exposed often enough to some rarer words (e.g. politics or economy), or learning words in groups can help you better understand them. And flashcards can help not forget them. But don’t learn words with flash cards, learn them in context.
The problem with proof is that it’s a bit hard sometimes to convey the value. The point is not to find bugs, but to prove that there are none (of a certain class; under certain assumptions; etc). But it’s a hard story to sell, so often the marketing is around “look at this bug we found”.
I personally thought to myself "written by AI" after this part:
... the restaurant was fully booked. No warmth. No conversation. Just a long wait and a closed door. In trying to humanise the process, he’d made it worse.
I'm sure some people write this way, but most don't. And AI writes this way.
My guess is they used # of PRs as a measure as it’s easy to obtain, while other measures are hard, may be due to other factors, etc.
FWIW I saw a similar number for myself, around 30% more PRs in the last 6 months, compared to the 6 months before that (I picked up agentic coding around at the start of the year). And a similar increase for closed issues.
In my case this clearly doesn’t translate to as much value for the organization, or rather, it’s hard to say, as many of those PRs were things I wouldn’t even have done without AI support. This means they were low priority. However, many were of the cleanup/refactor type, so they might result in speedups later.