The reason why I think bad AI writing bugs people so much - it is genuinely disrespectful to ask someone to put effort into reading something you couldn’t be bothered to put an ounce of effort into writing...
"Tech debt" is actually a great name because debt is not always bad. Sometimes it is good! We take on debt to buy a house, to buy a car, and in the most American tradition possible, sometimes we take on debt to ship software faster.
@bcardarella As long as they are in the lead they get to. If their lead slipped they will have to be backwards compatible with other companies standards in order to win users.
Internet explorer use to do wacky shit until they lost market share for example.
@IgorZIJ@swyx yes i have moved specific instructions into project scoped skills. and then globally useful instructions have moved to global skills.
there is a place for agents.md but it is slimmed down.
out of date agents.md/skills is another problem separate from the value of agents.md
@HamelHusain just wrote about this actually - with the thesis of being it is just so disrespectful. the focus is more on the work place than on social media though
https://t.co/3FtcfzpYrv
The reason why I think bad AI writing bugs people so much - it is genuinely disrespectful to ask someone to put effort into reading something you couldn’t be bothered to put an ounce of effort into writing...
@NoemiTitarenco I really like project specific skills. Use those all the time. Have specific use case skills I toggle on and off if the use cases arises. For general skills I had up to 10 but down to like 3. As the models have improved I have needed less.
Had they simply read the response before sending it, they would have realized the LLM was telling them they needed to do the tiniest bit of thinking themselves.
The reason why I think bad AI writing bugs people so much - it is genuinely disrespectful to ask someone to put effort into reading something you couldn’t be bothered to put an ounce of effort into writing...
```
The PR is well structured and handles X, Y, and Z effectively.
A few improvements could be made to Y by doing A or B.
The author is also asking for guidance on how to handle X. You'll still need to provide your own perspective on that.
```
@trikcode I think the best way is to just create artifacts as you go along. I get your point that it is hard to know when you will be rate limited.
I think the next best way might be to pull out your conversation history with one agent using another agent.
https://t.co/SeB9rTUa31
A software engineering truism that feels weaker now after AI codegen:
Brooks’s Law from The Mythical Man-Month - "Adding manpower to a late software project makes it later"
no longer is the equation:
1. Cut features (generally the right answer)
2. Delay release
But now there is a valid third option:
3. Throw another engineer at the problem.
If they are skilled and you develop in an AI friendly manner they should slot right in.