I do love one bot asking another bot to sign a CLA! - https://github.com/dotnet/runtime/pull/115732#issuecomment-2...
Well?? Did it sign it???
Not sure if a chatbot can legally sign a contract, we'd better ask ChatGPT for a second opinion.
Watching AI drive Microsoft employees insane
61–70 of 580 posts
Re: Watching AI drive Microsoft employees insane
#62With how stochastic the process is it makes it basically unusable for any large scale task. What's the plan? To roll the dice until the answer pops up? That would be maybe viable if there was a way to automatically evaluate it 100% but with a human in the loop required it becomes untenable.
The plan is to improve AI agents from their current ~intern level to a level of a good engineer.
I mean, I think this is a _lot_ worse than an intern. An intern isn't constantly going to make PRs with failing CI, for a start.
Re: Watching AI drive Microsoft employees insane
#63"fix failing tests" does never yield any good results for me either
Re: Watching AI drive Microsoft employees insane
#64Well, the coding agent is pretty much a junior dev at the moment. The seniors are teaching it. Give it a 100k PRs with senior developer feedback and it'll improve just like you'd anticipate a junior would. There is no way that FANG aren't using the comments by the seniors as training data for their next version.
It's a long-term play to have pricey senior developers argue with an llm
Re: Watching AI drive Microsoft employees insane
#65[flagged]
This analogy would only work if the electric light required far more work to use than a gas lamp and tended to randomly explode.
And didn’t actually provide light, but everyone on 19th century twitter says that it will one day provide light if you believe hard enough, so you should rip out your gas lamps and install it now.
Like, this is just generation of useless busy-work, as far as I can see; it is clearly worse than useless. The PRs don't even have passing CI!
Re: Watching AI drive Microsoft employees insane
#66GitHub is not the place to write code. IDE is the place. Along with pre CI checks, some tests, coverage etc. they should get some PM before making decisions..
This is the future envisioned by Microsoft. Vibe coding all the way down, social network style. They are putting this in front of the developers as take it or leave it deal. I left the platform, doing my coding old way, hosting it somewhere else. Discoverability? I don't care. I'm coding it for myself and hosting in the open. If somebody finds it, nice. Otherwise, mneh.
> I left the platform, doing my coding old way, hosting it somewhere else.
may you please let me know where are you hosting the code ? would love to migrate as well.
thank you !
Re: Watching AI drive Microsoft employees insane
#67Are people really doing coding with agents through PRs? This has to be a huge waste of resources. It is normal to preempt things like this when working with agents. That is easy to do in real time, but it must be difficult to see what the agent is attempting when they publish made up bullshit in a PR. It seems very common for an agent to cheat and brute force solutions to get around a non-trivial issue. In my experie…
Googles jules does the same (but was only published yesterday or so). I think it might be a good workflow if the agent is good enough. Copilot seems not to be in these examples and then I imagine it becomes quite tedious to have a PR for every iteration with the AI.
Re: Watching AI drive Microsoft employees insane
#68Earlier quoted context omitted.
As long as the resulting PR is less than 100 lines and the AI is a bit more self sufficient (like actually making sure tests pass before "pushing") it would be ok I think. I think this process is intended for fixing papercuts rather than building anything involved. It just isn't good enough yet.
As a matter of principle I don't use any network which is trained on non-consensual data ripped of its source and license information. Other than that, I don't think this is bad tech, however, this brings another slippery slope. Today it's as you say: > I think this process is intended for fixing papercuts rather than building anything involved. It just isn't good enough yet. After sufficient T somebody will rephrase…
> AI will never be a core and irreplaceable part of my development workflow.
Unless AI use becomes a KPI in your annual review.
Duolingo did that just recently, for example.
I am developing serious regrets for conflating "computing as a medium for personal expression" with "computing for livelihood" early on.
Re: Watching AI drive Microsoft employees insane
#69Clumsy but this might be the future -- humans adjusting to AI workflow, not the other way. Much easier (for AI developers).
Re: Watching AI drive Microsoft employees insane
#70Earlier quoted context omitted.
The plan is to improve AI agents from their current ~intern level to a level of a good engineer.
Seems like that is taking a very long time, on top of some very grandiose promises being delivered today.
I look back over the past 2-3 years and am pretty amazed with how quick change and progress have been made. The promises are indeed large but the speed of progress has been fast. Not defending the promise but “taking a very long time” does not seem to be an accurate representation.