Hacker Newsnew | past | comments | ask | show | jobs | submit | q8zd3's commentslogin

I enjoyed "old fashioned" hackathons, which you had 24h to build or play with some technology or API. It has lost its charm (for me)since it moved to be startup-ish like events that you need to pitch a product instead.

I think it will be considered a "blast from the past" at some point, due to the AI era we are getting into.


retro hackathons will replace retro arcades as gen x retires and millenials are 50+ and have money and kids out of the house. haha.


Some companies and OSS groups have a policy to inform if a patch's source code was AI generated, even if it was just parts of it. Committs messages are the "obvious place" to put it.


Honest question: Does anyone know about any quantitative study or analysis on productivity gains using code assistants? Asking for numbers comparing between the "pre AI era" and now.

Also, I have the impression that LLMs bring some gains or benefits for individuals but not relevant enough at the organization level.


I believe it is very hard to quantify „productivity“. I’m sure that for suitable definitions you can find gains from coding assistants. Personally I get more code written and more features implemented. Yet I’m very wary of coding assistants because I believe they deal a fatal blow to my ability to understand the system. All LLM generated code is (at best!) code that was written by an intern which I just helped with the design and reviewed (unless productivity expectations cut down my review time and I get LLM assistance for reviews too). My grasp on the inner working of that code is much more tenuous than had I written it myself. I will never become an expert by just reviewing code and prompting.

For a while this is not a problem: I can work with my current mental model. But every generated PR erodes my expertise a little bit. Eventually my mental model won’t fit anymore.

So how much of that model maintenance should I count into my productivity metric? Does that even matter or will the next model be able to reason well enough that my mental model doesn’t matter?


Here's the big one that was being passed around months ago, which nowadays usually gets dismissed out of hand because of when they did it (while ignoring the relative finding): https://metr.org/blog/2025-07-10-early-2025-ai-experienced-o...

Here's a slightly more recent one focused more on comprehension/learning than productivity: https://www.anthropic.com/research/AI-assistance-coding-skil...

Metr attempted to redo that first one to get trends over time, but couldn't recruit enough developers to get reliable results for it.


Followup question: Does anyone know about any quantitative study or analysis on productivity without using code assistants? (as a baseline)


It feels like a dirty secret, doesn't it?


Yeah, corps don't want you to know how to code, they want you to be a prompter...


Sometimes I squeeze in an hour or so a day to read. Living on the edge, looking for the next dopamine hit.


You guys have enough attention span left to read?

/s


I was not ready to start my day with a OS/2 Warp nostalgia feeling


Yeah, not a cult at all.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: