I feel like once Codex & Claude Code get proper conversation branching (like pi /tree) it's going to be a much bigger deal than subagents for solving context pollution
Linear conversations are the bottleneck
Anthropic engineer:
"You can build 5 assistants in one afternoon. Each one handles a task you've been doing manually every single day."
In 45 minutes he builds 5 focused agents from scratch on camera.
Most people are still doing code review, testing, and documentation by hand every single day
Watch the session, then save all templates below ๐
@satanacchio I suppose the real problem is on the tooling side. But that problem is pretty gnarly. Tsconfig has way too many options.
But.. If node could remove this restriction.. It makes typescript even more first class in node, and ts itself closer to "just a linter"
@satanacchio Used it yesterday. Worked like a charm. Exexuted node script from cli and again via a bash subshell FOO_ENV=$(./script.ts)
One thing i tripped on last month, was basically not being able to execute ts files from a package.
CTO posts bulletin about ai.
Does he know the vscode downloaded from jamf cant sign into github? The ai code review tool... Doesn't even work? Does he know devs are phoning in the test coverage requirements?
The scourge of "Claude said..." is rampant.
If you use "Claude said", it had better be in a sentence that is structured something like:
"When I did some digging, Claude said <thing>, so I looked into that and it looks like it might work based on <x, y, z>."
not
"Well, Claude said <thing> so you're wrong."
Claude says many things. Claude is very often wrong. Even when it's right, you should know why it's right.
Own your work. Don't delegate all thinking to AI.
The scourge of "Claude said..." is rampant.
If you use "Claude said", it had better be in a sentence that is structured something like:
"When I did some digging, Claude said <thing>, so I looked into that and it looks like it might work based on <x, y, z>."
not
"Well, Claude said <thing> so you're wrong."
Claude says many things. Claude is very often wrong. Even when it's right, you should know why it's right.
Own your work. Don't delegate all thinking to AI.
@kentcdodds If I shoe up in person, assuming you chose every word, I might have subtle assumptions/presumptions which don't exactly match your intentions.
@kentcdodds If we ignore confext, what were the final prompts asking for?
Typically when I read a pdf like this, I assume a human choose every word. Is that the case here?
If you could have worded something a little better thats one thing.
The AI ponzi scheme goes like this:
Everyone is generating all these long ass docs and then passing them off for others to read
Then the person receiving is like, wtf this is way too long, and hands that into an AI to read and summarize
Then they are generating a long ass response back
and this cycle goes like that forever. and we call this work now ๐
The token lords watch this from their towers nodding and grinning.
XState Store v4 is released ๐
Simple, small state management for stores (like Zustand/Redux) + atoms (like Jotai/Recoil)
Works everywhere: vanilla, React, Vue, Svelte, Solid, and more. Event-driven, type-safe, batteries included, *really* agent-friendly.
npm i @xstate/store
Some new features:
@kentcdodds Now if you happened to use some ai but weren't advertising your use of it, you implicitly give it the traditional full Kent C. Dodds signoff and stamp of approval.
@kentcdodds In a way I don't want to know that you are using ai. It feels cheap, I don't want to read something that I suspect is ai generated. I'd rather just read your prompt to create this, because I can then directly read your intentions. And you can't really back pedal -
Iโve decided to move back to sublime text.
- VSCode is a security liability
- I donโt edit code hardly ever, so I donโt need any IDE like features
- sublime is fast and comfortable to use
- it brings back good memories