With all the hype around using AI to improve developer efficiency, I haven't met anyone who is actually using AI to make a meaningful contribution to their developer productivity, quality of life, etc...<p>That said I'm sure there are people out there doing it. So I'm curious if there are members of this community that are using AI successfully in their org. What specifically are you using it for? What models/providers are you using?
I've been using co-pilot as well as Claude (Sonnet 3-3.5) and ChatGPT (4-4o) since the beginning of the year with somewhat mixed success. My general take is that, at least so far, the improvement is very much marginal. These tools save me a few seconds by quickly surfacing information that otherwise would require me having to google and search through docs/stackoverflow/etc, or by writing code that's generally tedious (like tests).<p>Co-pilot is mostly useful to stay in the zone, allowing me to focus on a larger task and letting it get some of the details right (which it does ~90% of the time).<p>On the other hand, I use ChatGPT/Claude for more open ended tasks (e.g. "I got this <insert obscure> error", "how do I configure this framework so that xxx") which previously I would have googled hoping to find a stackoverflow answer or a doc page somewhere. For this use case I'd say it's ~50% successful, but I often have to deal with hallucinations - some times just following up with "Are you sure?" does help, but it's hit/miss.<p>As I said a the beginning, mostly marginal improvement. It's definitely saved me time, but thus far nothing that I couldn't do myself by spending a little bit more time. Largely it is a nice to have, not a need to have.
AI has definitely boosted my productivity, although I can't quantify the time savings.<p>Usually using Raycast (a Mac app), I enter my prompts for GPT-3.5 Turbo or Llama 3 from anywhere without having to open a browser, go to an AI UI, log in, etc.<p>I did use CodeWhisperer for awhile in VS Code but found it kind of disorienting.<p>I use it several times a day for generating JS code snippets or boilerplate that I can tweak to meet my needs. I've found it surprisingly helpful for writing Python code that generates charts using Matplotlib or Seaborn (who can ever remember any of those functions and options?) – easily saving me hours of manual research, trial and error. And I use it to refine or optimize SQL queries.
Would be interesting to see people's effectiveness at these individual use cases and which models performed the best:<p>- Unit Tests
- Refactoring (classes, methods, and functions)
- Refactoring (cross-module or cross-interface)
- Translation (one language or run-time to another)
- Translation (one library to another - e.g. a ui kit)
- Code snippets or code completions while developing
- Building UIs from API
- Building APIs from UI
I've been using DialDoge to get insights and actions from transcripts and meetings with clients. If you're in an org that fails to document and is all meetings, it really helps to hold yourself and others accountable.<p>Like we already talked about that. Let's move on